Adaptive Strategies and Population Ecology of Northern Grouse
This page intentionally left blank
Adaptive Strategies and Populdtion Ecology of Northern Grouse
Arthur T. Bergerud and Michael W. Gratson, Editors
A Wildlife Management Institute-sponsored Book Published by the University of Minnesota Press Minneapolis
Copyright © 1988 by the University of Minnesota All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher. Published by the University of Minnesota Press 2037 University Avenue Southeast, Minneapolis MN 55414. Published simultaneously in Canada by Fitzhenry & Whiteside Limited, Markham. Printed in the United States of America. Library of Congress Cataloging-in-Publication Data Adaptive strategies and population ecology of northern grouse. "A Wildlife Management Institute sponsored book." Includes bibliographies and index. 1. Grouse. 2. Bird populations. 3. Adaptation (Biology) I. Bergerud, A. T. II. Gratson, Michael W. QL696.G285A33 1987 598'.616 86-19248 ISBN 0-8166-1469-5 ISBN 0-8166-1470-9 (pbk.: v. 1) ISBN 0-8166-1471-7 (pbk.: v. 2) The University of Minnesota is an equal-opportunity educator and employer.
This book is dedicated to those who walk in the Leopold tradition Robert McCabe J. J. (Joe) Hickey Al Hochbaum Fred Hamerstrom Jr.
Authors A. T. Bergerud, Biology Department, University of Victoria, Victoria, British Columbia, Canada R. G. Davies, Fish and Wildlife Branch, Ministry of the Environment, Nanaimo, British Columbia, Canada A. Gardarsson, Institute of Biology, University of Iceland, Reykjavik, Iceland M. W. Gratson, Biology Department, University of Victoria, Victoria, British Columbia, Canada J. E. Hartzler, Department of Zoology, University of Montana, Missoula, Montana R. A. Huempfner, Advanced Telemetry Systems, Inc., 23859 H. E. Highway 65, Bethel, Minnesota D. A. Jenni, Department of Zoology, University of Montana, Missoula, Montana D. H. Mossop, Yukon Wildlife Branch, Box 2703 White Horse, Yukon, Canada S. Myrberget, DVF, Viltforskningen, Eglesetergt. 10, N-7000 Trondheim, Norway R. E. Page, Research Branch, Ministry of Forestry, 1450 Government Street, Victoria, British Columbia, Canada R. K. Schmidt, I. D. Systems Limited, 966 Waverly, Winnipeg, Canada W. D. Svedarsky, Division of Agriculture, University of Minnesota, Crookston, Minnesota J. R. Tester, Department of Ecology and Behavioral Biology, University of Minnesota, St. Paul, Minnesota
Preface This book is about the biology of a familiar and important group of birds—the grouse, or tetraonids. These large-bodied birds of forest and field will be recognized by both the weekend hunting enthusiast and the rural resident. The sportsman will recall pleasant autumn days afield and the explosive flush and the blur of wings that got away. The rural citizen may recall the protective, hen matriarch shepherding her chicks across a backcountry road in July. And what early riser in May has not heard that mysterious, inboard-motor sound that starts in the bottomland when owls head home and roosters crow? Many will know this muffled roll call as the male ruffed grouse drumming on his favorite mossy log; like the cock rooster, he proclaims another dawn and that he is available if any hens are so inclined. This book is about the behavior of such individuals as the hen matriarch and the drumming male in the lowland. What are the behavioral tactics that the hen has adopted to choose a mate and a nest site, and to raise her chicks? How does the drumming male advertise for hens in a polygynous society yet elude the goshawk? For both the hen and the cock, the major problems in life are first to stay alive and avoid predators and second to breed and keep the family line intact. To accomplish these objectives grouse have developed a rich repertoire of behavioral patterns that this book discusses. But the book is also about groups of interbreeding individuals that live in distinct areas—populations. The major question here is what regulates the numbers of grouse in a population. If individual grouse are so well adjusted to their environment, why don't these collective adaptations result in continuous population increase? Why are we not inundated with grouse, as the hunter might wish? Why, as so many naturalists have asked since records were first kept, do many grouse populations fluctuate inexplicably in cycles of scarcity and abundance? Some populations show cycles of 3-4 years and others 10-year cycles. This book provides some insights.
vii
viii
PREFACE
The general question of what limits numbers is also sought by other specialists. The entomologist is concerned about insect outbreaks, the fisheries scientist and deer biologist wish to increase the stocks of fish and deer. Demographers and ecologists seek the answer for the human race—what will limit the human population explosion? This question of the limits of growth is the central issue in the important field of population ecology. Insight from grouse studies may be relevant to the larger fundamental problem. I have been involved in attempting to unravel the question of population limitation for grouse since 1955. First I studied willow ptarmigan, a grouse of the tundra, in Newfoundland during an entire 10-year cycle of change. Then I moved to British Columbia and investigated the behavior and dynamics of several blue grouse populations from 1970 to 1980. Concurrently, I cooperated with Rick Davies, who followed the fluctuations in numbers of a ruffed grouse population for 15 years, 1968 to 1982. From these studies, I concluded that the annual changes in the numbers of adult grouse breeding in the spring were caused primarily by yearly changes in nesting success and survival of young chicks in the previous summer. An alternative view to this production hypothesis is a widely supported and quoted theory that grouse regulate their numbers through spacing behavior; as populations increase, space for breeding becomes limited and birds that cannot compete successfully have a high death rate. This theory emphasizes the mortality of birds older than 4 months, whereas the production or breeding success theory stresses the size of the new generation that successfully navigated the early weeks of life. A great deal of research has been done on grouse to decide between these views. But much of the data was buried in unpublished theses and government reports. I felt that a synthesis for the breeding success hypothesis was possible by reviewing the literature and publishing the data from the blue and ruffed grouse studies together in one volume. Further, I solicited my colleagues with similar findings to provide chapters for other grouse species, based on unpublished research. Together we present the arguments here. As we assembled the chapters that dealt with both population ecology and behavior, we saw that a synthesis was warranted on the sociobiology of grouse. Sociobiology is a new field that attempts to explain the adaptive significance of behavior at the individual level. A major tenet of sociobiology is that a few environmental variables can explain much of the behavior of a species. For grouse, the variables mentioned time and again in the literature, and by the authors of this book, are food supplies and predators. This book offers some new hypotheses to explain why grouse perform certain behaviors such as roosting under snow, migrating, and making long movements as young broods. The role of food shortages versus predation is contrasted for many of these behavior patterns. An underlying theme is that much of what grouse do results from the need to remain inconspicuous to avoid predators.
PREFACE
ix
Grouse can remain inconspicuous to predators by seeking concealing plant cover, but they can also be inconspicuous if they are scarce in space. There appears to be a cover-space trade-off. One option for a grouse is to seek the best concealing cover, but if other birds do likewise an individual's inconspicuousness is compromised by numbers. A second option is to seek less-suitable cover where there is more space between birds. The first option could be called the densitytolerant mode, the second, the density-intolerant mode. Each mode appears to have its own distinct suite of behaviors. A major hypothesis of the book is that mutual attraction exists between birds of the same mode and that this assortment is a fundamental aspect of dispersal for noncyclic grouse and of population change for grouse that show 10-year cycles in numbers. The last chapter of the book discusses how to increase the numbers of grouse. One question posed is, Why do two separate populations exist at different mean densities over many years? This is not exactly the same question as why one population increases or decreases between years. Another question asked is, Does hunting reduce the size of next year's breeding population or does the harvest of birds simply take a surplus that would die anyway without breeding? This closing chapter also discusses the issue of whether food is in short supply and how to manage the habitat relative to the cover-space trade-off to improve nesting success. This work is divided into two major sections: "Population Studies" and "Theory and Synthesis." The empirical and original studies, Chapters 1-11, are arranged by a habitat-species classification (Hamerstrom & Hamerstrom 1961, de Vos 1979). North American forest grouse include spruce (Dendragapus [Canachites] canadensis), blue (Dendragapus obscurus), and ruffed (Bonasa umbellus) grouse; steppe grouse are the sharp-tailed grouse (Pedioecetes phasianellus) (more recently Tympanuchus phasianellus), prairie chickens (Tympanuchus cupido and T. pallidicinctus, if the latter is recognized), and sage grouse (Centrocercus urophasianus); and tundra grouse are white-tailed ptarmigan (Lagopus leucurus), rock (L. mutus), and willow ptarmigan (L. lagopus). Other classifications could have been used, but we wanted to emphasize the mold of the extrinsic environment—the habitat—as the prime mover in the evolution of behavior and adaptive strategies. Spruce and blue grouse, both of coniferous forests, go naturally together; their behavior and demography are similar, and recently the nomenclature of spruce grouse has been changed to Dendragapus canadensis. Ruffed grouse differ in that they generally lay large clutches and suffer high mortality rates. However, they also live under the forest canopy; their flocking, feeding, and breeding behaviors more closely resemble those of the blue and spruce grouse than those of the prairie grouse (sharp-tailed grouse and prairie chickens) that show demographic aspects similar to ruffed grouse. Of the steppe grouse, prairie chickens and sharp-tailed grouse are a natural pair in both behavior and population ecology; they commonly hybridize and now are considered to be congeneric, under Tympanuchus. Demographic properties of sage grouse
x
PREFACE
populations are akin to blue and spruce grouse (they have small clutch sizes and low mortality rates of adult individuals), but sage grouse live in open habitats and breed at communal advertising locations (leks) as do the prairie grouse. The three ptarmigan are found in the alpine and arctic tundras; each species is monogamous and shows other behavioral and demographic attributes in common with the other two. Many people have helped in the preparation of the book, taken part in the field work, and provided unpublished material. I thank you all, and your names are acknowledged by chapters in the Acknowledgments. I owe a special thanks to my coeditor, Michael Gratson. Michael worried over every page, especially those I wrote, trying to reduce the jargon and make the text more readable. He is a dyedin-the-wool grouser, who sees the birds in his dreams. One early morning, he pounded on my farmhouse door; it had come to him in his sleep, a hypothesis for a major question debated in the theory of sexual selection: Why do hen grouse choose as mates those males that tend to advertise in the center of the communal display grounds (leks), instead of peripheral males? Gratson's theory was that females mostly prefer central males because these males are more restricted and predictable in their movements, whereas males on the outside of leks reduce the inconspicuousness of females by unsolicited following and overenthusiastic courtship. One of the reviewers suggested that we could tone down the more novel hypotheses in the latter half of the book and present only a tidy summary of what is known. But then he said the book would not be half the fun it is now. I did want to put it together my way. Many of the hypotheses have a strong data base, but some involve a great deal of speculation where I have less experience. It certainly was an exciting winter putting the book together—sitting around the wood stove in the kitchen, arguing ideas with Gratson and sometimes Dave Mossop. I hope that some of the excitement can be shared with you. A. T. Bergerud
Acknowledgments The editors and authors thank the following individuals and organizations for their support. Chapter 1 J. F. Bendell, J. O. Anuk, K. R. Bunnell, K. R. Summers, T. B. Barnes, J. Theberge, A. T. Bergerud, K. R. Reid. Chapter 2 J. A. Redfield, L. M. Rotterman, H. E. Butler, R. Hunter, H. D. Hemus, J. Morris, J. Hedberg, T. Shields, R. Whitney-Ernst, J. F. Bendell, D. G. King, D. H. Mossop, F. C. Zwickel, B. Willie, R. L. Brown, Natural Sciences and Engineering Research Council. Chapter 3 D. R. Miller, R. Whitney-Ernst, R. L. Brown,D. Lowe, R. W. Ritcey, L. M. Rotterman, H. E. Butler, J. L. Donaldson, T. Shields, Natural Sciences and Engineering Research Council, British Columbia Fish and Wildlife Branch. Chapter 4 R. Scuster, R. Reichle, V. Kuechle, g. Erikson, S. Maxson, H. Archibald, P. Bonde, J. Jaeger, H. Tordoff, P. Stolen, U.S. National Institute of Health, U.S. Department of Energy. Chapter 5 R. K. Anderson, J. E. Toepfer, College of Natural ResourcesUniversity of Wisconsin-Stevens Point, Wisconsin Department of Natural Resources, The Prairie Chicken Society. Chapter 6 R. Seabloom, L. Oring, B. Youngquist, L. Rydell, The University of North Dakota-Grand Forks, Minnesota Agricultural Experiment Station, The Nature Conservancy, Minnesota Prairie Chicken Society, Minnesota Department of Natural Resources, U.S. Fish and Wildlife Service. Chapter 7 D. Pyrah, W. H. Bell, Montana Cooperative Wildlife Research Unit, Montana Department of Game, Fish, and Parks, Bureau of Land Management, University of Montana Zoology Department, Wildlife Unit. Chapter 8 R. Schmidt, Sr., C. E. Braun, R. A. Ryder, the late A. J. Pakulak, the U.S. Park Service—Rocky Mountain National Park, the Society of Sigma Xi.
xi
xii
Chapter 9
ACKNOWLEDGEMENTS
F. A. Pitelka, J. Davis, C. P. Huffaker, G. R. Miller, R. Moss, A. Watson, G. M. Christman, K. Geirmundsson, G. Gudmundsson, S. Stefansson, G. Svararsson, T. Thorsteinsson, U.S. National Science Foundation, Institute for International Education, University of California, University of Iceland. Chapter 10 G. Mossop. Chapter 11 B. Egeland, M. Bjaerum, P. Tallantire, R. Blom, K. E. Erikstad, and many others. Chapter 12 C. J. Walters, D. Chitty. Chapter 13 D. H. Mossop, M. W. Gratson, W. E. Mercer, S. J. Hannon, D. M. Keppie, R. E. Autenrieth, Natural Sciences and Engineering Research Council. Chapter 14 L. G. Sopuck, P. W. Herzog, R. E. Autenrieth, L. H. Kermott, L. M. Kirsh, D. M. Keppie, Natural Sciences and Engineering Research Council. Chapter 15 R. B. Weeden, J. B. Theberge, R. J. Hudson, H. Hristienko, T. W. Mussehl, M. Sexton, W. Wishart, S. R. Barber, J. F. Bendell, K. J. Szuba, W. E. Berg, R. L. Eng, G. D. Kobriger, D. M. Keppie, P. W. Herzog, K. M. Giesen, S. J. Hannon, M. W. Gratson, Natural Sciences and Engineering Research Council. Chapter 16 L. M. Kirsh, F. Hamerstrom, F. N. Hamerstrom, Jr., R. K. Anderson. J. W. Bradbury, R. M. Gibson, B. Tordoff,and others reviewed the book in its entirety; E. V. Hooke and P. G. Bressler drafted figures; S. McMillian typed many drafts; G. Gratson compiled the bibliography; grouse sketches are by D. H. Mossop. A special thanks to H. E. Butler and G. Gratson, and to J. E. Mclnerney, chair of the Department of Biology, University of Victoria, British Columbia.
Contents Authors Preface Acknowledgments
vi vii xi
Part I. Population Studies FOREST GROUSE Chapter 1. A Relation between Aggressive Behavior and Population Dynamics in Blue Grouse D. H. Mossop 1.1 1.2
Introduction Study areas and methodology 1.2.1 Population demography 1.2.2 Behavior 1.3 Demographic findings 1.4 Behavioral findings 1.4.1 Spontaneous behavior 1.4.2 Behavior in response to observer 1.4.3 Intraspecific behavior in census situation 1.4.4 Predator avoidance behavior 1.4.5 Behavior in the test arena 1.5 Discussion 1.51. Population trends 1.5.2 Aggressive behavior 1.5.3 Aggression and population dynamics 1.6 Summary
3 3 5 5 7 7 9
9 10 12 15 21 21 25 25 27 28
Chapter 2. Demography and Behavior of Insular Blue Grouse Populations A. T. Bergerud 29 2.1
Introduction
29 xiii
xiv
CONTENTS
2.2 2.3 2.4
Description of study areas Methods Demography 2.4.1 Survival and dispersal of founders 2.4.2 Survival of progeny 2.4.3 Mechanisms of population change 2.5 Behavior 2.5.1 Behavior of the founders 2.5.2 Behavior of the hand-raised chicks 2.5.3 Behavior of the males 2.5.4 Behavior of female progeny 2.5.5 Behavior of progeny at capture sites 2.6 Discussion 2.6.1 Demography 2.6.2 Behavior in space and time 2.6.3 Fitness and densities 2.7 Summary
31 31 35 35 39 43 46 46 51 51 55 61 65 65 67 71 77
Chapter 3. Demography and Behavior of Ruffed Grouse in British Columbia R. G. Davies and A. T. Bergerud 78 3.1 3.2
3.3
3.4
3.5
3.6
Introduction Study area and methods 3.2.1 Demography 3.2.2 Behaviour indexes Demographic Findings 3.3.1 Fluctuations in numbers and frequency of color morphs 3.3.2 Survival of juveniles and adults 3.3.3 Productivity Aggressiveness and approachability of birds 3.4.1 Individual differences 3.4.2 Basis of the differences in behavior 3.4.3 Behavioral differences in space and time Discussion 3.5.1 Role of the extrinsic environment in population change 3.5.2 Behaviour as a factor in the density of grouse populations 3.5.3 Male competition and population densities Summary
Chapter 4. Winter Arboreal Feeding Behavior of Ruffed Grouse in East-Central Minnesota R. A. Huempfner and J. R. Tester
78 79 79 82 83 83 88 92 98 98 100 105 108 110 117 119 121
122
CONTENTS
4.1 4.2 4.3
4.4 4.5
xv
Introduction Study area and methods Results and discussion 4.3.1 Morning and evening budding 4.3.2 Factors affecting frequency 4.3.3 Tree canopy position 4.3.4 Intensity of canopy use 4.3.5 Roosting and budding in "feeding trees" 4.3.6 Environmental and behavioral factors 4.3.7 Variations in evening budding among winters 4.3.8 Duration of budding 4.3.9 Seasonal budding rates 4.3.10 Tree species and number used 4.3.11 Total food consumed 4.3.12 Aspen food requirements and availability Conclusions 4.4.1 Foraging strategy Summary
122 123 125 125 127 128 133 136 137 140 140 142 142 148 150 152 153 155
STEPPE GROUSE Chapter 5. Spatial Patterns, Movements, and Cover Selection by Sharp-tailed Grouse M. W. Gratson 158 5.1 5.2 5.3 5.4
5.5
5.6
Introduction Study area Methods Results 5.4.1 Spatial patterns and dispersion of the sexes 5.4.2 Home range sizes 5.4.3 Daily movement patterns 5.4.4 Brood breakup and dispersal 5.4.5 Cover selection 5.4.6 Flocking Discussion 5.5.1 Prelaying ranges and activities of females 5.5.2 Separation of cocks and hens during nesting 5.5.3 Hen desertion and dispersal of juveniles 5.5.4 Daily movements as a strategy 5.5.5 Winter flocking, snow-burrowing, and the use of wetlands Summary
158 159 163 164 164 172 174 175 178 181 184 184 186 187 188 189 191
xvi
CONTENTS
Chapter 6. Reproductive Ecology of Female Greater Prairie Chickens in Minnesota W. D. Svedarsky 193 6.1 6.2 6.3 6.4
6.5
6.6
Introduction Study area Methods Results 6.4.1 Breeding chronology 6.4.2 Nesting chronology 6.4.3 Clutch size and fertility 6.4.4 Nesting success and renesting 6.4.5 Nest placement 6.4.6 Nest characteristics and success 6.4.7 Brood mortality factors 6.4.8 Female mortality factors 6.4.9 Movements 6.4.10 Home range 6.4.11 Habitat use and preference Discussion 6.5.1 Courtship and breeding 6.5.2 Preincubation movements 6.5.3 Preincubation habitat use 6.5.4 Nesting 6.5.5 Brood movements, habitat use, and disturbance 6.5.6 Brood mortality 6.5.7 Female mortality 6.5.8 Broodless-female movements and habitat use 6.5.9 Population density and possible limiting factors Summary
Chapter 7. Mate Choice by Female Sage Grouse J. E. Hartzler and D. A. Jenni 7.1 7.2 7.3
7.4
Introduction Study area and methods Territorial establishment and mating success 7.3.1 Fidelity of males to arena and territories 7.3.2 Fidelity and return of hens 7.3.3 Nonrandom mating of cocks 7.3.4 Fidelity of cocks compared to mating success Hen selection of mating centers or cocks 7.4.1 Hen clusters and mating centers
193 194 197 200 201 202 202 203 204 206 206 209 211 214 219 225 225 226 227 228 230 232 234 235 236 238
240 240 241 243 243 246 248 248 251 251
CONTENTS
7.5
7.6
xvii
7.4.2 Strutting frequency compared to mating success 7.4.3 Agonistic behavior compared to mating success Discussion 7.5.1 The site hypothesis for mating success 7.5.2 Male behavior and mating success 7.5.3 Conclusions Summary TUNDRA GROUSE
253 255 258 259 264 267 267
Chapter 8. Behavior of White-tailed Ptarmigan during the Breeding Season R. K. Schmidt 270 8.1 8.2 8.3 8.4 8.5
8.6
8.7
8.8
8.9
Introduction Study areas Methods Early territorial behavior Territories and territorial maintenance 8.5.1 Establishment and defense 8.5.2 Size, occupancy, and distribution Pair bonds 8.6.1 Establishment 8.6.2 Maintenance 8.6.3 Nesting period The brood period 8.7.1 Dispersal from breeding areas 8.7.2 Brood ranges 8.7.3 Brood maintenance and defense Male postbreeding flocks 8.8.1 Movements 8.8.2 Flock structure 8.8.3 Late-summer association with breeding areas Summary
270 271 273 273 274 274 280 285 285 288 289 290 290 293 293 296 296 296 297 297
Chapter 9. Cyclic Population Changes and Some Related Events in Rock Ptarmigan in Iceland A. Gardarsson 300 9.1 9.2 9.3 9.4
Introduction Study areas Methods Results and Discussion 9.4.1 Breeding density 9.4.2 Adult mortality in summer
300 300 302 304 304 306
xviii
9.5
CONTENTS
9.4.3 Production 9.4.4 Age ratios in the hunting season 9.4.5 Winter losses 9.4.6 The territorial stage 9.4.7 Territorial behavior and changes in numbers 9.4.8 An outline of behavior in flocks 9.4.9 Sex segregation in winter 9.4.10 Age-group segregation 9.4.11 Food Summary
306 308 310 310 316 318
321 323 325 329
Chapter 10. Winter Survival and Breeding Strategies of Willow Ptarmigan D. H. Mossop 330 10.1 10.2 10.3 10.4 10.5
10.6
10.7
10.8
10.9
Introduction Study area Methods Numbers, breeding success, and mortality The winter environment 10.5.1 Winter weather 10.5.2 Winter food survival and physical condition of birds 10.5.3 Abundance of predators Ptarmigan survival strategies for the winter environment 10.6.1 Flock formation 10.6.2 Snow-roosting 10.6.3 Synchronized, crepuscular behavior 10.6.4 Migration 10.6.5 Sexual segregation 10.6.6 Synthesis of winter survival strategies Spring spacing behavior and its effect on the population 10.7.1 The early return 10.7.2 Spring territoriality and social segregation 10.7.3 Mortality and condition of birds in waiting flocks 10.7.4 Test of the territorial self-regulation hypothesis 10.7.5 Cost-benefit trade-off in tactics for breeding Conclusions 10.8.1 Winter environment and ptarmigan mortality 10.8.2 Social segregation of the population 10.8.3 Population regulation Summary
330 331 333 334 337 337 337 340 345 345 345 348 348 354 355 355 356 360 368
370 371 372 372 373 374 377
CONTENTS
xix
Chapter 11. Demography of an Island Population of Willow Ptarmigan in Northern Norway S. Myrberget 379 11.1 11.2 11.3
11.4
11.5
11.6
11.7
11.8
Introduction Study area Materials and methods 11.3.1 Collection of population data 11.3.2 Definition of parameters 11.3.3 Estimates of plant production 11.3.4 Adult behavior General features of population variations 11.4.1 Breeding population 11.4.2 Adult summer mortality 11.4.3 Production and autumn numbers 11.4.4 Winter losses 11.4.5 Synthesis of changes in numbers Annual variations in production 11.5.1 Egg-laying date 11.5.2 Nesting failure 11.5.3 Clutch size 11.5.4 Loss of eggs 11.5.5 Failure of eggs to hatch 11.5.6 Chick mortality 11.5.7 Intercorrelation between productivity parameters 11.5.8 Comparison of ptarmigan production and plant growth 11.5.9 Hen behavior and production 11.5.10 Synthesis of productivity Hypotheses to explain changes in production 11.6.1 Clutch-size variation 11.6.2 Predation 11.6.3 Environment of chicks 11.6.4 Intrinsic phenotypic quality 11.6.5 Intrinsic genotypic quality 11.6.6 Synthesis Hypotheses to explain changes in breeding numbers 11.7.1 Winter food 11.7.2 Predation of adults 11.7.3 Spacing behavior 11.7.4 Juvenile production Summary
379 380 381 381 382 383 383 384 384 386 387 388 388 392 393 393 393 398 401 401 403 404 404 406 408 408 410 411 412 412 413 414 414 415 415 415 416
xx
CONTENTS
Part II Theory and Synthesis
Chapter 12. A Genetic Explanation for Ten-year Cycles of Grouse R. E. Page and A. T. Bergerud 12.1 12.2 12.3 12.4 12.5 12.6
12.7 12.8
Introduction Defining the assumptions The model Parameters of the simulation Realism of the assumptions Results 12.6.1 Basic factors 12.6.2 Other factors 12.6.3 Realism of the results 12.6.4 Falsification tests Conclusions Summary
Chapter 13. Mating Systems in Grouse A. T. Bergerud 13.1 13.2
13.3
13.4 13.5 13.6 13.7 13.8 13.9
Introduction The monogamy model 13.2.1 Size of the prelaying range 13.2.2 Dispersion of males 13.2.3 Effective avian predators 13.2.4 Conspicuousness of males 13.2.5 Mortality of males Female choice in the model 13.3.1 Differences among species 13.3.2 Female choice of conspicuous plumage Food shortage and monogamy The female view of the pair bond The male view of the pair bond Dispersed polygyny Clumped polygyny Summary
423 423 425 427 428 430 432 432 436 436 437 437 437
439 439 440 440 444 445 447 451 451 451 454 457 458 460 461 465 470
Chapter 14. Survival and Breeding Strategies of Grouse A. T. Bergerud and M. W. Gratson
473
14.1 14.2
474 474
Introduction Male advertising strategies
CONTENTS
xxi
14.2.1 Advertising displays 14.2.2 Fidelity as an advertising tactic 14.2.3 Advertising in the fall 14.2.4 Tactics of daily advertising 14.2.5 Advertising near females 14.2.6 Selection of safe advertising sites 14.3 Advertising tactics of yearling males 14.3.1 Early prospecting by juvenile males 14.3.2 Advertising in spring 14.3.3 Tactics for securing an advertising site 14.4 Female nesting strategies 14.4.1 Assessing predator abundance and nesting sites 14.4.2 Timing and inconspicuousness 14.4.3 Spacing and other behaviors to remain inconspicuous 14.4.4 Avoiding nest ambush 14.4.5 Philopatry of nesting females 14.4.6 Nest abandonment and renesting 14.5 The strategy of selecting a male for breeding 14.6 The strategy of improving chick survival 14.6.1 The trade-off: Optimal foraging versus antipredator tactics 14.6.2 Inconspicuousness of females and broods 14.6.3 Reducing interactions with predators 14.6.4 Defense of young against predation 14.7 Brood disbandment and migration 14.7.1 Brood disbandment 14.7.2 Fall migration 14.7.3 Movements between breeding seasons 14.8 Strategies for winter survival 14.8.1 Winter feeding tactics 14.8.2 Winter flocking 14.8.3 Winter roosting tactics 14.9 Polymorphic spacing strategy 14.10 Summary Chapter 15. Population Ecology of North American Grouse A. T. Bergerud 15.1 15.2 15.3
Introduction Percentage of hens nesting Clutch size in grouse 15.3.1 Maternal condition and clutch size 15.3.2 Clutch size determined by chick survival
474 476 477 479 482 489 494 495 496 499 503 506 511 517 524 526 527 533 537 537 542 545 546 548 548 550 555 555 556 561
564 567 575
578 578 579 581 582 584
xxii
CONTENTS
15.3.3 Clutch size determined by life span and predation risk of eggs 15.4 Nesting success 15.4.1 Frequency of renesting 15.4.2 Annual variations in nesting success 15.4.3 Nesting success and densities 15.5 Mortality of chicks 15.5.1 Mortality rates 15.5.2 Mortality factors 15.5.3 Density-dependence of mortality 15.6 Mortality of juveniles and adults 15.6.1 Mortality factors 15.6.2 Mortality rates 15.6.3 Bimodalism of mortality rates 15.6.4 Mortality rates and reproductive risk 15.6.5 Differential mortality of males and females 15.6.6 Mortality and density 15.7 Theories of population change 15.7.1 Threshold-of-security 15.7.2 Winter bottleneck 15.7.3 Switch-over of predators 15.7.4 Territorial self-regulation 15.7.5 Breeding success drives numbers 15.8 The 10-year cycle in grouse 15.8.1 Frequency- and density-dependent selection 15.8.2 The single-locus, density-dependent model 15.8.3 The female-choice, density-dependent model 15.8.4 Relevance of the density-dependence model to characteristics of cyclic populations 15.9 Limitation of breeding numbers 15.10 Summary Chapter 16. Increasing the Numbers of Grouse A. T. Bergerud 16.1 16.2
16.3
Introduction Mechanisms of density 16.2.1 Nesting success and density 16.2.2 The concept of a stabilizing density Control of hunting 16.3.1 Compensation principle
586 592 595 599 605 609 609 610 614 615 615 616 617 622 624 629 634 636 636 637 640 645 656 657 665 666 672 679 684
686 686 686 686 691 696 696
CONTENTS
16.4
16.5
16.6
16.7
16.8 16.9
xxiii
16.3.2 Is hunting mortality additive to overwinter natural mortality? 16.3.3 Compensatory aspects of hunting Control of food 16.4.1 Ruffed grouse and aspen buds 16.4.2 Prairie chickens and grain 16.4.3 Food is not limiting Control of cover 16.5.1 Cover during the breeding season 16.5.2 Cover during winter Predator control 16.6.1 Predator-control arguments 16.6.2 The return of prairie chickens in Wisconsin Control of space 16.7.1 Size and uniformity of space 16.7.2 Females space away from males 16.7.3 The big new space 16.7.4 Space and grouse introductions Successful management Summary
References Index
697 702 703 703 705 708 711 711 718 718 719 722 726 727 727 728 729 730 731 735 781
This page intentionally left blank
Part I Population Studies
This page intentionally left blank
1 Relation Between Aggressive Behavior and Population Dynamics of Blue Grouse D. H. Mossop
1.1 Introduction Blue grouse (Dendragapus obscurus} on Vancouver Island, British Columbia, have provided the material for one of North America's longest studies of grouse population ecology. In 1966 the investigation initiated by Fowle (1944, 1960) and carried on largely by Bendell (1954, 1955a,b), Zwickel (1965, 1967), and associates reached a crossroad that became graphic only in retrospect. These studies arrived at the general conclusion that breeding numbers were remarkably stable when compared with the large variations in natality and losses of young. This phenomenon of stable numbers was observed in several populations, although between areas large differences were observed (Bendell & Elliot 1967, Zwickel 1967). Once birds entered the breeding component of the population, they apparently disappeared at about 30% annually, a value that did not vary significantly from year to year. It appeared that production of chicks was always greater than this adult loss. Only a small, constant number of grouse returned in subsequent springs, in spite of a variable number alive until winter. Most of the work in the Vancouver Island study represented attempts to explain this apparent adjustment of the number of young that disappear over winter. Various hypotheses linked to environmental factors that may affect survival were tested, including disease, weather, and vegetative cover. None explained the numbers of breeding grouse observed (Bendell 1955b, Elliott 1965, Zwickel 1965). At the outset of the research for this paper, attempts to examine the breeding stocks themselves for factors to explain numbers were in their infancy. The works of Southwick (1955), Krebs (1966), Sadlier (1965), Chitty and Phipps (1966), 3
4
D. H. MOSSOP
and Healy (1967), all addressing small-rodent populations, were notable in showing that variations in social behavior could apparently result in numerical change. Similarly, Hornocker (1969) measured social behavior in the cougar (Felis concolor), investigating its relation to density. Carrick (1963) and Carrick and Murray (1964) implicated social interaction to explain magpie (Gymnorphina tibicd) and gull numbers; Tompa (1964) and Smith (1967) offered similar hypothesis to explain passerine population ecology. Working with grouse, Jenkins et al. (1963, 1967) and Watson (1964, 1967a,b) had suggested that the breeding density of red grouse (Lagopus lagopus scoticus) was directly related to competition by males for territories in the fall. Bergerud (1970a) and Mercer (1967) eventually showed that levels of aggression in willow ptarmigan (L. I. lagopus) were highest in a declining population, and suggested this resulted from selection for a more aggressive genotype. Bendell and Elliott (1967) studied territorial spacing of blue grouse. They removed breeding males in the spring and observed replacement, concluding that considerable interaction occurred and that at least some yearlings were apparently prevented from breeding. However, they felt that the largest and therefore critical elimination of recruits had already occurred before their field work commenced in the spring. Lance (1967), using radio telemetry, questioned whether dispersion on the breeding range could adjust the numbers of recruits. He found no evidence of this. Both studies concluded that if social interaction did regulate numbers of grouse, it must have been occurring outside the annual study period. Clearly a new tack was needed. Attempts to identify a process by watching relatively small study areas within large populations were not working. If the general idea, hypothesized by Wynne-Edwards (1962), that populations are regulated by social behavior was to be tested, processes had to be queried at the population level and had to apply throughout the annual cycle. Chitty (1960, 1964, 1967) was attempting to formulate a testable hypothesis that innate aggressiveness affected recruitment. He felt that populations differing in numerical trend were composed of individuals that also differed in aggressive tendencies. A null hypothesis thus had two parts. To disprove Chitty's (1967) theory: (1) no essential differences in behavior should be demonstrable between different areas regardless of the population demography of the areas, and (2) if differences did exist, increasing populations should not be composed of individuals showing less aggression than stable or declining ones. It was necessary to search out populations that exhibited the greatest variations in density and trend. There could be no significant mixing of birds between areas; comparisons among populations were emphasized and less importance was placed on studying plots in the center of large areas of homogenous habitat. One of the old, blue-grouse study areas on Vancouver Island was in slow decline. Another unstudied area was chosen from hunter returns, which showed a recent dramatic increase. A third, well-known, stable area was added.
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
5
1.2 Study areas and methodology Middle Quinsam Lake area (MQL) lies approximately 34 km northwest of the Comox Burn area (CB), and the Copper Canyon study area (CC) is about 136 km southeast of CB (Fig. 1.1). Table 1.1 presents a brief summary of the vegetation on these areas. The physical geography and vegetation of MQL and CB are further described by Elliott (1965) and Zwickel (1965); Copper Canyon is in a valley, intersected by drainage gullies, and is from 450 to 750 m above sea level.
1.2.1 Population demography Field work was carried out during the summers of 1967 and 1968. It was designed to measure breeding population density, age structure, survival of breeders, and annual production of chicks. Total counts of breeding males from repeated searches of known areas were made by lone observers, each accompanied by a trained pointing dog. Each grouse encountered was classified by age and sex, and its location was marked on a base map. Color banding facilitated the identification
Fig. 1.1. Blue grouse study areas on Vancouver Island, British Columbia.
6
D. H. MOSSOP
Table 1.1. Summary of vegetation on the three study areas on Vancouver Island, British Columbia, 1967-68 Study area
Size (ha)
Orig. vegetation
Hist, vegetation
Condition in 1966-67
MQL
337
Douglas fir (Pseudotsuga menziesii) Western hemlock (Tsuga heterophylla)
Clear-cut, burned about 1945, planted Douglas fir 1958
Open to very open, planted Douglas fir
CB
607
Douglas fir Western hemlock
Clear-cut, burned about 1960, planted Douglas fir 1961
Open to very open, planted Douglas fir
CC
664
Douglas fir Western hemlock
Clear-cut, burned 1959-68, replanted Douglas fir
Open to very open, planted Douglas fir
of individuals. Counts were aided by using recorded female calls, which induced singing in males and thus made them easier to locate. Listening stations were established at 100-m intervals along sections of logging roads to supplement total search and to allow a check of the density of territorial males over a larger area. Effective width of these transects was 100 m either side of the road. Stations were visited during May and June between 0530-0900 and 2000-2200 hours, when birds displayed territorial song. Overall, each transect averaged ten listening stations and was covered at least three times per year. After reaching each station and pausing for 10 minutes, a series of female blue grouse calls was broadcast. This method yielded an effective census of males along the roads and was especially useful in the high-density populatio (see Frandsen 1980 for a test of the methodology). Age structure and survival were determined from a pooled sample obtained by capturing unbanded birds in the study areas, by shooting birds in adjacent areas, and by observing the proportion of banded birds from previous years. Grouse were classified as adults > 2 years or as yearlings following Boag's (1965) procedures. The number of chicks produced by each population was estimated annually by determining the average clutch size, by estimating the number of successful and unsuccessful hens, and by recording the disappearance of chicks from broods over summer. Ages of captured chicks were determined by the method of Zwickel and Lance (1966). Because chicks were not captured from the majority of broods, a method was devised for determining age at a distance. This was based on feather and body development and allowed placement of any chick into
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
7
one of six, 10-day age categories when compared with a series of sketches carried by field workers. The B.C. Fish and Wildlife Branch cooperated by recording the number of birds shot per hunter per day, and the sex and age of each grouse harvested and reported at MQL and CC.
1.2.2 Behavior Blue grouse are secretive and cannot easily be observed in undisturbed situations. However, the song (hooting) of undisturbed males was recorded whenever heard. Hooting was classified by the number of syllables it contained and by how many songs were sung per minute over a 5-minute period. Most data, however, were obtained by presenting grouse with artifical situations such as encounters with field workers, and recording the birds' reactions. The response of a bird, whose location had been indicated by a pointing dog, to a slow stalk by the observer was measured by recording when and at what distance it flushed, displayed, and sang, and showed distraction or brood defense and other behaviors, and its reactions to other birds present. The response of hens to artificial chick "screaming" was also rated, as was the response of males to artificial song. The test arena was a situation to which only males responded. In a small, flat, clear area in the center of a male's territory, plate-glass mirrors and a stuffed, dummy hen grouse were placed. The mirrors were arranged in a triangle with their faces outward. The hen, positioned in a squatting pose, was placed opposite one of the faces of the mirrors and about 0.5 m from it. The speaker of a portable amplifier was placed in the center of the mirrors and the lead was run to a hide approximately 15 m away. Amplified hen calls were played from the speaker, concealed by the mirrors. When the hen call was broadcast, the cock was enticed to approach and court the dummy, thereby forcing an interaction with his mirror image. His response was tape-recorded as it occurred, and later quantified by measuring the time until specific behaviors were observed and by computing their frequency. Forty cocks were tested at the declining area (MQL), ten at the stable area (CB), and 40 at the dense, increasing area (CC). Differences in behaviors among populations were accepted when P < 0.05 using chi-square goodness-of-fit procedures. Statistical significance, unless otherwise stated, is at least at the 95% confidence level. 1.3 Demographic findings Densities of territorial male blue grouse on the areas were measurably different during the study (Fig. 1.2). One population of about 87 males/km2 was about six times the density of the lowest population, which had about 13 males/km2.
8
D. H. MOSSOP
Fig. 1.2. Breeding densities of blue grouse from three study populations (CC, MQL, CB) on Vancouver Island, 1962-68.
The numerical trends were obtained by incorporating the results of previous study (Zwickel 1965, 1967) at the two low-density areas and of a roadside sample from hunters at the high-density area. This showed MQL in a slow decline (about 10% per year), which appeared to have begun in 1964. Comox Burn was stable. The high population, CC, had apparently been increasing since 1964 and then stabilized during the 2 years of this study. This increase was about 12% per year. The assumption was that, given the adult mortality rate typical of blue grouse (30-31%), sufficient numbers of birds of the "increasing stock" (70% in 1967) were present throughout the study to make the test of the null hypothesis valid. Differences in age structure and mortality among populations were less convincing than differences in numerical trend (Table 1.2). A surplus of chicks over the number required for recruitment the next year was produced annually in blue grouse populations on Vancouver Island (Bendell 1955b, Zwickel 1965). It was a new idea that there may be innate differences between breeding stocks, which left open the possibility that parameters of productivity would vary with the type of breeding stock. Three such parameters are readily measurable for blue grouse:
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
9
Table 1.2. Age structure of females and mortality of banded male blue grouse on Vancouver Island, 1967-68 Study area CC CB MQL a b
Population high, increasing density low, stable density low, declining density
Age structure (% yearling)
n
Adult annual mortality (%)
n
37 41 27b
166 133 176
30-353 37 43
48 49
Estimated from unhanded birds. Significantly lower percentage than the others, P < 0.05.
nest loss, proportion of hens that produce no brood at all, and survival of chicks through the summer (Table 1.3). In general, the high population retained numbers in broods better during summer than did the declining population, and the stable area showed an intermediate loss (Fig. 1.3). The pattern of chick disappearance on these areas was basically the same—an initial high mortality until 10 days of age was followed by a slower, more constant loss. The difference in chick survival between the increasing CC and the declining MQL populations at 50-70 days was significant in both years. Moreover, within populations there was no significant difference in brood size between years.
1.4 Behavioral findings 1.4.1 Spontaneous behavior Hooting is the territorial song of the male blue grouse and the most obvious evidence of territorial behavior. There were no significant differences between years in the frequency of hooting recorded within areas, and in June there was no signiTable 1.3. Parameters of productivity of blue grouse on Vancouver Island, 1967-68 Study area
Population
Nest failure (%)
n
Broodless hens (%) a
n
Chick loss (%)
n
CC
high, increasing density
14
7
23
38/272
16b
239
CB
low, stable density
57
7
35
16/79
38
63
MQL
low, declining density
43
7
47
1 12/344
53
232
a
b
Corrected on the basis of resighting banded lone and brood hens to overcome differential observability (Zwickel 1965). Significantly higher or lower percentage than the others, P < 0.05.
10
D. H. MOSSOP
Fig. 1.3. Mean brood sizes at 10-day intervals, comparing chick loss among study populations, 1967-68.
ficant difference among populations (Fig. 1.4). However, in July the high, increasing stock showed significantly more hooting than either of the other types, and in August of both years 24 % of the males were hooting. No hooting was heard at the low, stable and low, declining populations. This tendency for males in the dense population to persist in hooting longer into the season agrees in part with the findings of Bendell and Elliott (1967). The rate of hooting and the type of song showed no differences among populations.
1.4.2 Behavior in response to observer In general, three measures of behavior were recorded during census: (1) approachability of grouse; (2) frequency of intraspecific behavior patterns apparently used toward us; and (3) rating of predator escape behavior used toward us. The number of birds that were seen before flushing, the number that afforded a "full view," and the number that could be captured were compared among populations (Fig. 1.5). There was no demonstrable difference between the two low areas. There was, however, a significant difference between the high, increasing stock and the other two in each case; the birds at CC were much less observable
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
11
Fig. 1.4. Percentage of males that engaged in spontaneous song (hooting) in the study populations, as the season progressed, 1967-68.
12
D. H. MOSSOP
Fig. 1.5. Percentage of birds that could be observed on the ground before flushing, and that could be "noosed" (both indexes of reluctance to flush from humans), 1967-68.
and approachable than those at the other areas. Roughly 60% of all birds encountered at the two low-density areas could be observed on the ground, and of these 30% could be captured. By contrast only 15% of the birds encountered in the high-density area could be observed, and only about 20% of those observed could be captured. Assuming similar field techniques I became impressed with the differing reluctance that birds in the three populations showed in flushing from us.
1.4.3 Intraspecific behavior in census situation Auditory signals were recorded while we stalked grouse, and the most frequent of these was the territorial song. Hooting often began in apparent response to human intruders. We hooted at 98 males and observed their responses (Fig. 1.6). In all, 25-30% made no obvious response. Responses of cocks at the two low-
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
13
Fig. 1.6. Comparison among study populations of the percentage of territorial males showing a particular response to human intrusion and artificial hooting, 1967-68.
14
D. H. MOSSOP
density areas were very similar. However, at the high, increasing area, significantly more birds either moved away or stopped singing. In addition, whereas at the two low-density areas 30-35 % of males either increased hooting or gave the "feather spread display," at the high, increasing area this never occurred. Another social signal of the blue grouse is a loud wing flutter, which may be made upon alighting. Bendell and Elliott (1967) described it as the "flutter flight" and suggested it was a territorial signal with significance comparable to song. Blackford (1963) described a similar sound in the dusky grouse (Dendragapus o. obscurus} and concluded that it was a form of "territorial drumming." I compared among populations the number of birds that gave the sound after being disturbed by us (Fig. 1.7). Both territorial males and brood females made significantly more flutters at the low, declining area than did males at the other two areas. Among silent males and lone females, less than 6% gave this sound, and no difference among populations could be demonstrated.
Fig. 1.7. Percentage of birds that displayed "flutter-flight" when disturbed by humans, 1967-68.
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
15
The threatening or "attack intention" calls of both sexes are usually observed only in highly aggressive birds immediately before attacking. These include the "growl" or "threat" call of the male (Stirling 1965, Stirling & Bendell 1970), the female cluck (Stirling 1965), and the "qua-qua" cry of the female, also described by Stirling. These sounds were heard only rarely during the study. At the low, declining area they were heard an average of 5.1 times/100 observations; at the low, stable area 4.3 times/100 observations; and at the high, increasing area only 0.4 times/100 observations. Blue grouse have many visual signals that they use in intraspecific encounters. As has been shown, grouse differed in the ease with which we could observe them. Some birds used behavior that seemed to have the function of reducing visibility. This behavior, called "crouch-and-run" (Mossop 1971), was also observed in yearling males when near territorial cocks. When stalked by humans, birds varied in their use of "crouch-and-run" (Fig. 1.8). There was no difference in the frequency of this behavior between grouse at the two low areas. In all cases, however, comparing sex and age categories separately, birds at the high, increasing area showed "crouch-and-run" significantly more often than at the other two areas. The "feather spread" or "sexual display" (Stirling 1965) is the most obvious visual signal of the blue grouse. Stirling described it and concluded that it functions principally in sex advertisement. The display seems to take over where song and flutter flights end in advertising a male's ownership of territory. There is ample evidence that males using this display toward us were acting more aggressively than males that did not (Lumsden 1965, Hjorth 1967). At the two low areas 35-37 % of all adult males gave the feather spread, whereas at the high, increasing area significantly fewer (11%) did so (Fig. 1.8). Gestures signifying direct aggression and attack intention were also observed. These included "neck stretching" and "head dipping"; both were highly aggressive patterns seen immediately before an attack was made on a mirror image or another grouse. Treating the two together, they were observed every 12-16 encounters at the two low areas, whereas at the high-density area they were observed only once every 30 encounters (see also Mossop 1971).
1.4.4 Predator avoidance behavior Flushing is the most obvious predator-avoidance behavior. The flushing distance, or distance between the observer and grouse when it flew, was the easiest feature of this behavior to quantify (Fig. 1.9). Again there was no significant difference between the two low areas. Flushing distances were longer at the high area in each sex and age category. A related phenomenon was the time it took for a grouse to flush after first being disturbed, and the distance it walked or ran before flushing. Once again in the increasing population, significantly fewer grouse took more than 2 minutes to flush and fewer moved 20 m or more than those at the other two areas (see also Mossop 1971).
16
D. H. MOSSOP
Fig. 1.8. Percentage of birds that showed "crouch-and-run" behavior to human intruders, and that showed advertising "feather-spread" displays, 1967-68.
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
17
Fig. 1.9. Flushing distances (m) of birds, compared among study populations (see Fig. 1.5 for symbol definitions), 1967-68. Lines show 95% confidence limits in this and latter figures.
Brood defense behavior, shown by brood hens, probably functions solely as a response to predators. Bendell and Elliott (1967) described it from the field. Two basic elements of the behavior were recognized: an attempt by the female to lead us away, and an intimidating rush toward us. Again the two low areas resembled each other; at the high area there was significantly less leading and rushing (Fig. 1.10). The response of hens during leading and rushing was given a rating on a predetermined scale. Three categories were devised for each display according to the nearness of approach, degree of feather spread, amount of vocalization, and general vigor of the response. The mean index for each population is shown (Fig. 1.10). This index of vigor did not vary between the two low types but was significantly lower at the high area. Hens in the increasing population apparently defended their chicks less vigorously than hens at either of the low, stable or low, declining populations. The chick behavior we observed was also principally a predator escape response. Critical evaluation of chick behavior was possible only at CC and MQL. The cohesion of undisturbed broods was measured by noting the distance of individual chicks from the hen when the broods were first discovered (Fig. 1.11).
18
D. H. MOSSOP
Fig. 1.10. Comparison of the percentage of brood females that showed "leading" and "rushing-at" behaviors, and vigor of brood-defense scores among study populations, 1967-68.
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
19
Chicks from the high-density population were consistently found farther from their hen than chicks at the low, declining area. The behavior of chicks at the moment of disturbance was classified into three types of response: "scatter-and-hide," "crouch-and-flush," and "flush immediately." Chicks were divided into three age groups, 1-2, 3-4, and 5-6 weeks and treated separately. At every age, differences occurred in the behavior of chicks
Fig. 1.11. Distance (m) between hens and their chicks from two study populations, 1967-68.
20
D. H. MOSSOP
from the two populations (Fig. 1.12). In general, chicks at the low, declining area showed a greater tendency to use the tactics of hiding or crouching and then flushing only when the "predator" (researcher) came very close. Chicks in the increasing population generally flushed immediately, vacating the area completely.
Fig. 1.12. Percentage of chicks in two study populations that showed "immediate flight," "hide," and "crouch" behaviors when disturbed by observers, 1967-68. Age 1 = 1-2 weeks; age 2 = 3-4 weeks; age 3 = 5-6 weeks.
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
21
1.4.5 Behavior in the test arena Three phases were recognized in the response of males to the test arena: exploratory, courting, and interaction. The cock exhibited exploratory behavior while seeking out the source of the hen calls being broadcast. Courting was behavior displayed when the dummy female was seen and before there was any reaction to the mirrors. Interaction was the reaction to the cock's mirror image. The exploratory phase was quantified by determining the speed of advance of the male toward the arena, and by noting the alterations in his song. Virtually all males, regardless of population, increased the volume of hooting when the female call was played. Also, the frequency of complete songs increased. There was high variation in the rate of advance; cocks at CC advanced slowest, those at CB most rapidly, and those at MQL with intermediate speed (Fig. 1.13). Courting was analyzed by the number of elements of behavior observed per minute. Ten patterns and poses were recognized (Mossop 1971). When all courting acts were totaled, the only two areas that differed were the increasing area and the declining area. These differences were largely a result of "aberrant" behaviors such as the "hard peck" delivered to the hen, a behavior seen only in the declining area. When these patterns were removed from analysis, cocks in all areas seemed to court with equal vigor. Interaction commenced when the male reacted to his mirror image. His response was quantified by measuring the time that elapsed from the moment of his arrival until one of the three most common aggressive behavior patterns was observed. The frequency of aggressive acts from the moment of first interaction was then calculated. No differences emerged among the areas in the time taken for a response to the mirrors. In addition, no differences were evident in the types of behavior observed in front of the mirror. However, the frequency of aggressive acts was different among populations (Fig. 1.14). The mean for the low, declining area (MQL) was significantly higher than that for the high area (CC), but there was no difference between MQL and CB or between CB and CC. Thus cocks at the low, declining area, compared with males at the high-density area, advanced more quickly toward the female call, courted the dummy hen with more acts that seemed aberrant, and made a more vigorous attack on the mirror. Cocks at the low, stable area reacted with intermediate vigor.
1.5 Discussion My null hypothesis predicted that no differences in behavior of birds should occur among the study populations. Clearly that part of the null hypothesis appeared false. Twenty measures of behavior were compared, and in 16, significant differences were recognized between areas. This was true for certain aspects of spon-
22
D. H. MOSSOP
Fig. 1.13. "Rate-of-advance" of males during arena tests, 1967-68.
taneous behavior, for the reactions of grouse toward people, and for the responses of males toward their mirror images (Fig. 1.15). The same differences were recognized in both years of the study by many observers. This was the first empirical inkling that blue grouse, observed at the population level, could not be considered standard entities. The immediate result was to cast doubt on several models, which as part of their assumptions, imply that individuals of the same species are essentially alike. Regulation of numbers solely
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
23
Fig. 1.14. Comparison among study populations of the number of aggressive acts performed by males in the test arena, 1967-68.
through mortality by environmental variables is such a hypothesis. Models of regulation through competition for some commodity (Nicholson 1933, Lack 1954), or through mortality by such environmental hazards as weather and disease, must at a minimum be modified to accept the possibility that significant variations may occur between different breeding stocks. These may seriously alter the way extrinsic factors interact with populations. The second contribution I had hoped to make with this work was to test for differences in aggressive behavior among breeding stocks with different numerical trends. Many models have used the idea that variants in natural populations could affect population density (Pimentel 1963, Wynne-Edwards 1966, Chitty
24
D. H. MOSSOP
Fig. 1.15. Summary and composite of behavioral responses shown by blue grouse, compared among study populations, 1967-68.
1967). The second part of my null hypothesis served to examine a prediction of one such model. Chitty (1967) suggested that population changes were a direct result of selection for different genetic types (genotypes) in a population. He predicted that "animals present in stationary or declining populations have been
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
25
selected for their superior ability to survive the effects of mutual interference." This means that "supposedly more aggressive individuals in the stationary or declining populations" are present. Two criteria must be met to test these ideas: there must be measurable numerical change, and critical measurement of aggressive behavior must be possible.
1.5.1 Population trends My confidence in the stated numerical trends of the three populations is based on long-term study prior to, and including, this research. At the increasing population (CC), the trend prior to the study was determined from relative success of hunters in fall, not from counts of breeding birds, as was the case for the other two areas. In the latter, we were able to confirm the trends as declining and stable through direct census. The same method did not confirm an increase over the 2 years of study at CC. However, it is unlikely the continued large increase in hunter success from 1964 through 1967 could be the result of anything but a population increase. In fact, the density of blue grouse breeding on the area at the beginning of the study was about as high as has been recorded. It seems most likely that the population underwent an increase which ended in 1968. Because breeders were dying on the area at about 30% annually, I feel confident that for the purpose of the test, most of the birds on the area during our field work would be of the "increasing stock," if such a type existed.
1.5.2 Aggressive behavior Interpretations of animal behavior are more difficult to defend. Behaviors measured included those known definitely to occur in interactions between grouse and for which the meanings are reasonably well understood, behaviors that when compared with those of other animals intuitively seem reasonable to interpret in an agonistic sense, and many behaviors that are probably not agonistic patterns at all. In every case, regardless of the situation, birds of the increasing population used patterns of direct-attack intention or high aggression less frequently than birds of the stable or declining areas. Conversely, known submissive display was most often used by individuals in the increasing population. Behavior in the second category refers in large part to those gestures that communicate higher social status and not only intentions of attack. In this sense, I accept the arguments of Hjorth (1967) and others, linking territorial and reproductive displays to aggression. These threatening patterns were virtually always used less by grouse in the increasing population than by those in the stable or declining ones. Spontaneous song apparently did not align with other behaviors of territorial ownership. However, unlike other behaviors, it not only was elicited by an intruder to which
26
D. H. MOSSOP
grouse were forced to react, but occurred without intrusion and was measurable, Bendell and Elliott (1967) stated (as I found) that grouse hoot more often in dense populations and explained it as simply a result of the birds stimulating each other more in dense populations. The third category —nonagonistic behavior —is important because the general model I tested proposes a single, unifying factor (aggressive behavior) to explain numerical change. Examples are the various chick behaviors, the measures of how approachable grouse were, the responses of male to the female calls, and the responses of males to the dummy female. Some of these measures showed striking differences among areas. Yet, interpretation as agonistic behavior seems unreasonable. This suggests that although birds showed great differences in agonistic behavior, which aligned against the null hypothesis, they also showed differences in other behavior that may not be related to aggression at all. It could be suggested that any interpretations of the behavioral data could be questioned on the basis that the data were simply artifact. Most observations were of situations quite removed from actual grouse-grouse encounters. It is difficult to tell exactly how behavior grouse use toward humans is related to behavior they show toward other grouse. Even the controlled situation, the test arena, suffers in the same way. This argument can be made to most of those (Lismann 1932, Noble & Vogt 1935, Tinbergen 1951, 1959, Marler 1955, McBride 1958, and others) who have presented animals with test situations. However, Theberge (1971) reviewed the use of mirrors as an ethological tool and concluded that it was a reliable method for studying animals in intraspecific encounters. Obviously, test situations will never be exactly like natural ones; however, there is ample evidence that they can usually be used to predict what will happen in the natural situation. One feature of our results that made artifact unlikely was the number and kind of observations that led to the same conclusion. Sixteen measures of behavior, albeit not independent of one another, contrasted the increasing, high-density population with the others. Significantly, the six or seven that could be interpreted with some certainty as agonistic behavior likewise contrasted the increasing stock with the others. The results are open to empirical testing by replication. Bergerud and Hemus (1975) and Bergerud (Chap. 2) made a major contribution in this regard. Cooper (1977) used captive grouse from different areas to the same end. All demonstrated persistent behavioral differences in spacing and aggressive behavior between breeding stocks. Frandsen (1980) studied 17 breeding areas with significant differences in density. He was able to show significant differences in the behavior of birds correlated with density. Although these subsequent studies find some areas of disagreement with early investigation, the thread of qualitative differences between various breeding stocks has remained intact, and these differences are persistently linked in large part to aggressive behavior.
AGGRESSIVE BEHAVIOR OF BLUE GROUSE
27
1.5.3 Aggression and population dynamics The findings presented here support the idea that social interaction may be potentially more severe in some populations, notably in those that are declining or stable in comparison with those that are increasing. They cannot, however, be used to suggest that the differences in behavior were causing densities to change. Similarly, a number of authors have reported changes in behavior that relate to changes in density (Kluyver & Tinbergen 1953, King 1955, Tompa 1964, Sadlier 1965, Watson 1965). These authors suggest that the intensity of elimination of potential breeders was determined by the territorial behavior of other individuals, but their data cannot be used to support this. Other researchers examined behavior and social interaction in general, concluding a correlation with population density. Errington (1951) showed that fighting was greater in years of declining muskrat (Ondatra zibethicd) numbers. Mercer (1967) demonstrated more aggression as willow ptarmigan numbers declined. Watson (1964) showed that as red grouse became sparse they defended larger territories and were more aggressive. Theberge and Bendell (1980) concluded that chicks taken from a declining rock ptarmigan (L. mutus) population were more aggressive when tested in the lab than chicks from an increasing one. If we do assume that behavior and population dynamics are related in a causeand-effect relationship, Chitty's general theory would contend that less aggressive grouse are excluded from the population. In populations where habitat is freely available, the "contest" is mild and most birds are able to survive. This contest may have a two-part effect: it may eject some potential recruits, and it may act as a selective force favoring aggressiveness. Presumably, in an increasing population, as the habitat availability decreases, the elimination and therefore the selective force becomes more severe until the population stabilizes. Several problems arise when trying to interpret the blue grouse data in this manner. First, if selection affects agonistic behavior, it apparently affects other types of behavior as well. It is possible that these other types are also selected for in the absence of selection for aggression (i.e., "fitted to withstand all other hazards of their environment" [Chitty 1967: 64]). But this needs testing. Alternatively, these behaviors may be the important ones to measure. The brood defense patterns showed striking differences between populations, and it is not difficult to speculate how such behavior could be directly linked to population performance. Jenkins et al. (1967) showed a variation in brood defense behavior that seemed to be related to changes in population size in red grouse. In the present study, brood defense was far more conspicuous in the declining and stable populations. Production of young was the one summer population parameter that did prove different between areas and thus correlated with measures of aggressive behavior between breeding stocks. The difference centered on survival from the nest to the time of migration, although in 1 year it also included a difference in the number
28
D. H. MOSSOP
of unsuccessful hens. Populations of red grouse (Watson 1965), willow ptarmigan (Mercer 1967, Bergerud 1970a), and rock ptarmigan (Theberge 1971) have also shown varying productivity that correlated with numerical change.
1.6 Summary This study tested the null hypothesis that (1) no differences in blue grouse social and aggressive behavior should be demonstrable between different populations, regardless of their demography, and (2) if differences exist, increasing populations should not be composed of individuals displaying less aggressive behaviors and intensities than stable or declining ones. Three natural populations on Vancouver Island, British Columbia, provided demographic and behavioral data for this experiment, a test of two consequences of the Chitty hypothesis (1967). Males in the increasing population (35 territorial males/40.5 ha, 12% annual increase) sang longer into the breeding season, hooted less frequently in response to humans, and reacted less aggressively to artificial song than did those in the declining (6 males/40.5 ha, 10% annual decline) and stable (5 males/40.5 ha) populations. Grouse in the increasing population were less frequently observed on the ground, were harder to capture, had longer "flush distances," and flushed more quickly than did the other populations. "Flutter-flights" and "feather-spread" displays, which probably serve as threats, were more frequent in the declining and stable populations. "Crouch-and-run," a nonaggressive pattern, was more frequent in the increasing population. Hens more vigorously defended their broods in the declining population than in the increasing one. Chicks were farther from the hen in the increasing population, flew more readily, and flew farther when flushed. Artificial hen calls caused territorial males in all populations to alter their songs similarly, but males in the declining and stable populations advanced more quickly toward the sound. Cocks in all populations courted a dummy female similarly. However, when reacting to their mirror image, males in the increasing population showed fewer aggressive acts than did those in the other two populations. Overall, 16 of 20 measures of behavior were significantly different between or among populations, and for six to seven indexes of aggressive behavior, birds in the increasing population consistently showed less intensity and/or frequency. Although no causal relationship was demonstrated between behavior and population demography, I concluded that the test hypothesis should be rejected. Hostile interaction between individual grouse ("mutual interference" as per Chitty 1967) has the potential of being less severe and less frequent in increasing populations. This may have in fact initiated their increase.
2
Demography and Behavior of Insular Blue Grouse Populations A. T. Bergerud
2.1 Introduction The Chitty hypothesis (Chitty 1967) of population regulation predicts that the phenotypic and genotypic qualities of animals in increasing and decreasing populations differ in fitness components. Individuals in declining populations have been selected for their competitive abilities during peak numbers; individuals from increasing populations, although inferior in competitive abilities, are genetically better suited to colonize vacant habitats and are superior in reproduction as a result of selection pressures at low densities. In Chapter 1, Mossop partly tested the Chitty hypothesis by comparing the demography and behavior of birds in an increasing population of blue grouse (Dendragapus obscurus} at Copper Canyon (CC) with birds in a decreasing population at Middle Quinsam (MQ) and with those in a stable population at Comox Burn (CB). His results suggested the predicted differences, but the differences in habitats or densities among CC, MQ, and CB could confound the distinction between phenotypic and genotypic explanations. To clarify the roles played by the environment and heredity in behavior and demography of blue grouse, I transplanted birds from these three stocks (CC, MQ, CB) to the Gulf Islands of British Columbia, where there were no native grouse (Fig. 2.1). On three islands, Sidney, Portland, and Stuart, I released birds of single stocks to prevent outbreeding and to maintain the genetic quality of the founders. Moresby Island received birds of mixed stock from CC, MQ, and CB. I then compared the behavior and demography of birds in the single-stock popula-
29
30
A. T. BERGERUD
Fig. 2.1. Capture sites (MQ, CC, CB) of blue grouse on Vancouver Island and release sites on the Gulf Islands. Stuart Island-MQ stock; Sidney Island-CB stock; Portland Island-CC stock; Moresby Island-CC, CB, and MQ stock.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
31
tions with those of birds in the mixed-stock population to evaluate competition among stocks as an influence on parameters of demography and behavior. In addition, because the founders and their progeny were unable to disperse, I compared their behavior with those of birds at the original capture sites on Vancouver Island, which were of course free to disperse. This was the first blue grouse study to control dispersal as a demographic variable. Redfield (1975) argued in favor of an experimental approach to the study of dispersal and genetics in the regulation of blue grouse numbers in which plant succession was held constant while bird densities changed. In this study plant succession was negligible as a factor in changes in behavior and population dynamics. The islands consisted of second-growth forests of large Douglas fir (Pseudotsuga menziesii), and Douglas fir regeneration was mostly absent under the older trees.
2.2 Description of study areas The locations of the capture sites and release sites are shown (Fig. 2.1). The three capture sites are similar in climate and native forest cover, and are at the junction of the Coastal Western Hemlock and Coastal Douglas Fir bioclimatic zones (Krajina 1965). All three areas had been logged, followed by slash burning and wildfires, and replanted with Douglas fir. The demographic history of the three source populations is shown (Fig. 2.2). In contrast to the capture sites, the release sites were covered with an older growth of Douglas fir. Some trees were 1.5 m dbh and over 40 m in height. Sections of Moresby Island had been selectively logged but little regeneration had occurred. There were open forests without shrubs on all the islands, where soils were thin and bedrock protruded. Such open forest was apparently preferred habitat for blue grouse. In those forest stands with a closed canopy a broadleaf evergreen salal (Gaultheria shallon) dominated the understory and provided a dense shrub layer often a meter in height. Such sites were seldom used by territorial males and even less so by females with broods (Donaldson & Bergerud 1974, Bergerud & Hemus 1975). Moresby Island had the most preferred habitat (Table 2.1, Fig. 2.3).
2.3 Methods Moresby Island received 40 banded birds from each of CC, CB, and MQ —60 males and 60 females for a total of 120 birds. Portland Island received 40 banded CC birds, Stuart Island 40 banded MQ birds, and Sidney Island 40 banded CB birds, in equal ratios of males and females. Sidney Island was deleted from the study when few males could be found in 1971 and 1972. Unbanded males on the release islands were captured with a noose pole and
32
A. T. BERGERUD
Fig. 2.2. Density of advertising males at the capture sites (Mossop 1971, Zwickel et al. 1983).
banded whenever possible. Hens with young broods were captured and marked to facilitate censuses, movement studies, and estimates of production and measurements of behavior. A majority of the hens were banded on Moresby from 1971 to 1976. Productivity was estimated by counting the number of chicks in broods in late July and August. These counts were made each year on Moresby, on Portland in 1972 and from 1974 to 1977, and on Stuart from 1975 to 1977. Large chicks were captured and banded to evaluate overwinter survival. Behavior of hooting males was quantified by recording the distance between an observer (or dog) and the bird when the bird flushed, the incidence of display in the presence of an observer (postures, calls, etc.), and length of time it took for the bird to resume hooting after observers disappeared. A major index was the hooter response score; birds that flushed at long distances when minimally disturbed were scored as 1, whereas birds that refused to flush and continued to advertise when severely disturbed received a score of 5. Details of this index can be found in Donaldson and Bergerud (1974). When a male was heard hooting, a test arena was often set up to allow us to observe the interaction of the bird with
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
33
Table 2.1. Comparison of environmental factors among blue grouse study areas on and near Vancouver Island, British Columbia Open/closed Competition system among stocks
Study area
Age of forest
Forest succession
Density of stock
Capture sites Vancouver Island Copper Canyon (CC) Comox Burn (CB) Middle Quinsam (MQ)
young young young
rapid rapid rapid
high medium medium
open open open
none none none
old old old old
little little little little
low low low low to medium
closed closed closed closed
none none none three stocks
Open fir area (ha)
Dense fir area (ha)
Unsuitable 3 area (ha)
Total area (ha)
251 48 323
261 125 250
205 62 137
717 235 710
Release sites Gulf Islands Stuart Island Portland Island Sidney Island Moresby Island
Island Stuartb Portland Moresby a b
For blue grouse. Excludes one peninsula.
its mirror image (see Chap. 1, Bergerud & Hemus 1975). A "female" dummy was placed in front of a triangle of mirrors, and males were attracted to the arena by recorded female "whinny" or "cackle" calls (Bergerud & Butler 1985). As the male approached the dummy and encountered his image in the mirror, he attacked the male image, moved away from it, or proceeded to mount the dummy. We measured the intensity of distraction behavior of hens with broods and recorded their flushing distance and the distance they flew after flushing. Our indexes included: whether the female "led" us away or approached us; the distance moved while displaying; the number of times a hen flew by us (passes) when we played chick distress whistles; and the length of time she remained in the area. We had no measurement of female aggressiveness until 1975, when we recorded a cackle call of a hen attacking a dummy in a mirror arena on Stuart Island (Bergerud & Butler 1985). The cackle call apparently had an aggressive function and was later played to measure answering responses as an index of aggressive behavior. In 1970 I raised a sample of chicks from each of the three Vancouver Island populations. Some chicks were hatched from eggs, and others were captured at
34
A. T. BERGERUD
Fig. 2.3. Topography and vegetation cover on Moresby Island. Most male territories were adjacent to the highest hill on the island and were in preferred habitats of scattered Douglas fir and bedrock.
1 to 4 days of age. I quantified the interactions of each chick with its mirror image weekly from the age of 1 week to 2 months, by placing a mirror in the brood box for 5 minutes and recording the frequency of hard pecking and shadowboxing for all members of the brood (Theberge & Bendell 1980). In 1971 and 1972 approximately 26 males and 39 females resided on Moresby Island. At that time, many of the hens had not nested, suggesting a female-
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
35
dominated, monogamous system. To test this hypothesis, 20 males and five replacement yearlings were removed from their territories for the Moresby population in 1976. I predicted that a decrease in the male population of a femaleenforced, monogamous system would result in a decrease in the number of females nesting in 1977. I also hoped to test the hypothesis that so-called silent males on Moresby were prevented from holding a territory by more aggressive males. The males removed included all the MQ and CB male founders still alive. These males and the others removed were the most aggressive on the island; nearly all hooted in our presence or had attacked their mirror images. If the more aggressive males were preventing other males from holding territories, I expected the less aggressive males to replace those that had been removed. I also removed 12 females that aggressively defended their chicks. Spacing between the more docile birds that remained was expected to be closer, and the population was expected to increase. Further, the selective removal of only aggressive males and females should result in a decrease of aggressiveness in the progeny of remaining birds, if aggressive behavior is inherited as a dominant characteristic. Redfield et al. (1972) determined the frequency of the Ng genotypes for most of the founders introduced. They recognized three alleles at an autosomal locus, NgM, NgS, and NgF. The common genotypes were NgS/NgM, NgM/NgM, and NgM/NgF.
2.4 Demography 2.4.1 Survival and dispersal of founders The introduction of the new populations to the islands was successful everywhere except on Sidney Island. The survival of males from spring 1970 until spring 1971 ranged from 40 to 50% for all the populations, except on Sidney Island, where only 15% survived. Survival of females was determined only on Moresby; 13 CC females (65%), 10 CB females (50%), and 16 of 20 MQ hens (80%) were alive in 1971. Yearlings did as well as adults on Moresby; 52% survived (14/27) compared to 55% survival of adults (51/93). Survival did not differ among genotypes (Table 2.2). Males generally established territories in open, Douglas fir forests, usually on elevated bedrock sites with little vegetation at ground level on all three islands. On Moresby, CC males scattered the least from the release sites, settled closer to other males, and had smaller territories than did either MQ or CB stock (Figs. 2.4, 2.5). Differences in dispersal distances among stocks on Moresby are not related to the dates of release: CC birds were released first, from 14 to 17 May, and moved the shortest distances; CB birds were released second, from 17 to 18 May, and
36
A. T. BERGERUD
Table 2.2. Survival and dispersal of colonizing birds of different genotypes
Genotype3
% of birds released
% of birds that survived 1 year
% of survivors that dispersed (<3d ) or did not nest ( 9 9 )
NgS/NgM NgM/NgM NgM/NgF
23 60 17
24 61 14
50 43 78
a
As determined by electrophoretic banding patterns (Redfield et al. 1972).
moved the greatest distances; and MQ birds were the final ones released, on 19 May, and moved intermediate distances. The release site on Moresby had been selected as a "preferred" habitat, a high elevation with an open, Douglas fir forest. The long dispersals of the CB and MQ males took them away from the higher land and from the most suitable habitat for territories (Figs. 2.3, 2.4). Bird CB 3717, released after all the CC birds, established its territory at the release site. Bird MQ 3453, released after all the CC and CB stock, was the bird second-closest to the release site (Bergerud & Hemus 1975). Unable to disperse farther, eight of the CB and MQ founders settled on the shoreline of Moresby Island (Fig. 2.4) in forested habitats with low elevations and limited visibility at ground level (Bergerud & Hemus 1975). Verification of the inferior quality of these eight territories was strengthened in later years when five of the six birds still alive shifted slightly inland (Fig. 2.4). Also when these males died, their original territories were not reoccupied by replacement males. Because females were located primarily in the center of the island (Fig. 2.4), the peripheral males may have sacrificed breeding opportunities by dispersing. As will be shown, these males were the most aggressive and presumably could have competed successfully for the better habitats near the release site. The 39 females that survived on Moresby Island also showed differences in dispersal with respect to their origin. Fifteen females settled in the center of the island and raised chicks; eight were from CC and only seven were from CB or MQ. A second group of eighteen females traveled around the island in a "hen party," and at least 13 did not nest (Fig. 2.4, Table 2.3). We saw these females repeatedly in April and May of 1971 on the south tip of the island (Fig.2.4). Often they sat together in the same tree and gave soft whinny calls. On several occasions individual females left the group and flew out over the ocean on a departing course, only to reverse direction and return. This hen party was composed of mostly CB and MQ females, many of which did not nest (Table 2.3). Some of these females also tried to leave the island in 1972 and 1973, and many failed to nest (Table 2.3). Again, the CC females showed the least tendency to wander and
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
37
Fig. 2.4. Top: Dispersal of founding males on Moresby Island. Several males shifted to the edge of the island. These males were usually aggressive and fought their mirror images. In 1971-72 a number of hens, mostly of MQ and CB stock, did not nest and dispersed to the south end of the island. Bottom: Several aggressive males shifted their territories inland in later years and were thus closer to females.
38
A. T. BERGERUD
Fig. 2.5. Comparison of disperal distances, distance to nearest neighbor, and territory size among founding stocks. Territory size of a male was calculated as the distance between the two sightings farthest apart.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
39
Table 2.3. Comparison among stocks of founding females that attempted to leave Moresby Island and females observed with broods
a
Year and stock
No. of different females observed
% of hens that roamed island
% of hens that raised brood
1971 CC MQ CB
9 14 9
11 64 55
78a 14 45
1972 CC MQ CB
9 11 6
22 45 33
78 72 67
1973 CC MQ CB
3 6 3
0 17 33
100 83 67
Percentages do not necessarily total 100%; for example, in 1971 one CC female roamed the island but two were observed without broods.
were more inclined to nest than were the more aggressive MQ or CB females (later section). The dispersing females that did not localize and nest, and the males that moved to the edge of the island, included all the genotypes (Table 2.2); however, the number of dispersing birds of the NgM/NgF genotype was larger than expected by chance (Table 2.2; P < 0.05). The initial, first-year survival rate favored females on Moresby; 39 of 60 females survived, compared with 26 of 60 males. After these initial losses there was no difference in the rate of survival between males and females or among birds of different origin. The annual survival series of males alive on Moresby Island for the 6 years beginning in 1971 is 26-21-12-8-5-3; for females, 39-31-21-19-13-7. The annual survival series for males on Stuart beginning in 1971 is 10-8-6-3-2-1. These figures are similar to the roughly 30% annual mortality rate observed in other blue grouse populations on Vancouver Island (Fig. 2.6).
2.4.2 Survival of progeny Nesting success was high; 93% (n = 221) of the adult and yearling females seen initially in June on Moresby had chicks from 1974 to 1977. Nesting success did not decline in 1977 as a result of the unbalanced sex ratio (34 males : 80 females) caused by the removal of 25 males in 1976. Based only on banded hens, the mini-
40
A. T. BERGERUD
Fig. 2.6. Mortality and survival rates of grouse on Moresby Island.
41
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
mum nesting success from 1974 to 1977 was 94% (n = 143). It did not appear that nonbreeding yearling hens were overlooked; adult females tagged as chicks and not seen as yearlings constituted only 12.5% (3 of 24), which is close to the percentage of tagged adult females who were alive but not seen in any given year-11% (11 of 103). The high nesting success may be related to the fact that there were few mammalian predators on Moresby. I estimate that nearly every hen attempted to nest in 1974-1977, and that the 6-7% of the hens seen without chicks in July and August had nest failures. One hypothesis that I tested for was that there would be r-selection for an increased reproductive rate because of the low initial density and the large area for colonization. Brood sizes in late July and August on Moresby showed little variation from 1971 to 1977; they averaged 2.2-2.7 chicks per brood (Table 2.4). Brood size did not vary among founding females from different populations, between yearlings or adults, or among annual cohorts of hens hatched on the island (Table 2.4). Broods of CC hens on Portland were generally smaller than those of CC hens on Moresby: 3.5 chicks in 1972 (n = 4 broods), 1.5 chicks in 1974 (n = 15), 1.9 chicks in 1975 (n = 11), one chick per brood in 1976 (n = 5), and in 1977 (n — 6). The brood counts of MQ stock on Stuart were similar to tallies on Moresby: 2.8 chicks in 1975 (n = 9), 3.2 chicks in 1976 (n = 6), and 2.2 chicks in 1977 (n = 8). A number of hens without chicks were seen on both Stuart and Portland in all years. Overwinter mortality of juveniles (August chicks) has not previously been Table 2.4. Sizes of broods in late July and August on Moresby Island Origin of female founders or cohort
n females
CC MQ CB 1971 1972 1973 1974 1975 1976
30 28 19 10 11 13 28 9 3
Mean n
No. of chicks per brood female 1971 1972 1973 1974
2.7 4.0 2.0 2.7 (13)
1975 1976 1977 Total
1.0 3.8 2.4 2.4 2.9 2.3 2.8 3.5 3.3 (2.5)a 3.0 2.3 - (2.7) 3.0 - (2.3) -
2.8 1.3 1.0 2.0 1.2 2.3 (2.9) -
1.3 1.5 2.7 1.0 2.0 2.2 2.0 (2.9) -
_ 2.0 3.0 2.0 3.0 (2.0)
2.4 2.4 2.7 2.3 2.1 2.3 2.5 2.9 2.0
2.4 (15)
2.4 (34)
2.2 (38)
2.3 (8)
2.4 (151)
2.7 (13)
a Parentheses indicate the brood female is a yearling.
2.6 (23)
42
A. T. BERGERUD
documented for a closed population of blue grouse. Most mortality of males occurred in the juveniles' first winter, whereas females had similar mortality rates in their first and second winters (Fig. 2.7). The total mortality rate for young birds up to their third summer was about 63 %, and was similar for males and females. Mortality rates of these juveniles were not related to social interactions implicit in recruiting to the breeding population as argued by Watson and Moss (1972) and Watson (1985) for the red grouse (Lagopus 1. scoticus} because most males died in their first year of life, before they had competed for territories as adults. A similar mortality sequence has been documented for blue grouse yearlings on Hardwicke Island (Hines 1986). Furthermore, after 20 of the most aggressive males were removed in 1976, the number of males advertising in 1977 declined (Table 2.5), despite the fact that there were approximately 70 juvenile males in the August 1976 population. The survival rate of these males did not improve, in spite of the reduced potential for competition from adults or the increased number of vacant territories. Females had similar mortality rates between their first and second springs, before breeding as yearlings, and between their second and third springs, after breeding.
Fig. 2.7. Age-specific mortality rates of birds up to 2 years old.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
43
Table 2.5. Composition of the population on Moresby Island No . of females
No. of males
a b e d
Year
Display hooting
Do not hoot
1971 1972 1973 1974 1975 1976 1977 1978 1979
20 22 (4)d 21 (3) 28 (0) 29(1) 44(7) 34(?) 40(?) 41 (?)
6a 4 9 9 10 5+ 9
With chicks 15 21 19 35 50 59 67
(0) (3) (5) (7) (15) (9) (?)
Without chicks
Chicks unknown
Minimum no. of chicks (Aug)
19b ?b
5C 6 10 4 6 3 7+
35 55 60 94 130 140 152
4b 1 3 5 6
9
9
9
9
9
9
9
9
9
9
All 6 are adults. All are adult founders in 1971 and 1972, 3 of 4 are founders in 1973. Banded females seen in later years not seen in year listed. Four of 22 are yearlings.
2.4.3 Mechanisms of population change The number of territorial males on Moresby increased little from 1971 to 1973 (Fig. 2.8). The slow initial growth was a result of the nonbreeding founders from CB and MQ stock (Table 2.3). Rates of increase were highest in 1973-74, 1975-76, and 1977-78 (Fig. 2.8). The number of territorial males declined from 1976 to 1977 following the removal of 25 males in 1976. Counts of territorial males on Portland showed that the population increased between 1972 and 1974, owing primarily to silent founders, four in 1971 and three in 1973, that advertised in the later years. The number of males on Stuart was essentially stable until 1975; one founder did not advertise in 1972. After 1977 the population declined and poaching was reported. The failure of the Portland and Stuart populations to show consistent increases cannot be related to social exclusion of potential recruits by territorial owners. Silent yearling males, as well as vacant territories that had been or were later occupied by other hooting birds, were observed in all years (Fig. 2.9). For example, four territories established on Portland by adults in 1974 were unoccupied and available the year before, in 1973, to six silent yearlings. These yearlings simply waited an additional year before advertising. Lewis and Zwickel (1982) have shown that a number of available territories were unoccupied each year at Comox Burn, Vancouver Island, where some blue grouse males did not even advertise as adults. Changes in the population size on Moresby Island corresponded significantly
44
A. T. BERGERUD
Fig. 2.8. Numbers of females and territorial males on Portland, Stuart, and Moresby islands.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
45
Fig. 2.9. Locations of male territories on Stuart and Portland islands and number of years the territory was occupied. Males on Stuart were located in 9 years; their locations are shown with circles because they seldom moved from specific sites on hilltops. Males on Portland were located in 8 years. Territories on both islands were often vacant.
46
A. T. BERGERUD
to the production of juveniles—those alive in August (Fig. 2.10). When there were more than 1.3 juveniles per hen in August, the population of hens increased the next year and the population of displaying males increased 2 years later (Fig. 2.10). The dependence of population increase on production can be approximated from composition and mortality statistics. Let us assume that of a population of 100 birds in which the number of males and females is equal, yearlings represent 30% of the population. Survival and mortality statistics needed to stabilize the number of territorial males 2 years hence are: (a) Recruited:
(50 females rear 1.3 chicks per hen (Fig. 2.10)/2) x 0.35 survival of male chicks to 2 years (Fig. 2.7) = 11 recruited adult males.
(b) Died:
35 adult males x 0.32 mortality (Fig. 2.6) = 11 males died.
The population is stable with the 11 recruited adult males replacing the 11 who have died. The population generally increased during the study when there were 2.4 chicks per brood (Table 2.4), and when 85% of the banded females had broods in August, 1974-77, (n = 145). The dynamics are as follows: (a) Recruited: (50 females x 0.85 rearing hens x 2.4 chicks per brood/2) x 0.35 survival of male chicks to 2 years = 18 adult males recruited. (b) Died:
35 adult males x 0.32 mortality = 11 males died.
The population has increased; 11 males have died and 18 have recruited. The rate of increase of the population in 1974-76 (after most hens were nesting and before the removal in 1976) —from 28 to 44 displaying males —is r = 0.226, or based on the above calculations, (35 - 11) + 18 = 35 in year 1 to 42 in year 2 or r - 0.182.
2.5 Behavior 2.5.1 Behavior of the founders My intention was to quantify behavior that might relate to the spacing of individuals and hence to the regulation of numbers, and to observe how differences among individuals or founding populations relate to environmental and heritability differences. An earlier investigation showed behavioral differences among the founding
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
47
Fig. 2.10. Annual rate of population increase on Moresby Island regressed against estimated number of chicks per hen, based on August brood counts 1 year earlier (for females) and 2 years earlier (for males).
stocks (Bergerud & Hemus 1975). The CC birds showed more avoidance behavior and displayed less frequently in the presence of observers than did CB or MQ birds. Again CC males interacted less vigorously with their mirror images than did CB or MQ males. This pattern was documented at the capture sites in 1967, 1968, and 1970, and on the islands in 1971 and 1972 (Chap. 1, Bergerud & Hemus 1975). These differences persisted despite the change in environment from the newly regenerating forests at the capture sites to the mature forests at the release sites. The best measure of male aggressiveness was the mirror test. Comox Burn and MQ founders on Moresby attacked their images in 9 of 21 tests (43%). Five of these CB and MQ birds were males that had dispersed the farthest and settled on the edge of the island (Fig. 2.4). There was no image fighting in the nine tests of CC males on Moresby. Middle Quinsam males on Stuart Island attacked their images more frequently than did CC males on Portland Island (Bergerud & Hemus 1975). Most MQ and CB birds hooted when we were close by and standing in plain view, whereas CC males tended to stop hooting when we were still too far away to see them (Fig. 2.11).
48
A. T. BERGERUD
Fig. 2.11. Mean hooter-response scores of territorial male founders and progeny for the different stocks, from 1971 to 1979. Nonaggressive males score 1 and most-aggressive males score 5.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
49
The mean hooter response score for 12 birds (founders and progeny) that fought their mirror images was 4.5 (range 3-5), compared with 3.0 (range 1-5) for 23 birds that did not fight their images (P < 0.05). Thus the hooter response score appears to be an indirect index to aggressiveness (cf. Donaldson & Bergerud 1974). Behavior remained relatively constant in founding birds of all stocks throughout their lifetimes, although most index averages increased slightly as the years passed (Table 2.6, Fig. 2.11). In later years, MQ and CB males usually continued to hoot even when we yelled at them, threw sticks or stones, and struck the trees from which they were hooting. Birds from CC stock on Moresby became silent or flushed on sight, scores of 1-2, whereas the CC males on Portland usually became silent but did not flush when they heard or saw observers, scores of 1-3. This behavior was stable over time (Fig. 2.11). The MQ and CB birds on Moresby continued to flush at closer distances and to display more frequently than CC birds throughout their lives (Table 2.6). In addition, the CC males on Moresby were more secretive and less frequently seen in all years than CB or MQ males there, even though we searched more frequently for them (Fig. 2.12). Three CC males were overlooked in 1971 when they did not advertise, one was missed in 1972 and another in 1973, for a total of five missed of 26, or 19%. None of the 45 MQ or CB males was overlooked in any year (for other differences in behavior, see Bergerud & Hemus 1975). When we measured distraction display in 1971 and 1972, we found that CB and MQ hens defended their broods more vigorously than did CC hens (Bergerud & Hemus 1975). These differences in behavior remained constant during the remainder of the study until all founders had died. Like the males, CC hens on Moresby were much more difficult to locate than CB and MQ hens (Fig. 2.12). Fifteen of 49 CC hens (31 %) were not found in some years, but only six of 84 CB or MQ hens (7%) were overlooked in any year (P < 0.05). Table 2.6. Behavior of founding males on Moresby for 4 years, 1971-74 Flushing distance3 (m)
a b
Year
CC males
MQ and CB males
1971 1972 1973 1974
13 (16)b 18 (18) 21 (4) -
7(40) 8(11) 8(14) 12 (2)
% of birds that displayed to humans CC males 0 7 0 0
(16) (30) (5) (3)
Distance from human observer when the bird left its perch and flew into the air. Sample size in parentheses.
MQ and CB males 32 (73) 22 (65) 13 (45) 60 (5)
50
A. T. BERGERUD
Fig. 2.12. Comparison of the observability of different founding stocks and tagged progeny on Moresby Island from 1971 to 1976.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
51
2.5.2 Behavior of the hand-raised chicks Figure 2.13 shows a summary of the reactions of young chicks to their mirror images. Chicks raised from eggs interacted less with their images than did chicks hatched in the wild (Fig. 2.13); however, chicks from CC and CB, regardless of age, showed less interaction than did MQ chicks (Fig. 2.13; P < 0.05).
2.5.3 Behavior of the males The behavior of the first six tagged yearling males observed on Moresby in 1972 ranged from aggressive to nonaggressive. One yearling held a territory, fought his reflection, and would not stop hooting when disturbed. A second territorial yearling ignored his reflection, copulated with the dummy, and scored 3 in hooter response. Another yearling, tagged as a chick in 1971 (MQ mother), approached
Fig. 2.13. Frequency distribution of chicks' responses to their mirror images. Chicks raised from eggs interacted less vigorously than chicks captured at 1 and 2 + days of age (table upper right). Sample size is in parentheses.
52
A. T. BERGERUD
a mirror arena set up for a territorial founder and chased off another yearling that had been attracted. Exemplifying nonaggressive behavior was a yearling tagged as a chick (CB mother) who would neither approach the mirrors nor hoot. Two yearlings held territories on Stuart in 1972; both hooted in the presence of disturbance (Table 2.7), and both fought their mirror images. A third yearling, who was not heard hooting, entered a mirror arena, dominated an aggressive territorial founder (R/S, R/Y; Table 2.7), and copulated with the dummy. Table 2.7. Hooter response scores from banded males from Stuart Island through the years (Y = yearling, A = Adult) Leg bands
1971
1972
1973
1974
1975
1976
Founders R/S R/Ya R/S G/R R/S Y/BK R/S BK/R R/S B/Y R/S W/Y R/S Y/BK R/S G/B
5 4 4 5 nt 5 5 5
4 3 nt 5 4.5 nt nt 5
5 ntc nt 4 4 4
4.5" nt nt gone gone gone
5 5 5 — —
4.5 gone 5
4.3 — -
— —
— —
4.7
4.3
Mean Progeny G/BK/B/S B/BK/S/B B/S B/R G/S R/Y S/R B/G Y/B S/R B/G B/BK G/B S/G Y/B G/Y G/G G/R R/B BK/B G/Y S/Y BK/Y Y/R
Mean
-
5Y 4.5Y
gone
_
5
gone
—
—
—
4.3
4.5
5.0
4.7
4.3
gone gone 5Yb
— — gone 5A 5A 5Y
— —
~
~ 4.5 4.3 5.0
5A
5 4.5 4.5 4.8 5 4.8
_
4.5A
4.4 4.6
5Y
gone
4.8
4.6
— -
4.5Y 4.5Y _
_
5.0
5.0
_
4.8
a Bands are red over silver left leg, red over yellow right leg. b
c
5 gone gone
5A 5A
4.8
Bird tested more than once, mean score, nt = not tested.
1977
4.8
gone gone
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
53
The most startling yearling occupied an isolated territory on Portland in 1972. This bird approached the mirror and attacked his image, then turned and flew into my face; I could not easily drive him off. All founding stock released on Portland were of the docile phenotype, yet one of its progeny was the most aggressive bird encountered in 10 years. Throughout the study it was a common occurrence for yearling males who had not been hooting to be attracted to the mirror arena by broadcasting taped female whinny calls. One such test on Portland attracted four yearlings and the territorial adult male. Usually young males did not hoot and "held back" if the territorial adult was present; but in his absence, they commonly mounted the dummy. Nonterritorial yearlings were definitely interested in breeding females. One cannot conclude that yearling males that do not advertise do not attempt to breed. Both docile and aggressive phenotypes were common in the male progeny on Moresby. Hooter response scores averaged 3.3; this was midway between the CC founders' mean score of 2.0 and the MQ/CB mean score of 4.2 (Fig. 2.11). There was no noticeable decrease in hooter response scores in 1979 as a result of the removal of aggressive males in 1976 (Fig. 2.11); nor could I detect any significant trends in the aggressiveness of progeny with respect to time or cohorts. The first generation of native birds tested in 1972 did not differ markedly in behavior from progeny tested in 1979 (Fig. 2.11). The limited evidence suggests that distinct docile and aggressive phenotypes continued to exist in the untagged Moresby progeny. There were progeny with hooter response scores of 5 who fought their image, and those with scores of 1 who would not approach the arena (see Fig. 2.21). There were also progeny with both long and short flushing distances (Fig 2.14). The male progeny on Stuart were all aggressive; hooter response scores were initially high and remained so throughout the study (Fig. 2.11, Table 2.7). However, on Portland many of the progeny were more aggressive than the docile founders (Figs. 2.11, 2.15, also see 2.21). This aggressive phenotype appeared fully developed in the first generation, which was raised on the island in 1971 and tested in 1972. The apparent trend toward increased aggressiveness and reduced avoidance on Portland (Table 2.8) is misleading. The founders were docile, and the progeny were either aggressive or docile. But the death of the docile founders resulted in a mathematical increase in mean scores for the entire population even though there was no apparent increase in aggressiveness of the progeny after the first generation (Table 2.8). Unfortunately with such an original docile population I seldom could see bands, and scores were combined for all birds seen. The hooting behavior of five territorial males on Moresby was compared with the distraction display behavior of their mothers. One territorial male reared by a docile CC hen commonly hooted in sight of the observers and would not flush when we yelled and threw stones. The four others, who had been raised by aggressive CB and MQ mothers, scored only 2-3 in hooter response. Thus there
54
A. T. BERGERUD
Fig. 2.14. Frequency distribution of distances male birds flushed from field-workers on Moresby Island. Flushing distance of progeny was intermediate between the close distance of MQ and CB and the longer distance of CC males.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
55
Fig. 2.15. Comparison of the behavior of male founders and progeny on the islands among stocks and also with birds tested at capture sites in later years.
56
A. T. BERGERUD
Table 2.8 Behavior of all males on Portland Island (CC stock) compared with males on Stuart Island (MQ stock)
Flushing distance (m)
a b
% of birds that would not flush
Year
Portland3 Stuart
Portland
Stuart
1971 1972 1973 1974 1975 1976
26 28 25 14
(7)b 15 (6) 9(16) (11) 9 (2) (2) 9 (1) (9) 5 (2) 16 (10) 10 (2) 1 1 ( 1 2 )
13 (8) 29 (17) 67 (6) 61 (23) 78(51) 86 (14)
58 43 75 86 82 75
(12) (21) (8) (7) (11) (4)
% of birds that displayed to humans
% of birds that fought image
Portland
Stuart
Portland
Stuart
0 (3) 13(15) 40 (5) 50 (16) 50 (6) 29 (7)
42 43 38 14 25 46
no test 17 (6) 20 (5) 25(12) 33(12) 33 (9)
no test 80 (5) no test 50 (4) 40 (10) 40 (5)
(12) (21) (8) (7) (4) (35)
Percent of territories on Portland *occupied by founders by year: 100 86, 45, 27, 25, 9. Sample size in parentheses.
appeared to be no clear association between the degree of aggressiveness of hooting males and the intensity of distraction display of their mothers. Originally several of the most aggressive male founders on Moresby, all from CB or MQ stock, dispersed to the periphery of the island, where their nearest neighbors were quite distant. As these founders died, progeny did not occupy the vacant territories. The new territorial males selected sites between other males, and even in 1976, when the population had doubled, vacant spaces remained between territories (Fig. 2.16). The birds that continued to hoot in the presence of disturbance by observers were no farther from their closest neighbors than were the birds that flushed or became silent (Fig. 2.17).
2.5.4 Behavior of the female progeny The only measurement of aggressiveness by females, relative to spacing, was the frequency of response to the taped cackle call. The call was played at fixed locations along roads and trails and at male locations early in the morning in April 1975-77. Frequencies of females responding to the call with their own calls were: Stuart Island Stuart Island Stuart Island Portland Island Moresby Island
1975 30% (n = 10 tests) 1976 18% (n = 90) 1977 19% (n = 121) 4% (n = 26) 1976 1% (n = 151) 1977
The MQ descendants on Stuart were significantly more responsive than the progeny on Moresby or Portland. Probably most of the females heard the calls, because females commonly responded from distances of several hundred meters.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
57
Fig. 2.16. Distribution of male territories on Moresby Island. Territories became smaller and more uniformly spaced as CB and MQ founders died. Note the large spaces between territories.
58
A. T. BERGERUD
Fig. 2.17. Comparison of the nearest-neighbor distances between advertising males that scored 1 and 2 in hooter-response scores with birds that scored 4 and 5. As population density increased and peripheral founders died, the population became more closely and uniformly spaced.
The percentages of hens that should have heard the call and that responded were: Stuart 80%, Portland 18%, and Moresby 1%. The mean values of distraction displays of hens hatched on Moresby were intermediate between those of their founding mothers (Fig. 2.18). There is a suggestion that the distraction displays abated somewhat with time, possibly because of habituation, but only one index showed a significant change (Fig. 2.19). Distraction behavior of the CC descendants on Portland was not markedly different from that of the founders (Fig. 2.18); the approachability of progeny decreased in three indexes and increased in two. The MQ female progeny on Stuart showed noticeably greater approachability and distraction display behavior for all five in-
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
59
dexes than did either the CC founders or progeny on Portland, or the CC founders on Moresby (Fig. 2.18).
Fig. 2.18. Comparison of distraction behavior of hens among founding stocks and progeny.
60
A. T. BERGERUD
Fig. 2.19. Top: Distraction behavior of hen progeny compared by annual cohorts with that of founding hens. Bottom: Distraction behavior of all hens compared between years.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
61
Finally, I compared the intensity of distraction display of nine hens with that of the mean values of the founding stocks and the behavior of their mothers or grandmothers (Fig. 2.20). I used only hens whose behavior had been measured on several occasions. The behavior of two hens, one with a CC mother and one with a CC grandmother, was closer to the mean behavior of the CC line than to the MQ and CB in six of seven comparisons. The behavior of seven hens that had had a CB mother or grandmother was nearer to the behavior of their maternal MQ and CB lines than the CC line in only 11 of 21 cases (Fig. 2.20). A progeny's behavior was closer to its mother's behavior than to the mean score of all other progeny in 13 cases; the reverse was true in 18 cases. However, aggressiveness of the progeny generally increased in parallel with the mother's aggressiveness (Fig. 2.20). Note that the mothers in Fig. 2.20 are ordered from left to right by increasing aggressiveness. Further, as the population on Moresby aged, its new cohorts of female progeny may have become progressively more docile (Figs. 2.18, 2.19). Originally the CC stock had a reproductive advantage in 1971 and 1972. However, the founders retained their distinct behavior, which indicates that the increased docility was not simply a response to increased density.
2.5.5 Behavior of progeny at capture sites I returned to the capture sites at MQ, CB, and CC on Vancouver Island 4 to 10 years after the founders were captured to study the behavior of progeny in open populations, where they were free to disperse. Behavior of hooters at MQ, CB, and CC in later years was similar to that of the founders in 1970 (Fig. 2.21, Bergerud & Hemus 1975). In 1980 the birds at Copper Canyon still flushed at greater distances than MQ males, failed to fight their images, and stopped hooting when approached (Fig. 2.21). Mean flushing distance of CB males at Comox in 1980 was 8.2 m (n = 32), similar to what Mossop (1971) observed 13 years before (Fig. 2.22). Mean hooter response of 31 CB males in 1980 was 2.7, still intermediate between CC and MQ scores (cf. Chap. 1). Relative behavioral differences between populations had been relatively constant throughout the period of study at the capture sites. Behavior of hens at the capture sites did not change in later years. The MQ hens were more approachable and performed more intensive distraction displays than did their counterparts at CC. At MQ, 14 hens responded to the cackle call in 60 tests in 1976 (23%), and four females entered a mirror arena and fought their reflections in 1980. At CC, three females answered in 47 tests in 1975 and 1980, but none approached the mirror arena. Scores of the hens at CB were intermediate between CC and MQ in 1980, with three replies in 25 tests (12%). Although aggressiveness was not measured with cackle calls before 1976, it appeared that the behavior of females, like that of males, had been constant in the mainland study areas since 1967 (13 years). The CC females were the least aggressive, MQ females the most aggressive, and CB females were intermediate.
62
A. T. BERGERUD
Fig. 2.20. Distraction behavior of hen progeny compared with that of their mothers or grandmothers. Mean values of the founding maternal stock and the mean for all progeny are represented by horizontal lines. Progeny, mother, grandmother are identified by stock and tag numbers at bottom. Number of observations of progeny below diagonal lines at bottom.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
63
Fig. 2.21. Frequency distributions of hooter-response scores of founders on the islands, progeny on the islands, and residents at CC and MQ in later years (all data pooled). Resident birds at capture sites were not tested in 1970.
64
A. T. BERGERUD
Fig. 2.22. Flushing distances (bird-to-observer) of resident birds at the capture sites from 1967 to 1980 (Mossop 1971, Zwickel et al. 1977).
I had not expected to find the behavior of males and females at the three capture sites to be constant throughout the 13-year period. Since Mossop (Chap. 1) had first quantified their displays, all the original birds had died, and both the population density and the vegetation had changed. The population had declined greatly at CC, increased at CB, and declined at MQ (Fig. 2.2). In all three areas suitable breeding space had declined as the forests regenerated. The aggressive phenotype was expected to be favored in all three areas as space became more limited, but this did not occur. To summarize, the behavior o.f progeny hatched on Stuart was the same as that of their parents, while the behavior of some of the progeny on Portland and Moresby Islands differed from that of their founding stock (Figs. 2.11, 2.21). At the capture sites, resident birds displayed the same intensity of behavior in 1980 as they had shown in 1970, which was similar to what Mossop (Chap. 1) observed in 1967-68. In short, behavior in populations in which birds were free to disperse remained constant, whereas behavior in two of three insular populations had changed (Fig. 2.23).
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
65
Fig. 2.23. Summary of relative aggressiveness of founders on the islands, the first generation, subsequent progeny, and resident birds at the capture sites tested in 1975 and 1980.
2.6 Discussion 2.6.1 Demography The density of breeding birds increased following years in which chick production was greater than 1.3 chicks per female in August. If production was less than this, populations usually declined. The increase in the number of females appeared the year following good production, and in the number of territorial males 2 years later. The average brood size of 2.4 chicks on Moresby was lower than that on Vancouver Island, but the low adult-mortality rates were similar. The population was able to increase because nearly all hens nested and because, apparently, at least on Moresby, nesting success was extremely high: more than 94% of the hens appeared to have hatched nests, and 85% of broods had at least one survivor in August. An interesting finding is that it was the aggressive females that did not nest in 1970 and 1971. These birds could not have been excluded by the docile hens. It appears that their commitment to disperse and show philopatry was stronger that that of the docile birds, which were more prepared to accept the new island habitat.
66
A. T. BERGERUD
Other researchers have argued that not all yearling females in increasing, decreasing, or stable populations will breed. They reached this conclusion because yearlings will replace adults and yearlings that have been removed; hence, hens are thought to be excluded through social interaction (Zwickel 1972, Hannon 1978, Zwickel 1980). Hannon and Zwickel (1979) found that all yearling females collected had developed follicles, although 44% would not have laid their first egg until 4 June or later. This 44% was arbitrarily classified as "nonbreeders." Exclusion of these late-developing birds skewed the normal distribution of nesting dates to the left. However, hatching curves, excluding renests, are not skewed (cf. Mercer 1967). Furthermore, hatching distribution of yearling females with broods in Zwickel and Hannon's study area showed that 14% had hatched their clutches during the period when the so-called nonbreeders would have hatched their young (Hannon 1978). Hannon and Zwickel's sample may have included some yearlings en route to other habitats, and breeding might have been later in such birds. When Sopuck (1979) radio-tracked yearling females in the same area in 1976 and 1977, he found that 39 of 46 yearlings (85 %) definitely nested, and another five may have laid eggs. This evidence leads me to conclude, as did Zwickel and Bendell (1967) previously, that most yearling hens in blue grouse populations attempt to nest. It was never shown in the removal experiments that the potential replacement stock would not have nested elsewhere in the absence of removals. To document selfregulation it is necessary to demonstrate that the replacement birds would not have nested elsewhere in the absence of the experiment (Brown 1969, Watson & Moss 1970). Study areas where the removals took place were in optimum habitats traversed by large numbers of potential colonizers (Zwickel 1972, Bendell et al. 1972). One would expect that these yearlings would move to such optimum territories, but if the preferred sites were unavailable, they could still have moved on to the adjacent, inferior habitats and nested (Sopuck 1979). That the number of chicks alive in the fall exceeds that needed to replace the natural losses of yearlings and adults is also used to support the argument for selfregulation in blue grouse (Zwickel & Bendell 1967, Zwickel 1982, Zwickel et al. 1983). It is presumed that such surplus birds are expelled when they return to the breeding range in the spring. In 1976 I removed 20 of 44 territorial males and five yearling replacements in order to determine whether the surplus chicks would compensate for the losses. At a minimum, 140 chicks were alive in August 1976, and 70 of these likely were males, potentially capable of taking territories in 1977. However, the number of territorial males declined from 44 to 34 in 1977, a decline of 23%. Such a decline contradicts the assumptions of the selfregulation hypothesis, but supports the view that adult mortality is relatively constant and chick production determines population density. The calculations for the latter are:
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
67
(a) 130 chicks (August 1975)/2 = 65 males x 0.35 (survival to adult) = 23 adults — 5 shot as yearlings = 18 available replacements as adults. (b) 44 territorial males - 20 removed + 8 dying naturally (44 - 20 x 0.32 natural mortality) = 16 alive in 1977. (c) 18 + 16 = 34 territories in 1977, as observed. The higher rate of overwinter mortality of juveniles than of adults shown must be considered before determining the number of "surplus" chicks in August. Overwinter mortality cannot be estimated by comparing autumn production with the number of territorial birds the following year. Because blue grouse do not always recruit to the ranges they used as chicks (Boag 1966, Bendell & Elliot 1967), some of the previous year's cohorts may have settled elsewhere. New generations redistribute according to breeding opportunities in preferred habitats, in newlylogged and burned areas (Redfield et al. 1970), and in secondary habitats (Sopuck 1979). There will be more yearlings in secondary habitats following a year of high production (Sopuck 1979); populations in the preferred habitats will be consistently more stable. This buffering response of habitat was shown 33 years ago in titmice (Parus spp.) populations (Kluyver & Tinbergen 1953). Blue grouse populations on a regional scale, which includes both optimum and secondary habitats, rise or fall from one year to the next (Redfield et al. 1970). Because adult mortality rates are relatively constant, differences in chick survival must determine these fluctuations. This study failed to document a higher chick survival of CC stock than MQ stock, contrary to the results of Mossop in 1967 and 1968 at the capture sites (Chap. 1). My evidence shows that chick survival is explained by differences in the extrinsic environment rather than by the availability of the intrinsic stock. I could not pinpoint the factors that contributed to the high mortality rates of chicks on Moresby. Few ground predators were found there, most females hatched their clutches, and the weather in June was warm and dry; yet approximately 50% of the chicks disappeared in July, a mortality rate similar to that found in other studies (Bendell & Elliot 1967, Zwickel & Bendell 1967, Zwickel et al. 1983). Chick survival was even lower on Portland, only 1.5 km away. Ground vegetation was considerably denser there than on Moresby, and it is possible that the chicks had an insufficient diet of insects in the relatively old, GulfIsland forests.
2.6.2 Behavior in space and time Founders from a high-density population (CC) settled in close proximity on Moresby and Portland; founders from low-density populations (CB and MQ) dis-
68
A. T. BERGERUD
persed farther from the release sites and were spaced at greater distances from each other. The male founders of high-density origin were less approachable, displayed less frequently, and interacted less with their mirror images; the opposite was true for founders of low-density origin. For brood females, founders from a high-density population (CC) were less approachable, less vigorous in distraction displays, and less aggressive than were their counterparts from low-density populations. These behavioral differences were unaffected by the transplantation of the birds from the young, regenerating forests of the capture sites to the second-growth forests of large trees at the release sites. Furthermore, all three stocks retained their original behavior patterns while sharing identical conditions on Moresby. One might argue that the density of the original population imprinted certain behavioral characteristics on maternal or genetic lines that persisted in the new locations. Behavior of progeny on the islands indicates that this was not the case. Some CC young hatched on Portland were nonaggressive, like their parents (high-density origin); others were aggressive, similar in behavior to the lowdensity MQ stock. On Moresby, density of hooting males was low, comparable to the MQ capture site, and the mean behavior score of progeny was intermediate to that of the CC and MQ/CB founders. As the density of males doubled, the behavior of cohorts remained constant. Behavior of birds that did not disperse in the populations at the capture sites was constant as well, despite a decline in density at CC and slight changes at CB and MQ, and despite the fact that the populations had completely turned over by the end of the study (Fig. 2.2). The evidence indicates that behavior was intrinsic and independent of density. A graph of the flushing distances of males illustrates these changes in behavior (Fig. 2.24). Males at MQ in 1970 flushed an average of 6.0 m, and males seen in the same vicinity in 1975-80 behaved in the same way. Copper Canyon males flushed at the significantly longer distances of 9.9 m in 1970, 9.4 m in 1974, and 9.4 m in 1980. When the founders were released on Moresby, all stocks flushed at greater distances, probably because of the lack of ground cover; however, flushing distances of CC stock were still the longest. At Stuart, short flushing distances were maintained by founders and progeny alike. The mean flushing distance of male progeny on Moresby was intermediate between that of the CC and MQ/CB stock, although individual birds had both long and short flushing distances (Fig. 2.14). Portland progeny flushed at half the distance of their fathers, or at about the same distance as their peers on Moresby (Fig. 2.24). The length of the bars in Figure 2.24 provides a concise summary of the diverse modes of behavior: the shorter the bar at a given level, the greater the hypothesized level of aggressiveness and approachability. Many progeny on Portland exhibited behavior different from that of the founding stock, whereas the behavior of all Stuart progeny was similar to that of the parents. This is consistent with the view that behavior cannot be fully explained by environmental influences. Maternal imprinting could not explain the behavior of several of the five males and much of
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
69
the behavior of eight females whose mothers' behavior was known. Aggressive and nonaggressive modes appeared in the first generation on Portland and Moresby, where habitats were similar, whereas only aggressive progeny appeared on Stuart, although both habitat and density were similar to those of the other islands. For the three open populations at the capture sites, behavior remained relatively constant between population for 10 years, despite changes in
Fig. 2.24. Summary of relative aggressiveness of founders and progeny on the islands and residents at the capture sites. Bar length represents flushing distance. Positive ( + ) values indicate a significant increase in flushing distances and negative ( —) values indicate a significant decrease in flushing distances. Founders flushed farther on the islands than did residents at the capture sites in 1970 because of reduced ground cover.
70
A. T. BERGERUD
habitat and density (Figs. 2.23, 2.24). Thus, neither habitat, density, nor the imprinting of behavior provides an adequate explanation for differences in the behavior of progeny: (1) Similar habitats:
Moresby, Portland, and Stuart—behavior of progeny on Moresby and Portland different from that on Stuart.
(2) Similar densities:
Moresby and Middle Quinsam; Portland and Stuart—behavior of progeny different.
(3) Similar mothers:
Portland and Moresby—behavior of many mothers similar, behavior of many progeny different.
(4) Different habitats:
Stuart and Middle Quinsam; MQ 1970 versus 1980; CC 1970 versus 1980; CB 1970 versus 1980behavior of residents constant.
(5) Different densities: CC 1970 versus 1980; CB 1970 versus 1980; Moresby 1973 versus 1980 —behavior of residents constant. (6) Similar upbringing: All hand-reared chicks —CC and CB behavior different from that of MQ. The very rapid change in behavior in the first generation supports a genetic explanation of behavior. Many of the progeny on Portland were aggressive, unlike their parents, which again supports a genetic hypothesis. The lack of nonaggressive progeny on Stuart suggests homozygosity for aggressiveness in MQ stock. The variability of behavior in CC progeny suggests a heterozygous basis for the nonaggressive CC stock and that nonaggressiveness was dominant. Behavior of the hand-reared CB chicks resembled that of CC chicks, whereas behavior of CB adult founders was closer to that of MQ founders in 1967-68 and in 1980 (Chap. 1, Bergerud & Hemus 1975). The behavior of CB stock was closer to the docile mode of CC before dispersal as chicks, but closer to the aggressive MQ mode after dispersal. In reviewing the Chitty hypothesis of population regulation in 1978, Krebs stated that the most convincing test would be as follows: by introducing or removing certain genotypes, one could cause the population to rise or fall at will. In 1976 we removed 50% of the territorial males. Birds removed were the most aggressive territorial males and 12 aggressive females. These aggressive birds had the largest territories (Bergerud & Hemus 1975). This selective removal should
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
71
have reduced mutual interference, allowing docile surplus birds to take territories. Because these birds would take smaller territories, densities should have increased. However, the population did not increase following this removal, but decreased, as might have been predicted by the removals. One of the most general predictions of the Chitty hypothesis is that "spacing" behavior will be less common or less intense in increasing populations than in declining ones (Krebs 1978). We now have data on the behavior of blue grouse in increasing and decreasing populations at both high and low densities (Chap. 1, Bergerud & Hemus 1975, this study). The data do not show predicted changes in aggressiveness, either for the various stages of population change or for different densities (Fig. 2.25). I therefore reject the original hypothesis of this research that spacing behavior based on genetic differences is a fundamental driving component in causing blue grouse populations to increase or decrease. I believe the different density phenotypes proposed by Chitty are important in the demography of cyclic populations (Chap. 12). But for these noncyclic populations of blue grouse their chief densitydependent response of the phenotypes relates not to dynamics but dispersal in space (next section).
2.6.3 Fitness and densities Spacing of blue grouse is determined largely by the areas where yearlings localize in their second spring (Hannon 1978, Sopuck 1979, Jamieson & Zwickel 1983, Hines 1986). Adults of both sexes are largely locked-in to the sites they selected as yearlings; they make little use of newly created habitats after their first year of colonization (Redfield 1973a). Presumably they increase long-term fitness by returning to familiar areas and excluding other birds. On the other hand, yearlings are free to colonize new areas whether such locations be newly created habitats (Redfield 1972) or areas where resident adults have been removed (Zwickel 1972, 1980, Bendell et al. 1972). Two density-dispersal modes of behavior can be seen for male and female yearlings: those that travel a wide area and commonly settle in a low-density situation, and those whose movement in the spring is more localized and that commonly settle in high-density areas (Sopuck 1979, Jamieson 1982). I suggest these two types represent two strategies for maximizing fitness and are not after-thefact results of social exclusion or repulsion. Roff (1975) has generated a model demonstrating that a stable polymorphism in dispersal tendencies can be generated in heterogeneous environments. Male blue grouse reside in habitats with dense to sparse conifer growth, but they must have open ground for breeding. Subalpine and montane habitats, with scattered conifers and sparse undergrowth, meet these requirements. Such sites show little variability in time and space. The breeding requirements for blue grouse are also met in recently burned areas, or logged areas where the slash has been burned; however, in these sites there is
72
A. T. BERGERUD
much variability in time and space. A polymorphism in dispersal behavior might be expected with such heterogeneity and would enhance population persistence (Roff 1974, 1975). If a yearling male selects a high-density site, he will have many potential females as an adult but also more male competition for these females. If he selects
Fig. 2.25. Relative aggressiveness of different stocks in relation to values predicted by the Chitty hypothesis at various stages of the population cycle (Krebs 1978a, 1985). Relative densities are shown at top and relative aggressiveness at bottom; lines connect density and aggression.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
73
a low-density site, such as a newly created burned area or a second-growth forest he will find fewer females the next year as an adult but also fewer male competitors. In the low-density environment he may even be able to attract hens as a yearling and thus improve his fitness. Whether or not he advertises by hooting and establishes a bonafide territory may depend upon stimulation from hens who localize in his vicinity (see Lewis 1984). Fitness of the female depends on her ability to successfully hatch eggs and raise young. She can seek a low-density site where she will have her pick of nesting habitats and increased space in which to avoid predators. If the area has been burned recently, nest and brood cover may be scarce. Or, she can select a highdensity situation with more protective nesting cover and an interspersion of nest and brood habitat. With many males present, she has a wider choice of mates than in low-density situations. I would characterize these two tactics as density-tolerant and density-intolerant modes. Segregation would result in a positive assortment of high-density, tolerant types together, with small home ranges, and low-density, intolerant types together, but farther apart and with large home ranges. I hypothesize that both types have a genetic basis and that in open populations dispersal is the means of positive assortment. The model I have constructed to explain the diverse behaviors described in this study distinguishes between: (1) open populations in which birds are free to disperse versus closed systems in which they are not, and (2) density-tolerant and density-intolerant phenotypes with a genetic basis (Fig. 2.26). Tolerant birds have evolved the strategy of selecting preferred habitats with many competitors. They have been selected for their inconspicuous behavior, they are less aggressive and less approachable, and their genotype is more heterozygous. Intolerant phenotypes have evolved the tactic of colonizing less-preferred habitats with fewer competitors. If this strategy is to be successful, these birds must be more conspicuous and wide-ranging, hence more aggressive and approachable. This genotype may be homozygous, recessive. The density-tolerant mode was represented by the birds from CC and the founders from these populations on Moresby and Portland. Densities of birds were extremely high at CC in 1967-68 (Fig. 2.2). Birds at CC and founders on Moresby and Portland were nonapproachable and nonaggressive, with small space requirements. Density-intolerant progeny (aa) could disperse away at CC in search of less-dense populations, leaving the density-tolerant (Aa), nonaggressive birds behind. If density-tolerant birds (Aa) dispersed into CC to replace losses, this would result in positive assortment. But on Portland Island, the intolerant aggressive phenotype (aa) could not disperse and leave the area as at CC, and thus I saw these aggressive birds, even by the first generation on the island but never at CC, where they dispersed away (Fig. 2.26) The density-intolerant phenotype was represented by the founders from MQ,
74
A. T. BERGERUD
who were generally aggressive, in 1967-68 (Chap. 1); in 1970 (Bergerud & Hemus); in 1971-72 (Low 1975); and in 1975 and 1980. On Stuart both founders and progeny from this stock were aggressive. In open populations, aggressive stock dispersed, especially because the birds required large space and thus settled in less-suitable habitats. Behavior of birds remained constant at MQ from 1967 to 1980, because of positive assortment of homozygous-recessive birds in seeking
Fig. 2.26. Proposed model of low, intermediate, and high densities composed of tolerant and intolerant genotypes. In open systems, genotypes are free to disperse away from the study area (arrows leaving circles). In closed systems, density-tolerant and intolerant genotypes must remain in the study area. Density-tolerant and intolerant phenotypes show positive assortment in open systems. Thus, selection maintains the constant observed phenotypes within study areas even though dispersal results in genotype sorting within the population as a whole.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
75
low-density environments. The unchanging behavior in the closed system at Stuart is my primary evidence of homozygous inheritance. Blue grouse have a long life span and show philopatry once established. Positive assortment could explain the constant phenotypes over time, even in the face of deteriorating habitats and declining densities. The birds at CB are an intermediate behavior type. Regrettably, the least information exists on the CB stock because the release on Sidney Island failed and I did not visit CB until 1980. Flushing distances for CB birds suggested that by 1980 they were still intermediate between CC and MQ stock, although closer to CC than they had been in 1967-69 (Fig. 2.22). The decline in flushing distances in MQ and CC from 1967 to 1980 (Fig. 2.22) as both populations declined may have been caused by increased forest cover. The slightly increased or equal flushing distances at CB over time (Fig. 2.22), despite the decrease in advertising space, may have resulted from a higher proportion of the more tolerant phenotype, because numbers increased there after 1970 (Fig. 2.2). As stated earlier, hand-reared chicks from CB showed behavior more similar to that of CC than MQ chicks, whereas the adult CB birds, after dispersing, behaved more like MQ than CC birds. Cooper (1977) also hand-reared chicks from MQ and CB. He tested them as yearlings and adults, and found the MQ stock more aggressive and more active ("pacers"). His captive stock, like the chicks I tested, could not disperse. When Hannon (1980) conducted cackle tests for females at CB, she recorded a greater answering response than did I at CC, but none of her birds approached or fought the hen dummy as did MQ hens in my tests. The data of Cooper and Hannon are consistent with my model of an intermediate position of CB birds between birds of MQ and CC, and with the idea that dispersal may sort behavior types within open populations. Blue grouse have been able to colonize a wide range of coniferous habitats — from nearly open sagebrush (Artemisia spp.) with scattered conifers to closed canopies. Densities can range from 0.2 to 90 territorial males/km2 (Redfield et al. 1970, Weber 1975, Frandsen 1980). The habitat is the template for life-history parameters (Southwood 1977) and for adaptive strategies (Wilson 1975). The density of birds must also be a major component of such a template. Chitty (1970) theorized that there could be inherent differences among young that might determine whether they settled in low-density breeding populations or remained to compete for breeding sites in crowded habitats. Redfield (1975) showed that birds with different genotypes at the Ng locus did select different density-habitat situations. This study has provided additional evidence that behavioral phenotypes appear to have intrinsic preferences for associating with birds of similar behavior. My model assumes a Mendelian genetic explanation. The behavior patterns may be continuous, contrary to this alternative inheritable premise. However, flushing distance and response to mirror images appeared bimodal (Fig. 2.14, Bergerud & Hemus 1975). These behaviors in discrete populations of blue grouse
76
A. T. BERGERUD
were more discontinuous than were the same behavior patterns for red and gray color morphs of ruffed grouse (Bonasa umbellus) living in the same population (Chap. 3). Parsons (1983) states that the genetic architecture of ecobehavioral phenotypes is composed mainly of a few additive genes of relatively large effect. This multi-loci genetic basis is contrary to my model but may hold for blue grouse. However, multi-loci effects are difficult to model. The genetic assumption underlying Maynard-Smith's ESS theory is that of a haploid organism. I am making a similar but less restrictive assumption of a diploid organism with 1-locus alternative segregation. This approach permits one to see whether a Mendelian genetic explanation could explain the variation in behavior—that the model fits does not mean it is true, but simply that it works. I return to the initial question of the relationship between densities and genetically programmed behavior. Densities result from the spacing of birds for breeding. This space is modified both by the suitability of habitat, as recognized by the species for nesting and advertising, and by the number and behavior of competing individuals. Through competition and avoidance of occupied territories, birds select those density habitats that best fit their behavioral phenotype. This selection is not an "ideal-free-choice" but an "ideal-dominance-choice" (Fretwell & Lucas 1969). These intrinsic differences between genotypes have been selected by the habitat in an ultimate manner (Williams 1966, 1975) in response to survival and fitness (Fig. 2.27). In turn, individual birds select in a proximate manner the environment where their specific behaviors can best maximize their fitness; therein is generated a feedback cycle of continuous adaptation (Fig. 2.27).
Fig. 2.27. It is proposed that density drives behavior and genetics in an ultimate manner and that individual phenotypes select density habitats that they are intrinsically adapted to as a proximate adaptation.
DEMOGRAPHY AND BEHAVIOR OF BLUE GROUSE
77
2.7 Summary Three blue grouse populations on Vancouver Island (CC, CB, MQ) that differed in avoidance and aggressive behavior, and in productivity and density, were transplanted to islands in the Gulf Islands, British Columbia, which previously had no grouse. Productivity of the founders from the three populations was not significantly different in the new environments, and each stock's specific behavior patterns persisted despite differences in densities and in coniferous plant cover. The high-density population (CC) dispersed shorter distances from the release sites and settled closer together on two islands than did the low-density stocks. Progeny of the MQ stock on Stuart Island remained aggressive and conspicuous, similar to their parents, for the 7 years they were monitored. The first generation of progeny of CC stock on Portland Island ranged in behavior from aggressive to docile. Both phenotypes were present in progeny on Moresby Island, where there were founders from all three introduction stocks. Birds resident at the capture sites on Vancouver Island were free to disperse and generally retained behavior patterns characteristic to each population over a 13-year period, even though their densities and the availability of suitable habitats declined. It is proposed that density-tolerant and density-intolerant phenotypes in these populations have a heritable basis. These behavior types effect positive assortment during dispersal by seeking those density-habitats to which they are intrinsically best adapted.
3
Demography and Behavior of Ruffed Grouse in British Columbia R. G. Davies and A. T. Bergerud
3.1 Introduction A basic problem in ecology is to determine the limiting factors that prevent natural populations from continuing to increase. Especially intriguing also are the wide and regular fluctuations (cycles) in the numbers of some grouse species in Canada (Keith 1963). Chitty (1967) recommended that students study the demography of fluctuating rather than stationary populations to understand the mechanics of population change. Chitty's (1967) hypothesis of density-dependent selection emphasizes the behavior and genetics of animals in increasing and decreasing phases of population growth. We elected to study the demography of ruffed grouse (Bonasa umbellus) in central British Columbia to investigate this and competing hypotheses. At the beginning of our study, ruffed grouse in central British Columbia appeared to exhibit a 10-year cycle in abundance; a high harvest occurred in 1958, and the population was again relatively large in 1968. Ruffed grouse have a polymorphism in body color (Bent 1932). This type of polymorphism in body color generally results from simple Mendelian segretation of alternate alleles at an autosomal locus (Cooke & Mirsky 1972, Innes & Haley 1977, Sassaman & Garthwaite 1980, Vakkari 1980, Owen & Plowright 1980, Fogleman et al. 1980, Zweifel 1981). However, different color morphs sometimes result from chromosomal polymorphisms (cf. Thorneycroft 1975, Christensen & Pederson 1982). We obtained the frequencies of the red and gray color morphs for ruffed grouse and used this overt genetic marker in comparisons with parameters of demography and behavior. We did not presume that correlations between color and behavior or demography necessarily meant that these lat78
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
79
ter parameters were inherited in a similar manner as the color morphs. Gullion and Marshall (1968) documented for their ruffed grouse population in Minnesota that the frequency of the red color morph increased during a cyclic increase. For our investigation, the null hypothesis was that there would be no differences in the behavior and demography of gray- and red-phase birds as the population fluctuated. Field work commenced by counting displaying males in May 1968, and finished 14 years later with a similar count in May 1982.
3.2 Study area and methods Our study area was at Watch Lake, central British Columbia (Fig. 3.1). Vegetation of the general region has been described as a montane-Douglas fir zone by Halliday (1937), and as caribou aspen-lodgepole pine-Douglas fir parkland by Krajina (1965) and Beil (1969). It was essentially a mosaic of mixed and pure stands of aspen (Populus tremuloides), lodgepole pine (Pinus contorta), Douglas fir (Pseudotsuga menziesii), and white spruce (Picea glauca), with numerous pastured and willow (Salix spp.) swamp openings. The vegetation pattern has been strongly influenced by past disturbance: selective logging for Douglas fir in the 1950s, patch-burning in the 1890s and 1930s, and seasonal grazing and pole cutting. The climate of this area is characterized by long, cold, and relatively dry winters with snow on the ground from November to April; summers are short, hot, and dry. Frost-free days range from 50 to 70 per year. Annual precipitation is 50-65 cm, most of which falls as snow. The seasonal distribution of precipitation shows a regular pattern: March and April are typically dry, and June and December-January are periods of maximum precipitation. Seasonal changes are abrupt. Our study area was on southeast slopes extending from low ridges (1300 m above sea level) to flat meadow land (1100 m). Topography was gently rolling. The chief natural predators of grouse in the area were the coyote (Canis latrans) and goshawk (Accipiter gentilis). Goshawks were especially common in 1971-73 and 1980-81. Lynx (Lynx canadensis) and red fox (Vulpes vulpes)were uncommon. Our statistical analysis follows Sokal and Rohlf (1969). Indexes of behavior were tested for normality by rankit plots before using parametric ANOVA and Mests.
3.2.1 Demography We counted advertising males annually by searching for logs used by drumming males. An advertising male was defined as a drumming male occupying a restricted advertising range of one or more drumming sites. The search area was restricted to 109 ha (Study Area I; Fig. 3.1) in 1968 and 1969. In 1973 we en-
80
R. G. DA VIES AND A. T. BERGERUD
Fig. 3.1. Ruffed grouse study area at Watch Lake, British Columbia, showing drumming transect and listening posts.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
81
larged Study Area I to include 376 additional ha. The population in Study Area I had shown little change from 1968 to 1971, and we suspected that the area was of optimum habitat and would not reflect general population trends. The enlarged Area I was censused until 1982. A second study area (Study Area II) of 129 ha was also included from 1970 through 1975 (Fig. 3.1). Data should reflect complete counts of advertising males; we repeatedly searched for birds throughout May in the early years, 1968-74. In 1975-82, we searched for approximately 1 week during the peak of advertising (the first week of May). By that time we knew the location of logs used in past years. Advertising sites were checked first for droppings during midday, then during the drumming hours to verify the presence of drumming males and to search for birds advertising at new locations. A second census method included a drumming transect conducted from 27 April to 10 May along a 24-km section of the Watch Lake and Bridge Lake roads (Fig. 3.1). Listening posts were at 1.6-km intervals. At each post the number of drumming males was recorded during a 3-minute listening stop, between 0500 and 0800 hours. Previous counts were conducted from 1957 to 1967 by the British Columbia Fish and Wildlife Branch (there were no counts in 1962 and 1963) and were continued by the Wildlife Branch until 1982. These transects were repeatedly sampled, with an average of 4.6 censuses per year. A primary reason for establishing Study Area II was to measure habitat factors that might influence the density of grouse. Generally there was a higher density of drumming males in Area II than in Area I. We cover-mapped forest associations in both areas and also recorded tree stocking, age, and size. Indexes of productivity were (1) the percentage of juvenile birds in the fall harvest and (2) the average number of chicks per brood during the summer. In the latter we used dogs to find flying-age broods mostly in July and August and the first week of September. Broods were not counted in 1972 or 1981. We recorded summer temperatures in the 6-year period from 1968 to 1973 to evaluate the influence of temperature on chick survival. The juvenile/adult ratios were based on determining the age from wing and tail feathers collected by the B.C. Fish and Wildlife Branch at Cache Creek check station (Fig. 3.1). These were from birds killed by hunters in the wildlife management region in which the study was conducted. We determined the age and sex of birds by measuring tail and wing feathers as outlined by Dorney (1966) and Gullion (1964). The wing and tail sample in 1972 was destroyed before we determined the sex of birds, and most of the sample from 1976 was considered inaccurately measured and was thus disregarded. Davies made the wing and tail measurements in all the other years. We distinguished gray birds from red birds in all wing/tail samples and for the birds that we flushed in the summer. A red bird was defined as any grouse whose two central rectrices were red; a gray bird was any other grouse. Survival rate was measured by capturing and banding males on drumming logs
82
R. G. DAVIES AND A. T. BERGERUD
in May and returning to see if they were alive the following spring. Males were tagged for 7 years, from 1968 to 1974. Captured birds were weighed, and their age and color phases were noted.
3.2.2 Behavior indexes We were concerned with establishing indexes to aggressiveness. Other workers have noted that differences in mean levels of aggressiveness are correlated with demographic parameters (Jenkins et al. 1963, Mossop 1971, Moss et al. 1985). Our primary method to obtain measurements of aggressiveness was to present advertising males with their mirror image and record the hard-pecks and jumpattacks directed at the image. The procedure in 1970 and 1971 was to place a covered mirror on the drumming log adjacent to the drumming stage in the evening. In the morning, 0400 to 0700 hours, we would slip into a blind and pull the cover off the mirror with a drawstring. We tried to test all birds in Study Areas I and II in these 2 years. If a bird flushed while we were getting into the blind or pulling off" the cover, or if he failed to see his image, we termed the test unsuccessful and repeated it. A successful test was one in which the bird interacted with his image for at least 30 minutes. After a successful test, we attempted to trap and weigh the bird. Approximately 10 days after capture we retested the same individual. Nearly all tests were restricted to May. In addition, five males were tested six times in the same month to assess consistency in behavior related to their experience and the season. We again tried to measure hard-pecks and jump-attacks in 1973, after the population had declined, but the birds had become wary and only six birds were successfully tested in 143 attempts. To remedy this situation, we built automatic recording devices consisting of a mirror and tape recorder. As a bird approached his mirror image, he stepped on a treadle that activated the tape recorder. A microphone was taped to the back of the mirror. We could recognize from the tape recordings hard-pecks (no foot action), pecks with foot kicks (pecks with minor foot movement), peck thrusts and clashes (a hard thrust with or without wingbeats, but a major displacement of the feet), exploration pecking, and a number of vocal sounds —coos, chirps, or growls. An automatic recorder was placed on the log the night before the test, and the bird generally first saw its mirror image as it became light in the morning and when he was on the log. We watched and timed a sample of behavior sequences to better interpret the sounds. Birds were tested with automatic recording devices in 1974 and 1978. Our second method was to measure the distance at which a male ran or flushed as we approached him when he was near or on his drumming log. We recorded the color phase of the bird as it flew and checked this against the color-phase record of the same bird when captured. Birds that flushed away from display sites were singles, and were mostly nondi splay ing males or females.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
83
During the summer, we noted whether or not a hen with a brood "rushed" toward the observer, attemped to "lead" him away, or simply flushed. The intensity of these actions was rated as follows: 2 for vigorous action; 1 for a medium to weak response; and 0 for no response. The age of the brood was estimated according to the procedure outined by Mossop (1971).
3.3 Demographic findings 3.3.1 Fluctuations in numbers and frequency of color morphs The number of advertising males in Area I remained constant at 10 males in 1968, 1969, and 1970, and increased to 13 males in 1971. The population declined to only 6 drumming males in 1975, reached a second peak of 14 in 1979, and declined to 8 in 1982 (Fig. 3.2). As the population increased, birds established new advertising sites on logs not previously occupied (Fig. 3.3). Four sites, numbered 2, 3, 7/10, and 8, were termed "permanent" logs and were occupied in nearly all years. The density of advertising males was higher in Area II than in Area I. Area II's population also increased from 1970 to 1971 and then declined, from 28 birds in 1971 to 24 in 1972 and down to only 18 in 1973. This population followed the same sequence as did that of Area I, though we removed 5 males in May 1970 and 11 males in May 1971 (Fig. 3.2). The enlarged census area (484 ha, including Area I) had 43 advertising males in 1972; it declined to 22 in 1974, increased to a peak of 51 in 1979, and declined a second time to 20 males at the final count in May 1982. The morning drumming transect count along the Watch Lake and Bridge Lake roads showed considerable variation between counts within years. However, the annual maximum counts agreed with the trend indicated by the log-search method. A peak in the drumming count occurred in 1971 when 41 males were heard, a low occurred in 1974 when only 22 males were heard, and a second peak was reached in 1979 when 39 males were recorded as drumming. The drumming transect data further suggest that past peak populations, preceeding our intensive study period, occurred in 1967 and 1958 (Fig. 3.2). We conclude that the ruffed grouse population increased to a high in 1971, declined to a low in 1974, and reached a second high in 1979. The percentage of red birds in the population generally increased from 1968 to 1971-72 (Fig. 3.4). This increase was significant for juveniles and adult females in the harvest, 1968-71, and for red males advertising on logs, 1968-1972. We located 96 brood hens in 1970, of which 21 % were red, whereas in 1971 34% of 127 were red (P < 0.05). The percentage of red birds generally declined as the population decreased from 1971 to 1974-75 (Fig. 3.4, Table 3.1). The decline was most drastic for
84
R. G. DA VIES AND A. T. BERGERUD
red advertising males between May 1972, when 12 of 39 (31%) were red, and May 1973, when only 1 of 50 (2%) was red, a decline of 94% in 1 year. We excluded from the sample of red birds in 1973 three drumming males that had red bodies but gray central rectrices. There was also one red-bodied-gray-tailed drumming male in 1974 of the 23 advertising birds classified. The only other redbodied-gray-tailed birds we saw in 15 years were 10 juveniles in the 1973 harvest. As the population increased from the low in 1974, the percentage of red males advertising on logs slowly recovered from the 1972-73 crash, and showed a sig-
Fig. 3.2. Top: Trend in male population based on search of advertising locations. Bottom: Count of drumming males heard from listening posts.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
Fig. 3.3. Locations of advertising males in Study Area I, 1968-82.
85
86
R. G. DA VIES AND A. T. BERGERUD
Fig. 3.4. Fluctuations in the number of red-phase birds from 1968 to 1982 compared with the stage of population increase or decrease. Top: Percentage of red-phase juvenile birds in the fall harest. Middle: Percentage of red-phase adult females in the harvest compared with the percentage of red-phase juveniles and adults in the harvest the previous year. Bottom: Percentage of red-phase males advertising compared with the percentage of red-phase juveniles and adults in the harvest the previous fall.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
87
nificant positive regression on year (1973 to 1981) (Fig. 3.4). Red males in the harvest may also have increased. The percentage of red females showed no clear trend from 1975 to 1981 (Fig. 3.4). A second major decline of displaying red males occurred just as we terminated the study. Five red advertising males were classified on 23 sites in 1981 (22%) but we could find only one red male on 23 sites in 1982 (4%) (P < 0.05), a decline of 81 %. Actually there were no red males drumming in the census block in 1982 (20 males), but we continued our search until we found one advertising on the 23rd site.
Table 3.1. Population parameters of ruffed grouse at Watch Lake, 1968-82 Production
% red birds of total Sex ratio, in harvest (%) b
in harvest"
No. of No. of juveniles chicks per adult c per brood
Density /km 2
on Logs
Adult
Juvenile
Total
Adult
Juvenile
1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982
8.2 9.9 8.9 6.2 4.5 5.2 6.8 8.9 9.3 10.5 6.0 4.7 4.1
20 (10)d 20 (10) 22 (36) 27 (45) 31 (39) 2(50) 13 (23) 15 (27) 15(13) 19 (32) 21 (29) 23 (31) 25 (20) 23 (20) 4(23)
18 20 25 30 34 14f 21 22 17 18 21 24 -
12 11 16 29 21 26 15 18 21 17 19 35 26 -
14 13 19 21 25 18 17 19 19 18 19 24 26 -
54 53 53 64 63 65 65 60 59 56 62 59 61 -
55 52 56 52 57 50 46 56 57 49 51 48 -
2.9 5.8 3.5 1.7 2.1 1.9 2.1 2.3 2.3 2.7 1.5 1.3 2.0 -
4.1 (?) 4.9 (28)e 4.9 (41) 4.0 (68) 3.5 (25) 2.9 (22) 3.3 (20) 2.8 (10) 3.4(12) 4.2 (9) 2.2 (5) 2.0 (2) -
Unweighted means
7.2
18.7
21.0
20.0
20.0
60.0
52.0
2.5
3.5
Year
a
b c
d e
f
Percent of red birds in harvest: adults, n = 1,049; smallest sample of adults 1980, n = 17; juveniles, n = 1,926; smallest sample of juveniles 1980, n — 20. Sex ratios: adults, n = 813; juveniles, n = 1,753. Juveniles per adult: n = 3,149. Color of male determined for 10 advertising males. 28 broods observed. 4% red males (n = 26) but 31% red females (n = 16).
88
R. G. DAVIES AND A. T. BERGERUD
3.3.2 Survival of juveniles and adults Survival of banded advertising males from one drumming season to the next averaged 51% from 1968 to 1974 (n = 77; Table 3.2). In those 7 years, juveniles on logs accounted for 83 of 157 birds, or 52% of the population. The survival rate was significantly less in 1972-73; only 28% of the males lived between these springs (P < 0.05; Table 3.2). During this year we saw the 94% decline in the red-phase birds (Fig. 3.4). Mortality of banded birds from May 1972 to May 1973 was 100% for six red males but only 58% for 12 gray males. Annual survival rates (1968-74) were not correlated with the spring body weights of males nor with winter snow depths (Table 3.2). Harvest data also suggested differences in the survival of red and gray males in some years. The total percentage of red juveniles in the harvest was 19% in 13 years (171/913), similar to that of red adults for all years, 20 % (92/467). However, during the increase of 1968-70 there was a significant increase in the percentage of red males between the fall harvest (18%) and red males advertising the next year on logs (24%), and red adults in the harvest (26%) (Table 3.3, Fig. 3.4). During the decline of 1971-73, there was a decrease in the percentage of red males, between the fall harvest 1971-73 (21 %) and adults and yearlings advertising on logs in subsequent years 1972-74 (14%), and in the bag 1972-74
Table 3.2. Snow depths, weights, and percentages of yearling males advertising in relation to May-to-May survival of advertising males Spring weights of males (g)b
Period
Mean monthly end-of-iiioiilli snow depth (cm)a
1967-68 1968-69 1969-70 1970-71 1971-72 1972-73 1973-74
12 43 14 47 40 15 45
533 537 548 541 555 540 542
Unweighted means
31
542
a
b c
d
Adults c ± ± ± ± ± ± ±
10 (7)d 10 (8) 7 (7) 8 (8) 11 (7) 6 (18) 10 (6)
% advertising males
Juveniles 0 516 523 532 525 510 526 515
521
± ± ± ± ± ± ±
11 (4) 13 (5) 1 (13) 4 (22) 8 (5) 5 (18) 18(10)
That are yearlings
That survived
36 (11) 33 (9) 62 (21) 73 (41) 36 (14) 48 (42) 42 (19)
55(11) 44 (9) 53 (15) 62 (13) 28 (18) 64 (11)
47
51
End-of-month snow depth at Williams Lake summed for November to March and divided by five. Includes only birds trapped and not those shot. Weights in May 1968. Sample size in parentheses.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
89
Table 3.3. Changes in the frequency of the red color morph with phases of population growth % red-phase birds of total
Type of bird
Increase (1968-70)3
Females J + A in fall harvest (Yi) Adults in harvest (Yz) With brood Y2 Males J + A in fall harvest (Yi) Adults in harvest (Yz) Advertising Yz a
b c
Decrease (1971-73)
Increase (1974-78)
Decrease (1979-81)
18 (375) 31 (130) 29 (200)
30 (233)c 32 (25) 13 (89)
20 (355) 24 (125) 16 (51)
22 (197) 23 (26) no data
18 (469) 26 (185) 24 (91)
21 (289) 9 (43) 14(112)
17 (389) 16 (191) 19(132)
22 (233) 20 (40) 15 (65)
b
Harvest year, total 1968 + 1969 + 1970 (Year 1) compared with adults in harvest and with broods, total 1969 + 1970 + 1971 (Year 2). Lines connect significantly different percentages. 233 birds in sample.
(9%) (Table 3.3). Hence in two fluctuations grays disappeared more during an increase than did reds but survived better during a decrease in numbers. No females were banded so we could not determine their mortality rates. However, the sex ratio of adults in the harvest often was weighted toward males more than was the sex ratio of juveniles (Fig. 3.5). These data indicate that generally more females than males died between hunting seasons (Table 3.1). For all years, the percentage of juvenile females was 47% (815/1,752), and the percentage of adult females was 41 % (330/813) (P < 0.05). Females survived as well as males in 1968 and 1969. The percentage of juvenile females was 47% (n = 90) in 1968 and 1969, and the percentage of adult females in the falls of 1969 and 1970 (includes 1968 and 1969 juveniles) was 47% (n = 192). Heavy differential loss of females first occurred in 1970-71 as the population reached a peak. Juvenile and adult females made up 45% of the 1970 fall harvest (n = 144, P < 0.05). Females continued to die at higher rates than did males for the remainder of the study. The percentage of juvenile females from 1973 to 1981 was 48% (n — 860) whereas for adults it was 39% (n = 464, P < 0.05). The percentage of females declined the most when the weather was cold in June (Fig. 3.6). This suggests that the greater mortality of females was associated more with nesting and the early brood season than with factors related to the overwinter period. There were also differences in mortality rate between red and gray hens. The total percentage of juvenile hens that were red in the autumn for 13 years (1972 missing) was 20% (167/841), whereas the percentage of adult hens that were red
90
R. G. DA VIES AND A. T. BERGERUD
was 27% (85/310; P < 0.05). Gray hens probably had significantly greater mortality than did red hens during the period of increasing numbers, 1968-70 (Table 3.3, Fig. 3.4). The greatest differential loss between color morphs probably occurred between 1970 and 1971 as the population reached a high and the percentage of females declined (Table 3.1). The percentage of adults plus juvenile females in the 1970 harvest that were red was 19% (n = 324), but the percentage of adult females in the 1971 harvest that were red was 37% (n = 52, P < 0.05) (Fig. 3.4). In later years when more females than males died, there was little change in the proportion of adult plus juvenile red hens in the fall and that of adults the next year (Table 3.3). Apparently in less favorable years, when more females died, the survival of red and gray hens was quite similar.
Fig. 3.5. Top: Annual changes in productivity based on the number of juveniles killed per adult in the harvest, and the mean number of chicks per brood seen in August and September. Bottom: Percentage of adults that were males compared with percentage of juveniles that were males in the harvest, 1968-81.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
91
Fig. 3.6. Top: Differences in the percentage of juvenile females (year 1) and adult females (year 2) in the harvest compared with the mean maximum June temperatures — e.g., 47 % adult females in 1969 minus 45% juvenile females in 1968 equals +2, and June 1969 the mean maximum temperature was 26°C. Bottom: Regression of the percentage of adult females in the harvest and the mean maximum June temperature °C (temperatures from Watch Lake, 1968-73, and Williams Lake, 1974-81, corrected on the basis of Watch Lake temperatures).
92
R. G. DA VIES AND A. T. BERGERUD
3.3.3 Productivity Each spring most chicks hatched in the first and second weeks of June. The earliest year was 1978, when the mean date of hatch was 6 June, and the latest year was 1975, when the average was 17 June (Fig. 3.7). The mean date of hatch was weakly correlated with the end-of-March snow depth at Williams Lake (r — 0.577; n = 11, excluding 1976). In the earliest year, 1978, there was no snow on 31 March, and in 1975 the snow depth on 31 March was 43 cm. The hatch in 1976 was early, 7 June, but there was reported 33 cm of snow on the ground at the end of March. Unfortunately, Williams Lake, 93 km from our study area, was the nearest station that kept continuous records of snow depths. The survival of chicks was higher in early-hatch years than in late-hatch years (Fig. 3.7; P < 0.05, one-tailed test). Years with the highest juvenile/adult ratios in the fall were those when approximately 40% of the chicks hatched in a 1-week period in the first (1978) or the second week of June (1969, 1970). Poor chick survival occurred when many chicks hatched in the third week of June or later, and when the total hatch was spread over several weeks (Fig. 3.7). Hatching histograms suggest that significant renesting may have occurred in 1968 and 1974 (Fig. 3.7). We have no data on the percentage of hens that successfully hatched their clutches. The only index of nesting success was the ratio of females with broods versus the number of single adults seen in 1970, 1971, 1973, and 1974. These single adults could be males or unsuccessful hens. We used the fall ratio of males in the harvest to estimate the proportion of singles that were males. Singles unaccounted for by the number were called unsuccessful females. This ratio may be seriously biased if unsuccessful hens seek denser cover than do successful females (Maxson 1974). Estimated nesting success was high, exceeding 60% in all 4 years. Nesting success appeared density-dependent, with reduced success when numbers were high in 1970 and 1971. When the population declined in 1973 and 1974, nesting success was higher (Table 3.4). Red hens appeared less successful than gray hens in several years. The percentage of adult hens that were red in the harvest from 1973 to 1979 was 25% (n = 150), but of the females that had broods in August-September from 1973 to 1979, only 14% (n = 140) were red (Table 3.3; P < 0.05). The regression of juveniles/adult ratios in the harvest on the mean brood sizes in August-September was highly significant for 11 years (r = 0.773). However, in 1971 there were significantly fewer juveniles in the harvest than predicted from the regression of the juveniles/adult ratios on brood size (Table 3.1). The regression (Y = -1.0605 + 1.0067 X) gave a predicted value of 3.0 juveniles/adult, but the actual ratio was 1.7 (Table 3.1). The adult sex ratio of the population changed markedly between 1970 and 1971 (Table 3.1). The sex ratio (juveniles and adults combined) in fall 1970 was 55 males:45 females (n = 752). The ratio
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
93
in fall 1971 was 64:36 for the adults (n = 144). We may have failed to see many unsuccessful females in some years because in fact these hens and their chicks had died. Generally the sex ratio difference between juveniles and adults in the harvest was correlated with the June maximum temperatures (Fig. 3.6). This correlation further suggests that increased mortality of females occurred in colder Junes.
Fig. 3.7. Hatching histograms for 12 years comparing early and late nesting years. Early years resulted in significantly greater proportions of young in the fall harvest (t = 1.95, P < 0.10).
94
R. G. DAVIES AND A. T. BERGERUD
Table 3.4. Estimated nest success based on single adults, sex unknown, seen during high, peak, and low populations. Population size for month and year
Total aggregations seen
Adult males in Calculated3 harvest (%) nesting success
Broods
Singles
High population June 1970 July 1970
35 25
61 64
53 (177) b 53 (177)
0.78 0.60
Peak population June 1971 July 1971
28 31
97 112
64 (144) 64 (144)
0.62 0.60
Low population June-July 1973 July 1974
33 16
54 22
63 (49) 65 (26)
1.00 1.00
a
b
For example, for June 1970, 94 males of 177 adults harvested in fall = 0.53: 0.53 x (35 + 61) = 51 males expected in 35 + 61 aggregations; (61 - 51 males)/(10 + 35) = 0.22; 1.00 - 0.22 = 0.78 nest success. Sample size in parentheses.
The two annual indexes of production, the juveniles/adult ratio in the harvest and the mean August-September brood size, were also correlated with the mean maximum temperature in June (Fig. 3.8). There was no correlation between breeding success or the sex ratio and mean May temperatures as reported by Dorney and Kabat (1960) for ruffed grouse indexes in Wisconsin. In 1970 and 1971 there were sufficient brood classifications to compare the breeding success of red and gray hens. Gray hens raised significantly more chicks than did red hens in 1971. The mean brood size for gray hens, in August-September was 4.3 ± 0.15 (n = 47); for red hens, 3.3 ± 0.21 (n = 21; t = 3.99, P < 0.01). In 1970 we saw only 17 red hens, but there was a significantly more rapid decline in the number of chicks in broods of red hens during the summer than in those of gray hens (Table 3.5). There were also significantly more chicks with gray hens than with red hens when we combined all broods from 1973 to 1980 (Table 3.5). Gray hens commonly raised about 0.8 more chicks per brood from 1970 to 1980 than did red hens. Red hens also either nested less frequently than gray hens or lost their clutches more often. The percentage of red hens seen with broods was 19.2 + 2.40% (n — 8 years 1970-78 and 335 broods, see Table 3.5), whereas the percentage of red adult hens in the harvest for the same years was 28.0 + 1.96%. In each of the 8 years there were more red hens in the harvest than were seen with young, despite the fact that red hens were more easily seen in distraction displays (Sec. 3.4.2).
95
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE Table 3.5. Survival of chicks of gray hens compared with that of red hens. Mean chicks per brood Gray hens
Year and month
Difference gray > red
Red hens
1970 June July August September
7.6 5.8 4.8 5.5
± ± ± ±
0.38 0.36 0.33 0.60
(ll)a (20) (20) (11)
7.0 5.8 4.0 4.3
+ ± ± ±
0.91 0.86 0.90 0.50
(4) (5) (4) (4)
0.6 same 0.8 1.2
1971 June July August September
6.6 5.4 4.4 4.2
± ± ± ±
0.30 0.36 0.19 0.30
(16) (18) (34) (13)
5.1 4.3 3.3 3.4
± ± ± ±
0.63 (7) 0.50 (12) 0.24 (16) 0.50 (5)
1.5 1.1 1.1 0.8
1973 June July August September
6.0 5.9 4.5 3.5
± ± ± ±
0.58 (10) 0.65 (18) 0.50 (2) 0.35 (17)
4.5 ± 2.5 4.7 ± 1.45 4.0 3.5 ± 0.50
(2) (3) (1) (2)
1.5 1.2 0.5 same
1974 July September
3.1 ± 0.38 (13) 3.5 ± 0.35 (17)
1.5 ± 0.50 3.5 ± 0.50
(2) (2)
1.6 same
1975 August-September
3.3 ± 0.50 (13)
3.0 ± 1.0
(2)
0.3
1976 September
3.4 ± 0.57
(8)
2.0 ± 1.0
(2)
1.4
1977 September
3.5 ± 0.43 (10)
3.0 ± 1.0
(2)
0.5
1978 September
4.4 ± 0.37
(7)
3.5 ± 1.5
(2)
0.9
1979b September
2.2 ± 0.58
(5)
-
1980 September
2.0
(1)
a
b
Number of broods in parentheses. No broods with red hens observed in 1979.
2.0
-
(1)
same
96
R. G. DAVIES AND A. T. BERGERUD
On our study area, annual changes in production of young was a major factor explaining annual changes in spring numbers of advertising males (Fig. 3.9). The population the next spring could be expected to increase when the juveniles/adult ratio in the harvest exceeded 2.3. By contrast, brood sizes were not well cor-
Fig. 3.8. Regression of juveniles harvested per adult and mean number of chicks seen in broods on mean maximum temperature in June.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
97
Fig. 3.9. Ratio of population change in advertising males between year 1 and year 2 compared with annual productivity indexes.
98
R. G. DA VIES AND A. T. BERGERUD
related with subsequent changes in spring numbers of males (Fig. 3.9). Again, this suggests large annual differences in the losses of clutches or of entire broods.
3.4 Aggressiveness and approachability of birds Our interpretation of the hypothesis of self-regulation of numbers (Chitty 1967) is that genotypes should vary in aggressiveness during population fluctuations. Aggressive genotypes should be favored when a population is high and space is limited, whereas more docile genotypes should have the selective advantage in raising viable young when a population is low and space is not limited. Our indexes to behavior include the pecking and attacking by males of their mirror images at advertising sites, flushing distance of males near logs, catchability and testability of advertising males, and intensity of distraction behavior of females with broods. The majority of data was gathered in 1970 and 1971, when the ruffed grouse population was at a high density (Fig. 3.2).
3.4.1 Individual differences There were intrinsic differences in the aggressiveness of males. Five males were tested in the same year with their mirror images on six occasions in May. Each pecked and attacked his image in the six tests with a frequency more consistent than expected from the variability documented between males (Table 3.6). Individual birds that most frequently pecked their mirror images also attacked them more often (Fig. 3.10). Further, 40 males that were tested before capture and again after capture and banding showed no significant difference between the two tests in peck frequency; they averaged 507 + 36 pecks per hour before capture and 474 + 34 pecks per hour after capture. Repeat tests within years include a time variability. The repeat tests were conducted later in the month of May. We considered 1 May the time of peak advertising; thus testing generally began at that time. Therefore, birds tested a second time after capture, or tested repeatedly, were measured at times further removed from the maximum display date. Still, birds tested repeatedly showed no decline in the intensity of interaction with their mirror images. Another measurement of the consistency in behavior between individuals resulted when we tested the same birds using the same logs in both 1970 and 1971, compared with different birds using the same logs in 1970 and 1971. The same birds tested in both years showed a consistent behavior pattern between years, whereas scores from the different birds using the same log in each year showed significantly greater variability in aggressiveness (Table 3.7). The aggressiveness of males when they fought their mirror images was also correlated with approachability; the more aggressive males had shorter flushing distances as we approached them near their advertising sites (Fig. 3.11). Simi-
99
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
Table 3.6. Behavior of five males in six repeat mirror-tests conducted in May 1970a Test number Bird # Behavior
1 33 24 2 5
a
Pecks/hr Attacks/hr Pecks/hr Attacks/hr Pecks/hr Attacks/hr Pecks/hr Attacks/hr Pecks/hr Attacks/hr
1
2
3
4
5
6
178 2 308 1 617 4 304 3 381 3
171 1 352 3 629 6 316 1 364 6
186 1 301 3 616 2 281 3 398 4
180 3 312 2 600 4 300 2 351 2
161 1 306 1 564 2 308 2 359 4
177 1 282 4 601 5 284 3 372 3
Statistical tests—Pecks/hr: between days F = 1.440. Attack/hr: between daysF = 1.106.
F = 757.224, same F = 3.226, same
Means 176 1.5 310 2.3 605 3.8 299 2.3 371 3.7
+ ± ± + ± ±
± ± ± ±
3.50 0.34 9.40 0.49 9.65 0.65 5.60 0.33 6.90 0.56
between between
Table 3.7. Comparison of mean number of pecks per hour between the same and different males occupying the same advertising sites in both 1970 and 1971 Color and agea
Pecks per hour
Area and log
1970
1971
1970
1971
1971 > 1970
Different birds Area I — log 2 Area I — log 3 Area I — log 1 1 Area II — log 3 Area II — log 7 Area II — log 8 Area II - log 10
G/J a R/J G/J G/J G/A G/J G/A
G/J G/J G/J R/J R/H G/J G/J
347 199 617 278 200 32 637
611 46 475 687 742 40 16
+264 -153 -142 +409 + 542 + 8 -621
330
374
306"
697 317 331
792 276 351
+ 95 - 41 + 20
448
473
52b
Means Same birds Area I - l o g 5 Area I - Iog8 Area II - log 1
R/A G/J G/A
Means a b
G = gray; R = red; J = juvenile; A = adult. Sign ignored.
R/A G/A G/A
Difference
100
R. G. DA VIES AND A. T. BERGERUD
larly, success in having birds remain on the logs while we entered the blind and pulled the cover, increased with mirror scores of these birds (Fig. 3.12). We have sufficient data on the consistency of behavior of these birds to reject the view of McNicholl (1978, 1979) that variability in behavior of individuals exceeds that between individuals. Some males in our study areas were clearly more aggressive than others.
3.4.2 Basis of the differences in behavior Intrinsic differences could be related to age, physical conditions, experience, or genotypes of individuals. There was no significant difference in the mirror score
Fig. 3.10. Correlation between attacks per hour and pecks per hour for combined 1970 and 1971 data.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
101
of juveniles compared with that of adults, nor between birds classified as large, greater than 536 g, and those classified as small, less than 536 g, in 1970 or 1971 (Table 3.8). Mean scores of pecks per hour of red-phase birds were generally greater than those of gray males in all years, and significantly so in 1971 and 1974, when our sample of red birds consisted of only 7 and 5 individuals respectively (Table 3.8). Red-phase hens with broods responded more vigorously than gray hens in distraction behavior, especially in June. Red hens rushed and led more often than gray hens, which more frequently flushed and left their chicks (Fig. 3.13). The
Fig. 3.11. Correlation between pecks per hour and flushing distances for advertising males in Study Area I and Study Area II, combined 1970 and 1971. Flushing distances of gray males are normally distributed; those of red males were skewed to the shorter distances. All 51 males are different males.
102
R. G. DA VIES AND A. T. BERGERUD
Fig. 3.12. Attacks and peck per hour of individual birds compared with the percentage of unsuccessful attempts to test these birds in 1970 and 1971.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
103
Table 3.8. Peck and attack frequency for the mirror tests compared among color phases, years, areas, ages, weights, and time of test.
Variable Year and color 1970 Red Gray 1971 Red Gray 1973 Red Gray 1974 Red Gray 1978 Red Gray
Mean pecks per hour (n)
Mean attacks per hour (n)
410 + 149 (3)a 356 + 53 (17)
5.3 ± 0.67 5.8 ± 0.66
(3) (17)
590 + 70 353 ± 49
(7) (24)
5.7 + 0.64 3.3 ± 0.38
(7) (24)
582 + 230 370 ± 125
(2) (4)
16 6.0 ± 3.8
(D (4)
235 ± 105 (5) 97 ± 29 (12)
-
306 + 31 205 ± 65
(2) (10)
-
479 ± 65 270 ± 60
(9) (11)
6.4 ± 0.82 (9) 5.1 ± 0.76 (11)
509 ± 74 341 + 51
(12) (19)
4.4 ± 0.73 3.4 ± 0.39
(12) (19)
Age (1970 & 71) Juvenile Adult
352 ± 38 489 ± 62
(37) (13)
4.4 ± 0.40 5.0 ± 0.70
(37) (13)
Weight (1970&71) High > 536 Low < 536
411 + 42 371 ± 53
(25) (25)
4.8 ± 0.44 4.2 ± 0.53
(25) (25)
Capture (1970 & 71) Before After
507 + 36 474 ± 34
(40) (40)
5.3 ± 0.46 4.8 ± 0.42
(40) (40)
Year and area 1970 Area I Area II 1971 Area I Area II
NOTE: Probability of F values in two 4-way ANOVAS. First ANOVA: (a) year (1970 vs. 1971) P < 0.801; (b) age (adult vs. juvenile) P < 0.353; (c) weight (high vs. low) P < 0.751; (d) Area I vs. II P < 0.0082. Second ANOVA: (a) year P < 0.663; (b) area P < 0.012; (c) color (red vs. gray) P < 0.014; (d) age P < 0.293. a Sample size in parentheses.
104
R. G. DA VIES AND A. T. BERGERUD
Fig. 3.13. Comparison of brood defense behavior of hens between color phases, months, and years.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
105
statistical probability of differences between color phases was 0.122 for rushing, 0.023 for flushing, and 0.059 for leading (Davies 1973). Because red males and females were more approachable than gray, we reasoned that the vulnerability of red birds to hunting should be greater. The only years with large samples of birds killed by hunters and also a high proportion of red birds in the population were 1970 and 1971. In both years, the percentages of juveniles in the harvest that were red increased as the hunting season progressed (Fig. 3.14). The hunting season generally opened in the first week of September when birds were still in broods. Young males probably dispersed from broods in mid-September and females later, in October (Eng 1959). The increase in red-phase birds harvested thus coincided with increased mobility of juveniles. Also, individual vulnerability may have been masked when juveniles flushed with siblings in broods during the early part of the hunting season. Thus, red juveniles, like adults, appear to have a greater tolerance for being approached than do gray birds. Our data do not support the view of a discontinuous polymorphism between the behavior of red and gray color morphs. All nine males with the lowest rates of pecking at the mirror were gray, but the two males who fought the mirror the hardest were also gray (see Fig. 3.10). There was a continuum of aggressive behavior; hence, behavior was quantitative, not Mendelian, in character. Red males were scattered along this normal distribution, but their mean scores were skewed statistically to the aggressive end of the scale; thus the color marker was useful in distinguishing intrinsic differences in behavior based on a genetic marker in color. 3.4.3 Behavioral differences in space and time The aggressiveness of males varied, not only with their color phase, but also between study areas within years (1970 and 1971). Males in Study Area II fought their mirror images significantly less than did males in Study Area I, and also had longer flushing distances (Table 3.8, Fig. 3.15). Critical to the Chitty hypothesis is that changes in aggressiveness occur between years with phase of population growth and are caused by differences in the frequency of aggressive and nonaggressive genotypes. Our major index to aggressiveness was the annual mirror test scores, but there were two serious problems. First, in 1973 we found that we could obtain test scores on only 16% of the males by the methods used in 1970 and 1971. In 1970 we had tested 59% of the males (20/34), and in 1971, 81% (30/37). This led to the second problem: if we could test only a few birds, we were likely to test the most approachable, and thus the most aggressive. We changed our measuring technique in 1974 by building an automatic recorder. This increased the success rate in 1974 to 68% (17/25), but the success rate again declined in 1978 to only 33% of the birds in
106
R. G. DAVIES AND A. T. BERGERUD
Fig. 3.14. Regression of percent juvenile and adult red-phase grouse on days of the fall hunting season in 1970 and 1971. Weekly samples obtained from the British Columbia Fish and Wildlife check station at Cache Creek, B.C.
the census area. Thus, in the mirror tests, we had considerable experimental variability. The two methods were dissimilar, and different proportions of the birds were left untested. The only significant difference in mirror scores between years was that using automatic recorders in 1974 and 1978. The birds in 1978, near the population high, may have been more aggressive than the birds at the low population in 1974 (Table 3.8). Birds were definitely wilder in 1973 than in 1970 and 1971. This is reflected in the percentage of successful tests, flushing distances, and the trapnights necessary to capture birds, using mirror traps (Table 3.9). Davies did all the trapping in 1970, 1971, and 1973; if trappability of the birds remained stable, his success rate should have increased as he gained experience, but instead it
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
107
Fig. 3.15. Comparison of flushing distances of males between color phases, areas, and years, and between aspen and aspen/conifer cover. Years 1 and 2 are 1970 and 1971. Large circles indicate number of males per study, area. Shaded portion of circles represents percentage that are red-phase. Probability of differences between Area I and Area II in aspen F = 43.361, P < 0.001, and color phase F = 35.165, P < 0.001; between Area I and Area II in aspen conifer F = 4.984, P < 0.033.
108
R. G. DAVIES AND A. T. BERGERUD
Table 3.9. Three behavior-related indexes reflecting changes in grouse population in 1970-73
Year
1970 1971 1972 1973 a
Point in population cycle One year before high High One year after high One year after crash
% of mirror tests unsuccessful
Trap days per bird captured
Flushing distance of log birds (m)
80 (160)a 73 (256) 96 (143)
4.0 (93) 3.2 (77) 1.8 (24) 6.9 (188)
14 (34) 13(41) 19 (37) 28 (26)
Number of tests in parentheses.
declined. This agreed with our subjective assessment that the birds were considerably wilder in 1973 than in 1970 and 1971. The best index of annual changes in behavior was flushing distances. Birds flushed at shorter distances when the population was high in 1971 and 1979, and at longer distances when the population was low in 1974 and 1982 (Fig. 3.16). Flushing distances of males at logs were correlated with spring numbers and with population trends (Figs. 3.2, 3.16). The correlation between flushing distances and the numbers of birds advertising from the high in 1971 to the low in 1974 was r — - 0.994 (n = 4); from the 1974 low to the next high it was r — - 0.904 (n = 6). Single birds, most of which were probably females, also flushed at shorter distances when the population was high than when it was low (Fig. 3.16). Red males flushed at significantly shorter distances than gray males from 1970 to 1976, and both color phases showed the same increase and decrease pattern through time. Both color phases tracked a cycle in behavior. Approachability (flushing distances) and aggressiveness were correlated in individual males (Fig. 3.11). Hence, changes in flushing distances can also serve as an index of changes in aggressiveness. Birds were more aggressive when the population was high in 1971 and 1979, than in the adjacent years- 1970, 1972, 1980, and 1981. These annual changes in behavior were correlated with percentages of advertising males that were red-phase (Fig. 3.16), consistent with the view that our color marker was useful in distinguishing changes in spacing behavior.
3.5 Discussion This study provides insight into the mechanics of population change and the causal extrinsic factors. In addition, we have evidence of intrinsic differences in behavior correlated with a genetic marker. Here we comment on the influence of the extrinsic environment on population numbers and the behavioral differ-
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
109
Fig. 3.16. Top: Changes in flushing distances of males and single birds (sex unknown) between increasing and decreasing phases of population growth. Middle: Flushing distances of red and gray males. Note that both color morphs track the cycle. Bottom: Annual percentage of advertising population that consisted of red males.
110
R. G. DAVIES AND A. T. BERGERUD
ences between individuals that may affect breeding densities and moderate the impact of extrinsic environmental factors as they relate to numerical fluctuations.
3.5.1 Role of the extrinsic environment in population change There were annual changes in the survival of young till autumn; juveniles/adult ratios were correlated with changes in the number of advertising males between years (Fig. 3.9). Generally, breeding males in the population increased if there were more than 2.3 young/adult in the fall harvest. Survival of August young explained 50% of the changes in spring numbers (r2 = 0.457). Mean brood sizes were not significantly correlated with population changes (Fig. 3.9). These indexes, unlike fall age ratios, did not account for changes in adult mortality or nesting success, which affect spring numbers. There were more hens without chicks when the population was high in 1970 and 1971 than when the population was low in 1973 and 1974 (Table 3.4). Adult mortality also varied between years. Survival of chicks was higher in years when the weather was warm in June, the time when the chicks were only 2-3 weeks old. Similarly, Ritcey and Edwards (1963) noted that a ruffed grouse population in Wells Grey Park, B.C., increased coincident with warmer June temperatures. Warm springs probably resulted in favorable insect supplies. Several studies of grouse have now shown that growth is improved with warm springs (Mercer 1967, Myrberget et al. 1977, Redfield 1978, Erikstad 1982). Juvenile ruffed grouse in Alberta grew the slowest in 1968, when they hatched the latest in 7 years (Rusch et al. 1984). Late hatches should occur in late phenology years that should reduce the quality of food, resulting in increased energy expenditure, reduced growth, and increased mortality (Jorgensen & Blix 1985). Chick survival was not correlated with mean or mean-maximum temperatures in May. Dorney and Kabat (1960), however, reported that reduced percentages of juveniles and fewer hens in the harvest occurred when the weather was cold in May, the time when hens in Wisconsin were incubating. They suggested that females may have died during incubation from stress, yet their data show a strong correlation between the size of broods and the relative numbers of juveniles shot by hunters (r = 0.855). Thus weather factors after hatch are an alternative explanation for annual differences in the survival rates of chicks they saw in Wisconsin. Changing adult mortality rates may account for the remaining variability in the change in breeding numbers not explained by annual production indexes. The population at Watch Lake declined from 43 to 30 males from 1972 to 1973. This was a significant departure from the regression of population change on breeding success (Fig. 3.9). Between May 1972 and May 1973, only 28% of our banded males lived; this was the lowest survival rate in 6 years (Table 3.2). In this winter, mortality appeared to more important than production in altering spring numbers.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
111
Moreover, there was considerable variation in the survival of hens between years, based on the percentage of females in the juvenile population (approximately 48 % females) compared with the percentage of females in the adult population (40% females). The years of greatest hen loss were 1974-75, 1975-76, and 1978-79 (Fig. 3.6). These variable mortalities should affect population status. The annual harvest of ruffed grouse in British Columbia since 1950 showed approximately 10-year fluctuations in abundance (Fig. 3.17). Major peaks occurred in 1958, 1969, and 1978. These peaks were correlated with high mean maximum temperatures in June recorded at Williams Lake (Fig. 3.17). This suggests that population peaks resulted from changes in production, rather than from a 10-year cycle in the spring density of adults. We estimated the proportion of adults and young in the reported total kill (1950 to 1981) by: (1) taking the mean maximum June temperatures from Williams Lake and regressing them on the mean maximum temperatures recorded at Watch Lake in 6 years, from 1968 to 1973, and (2) using the corrected Williams Lake temperatures to generate expected juveniles/adult ratios in the harvest, based on regression of juveniles/adult ratios on 13 June temperatures (Fig. 3.8). The resultant partitioning of the total kill into juveniles and adults resulted in peaks in juvenile production coincident with peaks in total grouse (Fig. 3.17). The annual plot of adult numbers showed little evidence of a 10-year cycle of abundance. Peak adult numbers occurred in 1957-58, 1971, and 1979 (Fig. 3.17). These peaks agreed with annual spring census results (Fig. 3.2) and suggested that constructed juvenile and adult totals had reality with field observations. It seemed evident that fall populations and hunting success could be predicted based on June temperatures and that production was more important than breeding numbers in determining fall numbers. Next we compared the annual changes in our calculated adult harvest totals (1950 to 1981) with the expected values based on those predicted from the constructed juveniles/adult proportions in the populations. In a numbers of years the calculated adult population increased even though productivity the previous year was well below 2.3 young/adult (Fig. 3.9). These years were 1955-56, 1956-57, 1966-67, 1975-76, and 1976-77 (Fig. 3.17). There were also years when the adult segment declined more than expected, based on production in the previous year, as in 1951-52, 1958-59, 1959-60, 1961-62, 1962-63, and 1972-73. Thus 11 of 30 years deviated from what we would expect, based on production alone. These were years when adult mortality should have varied considerably from the average. Our hypothesis is that changes in the number of adults that could not be accounted for by production resulted from variable adult mortality caused by predators "switching" to grouse after hares (Lepus americanus) declined (cf. Keith 1974). Our field notes indicated that snowshoe hares reached high numbers in 1970 and 1971. We saw no hares in 1972 during the spring grouse census. Hares
112
R. G. DAVIES AND A. T. BERGERUD
were again common in 1979 and 1980, but we saw none in spring 1982. Goshawks were unusually common on the study area from 1971 to 1973, and we noted a large number of grouse killed. Goshawks again returned in 1981. Goshawks were generally common in B.C. in 1971-72, 1975, and 1981 (B.C. Provincial Mus. unpubl. data). Our grouse population rapidly declined in 1972-73 and 1981-82 during the winter, coincident with the reduced hare population and the increase in goshawks. Males banded in 1972 had a significantly greater mortality than in 5 other years (Table 3.2). If hare peaks occurred before our intensive study, at 10-year intervals, there would have been peaks in hares about 1950-51 and 1960-61. Thus the years of unexpected increases in adults (Fig. 3.17) occurred midway between presumed peaks of hares, and the years of unexpected decline in spring numbers occurred after hares had presumably crashed (Fig. 3.17).
Fig. 3.17. Total estimated kill of ruffed grouse for northern British Columbia compared to June temperatures for Williams Lake. Total harvest calculated from statistics from Cache Creek station. The proportion of adults is calculated from Fig. 3.8.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
113
In summary, results support a hypothesis that approximately 50% of the population change was caused by spring weather and chick survival, and 50% by changing, adult survival rates than may result from predation. Ruffed grouse use snow as cover to escape goshawks (Chap. 4). Adult mortality rates should be higher in winters with only enough snow to make brown grouse conspicuous, yet not enough for escape plunges from goshawks and owls. A year with inadequate snow and abundant goshawks should result in high mortality rates. Red males declined 96% in the winter 1972-73, but only 81 % in 1981-82. Drumming males declined from 43 to 30 (30%) in 1972-73 but only 23 to 20 (13%) in 1981-82. Prior chick production was approximately equal in both 1972 and 1981 (2.1 compared to 2.0 juveniles/adult). The greater decline occurred in the winter of 1972-73, which had the least snow, consistent with the explanation of the need for snow by ruffed grouse to escape goshawks. Adult mortality rates showed some correlation with color phenotypes. In the winter of 1972-73 we lost nearly all of our red-phase birds (Fig. 3.4). All six banded, red males disappeared, but only five of 12 gray males died. Red-phase males also disappeared between May 1981 and spring 1982. These birds disappeared between fall and spring, because red males were common in the harvest in both 1972 (25%, n = 218) and 1981 (27%, n = 73). Also, there were few red males in the harvest in 1973 (4%, n = 26), which counters the argument that red males were still alive but had simply stopped advertising. We have documented that red males usually have a shorter flushing distance than gray males (Figs. 3.15, 3.16). Red males in 1972 and 1981 flushed at approximately 15m, whereas gray males flushed at 20-23 m. We have also shown that red-phase birds are more vulnerable to hunting (Fig. 3.14). We hypothesize that red males are more susceptible to predation than are gray males. Data are insufficient to determine what proportion of susceptibility is caused by behavior versus color. The short flushing of red birds should increase their vulnerability in the absence of a cryptic background. In the two winters that reds disappeared (1973 and 1982), no snow was on the ground at the end of March. In some years the percentage of red adult hens in the harvest increased over that expected based on the number of red juveniles plus adult hens in the prior autumn. This was especially true in increasing years 1968-69, 1969-70, 1970-71, and 1977-78 (Fig. 3.4; Table 3.3, 3.10). In the years that red hens were most abundant, mostly the decrease years (1971-74), the sex ratio of adults was commonly 64:36, yet the sex ratio of juveniles was more equal, 53:47 (Table 3.1). Furthermore, the proportion of adult females increased as the mean maximum June temperatures increased (Fig. 3.6). Thus, in warm springs there were generally more females, and of those that were present, a greater percentage was red than expected from the color-phase ratios found in juvenile birds (Table 3.10). The percentage of red juveniles in the harvest increased subsequent to low temperatures in June and reduced production indexes (Fig. 3.18). This rinding
114
R. G. DA VIES AND A. T. BERGERUD
is consistent with a differential death rate of gray hens over red hens in cold springs. The proportion of red progeny increased when the percentage of red adults increased in the population (Fig. 3.4). The correlation between the proportions of red adult females and red juveniles in the harvest, for the 8 years when there were more than ten adult females per sample, was significant, (r = 0.625, P < 0.10). Heritability ratios of the red and gray colors in ruffed grouse have not been quantified; we have noted red chicks in broods with gray hens, but red hens may produce disproportionately more red chicks than gray hens. The annual changes in red progeny were more closely associated with the juveniles/adult ratios than with mean brood sizes (Fig. 3.18), which again suggests differential mortality of hens by color phase. Table 3.10. Comparison of percentage of red hens in the population and with brood with the density of advertising males and the mean temperatures in June % hens in harvest that are red
Year 1 (Yi)
and Year 2 (Y2)
km2 Y2
1968-69 1969-70 1970-71 1971-72 1972-73 1973-74 1974-75 1975-76 1976-77 1977-78 1978-79 1979-80 1980-81
<8.0
Unweighted means a b c
d e f
g
d<J
June Temp.
A + Ja
A + Yb
Y2
Y,
Y2
6.0 4.7
26.0 23.5 17.8 17.2 17.8 19.9 17.3 14.8 20.7 22.7 18.8 17.5 14.7
7.5
19.1
8.2 9.9 8.9 6.2 4.5 5.2 6.8 8.9 9.3 10.5
(15) 29 (7) (36) 27 (71) (324)e 37 (52)e (202) 31 (16) 28 (46) 33 (9) 24 (34) 23 (13) 19 (53) 21 (47) 21 (138) 29 (24) 21 (97) 23 (39) 20(117) 14 (7) 27 (15) 26 (19) 13 14 19 28
21f
27fg
% hens with Difference broods that Y, - Y2 are red (Y2)c + 16 + 13 + 18
% difference in reds harvest (Y2) and broods (Y2)d
- 6 - 6
8 2 6 1
21 31 14 11 13 20 17 22 -
+ 6
19g
-10
+ 5 - 1
+ + -
-17 -22 -10 - 4 - 7
-
Adult plus juveniles. Adults plus yearlings. Broods in August and September, sample sizes in Table 3.5. Difference (% red adults plus yearlings) minus (% red with broods). Difference: t = 2.761, m = 60/324, nz = 19/52. Difference weighed means: (adults + juveniles) + (adults + yearlings), / = 2.109, n = 231/1077 and H 2 = 83/304. Difference weighted mean: % red adults + yearlings vs. % red with broods, paired t = 4.061.
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
115
Hypotheses for the different mortalities of gray hens in colder springs are: (1) physiological stress during incubation and early brood rearing and (2) differential mortality relative to predation. Dorney and Kabat (1960) found fewer female ruffed grouse in years with cold springs than in years with warm ones and felt that physiological stress was a causal factor. Their significant correlation depends on one cold spring, 1954, and they do not report whether June 1954 was also cold. Also in Wisconsin, Wagner et al. (1965) reported that the annual sex ratio of pheasants (Phasianus colchicus) changed in relation to summer weather; there
Fig. 3.18. Percentage of red juveniles in the harvest was greatest in years with a cold spring and in years of low production.
116
R. G. DA VIES AND A. T. BERGERUD
were fewer hens in dry, hot summers. They, too, suggested physiological stress as an explanation. By contrast, Dumke and Pils (1973) (again Wisconsin) found by radio-tracking that female pheasants were not dying from stress in the dry, hot years, but from predation. Weather not only may affect birds directly but may also affect plant growth and hence concealment from predators. We found no difference in the weights of male ruffed grouse between years or between red and gray males. Maxson (1974) found that brood females died at higher rates from predators than did hens without broods, and those that died during incubation were killed by predators. Anglestam (1984) reported that black grouse (Tetrao tetrix) hens died from predation during egg-laying and incubation. He thought that females might be less cautious when they had high energy demands and when feeding bouts had to be brief. The ruffed grouse in Wisconsin hatched about 1-2 weeks earlier than our birds (Hale & Wendt 1951). Thus, Dorney and Kabat's (1960) finding of fewer females in years with cold Mays and our finding of fewer females in years with lower June temperatures could both be explained by the influence of weather on food and cover and vulnerability to predation. We propose that gray hens are more susceptible to predation than are red hens because of spacing behavior. If ruffed grouse females space their nests, as do spruce (Dendragapus canadensis] and blue grouse (Herzog & Boag 1978, Sopuck 1979), the less aggressive gray hens may settle in "second-best" habitats at high numbers. Moreover, if the spring is cold, plant growth will be retarded and insects will be less abundant; gray hens may then have to take broods long distances (Godfrey 1967) and may encounter predators more often in these lesssecure habitats. We suggest that there is a trade-off between the survival of nesting hens and investment in progeny, which is influenced by intrinsic behavior. Our color markers provided a partial index to this intrinsic behavior. Aggressive hens should invest more in the selection of nesting sites and exclusion of less-aggressive females. Red hens increased over gray hens at high densities when competition was keenest for safe nesting sites and in springs with less concealing plant growth. The greatest disparity should occur when both the population is high and the spring is cold. This occurred in 1971 when the percentage of red hens, adults plus juveniles, had increased from 19% (n = 324)) in autumn 1970 to 37% (n — 52) (adults and yearlings) (Table 3.10). This was the greatest increase in red hens we observed in 9 years where comparisons could be made. In all years red hens, when compared with gray hens, raised fewer broods (Table 3.10) and fewer chicks per brood (Table 3.5). The correlation between the proportion of broods accompanied by red hens versus population density was r — 0.853, n = 8 years. The disparity between the proportion of red hens in the population and those with broods was negatively correlated with density (r = — 0.829, n = 7) (Table 3.10). Although red hens always raised fewer broods and
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
117
chicks than did gray hens, red hens did relatively better (or gray hens did worse) when numbers were high (Table 3.10). Fewer red hens with broods need not be at odds with our statement that red hens should have the pick of the best nesting habitat. In Newfoundland, willow ptarmigan (Lagopus lagopus) deserted their nests more and renested less when the population was aggressive (Bergerud 1970a), a sequence also documented for red grouse (Jenkins et al. 1963). Aggressive hens in those populations also had fewer chicks per brood, which could have resulted from reduced brooding behavior and/or reduced chick viability. In Chapter 2, aggressive hens from MQ and CB nested less frequently than did docile hens from CC in 1971 and 1972. That red hens were aggressive in defending chicks does not necessarily mean they were superior mothers. Their distraction behavior may be linked to their aggressiveness in spacing interaction. No one has yet shown that such defense improves chick survival over that of hens that take a more secretive approach to distraction (but see Pedersen & Steen 1985). Hens in this study had a high mortality rate. The mean percentage of hens versus males of adults in the fall was 40:60 (Table 3.1). The juvenile:adult ratio in the autumn was 1:2.5, but the mean brood consisted of 3.5 chicks (Table 3.1). In a fall population of 100 adults there would be 40 females, but sufficient young for 71 broods (250 -=-3.5 =71), or a shortfall of 31 hens. The greatest mortality of hens occurs during nesting and brood-rearing duties (Maxon 1978, Anglestam 1984, Table 15.7). Hence an aggressive hen should be able to improve her chances for survival by reducing maternal duties or nest tenacity, renesting, and care of young. In contrast the docile phenotype would be prepared to invest more in progeny at the expense of survival and longevity.
3.5.2 Behavior as a factor in the density of grouse populations The density of grouse was higher in Study Area II than in Study Area I from 1970 to 1973, despite the fact that we removed five males from Area II in 1971, and 11 males in 1972. The aspen forest was older and covered a smaller percentage of Area II than Area I (Table 3.11). According to Doerr et al. (1974) and Gullion (1982, 1984a), it should have been less preferable than the younger more abundant aspen in Area I. Many workers have tried to explain the density of ruffed grouse by the vegetative characteristics found at drumming sites, yet we can see no clear generalizations from these studies. We propose a new hypothesis to explain the density of advertising males. Females choose their nest sites before they breed with males (compare Maxson 1974), and females seek nesting sites with herbaceous cover (Bump et al. 1947). Males should attempt to display near nesting females yet be safe from raptors. A common finding is that males display near edges (Gullion et al. 1962, Little 1978, Kubisiak et al. 1980). The edge location provides proximity to females yet some cover from predators. However, edges are dangerous places for drumming
R. G. DAVIES AND A. T. BERGERUD
118
Table 3.11. Characteristics of Study Areas I, and II Characteristic
Study Area I
Study Area II
No. of Advertising males/km 2 , 1970-72
11.0
19.6
Forest characteristics Age of aspen (years) Age of pine (years) Grazing Pure stands Rich sites
75 45 most most least
100 80 least least most
Forest types (as % of total study area) Aspen and willow (Salix spp.) Aspen and grass Aspen and conifer/closed canopy Aspen and conifer/open canopy Douglas fir and pine Lodgepole pine Pine and fir Pine and aspen (young brood habitat)
7 6 10 7 13 13 4 11
<1
30 41
18 34
15 14
23 23
Total aspen % Total conifer % No. of young broods seen 1970 1971 No. of chicks per brood June July
5.3 ± 0.6 (16) 4.2 ± 0.4 (13)
Male fitness index b a b
5.1
3 3 12 14
3 7 25
6.1 ± 0.3 (33) 6.2 ± 0.4 (10)a 5.5
Three broods no chicks counted. broods X July brood size/total males. Calculated:
by males compared to closed canopies (Gullion 1984b). Some males are less aggressive than others and these more docile males may be prepared to advertise for females at safer display sites at higher densities if they can display near areas selected by nesting females. The males in Study Area II were significantly less aggressive in 1970 and 1971 than those in Study Area I (Table 3.8). Study Area II had more herbaceous cover, less area grazed by cattle, and richer site characteristics (Davies 1973). There were more hens with young broods in or adjacent to Study Area II than Study Area I, and the hens in Area II raised more chicks (Table 3.11). Our hypothesis
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
119
is that males of differing intrinsic aggressiveness sort themselves in space to maximize breeding opportunities. More aggressive, density-intolerant males may seek sites where there are fewer advertising males, but these sites also have fewer females. An aggressive male in the low-density situation would attract fewer predators if predators search in a density-dependent manner than would males in the high-density situation. The intrinsic behavior of docile males better fits them for advertising closer to other males than does the intrinsic behavior of more aggressive males. If docile females also settle closer together, there could also be more hens. The reproductive fitness of both types of males may have been similar (Table 3.11). Male density can be more easily explained theoretically by the distribution of nesting hens and the suitability of nesting cover than by vegetative characteristics of male advertising sites.
3.5.3 Male competition and population densities The proportion of red males in the population increased as the population increased from 1968 to 1971 and from 1974 to 1979 (Fig. 3.4). This sequence has also been noted in Minnesota by Gullion and Marshall (1968). Furthermore, as the population increased, there was commonly a greater percentage of red males in the advertising segment than in the juvenile-plus-adult population in the previous fall harvest. This apparent increase in advertising reds cannot be explained as an artifact of harvest statistics biased to red males the prior fall. If the harvest was inflated to red males, and if both color phases lived equally well over winter, there would be fewer red birds advertising in the spring than predicted from the harvest. However, the contrary occurred. Nor can this increase in red males, correlated with increased density, be explained by overwinter mortality of gray adult males. More red adult males died in the 1972-73 and 1981-82 winters than did gray males. If predation was the common cause of winter loss throughout the study, the more approachable red males should have generally disappeared faster than gray males, but in fact the proportion of red males commonly increased between fall and spring among those seen advertising (Fig. 3.4). We propose that as the population increased there was an increase in the number of silent yearling males, and that those males were proportionally more often gray than red. Rusch et al., (1984) documented that the silent male segment in a ruffed grouse population they studied varied from 0 to 38% over 9 years. This silent-male component was positively correlated with density (r = 0.780, n = 9). Gullion (1981) found that the proportion of silent males in Minnesota varied from 0 to 30% and was correlated with density (Fig. 15.21). Because the proportion of silent males varies with density, it provides a means to change the genetic composition of the breeding population. As the population increased in our study, males selected logs that had not been previously used, i.e., they squeezed in between other males (Fig. 3.3). Yet these logs were not selected at low densities, possibly because they were not near nesting females and/or they were at sites less safe from predators.
120
R. G. DAVIES AND A. T. BERGERUD
Mortality of grouse can be significantly influenced by escape features of the advertising site (Meslow 1966). Gullion (1967) termed some sites "traps"; a bird advertising there had a short life expectancy. Mortality of blue grouse males can also vary depending upon advertising sites (Lewis & Zwickel 1981, 1982). Lessaggressive males may opt for increased safety in the face of effective competition for females. We do not propose that aggressive males usurp sites from docile males, but given an uncontested site of questionable safety, aggressive males more often than docile males may be prepared to compromise safety for display opportunities; their intrinsic behavior makes them more conspicuous. Males that used four sites in Area I, which were occupied nearly every year (permanent logs, Fig. 3.3), averaged 303 + 57 pecks per hour during mirror tests (n = 8), compared with 615 + 48 pecks per hour (n = 13) (P < 0.01) by eight males that displayed on logs used only in high years (transient logs). The survival rate of birds on permanent logs from 1969 to 1973 was 60% (12 of 20) and on the transient logs, 45% (17 of 38). Less-aggressive birds may have occupied the safer sites, which may have been near females, as shown for blue grouse (Fig. 2.4). We propose the scenario that as the population increased, the percentage of red males advertising increased, whereas gray males were more inclined to be silent. Red hens also increased, although in this case differential mortality was suspected. A larger, breeding red component produced more red progeny (Fig. 3.4), again producing an aggressive, more approachable population (Fig. 3.16). The genetic-behavior model we have described for ruffed grouse appears to have the means whereby density-dependent regulation occurs, as proposed by Chitty (1967). More-aggressive males have a selective advantage in obtaining advertising sites as the population increases, and more-aggressive hens may secure the best nesting sites when numbers are high, but these aggressive parents provide fewer chicks per brood. Eventually, total chick production may be less than that needed to maintain numbers. Herein, we feel, is a system in which increased aggressiveness is favored at the expense of reduced production. Superimposed on this intrinsic system is the mortality caused by the extrinsic environment. June weather varied randomly and affected the survival of hens and chicks. Adverse weather may have affected gray hens more than red hens, buffering the intrinsic population mechanism. In addition, goshawks appeared at 10year intervals and probably caused high mortality, which again was not evenhanded, but directed more at the aggressive birds. A population that had been gradually increasing in aggressiveness (more red birds) over several years was in one winter (Fig. 3.4) set back to a predominantly less aggressive population (Fig 3.16). This occurred both in 1972-73 and 1981-82. Our findings combine the limiting factors of: (1) lack of spring plant growth and insect availability as it affects production; (2) predators, as they "switch-over" and cause variable adult grouse mortality; and (3) differences in intrinsic behavior and viability between phenotypes. We believe that the relative importance of
DEMOGRAPHY AND BEHAVIOR OF RUFFED GROUSE
121
these environmental and intrinsic factors in the demography of other ruffed grouse populations would be determined by the variability in spring weather and the opportunity for predator switch-over to occur. If the extrinsic modifiers are attenuated or are relatively regular, changes in the outcome of competition for advertising and nesting sites between phenotypes as populations increase could provide an ultimate mechanism that prevents unlimited population increase.
3.6 Summary A ruffed grouse population living in aspen parkland in central British Columbia was monitored from 1968 to 1982. The population increased to 9.9 advertising males/km2 in 1971, declined to 4.5 advertising males/km2 in 1974, and again increased to 10.5 advertising males/km2 in 1979. When the population increased, the percentage of red-phase males and females generally increased. Survival of chicks was highest when temperatures were warm in June and gray hens raised more broods and had larger broods than red hens. During the sharp declines in 1972-73 and 1981-82, the percentage of red males declined drastically. In years of cold springs especially, when the population was high (e.g. 1971), the proportion of females in the population declined, which suggested mortality. Gray rather than red hens were affected most. Red males and females were more aggressive and more approachable than gray birds. The dynamics of the population included: (1) a change in chick survival depending upon June weather; (2) a change in adult overwinter mortality rates probably caused by goshawks switching to grouse after hares declined; and (3) intrinsic differences in the viability of chicks of red and gray genotypes. The argument is made that red males and females are better at securing advertising and nesting sites at high densities than are gray birds but that they produce less-viable young, which results in population decline.
4
Winter Arboreal Feeding Behavior of Ruffed Grouse in East-Central Minnesota R. A. Huempfner and J. R. Tester
4.1 Introduction A number of studies have shown the relative importance of aspen (Populus spp.) in the winter diet of ruffed grouse (Bonasa umbellus), especially in northern parts of their range where snow cover persists for 4 to 6 months (Bump et al. 1947, Korschgen 1966, Gullion & Marshall 1968, Huff 1970, 1973, Schemnitz 1970, Svoboda & Gullion 1972, Doerr et al. 1974, and others). Bailey et al. (1955) were perhaps the first to note the similarity between the ranges of ruffed grouse and of trembling Aspen (Populus tremuloides) in North America. Many questions remain, however, pertaining to criteria used by grouse in selecting preferred aspen clones, the degree of food resource utilization at these sites, and similarities or differences in feeding strategies as grouse populations increase and decrease. The primary objectives of this study were: (1) documentation of the species of trees utilized by ruffed grouse for winter feeding; (2) determination of feeding preference within three canopy levels (top, middle, and bottom); (3) evaluation of environmental factors and structural characteristics of aspen clones that may influence selection of feeding trees; (4) documentation of the seasonability of arboreal feeding; and (5) estimation of the number of trees required per grouse for the entire winter season and the proportion of the available, aspen food resource consumed. 122
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
123
4.2 Study area and methods Research was conducted from 1971 through 1974 on a 250-ha plot within the Cedar Creek Natural History Area (CCNHA) in east-central Minnesota. Proportions of ten habitat types within this 250-ha area, previously cover-typed by Huempfner and Erickson (1975), have been determined by Maxson (1978). In brief, the area consists of 37% sandy upland. Of this upland, 14% is open, abandoned farm fields and 23 % mature oak forest dominated by red oak (Quercus borealis) and northern pin oak (Q. ellipsoidalis). Another 6% makes up transitional zones between uplands and lowlands dominated by paper birch (Betula papyri/era) (3%) and aspen (big-tooth, Populus grandidentata and trembling) (3 % total). Fifty-seven percent consists of lowland peat soils dominated by marsh (10%), alder (Alnus rugosd) (16%), mixed shrubs (Alder, willow, Salix spp.), dogwood (Cornus spp.), bog birch (Betula pumild) (5%), tamarack (Larix laricina) (10%), white cedar (Thuja occidentalis) (6%), and mixed hardwoodscombinations of paper birch, yellow birch (Betula luted), aspen, red maple (Acer rubrum), American elm (Ulmus americana), slippery elm (U. rubra), black ash (Fraxinus nigra), and tamarack (10% total). A 2-km transect divided into 0.33-km sections was established along Anoka County Road 24 to monitor winter arboreal feeding. Observations were made from November to April, with each month divided into three data periods (DP). These were numbered consecutively from 1 (1-10 November) to 18 (21-30 April). Activity patterns of radio-tagged ruffed grouse were monitored to determine the timing of feeding activity. Methods of capture and telemetry techniques used were described by Archibald (1975), Maxson (1978), and Huempfner et al. (1975). Morning observations began 40-50 minutes before sunrise and usually ended 15 minutes after sunrise. Evening observations began 30-60 minutes before and continued to 30-50 minutes after sunset. Midday observations were 30-120 minutes in length from 1000 to 1400 hours on selected days. At least one circuit (4 km long) was completed by researchers before ruffed grouse were observed feeding. Observations were continued until the last individual had left the trees to roost. Each complete circuit provided two opportunities to observe budding at each clone along the route. Grouse canopy position was determined by visually dividing the living canopy into three equal, horizontal portions (Fl = bottom, F2 = middle, F3 = top). The total number of feeding sessions per tree or clone during each feeding period (morning, midday, evening) was calculated by using the maximum number of grouse seen at each location. It was assumed that an individual bird was seen at only one feeding location during the same period on a given day. The size of a feeding aggregation (the number of grouse at any given location) was calculated from the maximum number of birds seen at a given site at the same time. Activity for each bird was recorded as roosting (R; inactive with neck pulled in and
124
R. A. HUEMPFNER AND J. R. TESTER
feathers fluffed out), alert (A; inactive with neck stretched out and feathers pressed to body), or feeding (F; actively pecking at buds, catkins, and twigs, and/or changing location within the canopy). Individual grouse or feeding aggregations were observed from one to eight times on any given day during a single feeding period and thus provided a substantial data base for determining canopy preference over the entire duration of budding. Duration and rate of feeding were determined by stationing observers at favorite aspen clones for the entire feeding period. Observers concentrated on an individual grouse and recorded arrival and departure times, direction to and from the site, and number of bites taken in 1-minute sample intervals. Each bite was assumed to represent a single bud or portion of twig eaten. The Relative Production Index (RPI; Svoboda & Gullion 1972), an estimate of the abundance of male flower buds, was calculated for nine aspen clones at CCNHA in January and in spring 1972, and in fall before budding started in 1972 and 1973. The index was a visual average of the number of male buds on 20 twigs located throughout the canopy multiplied by the percentage of the living canopy that had buds. Binoculars were used for the assessment. An average of four buds per twig over 80% of the canopy would yield an RPI of 320. All sexually mature, trembling and bigtooth aspens (male and female) growing along the 2-km transect and visible from that route were counted during the winter of 1971-72. Paper birch that produced catkins and black cherry trees growing within sight of the observation transect were also counted. Crop samples from 588 ruffed grouse collected throughout Minnesota in fall and winter 1971 to 1975 were examined. Samples were air-dried, separated, and weighed to the nearest 0.1 g. Percentage dry matter represented by aspen, compared with all other food items, was determined. Nineteen grouse collected during all winters had fed exclusively on trembling aspen. Crop contents from these birds were separated into four classes: male flower buds only; twigs with male flower buds attached; twigs with vegetative buds attached; and twig fragments only (no buds attached to twig). The proportion of each, based on dry weight, was calculated. Mean weight of these food items (each assumed to represent one bite) was used to calculate the total weight of material budded from nine male clones at Cedar Creek, where clone attendance was documented by field observation. Fifteen sexually mature, male trembling aspen trees near Coon Lake, Minnesota, approximately 11 km south of Cedar Creek, were sampled on 17 March 1977 to determine bud and twig (current-year growth) weight, twig diameter, twig length, and number of male buds/twig for the three canopy levels. After trees were felled, the living portion of each canopy was equally divided into three sections perpendicular to the trunk. Twigs were collected throughout each section to provide 200-400 male buds from each of the three canopy levels. One-hundred male buds were then sampled. The 45 bud and 45 twig samples
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
125
were oven dried for 48 hours at 60°C and weighed. Twig diameters and lengths were measured after drying. Twig diameters (mm) were measured just above the attachment of the first male bud below the terminal bud. Twig length (cm) was measured by following the curvature of the dried sample. Average width and height of the three canopy sections were recorded to the nearest 0.1 m. These measurements were used to calculate average canopy volume and the proportion contributed by each level. 4.3 Results and discussion
4.3.1 Morning and evening budding The number of grouse observed budding at Cedar Creek was extremely high and probably reflected the limited (but not necessarily limiting) distribution of aspen on the study area. Some of the most heavily used clones were apparently established as a direct result of disturbance during road construction years earlier. Mean number of grouse seen per kilometer during the evening was 9.0 (1971-72), 1.8 (1972-73), and 2.2 during the three winters (1973-74) (Fig. 4.1). Comparable densities along observation routes at Cloquet from 1965-66 to 1971-72 ranged from 0.03 to 0.39 grouse/km during January and February (Svoboda & Gullion 1972). The 1 year of overlap between the Cloquet study and this investigation showed a 23-fold increase in the observation rate at Cedar Creek (1971-72; 9.0 versus 0.4 grouse/km). These data suggest that grouse were much more concentrated in winter at Cedar Creek than in the aspen-rich forest at Cloquet, located about 150 km north. Fewer grouse were seen budding during the morning than the evening period (Fig. 4.1). An average of 30.5% fewer grouse were seen in the moring (DP 7-14) in 1971-72 and 19.7% fewer in 1972-73 than in the evening periods of these 2 years (P < 0.05, 2-way ANOVA). Shorter budding periods during the morning, with generally poorer light conditions than in the evening, were probably the main reasons for these differences. Based on telemetry data, there was no evidence to suggest that differences resulted from using different clones in morning from those used in evening budding. The sharp increase in the number of budding grouse in December over those seen in November was related to the beginning of permanent snow cover. As snow depth increased during December, the number of budding grouse increased until a peak was reached in late December to mid-January. The dramatic and continual decline in the number of grouse observed budding during 1971-72 resulted from a number of factors. Grouse apparently converged on several clones in early December, especially at male aspen clone numbered "4170." Continual evening observations at that site from 25 to 31 December 1971 revealed that 11, 12, 16, 22, 6, 18, and 14 grouse
126
R. A. HUEMPFNER AND J. R. TESTER
Fig. 4.1. Mean number of ruffed grouse seen budding during morning and evening observation periods, within 10-day data period (DP) intervals, Cedar Creek, Minnesota, 1971-74.
were present during peak budding. The majority of these birds filtered-out to other feeding sites during the remainder of the winter. Some of these sites, based on telemetry location data, were too distant from the transect to be recorded. Predation accounted for a portion of the reduction. Predation of radio-tagged
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
127
grouse was high during late January, peaked in February, and decreased sharply in March (Huempfner et al., unpubl. data). The accelerated decline in the number of budding grouse seen from early March to late March did not, however, coincide with predation on radio-tagged grouse during that 30-day interval. The rapid decline coincided with an increase in snow crusting in late afternoon, evening, and early morning, and rapid melting of the entire snow profile during midday. Studies at Cloquet (Vanderschaegen 1970, Svoboda & Gullion 1972) have also shown decreased arboreal feeding and increased use of hazel catkins when crusting occurs, and during most of the winter period when shallow snow conditions persist. Another factor contributing to reductions in late winter budding could be an increase in aggression and reduced sociability owing to an increase in the production of sex hormones. This was suggested by Doerr et al. (1974), who observed reduction in size of budding aggregations in late winter in Alberta, Canada.
4.3.2 Factors affecting frequency Ruffed grouse populations declined substantially from 1971-72 through 1973-74. The number of territorial males present in spring on a 242-ha census area at Cedar Creek decreased 71.4%, from 28 drummers in 1972 to only eight in 1974 (Table 4.1). The fall flushing rate declined 37.1% from fall 1971 to fall 1973. A decreasing grouse population thus allowed comparisons of winter feeding behavior under a variety of densities. This same factor, however, also compli-
Table 4.1. Demographic changes in ruffed grouse at the Cedar Creek area, Minnesota, 1970-74
a
b
c
Year
No. of spring drumming males3
Grouse flushing rate/hour during fall hunting seasonb
Ratio of juveniles to adultsc
1970 1971 1972 1973 1974
23 30 28 18 8
2.3* 3.5 2.8 2.2 2.4
2.9 1.6 1.6 2.5 3.2
All known drumming logs from past studies and new sites where displaying males were heard were examined each spring during the peak drumming period. The area that was searched was determined by connecting all outermost logs and totaled 241.6 ha. Flushing rates provided by author and several interested sportsmen. Trained pointing dogs were used except as noted by asterisk. Flushing rates based on average of 228.3 hours of field hunting time each year (range 68-378 hours). Data obtained from tails and wings contributed by interested sportsmen, biologists, and the author. Ratios based on 1,736 samples over the years indicated.
128
R. A. HUEMPFNER AND J. R. TESTER
cated our interpretation of data relative to snow conditions and their affect on stimulating or decreasing arboreal feeding. The winters also varied substantially in amount, quality, and duration of snow cover. Winter 1971-72 could be considered a "good" winter for grouse, with deep, soft snow that allowed snow burrowing for 50-80 days (Fig 4.2). The following winter was the most severe, with shallow and/or crusted snow profiles. Burrowing was possible for only 5-10 days. Winter 1973-74 was intermediate to the other years and grouse could burrow roost for 35-45 days. When interpreting the information presented on snow depth and snow quality it is important to know that ruffed grouse will snow burrow when favorable snow depths are approximately 18-20 cm or more. In instances where lower-quality, granular snow has been covered by high-quality, fluffy snow to a depth of 10 cm or more, ruffed grouse can still burrow roost by scratching their way into the compacted, lower snow deep enough to support a snow roof. The mean numbers of grouse per km observed each day during evening budding were 9.0, 1.8, and 2.2 for the three winters, respectively. About half of this 82 % reduction can be explained by a decline in the ruffed grouse population based on fall flushing rates (Table 4.1). The remainder was apparently caused by variations in snow depth, snow quality, and duration of snow cover among winters, and the actual timing of winter predation. The generally deep and more favorable snow profile in 1973-74, compared with 1972-73, resulted in a slightly higher daily observation rate in 1973-74, even though the grouse population was apparently lower.
4.3.3 Tree canopy position A number of studies have revealed a preference by browsing birds for food resources available in the upper canopies of trees in winter. This trend has been shown for ruffed grouse (Huff 1970, 1973, Schemnitz 1970, Svoboda & Gullion 1972), blue grouse (Dendragapus obscurus) (Hoffman 1961), black grouse (Tetrao tetrix) (Seiskari 1962), and capercaillie (Tetrao urogallus) (Lindroth & Lindgren 1950). A nutritional basis for this preference was proposed by Huff (1970, 1973), who found that upper-canopy buds had higher percentages of protein and crude fat than mid- or low-canopy buds. Huff also reported that by late winter (March and April) the minimum available energy was similar in the upper and lower portion of the canopy. Canopy use was determined from 7,341 feeding observations (all canopy levels) for 3,987 ruffed grouse budding sessions in the three winters (Fig 4.3). An individual grouse was observed an average of 1.8 times (7,341 -=- 3,987) during any given feeding session. The rate increased from 1.3 to 3.2 and 3.7 over the three winters, respectively. We believe this increase resulted primarily from the declining grouse population, which allowed us to complete a full transect more rapidly.
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
129
Fig. 4.2. Snow quality and depth in relation to snow burrowing by ruffed grouse, Cedar Creek, Minnesota, 1971-74.
130
R. A. HUEMPFNER AND J. R. TESTER
Fig. 4.3. Percentage of ruffed grouse budding in the top, middle, and bottom canopy in the morning and evening during three winters, 1971-72 through 1973-74. One asterisk means a P < 0.05 difference within feeding levels between years, two asterisks mean P < 0.01. See text for explanation of F-value comparisons.
Grouse used the upper canopy most intensively during morning and evening budding in 1971-72 (Fig. 4.3). The middle level was used about half as often as the top, and feeding in the bottom third accounted for only 1.1% (morning) and 2.5% (evening) of all observations that winter. Morning and evening budding was concentrated in the top of the canopy in January and the first 20 days in February, with increased use of the middle level in late February and early March. Midday feeding (not shown in Fig. 4.3) was more variable but showed a slight seasonal preference for the top over the middle, and increased use of the bottom canopy compared with morning and evening observations that year. Both of the following winters were in sharp contrast to winter 1971-72 and showed greater fluctuation and increased use of the middle and bottom canopy
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
131
levels. Budding during the morning in 1972-73 was concentrated in the middle canopy level, and evening use was primarily in the top. Use of the bottom level was also substantially higher than the previous year. Evening budding in 1973-74 was about equal in the top and middle levels, with use of the bottom continuing to increase. Patterns of canopy use varied considerably among years (Fig. 4.3). Data from evening feeding for all years indicate that the high grouse population in 1971-72, during favorable burrowing conditions, utilized buds and twigs from the top canopy level to a much greater extent (61.9% of all observations) than did a low population in 1973-74 (43.2%). Intermediate use of the top canopy occurred in 1972-73 during a rapid decline in grouse numbers (57.1%). The percentages of observations that occurred within the three canopy levels of each of seven data periods were compared by one-way ANOVA (Fig. 4.3). Ruffed grouse fed in the upper canopy during morning budding significantly more often (P < 0.05) in 1971-72, during high grouse populations, than in 1972-73. The reverse was true for use of the bottom level, with significantly more use (P < 0.01) occurring during 1972-73. The middle level also received more use as the grouse population declined, but the means were not significantly different. There was significantly more evening budding in the top of the canopy during 1971-72 than in 1973-74 (P < 0.05, Fig. 4.3). The continual decline in utilization of that level over all three winters combined, however, was not significantly different (P > 0.05). Means for the middle canopy were similar, and no combination of years produced significantly different results. The proportion of budding that occurred in the bottom of the canopy remained relatively low in all years but significantly more (P < 0.05) budding occurred in this level as the grouse population declined (Table 4.1, Fig. 4.3). Chi-square was used to test for significant differences between observed and expected values of canopy use for the three feeding periods during all winters (Fig. 4.4). Expected values 1 and 3, corrected for measured canopy-volume differences, as well as uncorrected values (Expected 2, Expected 4) that assumed equal volume in each level, were used (see Huempfner 1981, pp. 43-44, 148, for more details). The comparison of observed and expected values for canopy use resulted in rejecting the null hypothesis of equal canopy use during budding (Fig. 4.4). Of the 12 comparisons that tested the observed and expected frequency of observations among the three canopy levels, observed values for the top exceeded the expected value in all cases (Fig. 4.4). Observed values for the middle were greater than expected seven of 12 times, while observed use of the bottom was less than expected in all 12 instances. For all winters the 7,241 buddings observations (morning, midday, and evening) were distributed among the three canopy levels in the following proportions: top, 55.3%; middle, 38.5%; bottom, 6.3%. These data show a consistent reduction in the use of buds and twigs from the
132
R. A. HUEMPFNER AND J. R. TESTER
Fig. 4.4. Observed (0, x-axis) and expected (1, 2, 3, 4, x-axis) number of arboreal observations in all canopy levels during morning, midday, and evening budding periods. Expected values corrected for volume differences (1, 3) or uncorrected and based on equal volumes (2, 4); see text for further details.
top canopy level during consecutive winters. Spring drumming census and fall flushing rates confirmed a population reduction over this period (Table 4. 1). The above findings are by no means conclusive, but strongly suggest that the wintering ruffed grouse population during a population low utilized the bud and twig
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
133
resources available in the tree canopy differently than during the periods of higher populations. Reduced use of the top canopy level during this study may be related to one or any combination of the following factors: (1) reduced bud production in the upper levels of preferred clones owing to heavy browsing in previous winter(s); (2) seasonal shifts in the distribution of fat, protein, and other nutrients affecting palatability and caloric value of buds and twigs within the canopy; (3) mean winter temperatures; (4) age and condition of feeding trees; (5) behavioral changes of "genetically different" ruffed grouse at the height of a cycle than at or near the low; (6) competition for a limited, aspen food supply which becomes more acute during high grouse populations; (7) varied response to potential predation by members of flocks (e.g., as populations declined and flock size decreased, fewer flock members "felt safe" feeding in the upper canopy), and increased number of grouse feeding as singles. Many of these and other variables will be discussed in the sections that follow. Data presented to this point and substantiated by other findings yet to be presented suggest that competition by ruffed grouse for aspen buds and twigs, modified by behavioral changes, may have been a critical factor that affected canopy-level utilization.
4.3.4 Intensity of canopy use Ruffed grouse fed progressively higher in tree canopies as the feeding period progressed. This was true for both the morning and the evening period, but was especially evident during evening budding. Schemnitz (1970) also observed this budding trend in ruffed grouse in Maine. To test this tendency, we organized morning and evening budding data into 5-minute intervals before and after sunrise and sunset. The percentages of all observations occurring in the top, middle, and bottom portions of the canopy by 5minute intervals are presented (Figs. 4.5, 4.6). Correlations were calculated for all consecutive intervals containing at least 1% of the total observations. This prevented biasing the results by including extremely small sample sizes, but still
134
R. A. HUEMPFNER AND J. R. TESTER
Fig. 4.5. Percent of evening budding occurring in each canopy level by 5-minute intervals. Includes observations from 1971-72 (n = 2,111), 1972-73 (n = 1,720), and 1973-74 (n = 945). Top correlations: percent of observations in each 5-minute interval versus the assigned number of each 5-minute interval. Data plotted and correlations (7) include intervals that contained 1.0% or more of all observations (intervals 1-15, inclusive, 13 df). Bottom correlations (2) include intervals 7-15, inclusive, 7 df. This interval includes 86.3% of all observations during three winters.
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
135
Fig. 4.6. Percent of morning budding occurring in each canopy level by 5-minute intervals. Includes observations from 1971-72 (n = 1,248) and 1972-73 (n = 1,036). Correlations: percent of observations in each 5-minute interval versus the assigned number of each 5-minute interval. Data plotted and correlations include intervals that contained 1.0% or more of all observatons (intervals 1-9, inclusive, 7 df).
136
R. A. HUEMPFNER AND J. R. TESTER
allowed use of 96.1 % of all data for the morning (n = 2,254 observations) and 98.7% for the evening (n = 4,714 observations). Budding in the top of the canopy was positively and significantly correlated with advance of the feeding period for both morning and evening sessions (Figs. 4.5, 4.6). As budding in the top increased, use of the lower levels decreased simultaneously. Significant negative correlations were found for the middle and bottom levels in the morning, and for the middle during evening. A high positive correlation (r = 0.99, P < 0.01) was found for percentage use of the upper canopy during the evening and consecutive intervals from 5 minutes before to 40 minutes after sunset (intervals 7-15). This was the principal budding period and included 86.3% of all observations. The bimodal pattern of budding in the upper canopy during the evening is particularly interesting (Fig. 4.5). It was uncommon for individual ruffed grouse to descend to the middle or bottom levels once they had moved to the upper canopy to feed. Arrival and departure information, however, indicated that this pattern was caused by a small proportion of grouse that arrived before sunset in the first two winters, moved to the upper levels, and then departed before the primary budding period began. This pattern did not occur during low grouse densities in 1973-74, and was most characteristic of their behavior in 1971-72 during a population high. These data indicate that at least a portion of the population fed in "shifts" during high and moderately high winter densities. We can only speculate that these early arrivals were juveniles and/or adults that were behaviorally prevented from feeding at peak budding periods. Still another contrasting explanation is that these were dominant males and females that would not tolerate the often large aggregations feeding at clones later in the afternoon.
4.3.5 Roosting and budding in "feeding trees" Ruffed grouse arrived at feeding sites by flying in from day roosts (most common) or by walking to a location near the tree and flying to the canopy at a sharp upward angle. Grouse normally flew to the lower or middle canopy, rarely flying directly to the top level. Birds most often landed near the trunk, remained there for 1 minute or less, then walked toward the outer edge of the canopy and began to feed. This inactivity at arrival and at other times during the budding period was termed "roosting" or "alert" (discussed later). Roosting grouse perched motionless on a branch, body feathers fluffed-out, neck and head appressed to the breast, with legs and feet covered with breast feathers. Of the 7,060 arboreal observations (morning and evening, all years), only 554 (7.8%) were classified as roosting. The remainder (6,506, 92.2%) were instances of active budding. Nearly identical proportions of morning (174/2,284, 7.6%) and evening (380/4,776, 8.0%) observations were classified as roosting behavior. Roosting for extended periods (5-10 minutes) occurred occasionally during
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
137
both morning and evening budding, but was much more common during midday. A total of 76 of the 281 midday observations were classified as roosting (27.0%). Midday budding activity was also more "casual," and the entire sequence of active-inactive behavior appeared to be a combination of comfort and feeding movements. Brander (1965) also suggested that daytime winter feeding was at least partly "pastime" feeding for ruffed grouse at Cloquet, Minnesota. Of the 630 nonfeeding observations (morning, midday, and evening), 87 (13.8%) were further classified as "alert" behavior. The alert posture was often observed when raptors were seen flying near or perched within sight of feeding trees. Ruffed grouse perched erect and motionless on branches, their legs and necks readily visible and body feathers depressed. Raptors were observed flying within 75 m of grouse in feeding trees on eight occasions during the study. Three observations were of great horned owls (Bubo virginianus) and five were of goshawks (Accipiter gentilis). Two grouse stopped feeding prematurely on one occasion, flying from the clone after a goshawk had passed within 35-45 m. The goshawk made no attempt to follow them. On the other seven occasions, ruffed grouse perched in the alert position for 0.5 to 3.5 minutes and then continued to bud in the same clone. During four observations grouse were perched near the top of the aspen canopy with little or no protection from surrounding branches. These observations suggest that predation at feeding trees was uncommon. However, raptors may have benefited over the long term by perching near intensively used clones and intercepting grouse as they arrived or departed (goshawks), or perching near grouse feeding sites (owls) in an attempt to take birds roosted near feeding clones. Raptor activity patterns determined by telemetry at Cedar Creek (Huempfner et al., unpubl. data) suggest that owls would benefit from this behavior during both morning and evening feeding periods and goshawks would concentrate on grouse at or near morning feeding locations. The density of owls at Cedar Creek was probably similar during all three winters of this study (Huempfner et al., unpubl. data). Migratory goshawks were also present during each winter, but may have peaked in density during the winter of 1972-73 following a peak fall migration that year through Duluth, Minnesota. We do not believe that casual observations of raptors during budding observation periods were indicative of relative densities of these predators at Cedar Creek during three winters.
4.3.6 Environmental and behavioral factors The pattern of morning budding was similar among winters (Fig. 4.7), with the earliest budding in both winters occurring 40-45 minutes before sunrise. Nearly 60% of all morning budding in 1971-72 occurred in a 15-minute interval from 15-30 minutes before sunrise. A similar proportion was concentrated in a 20-35minute interval before sunrise in 1972-73.
138
R. A. HUEMPFNER AND J. R. TESTER
Fig. 4.7. Percent of morning and evening budding by 5-minute intervals relative to sunrise and sunset during three winters of varying snow quality and depth. Each column represents one 5-minute interval. Only data from DP 7-14, inclusive, were used for each year to eliminate biases resulting from comparing different seasonal intervals.
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
139
Circadian shifts relative to the beginning and termination of activity and daily activity peaks are strongly influenced by seasonal cycles of illumination (Aschoff 1966). Radio-tagged ruffed grouse monitored during this study were normally active before sunrise and terminated activity after sunset in late fall and winter. Daily activity during periods of deep, uncrusted snow were usually tri-modal, with peaks occurring near sunrise and sunset and again over midday for 1 to 2 hours between 1000-1400 hours (Huempfner, unpubl. data, 1971-73). When snows were crusted, total daily activity was increased as grouse walked on the snow surface primarily in search of hazel catkins (Huempfner, unpubl. data, 1971-73). This pattern changed from April to June for both male (Archibald 1973) and female (Maxson 1974, 1977) grouse at Cedar Creek. Both sexes showed increased daily activity during this interval, even though they began activity later and terminated it earlier relative to sunrise and sunset, respectively. Our explanation for this change from an early morning and late evening activity pattern in winter is based on the elevated light levels in the woodlot in winter owing to snow cover and the absence of leaves. It may be that daily activity of ruffed grouse begins and ends at similar light levels year round, but that these levels occur at different times relative to sunrise and sunset based on the presence or absence of snow and leaf cover. Correlation coefficients were used to test the relationship between the median minute of morning budding and mean temperature for 10-day intervals. These variables were significantly correlated in 1971-72 (r = - 0.747, 6 df, P < 0.05) and showed even greater correlation in 1972-73 (r = - 0.90s, 6 df, P < 0.01). The negative correlation coefficients indicate that peak feeding occurred earlier in the morning as the mean temperature dropped. Both years combined also showed a significant correlation (r = -0.62s, 14 df, P < 0.01). Consistently cold temperatures in data periods 8-10 in 1971-72 (x = - 19.3°C, range - 16.5 to -21.5°C) appeared to cause equally consistent and early peak feeding (x = 26.2 minutes before sunrise, range 26.0 to 26.5 minutes). A warmup during data periods 11-13 (x = - 12.1°C, range - 10.2 to - 13.2°C) appeared to cause a shift in peak budding to almost 8 minutes later (x = 18.5 minutes before sunrise, range 18.0 to 19.0 minutes). A gradual warming trend in 1972-73 from data periods 11 (x = - 10.4°C) to 14 (x = +2.2°C) was also accompanied by a shift in peak budding from 32.5 minutes to 9.5 minutes before sunrise. It appeared that ruffed grouse may have shifted the period of peak morning feeding to compensate for a potentially large, negative energy balance during particularly cold portions of the winter. Seiskari (1962, pp. 63-69) noted that feeding activity of penned black grouse and capercaillie increased with cold weather and high barometric pressure. Seiskari's data suggest that earlier morning feeding appeared to follow low temperatures for each species. The earliest recorded evening budding occurred 55-60 minutes before sunset,
140
R. A. HUEMPFNER AND J. R. TESTER
the latest 50-55 minutes after sunset (Fig. 4.7). Peak budding occurred from sunset to 35 minutes after sunset in all winters. Unlike the morning period, peak evening budding was poorly correlated with temperature. Correlations of mean daily temperature and peak median values of budding showed a significant inverse relationship only in 1971-72 (r = — 0.866, 6 df, P < 0.05), indicating that later feeding occurred as the mean temperature decreased during this winter. Temperature and peak budding were not significantly related in 1972-73 (r = 0.705, 6 df, P > 0.05) or 1973-74 (r = - 0.050, 6 df, P > 0.05). All years combined also showed no significant relationship (r = - 0.154, 22 df, P > 0.05). We conclude from these data that winter temperature had a limited influence on the timing of peak evening budding by ruffed grouse.
4.3.7 Variations in evening budding among winters The shift to later and more concentrated evening budding over three consecutive winters was at best poorly correlated with the seasonal and environmental factors previously discussed. These differences may well be related to the changing composition and behavior of the grouse population as numbers declined. Competition at feeding sites may have played an important role in the timing of evening budding, in combination with differences in sociability, physiology, genetic makeup, age structure, and behavior. Age structure, determined from fall and early winter hunter-shot birds, changed dramatically from 1971 to 1973. Although the population was declining, the ratio of juveniles:adults in the hunting sample was much higher in 1973 (2.5:1) than the previous 2 years (1.6:1 for 1971 and 1972, Table 4.1). Juveniles may have responded differently from adults to the timing of evening feeding. Other studies have shown shifts to "less wary" individual ruffed grouse during peak densities (Keith 1963). This has also been shown for willow ptarmigan (Lagopus lagopus) (Bergerud 1972) and ruffed grouse (Marshall 1954) immediately after a cyclic peak. We have also noted a greater proportion of ruffed grouse "holding" for pointing dogs and thus flushing nearer the hunter when the populations were high, as opposed to low, in Minnesota. These data strongly suggest that factors other than environmental variables may have caused a postponement and shortening of the evening budding interval as the grouse population declined at Cedar Creek. Long-term studies are required to more fully understand the mechanisms of these changes.
4.3.8 Duration of budding A total of 102 complete arboreal budding sessions were observed over the three winters of this study. Thirty-three were recorded for the morning period, nine for midday, and 60 during the evening (Table 4.2). The mean duration of morn-
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
141
ing budding was 12.2 + 0.9 ( + SE) minutes, whereas evenings averaged 17.6 + 1.0 minutes. The average duration of morning arboreal feeding for ruffed grouse in Alberta, Canada (Doerr et al. 1974) was 16 minutes, with evening budding averaging 24 minutes. At Cloquet, Minnesota, morning budding averaged 19.0 minutes and evening feeding varied from 11.9 to 18.3 minutes (Brander 1965, Godfrey 1967, Svoboda & Gullion 1972). Ruffed grouse at Cedar Creek tended to bud for longer periods of time during both morning and evening as winter progressed. Variability was high, however, and monthly means were not significantly different (P > 0.05). Midday budding showed large variation, lasting from 12.0 to 61.0 minutes, with a mean of 31.3 + 5.5 minutes. Grouse paused often during this midday period to roost and seldom fed as rapidly as they did during morning or evening. Table 4.2. Duration of morning, midday, and evening arboreal budding by ruffed grouse, Cedar Creek, 1971-74 Duration of budding (minutes) 3 Sample sizeb
Max.
Min.
X
SE
Morning Dec. Jan. Feb. Mar.
7 14 9 3
18.0 18.5 29.5 13.5
7.5 3.0 5.0 12.0
10.9 11.8 13.5 13.0
1.4 1.3 2.4 0.5
Total
33
12.2
0.9
61.0 27.5
4.6
31.3
5.5
16.5 16.0 18.9
1.5 1.8 1.2
17.6
1.0
Month
Midday Jan. Feb. Total Evening Dec. Jan. Feb. Total a
b
1 8
48.0
12.0
9
2 25 33 60
18.0 39.0 28.0
15.0 8.0 2.5
Duration of budding recorded to nearest 0.5 minute and included only those periods when the principal activity was budding, or included only those observations where grouse were seen arriving and departing from arboreal feeding trees. Number of individual ruffed-grouse budding sessions. Data from all three winters combined.
142
R. A. HUEMPFNER AND J. R. TESTER
Schemnitz (1970) made only four observations of grouse budding during midday, and these were during or preceding a storm. This trend was not apparent at Cedar Creek, but most midday budding there in 1971-72 generally coincided with the coldest portion of that season. Similar increases in midday feeding related to cold weather have been noted for black grouse and capercaillie in Finland (Seiskari 1962).
4.3.9 Seasonal budding rates Budding rates were recorded in bites per minute for individual ruffed grouse during evening budding from 18 January to 4 March 1973. One-minute observation periods made during all portions of the feeding session at heavily used clones resulted in data for 181, 1-minute periods. These data were divided into 96 minutes of feeding by single birds and 85 minutes of birds feeding in flocks (two or more birds). The overall mean feeding rate was 25.1 + 0.7 bites/minute, with a range of 7-50. These rates were similar to the 30 bites/minute under usual conditions, and 47.4 bites/minute for birds during "almost frantic" evening feeding at Cloquet, Minnesota (Svoboda & Gullion 1972). Ruffed grouse feeding as singles seemed at a disadvantage compared with birds feeding in flocks. The mean seasonal feeding rate for singles was 23.1 + 0.9 bites/minute and that of flock members was 27.5+ 1.0(/ J <0.01). Grouse in two-bird flocks also budded significantly faster (x = 29.2 ± 9.9) than those in three-bird aggregations (x = 24.1 + 7.0, P < 0.05). Budding rates appeared to decrease as winter continued, a factor perhaps related to a dwindling bud supply. Single birds in January fed at the rate of 25.9 + 9.9 bites/minute, but those in February at the significantly lower rate of 21.5 ± 7.9 (P < 0.01). Laboratory experiments with starlings (Sturnus vulgaris) and tricolored blackbirds (Agelaius tricolor} and extensive field studies indicate that single birds increase feeding rates when they join flocks (Morse 1970, Powell 1974). This hypothesis states that the mean time an individual spends in surveillance for predators is substantially reduced and shared by all members of the flock, increasing the amount of time available for individual foraging. The same principle may be operating on ruffed grouse budding in winter at Cedar Creek, but we are unable to explain why two-bird flocks budded faster than three-bird flocks.
4.3.10 Tree species and number used We recorded budding observations for four tree species over three winters (Table 4.3). Of the 7,221 feeding observations, 83.6% were in male trembling aspen and 4.9% in female. The percentage for big-tooth aspen was 4.2% in male trees and 1.5% in female trees. Use of trembling aspen increased from early to
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
143
late winter in 1971-72 and 1972-73. The only significant use of big-tooth aspen was during 1971-72; a trend reversed from that of trembling-aspen use was revealed, with big-tooth used only occasionally in late winter (March) compared with heavier use of this species in January and February. These findings contrast with the sequence of winter budding in these aspen species at Cloquet, Minnesota. Huff (1970, p. 25) stated that the "decline in caloric value suggests that male flower buds of trembling aspen may be a lower quality food for grouse than those of big-tooth aspen during late winter." He found that eight ruffed grouse collected in February and March 1970 had crops containing 80% big-tooth aspen buds and twigs. Our observations suggest that big-tooth is used far less often than trembling aspen in late winter at Cedar Creek. We observed 73 complete budding sessions during 19 evenings in 1973 (3 January to 1 March) to assess the number of different trembling aspen trees used
Table 4.3. Abundance of 710 seed-producing feeding trees and their relative importance value (IV) based on morning, midday, and evening budding observations during data periods 4 (1 December) through 14 (20 March) inclusive Species and sex
% available
Trembling aspen (M)
b
1973-74
Combined
Use"
IV
b
Use
IV
Use
IV
Use
IV
34.5
77.0
2.2
89.8
2.6
91.1
2.6
83.6
2.4
Trembling aspen (F)
10.7
4.6
0.4
5.5
0.5
4.6
0.4
4.9
0.4
Big-tooth aspen (M)
3.5
8.5
2.4
none
4.2
1.2
Big-tooth aspen (F)
1.7
2.5
1.5
0.6
0.4
none
1.5
0.9
Paper birch
48.3
3.8
0.1
2.8
0.1
2.1
0.1
3.2
0.1
Black cherry
1.3
3.6
2.8
1.3
1.0
2.1
1.6
2.5
1.9
Sample size a
1972-73
1971-72
3,555
none
2,735
Percent of total budding occurring in each species and sex. % of budding observations in a species Importance Value = % of available trees of that species
931
7 ,221
144
R. A. HUEMPFNER AND J. R. TESTER
by individual grouse from arrival to departure at clones. We combined data from grouse budding as singles and as members of flocks. The minimum number of trees used during an entire evening budding session was one (13.7%, used by ten of the 73 birds), with a maximum of 11 trees used by one grouse on 10 January. The remaining observations were: two trees, 31.5%; three trees, 27.4%; four trees, 13.7%; five trees 6.8%; six trees, 2.7%; and seven trees, 2.7%. The mean for all 73 budding sessions was 3.0 ± 0.2 trees used per grouse per session. Seasonal comparisons showed that fewer trees were used during a given budding session in January (jc = 2 . 8 ± 0 . 3 , n = 45) than in February (x = 3.3 + 0.3, n = 25). This suggests that additional searching may have occurred as winter progressed. Competition for a limited (but not necessarily limiting) aspen resource may have accelerated the use of big-tooth aspen in 1971-72 during a time of high population. Reliance on this species was reduced as the population declined. It does not appear, however, that use of big-tooth aspen was initiated by a bud shortage. Relatively heavy use of big-tooth in January and early February 1971-72 occurred before buds at several trembling aspen clones were primarily exhausted. Some unanswered questions that may be directly related to the proportional use of these species are: (1) Was the high proportion of juveniles in the population during low grouse densities somehow related to the avoidance of big-tooth during that winter? (2) Is big-tooth aspen an alternate winter food primarily during population highs? (3) Were big-tooth aspen buds and twigs in 1971-72 unusually high in nutrients compared with trembling aspen, thus stimulating relatively heavy use that year compared with the other two winters? Mature paper birch trees were the most commonly available food resource along the transect. Nearly half of all trees were birch (Table 4.3), yet only 3.2% of budding observations in all three winters occurred in this species. Birch had an importance value of only 0.1, indicating extremely limited use compared with its availability. Large, mature black cherry trees were uncommon along the transect. However, black cherry had an importance index of 1.9, second only to male trembling aspen as a winter food source (Table 4.3). A total of 184 budding observations were recorded in four of the nine trees found along the transect. Heaviest use was in 1971-72, when the importance value was 2.8, exceeding that for male trembling and male big-tooth aspen. Cherry, like paper birch, was used more intensively by grouse in the evening than morning. Only 1.4% of the morning observations was in black cherry, but 3.4% of the evening feeding occurred in this species. On days when cherry trees were used, the majority of feeding coincided with the beginning of arboreal budding and terminated when birds flew to nearby aspen clones to compete their budding. Winter use of black cherry and other Prunus spp. in northern regions of the
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
145
United States shows a great deal of regional variation. Schemnitz (1970) reported 4.4-6.4% of budding observations in Prunus spp. in Maine for data obtained over a 20-year period. Brown (1946) also considered cherry an important winter food in Maine. Bump et al. (1947) reported that Prunus spp. made up the greatest volumetric percentage in winter crop and gizzard samples (17.4%), followed closely by Populus spp. (16.4%). They also stated that only the cherries appeared in all four seasons in their samples. Woehr (1974), working in central New York, found that black cherry was the leading winter food and was found in every fecal sample examined in 2 years. Black cherry had an index of use (percent occurrence/percent available) of 19.4 during each winter, but aspen values were only 3.4 and 1.0 in 1973 and 1974, respectively. Fall and early-winter aspen bud indexes of approximately 100 or less can be considered low production years (Svoboda & Gullion 1972). The mean values for all clones combined at Cedar Creek were substantially above this level in all years (Table 4.4). The value for 1971-72 should be considered minimal because extensive budding had already occurred at several clones before its calculation in January. The actual fall Relative Production Index that year may have been in the range of 330-360. We concluded, based on these representative clones, that no overall production shortage of male buds occurred at Cedar Creek preceding the three winters being considered. However, a spring bud index following the winter of heaviest use (1971-72) indicated that the January bud resources at five of the nine clones had been reduced an average of 83.5% (clones 4333, 3982, 4170, 4007, and 2584; Table 4.4). The mean January RPI for these five heavily used clones was 275.0 compared with 45.4 in spring. The remaining four clones had bud reductions of 32.9%. Four of these clones (all except 2584) produced leaves in spring 1972 in a "clumped" or "patchy" pattern. This patchiness, which persisted all summer, indicated the loss of a large proportion of leaf buds and current-year twigs in addition to male flower buds. We did not observe patchy leaf arrangement at any aspen clone at Cedar Creek in spring 1973 or 1974. The mean RPI of these four heavily used clones was reduced an average of 42.1% in fall 1972 compared with only a 7.8% reduction at the other five clones monitored during the study. This suggests that extremely heavy budding by grouse in 1971-72 had a deleterious effect on the quantity of buds available during the winter of 1972-73. This reduction was not evident 2 years later when we counted buds in fall 1973. To determine the weight of aspen buds we cut two, 24-year-old, male trembling aspens from a 15-tree clone. This clone had been used by one to four ruffed grouse during at least three previous winters. Five 100-bud samples were collected from each of the three canopy levels and dried at 60°C for 48 hours. The mean weights of these ten samples from each level were as follows: top, x = 12.9 ± 3.4 g; middle, x = 8.7 ± 2.1 g; bottom, x = 6.8 ± 1.8 g. Buds from the
146
R. A. HUEMPFNER AND J. R. TESTER
Table 4.4. Productivity indexes of male flower buds at selected aspen clones used by ruffed grouse Productivity indexes
Clone no. and species3 e
4333 4158 3982 e 4170 3902 4090 4007 e 2670 e 2584 e
(TA) (TA) (BTA) (TA) (TA) (BTA) (TA) (TA) (TA)
Mean
No. of trees with male buds 16 RPIf 37 2 28s 17h 8 161 15 17j
Winter 1971-72
Winter 1972-73
Winter 3-year mean 1973-74 (spring indexes Fall" omitted)
Winter 6
Spring0
Fall"
125 310 405 195 150 475 225 375 425
7 212 47 30 122 210 43 335 100
85 372 245 105 275 309 115 386 258
168 246 325 193 205 175 125 325 125
126 309 325 164 210 320 155 362 269
298
123
239
210
249
a
TA = trembling aspen; BTA = big-tooth aspen. Indexes calculated on 10 January 1972 and do not include buds removed during budding in November, December, and early January that year. c Indexes calculated on 10 April 1972 after winter budding had ended. d Indexes calculated on 12 October each year before budding had started. e Traditional clones used during all three winters. f RPI = Relative Productivity Index. Procedure: mean number of male buds/twig x % of living canopy producing male buds (Svoboda 1972, Svoboda and Gullion 1972). g January 1972, 28 living trees; October 1972, 18; October 1973, 16. h January 1972, 17 living trees; October 1972, 12; October 1973, 12. i January 1972, 16 living trees; October 1972, 11; October 1973, 8. j January 1972, 17 living trees; October 1972, 12; October 1973, 8. b
top were significantly heavier than ones from the middle and bottom (P < 0.01), and buds from the middle weighed significantly more than those at the bottom (P < 0.05). There was a 33 % difference in bud weight between top and middle samples and 47% between top and bottom. These initial tests indicated that bud weight may have been an important factor dictating canopy, and perhaps clone, preference. It was doubtful that this variation in bud weight was characteristic of feeding trees alone. In 1977, 15 trembling aspens were cut in the same vicinity. Inspection of twigs suggested that the trees had not been used, or had received only limited use that winter by budding grouse. Five parameters were measured from these bud and twig samples (Fig. 4.8). The mean weight of 100 male bud samples from each canopy level indicated that buds at the top (i = 12.3 + 0.1g) were signi-
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
147
ficantly heavier than those from the middle (x = 9.9 ± 0.1 g) and the bottom (x = 8.9 + 0.1 g). Bud weights from the middle and bottom were not significantly different. However, three-way comparison was also significant (P < 0.01). In addition, twigs were larger in diameter at the top, with the overall comparison for the three levels significant at P < 0.05. Twigs at the top were also significantly longer and heavier than those from the middle and bottom levels (Fig.
Fig. 4.8. Comparison of five parameters measured for 100 male buds and ten twigs from three canopy levels (top, middle, bottom) for 15 trembling aspen trees. One-way ANOVA. Two canopy-level comparisons have 1, 28 df; three canopy levels 2, 42 df; *P < 0.05, **P < 0.01.
148
R. A. HUEMPFNER AND J. R. TESTER
4.8). Finally, there were also significantly more buds per twig in the top canopy than the bottom (P < 0.01); top-middle and middle-bottom comparisons showed no significant differences.
4.3.11 Total food consumed Nearly continuous observations of budding grouse along the transect during 1971-72 and 1972-73 provided sufficient data to estimate the actual weight of buds and twigs consumed at selected male clones. A working knowledge of five aspects of winter use of aspen was required to calculate total consumption: (1) mean dry weight of a single bite of food for crops collected from 19 ruffed grouse in winter (Huempfner 1981; (2) number of aspen trees required by each grouse to provide food for the entire winter; (3) number of total feeding sessions (morning and evening) at each clone; (4) mean duration of feeding; and (5) the mean rate of budding (bites/minute). The assumptions and resulting calculations are presented in Table 4.5. Total estimated consumption of aspen buds and twigs at each clone should be considered a minimum value because midday feeding is not included. We estimated that ruffed grouse consumed 44,270 g of food (dry weight) in 1971-72 from nine clones totaling 156 trees. This dropped to 10,719 g in 1972-73, only 24.2% the amount used the previous winter. As stated earlier, snow depth in 1972-73 was much reduced and a surface crust was present during nearly the entire season, thus allowing use of alternate food resources such as hazel catkins. Assuming equal budding of all trees within this 156-tree sample, ruffed grouse removed an average of 284 g (dry weight) of material from each tree in 1971-72 and 82 g the following winter. The fresh-weight equivalence, assuming 47% moisture as determined from crop samples, would total 536 and 155 g/tree, respectively. Field data on the duration and rate of feeding showed that 11.5 g of dry matter were consumed in the morning and 23.4 g during evening periods. Converting this to fresh weight, grouse consumed 21.6 g and 44.0 g, respectively. This is about half the fresh weight of crops from two grouse collected at Cloquet as these birds completed evening feeding (x = 97.5 g; Svoboda & Gullion 1972). The 44.0 g evening crop weight is, however, comparable to the fresh weight of crops from eight grouse collected during this study in January 1973, also near the termination of evening feeding (x = 37.2 + 4.1 g; range 28.1-59.0 g). Using the dry-weight value of 34.9 g (morning and evening combined), determined from budding observations in this study, the number of grouse-use days of feeding provided by all trees in this sample was calculated. Each clone provided from 1.7 to 30.4 grouse-use days per tree, a mean of 8.1 grouse-use days/tree in 1971-72 and 2.4/tree in 1972-73 (Table 4.5). A spring bud index at nine clones (Table 4.4) conducted after winter budding had terminated in 1971-72 showed that 47% of the buds present in early January were still present
Table 4.5. Estimated food consumed3 at nine male aspen clones during two winters of varying snow quality
Clone no. and speciesb
No. of trees with male buds
Budding sessions/ clonec
Dry weight (g) Consumption/ Consumption/ cloned tree6
Grouse-use days/treef
Winter 7977-72 (deep, soft snow) H333 (TA) 4158 (TA) g 3982 (BTA) Ml 70 (TA) 3902 (TA) 4090 (BTA) g 4007 (TA) 2670 (TA) 2584 (TA) Total
16 37 2 28 17 8 16 15 17 156
245 183 120 53 154 213 144 173
4,999 3,285 2,119 19,370 1,039 3,011 4,250 2,717 3,480
2,341
44,270
1,056
312 89 1,060
692 61 376 266 181 205 x = 284b
8.9 2.6 30.4 19.8
1.7 10.8
7.6 5.2 5.9 x = 8.21
Winter 1972-73 (shallow, crusted snow) 4333 (TA) 4158 (TA) 3982 (BTA) 4170 (TA) 3902 (TA) 4090 (BTA) 4007 (TA) 2670 (TA) 2584 (TA) Totals a
16 37 2 18 12 8 11 15 12
16 294 0 154 34 0 0 17 48
256 5,503 0 2,880 618 0 0 398 1,064
131
563
10,719
16 149 0 160 52 0 0 27 89
x = 82h
0.5 4.3 0.0 4.6 1.5 0.0 0.0 0.7 2.6
x = 2.41
Food consumed includes male flower buds, vegetative buds, and twigs. TA = trembling aspen; BTA = big-tooth aspen. c Only morning and evening budding were considered. In 1971-72, total budding sessions in December, when limited observatons were conducted, were corrected upward by assuming a similar proportion of budding increase indicated for this month in 1972-73 (Table 4.2) and 1973-74 (Table 4.3). d Morning consumption = (duration of budding, 12.2 minutes) (rate, 17.8 bites/minute) (mean weight of each food item, 0.053 g) (number of budding sessions) + evening consumption = (duration of budding 17.6 minutes) (rate, 25.1 bites/minute) (Mean weight of each food item, 0.053 g) (number of budding sessions) = total consumption/clone. e Consumption/tree = (dry weight consumed/clone -=- number of trees with male buds/clone). All trees assumed to be used equally. f Grouse-use days/tree = (dry weight consumed/tree) -=- (34.9, dry weight consumed/grouse/day during morning and evening budding). Consult formula in footnote d. g Approximately 80% of the male flower buds present were consumed at these clones in 1971-72. h Mean consumption/tree = (consumption at all clones) H- (number of trees at all clones). i Mean grouse-use days/tree = (grouse-use days/tree/clone) (Number of trees/clone) 4- (total number of trees at all clones). b
150
R. A. HUEMPFNER AND J. R. TESTER
in spring, indicating that an average of 53% had been removed by ruffed grouse during the winter. The minimum number of trees required per grouse for an entire winter was calculated by using data from four clones where grouse consumed 80-90% of the male bud crop in 1971 -72 (Table 4.5), and assuming that this rate of consumption was the maximum before grouse abandoned a clone. We also assumed an average of 120 days of permanent snow cover. In 1971-72, the four clones mentioned above provided an average of 14.2 grouse-days of feeding per tree (average weighted for the number of trees present at each clone). Using this figure as the maximum allowable grouse-days per tree, with the assumptions stated above, an individual grouse would require 8.4 mature aspen trees (120/14.2) to sustain it through the winter. This value also assumes an RPI of 260 and similar proportions of buds and twigs consumed as indicated by crops analyzed during the study. If the RPI was decreased substantially, and/or the ratio of buds to twigs increased, and/or the weight of male buds decreased, the total number of trees required could easily double or triple. Svoboda and Gullion (1972) estimated, based on consumption of male buds only, a maximum of 8 or 9 bird-use days per tree. At that rate, 13 to 15 trees would be required for a 120-day budding period per grouse per winter. Huff (1973 p. 137) suggested a "minimum number of 14 mature male aspens per grouse" were needed for the winter, with a range as high as 112. His study was also based on male buds only. We suggest that requirements lie within our range of 8.4 and 20 productive male aspen trees per grouse per winter. Seiskari (1962:81) suggested that 2.5-7.5 pines (Pinus sylvestris) were required for a single capercaillie in Finland in winter to "exhaust the expendable resource" of needles. He also estimated that 12 to 14 birch trees (Betula verrucosa, B. pubescens) were needed by an individual black grouse to provide adequate winter food.
4.3.12 Aspen food requirements and availability All available data pertaining to arboreal budding and aspen food production collection during this study were used to estimate total consumption and availability of the aspen resource at male trees used by grouse along the entire transect. Projected peak consumption assuming maximum numbers of budding grouse present during a 30-day period for each year were also estimated. Calculations predicted that 63.5 % of the available aspen food resource was actually consumed by ruffed grouse during 1971-72 (Table 4.6). Had peak numbers of grouse present in January of that year used these clones for the entire 120-day interval, nearly the entire food resource (93.8%) would have been consumed. The number of grouse-use days dropped substantially during the following two winters and may be characteristic of budding requirements during 5 to 7 years of each grouse fluctuation at Cedar Creek. Availability of the winter food re-
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
151
source also decreased during these years, caused by the loss of budding trees and reduction in the RPI. Only 25.7% (1972-73) and 37.6% (1973-74) of the available aspen resource was actually consumed (Table 4.6). Projected peak budding would still have utilized only 31.6 and 45.1 % of the available resource, respectively. These estimates, supported by substantial field data, suggest that the intensity of budding during times of ruffed grouse population highs at Cedar Creek can cause a pronounced reduction in food resources at preferred aspen clones. Earlywinter grouse densities probably could not have been sustained in 1971-72 by the resources available at preferred clones alone. The potential threat of an aspen "eat-out" was lessened in 1971-72 when radio-tagged grouse, and presumably untagged birds as well, moved to locations where they could not be seen from the transect during budding observations (Huempfner, unpubl. data). Predation also helped to reduce the winter population (Huempfner et al., unpubl. data). As the Table 4.6. Aspen food resource consumption, projected peak consumption, and availability during three winters at Cedar Creek, Minnesota Variable
1971-72
1972-73
1973-74
Estimated days required 3
1,894
452
545
Available days at used clones'3
2,982
1,795
1,451
% of available days consumed 0
63.5
25.7
37.6
Potential peak days required d
2,797
% of available days potentially consumed 6
93.8
a
b
c d
e
568
31.6
654
45.1
Includes all known and estimated budding during the morning, midday, and evening periods. Number of midday budding sessions in winter 1972-73 and 1973-74 calculated from proportional use observed in winter 1971-72. Morning budding in winter 1973-74 estimated from proportional use in winter 1972-73. Total estimated grouse-use days obtained from the above calculations were reduced by 14.5% (1971-72), 10.2% (1972-73), and 8.9% (1973-74), the percent of all known budding observations that occurred at trees other than male aspens. All male aspen trees (trembling and big-tooth) along the 2 km transect used by grouse in one or more winters were included in this estimate. Available grouse-use days calculated from number of living, male-bud-producing, and used trees during each winter and assuming 80% consumption of buds at each tree as a maximum. Available grouse-days were also corrected for changes in the RPI. Available grouse-use days: 1971-72 = (210 trees) (14.2 grouse-use days/trees); 1972-73 = (176 trees) (10.2 grouse-use days/tree); 1973-74 = (163 trees) (8.9 grouse-use days/tree). (Estimated days required) -^ (available days) = % consumed. Based on maximum number of grouse observed budding during three consecutive data periods and projected for entire winter each year (120-day projection). (Potential peak days required) -=- (Available days) = % consumed.
152
R. A. HUEMPFNER AND J. R. TESTER
grouse population declined the following 2 years, budding pressure on preferred aspen clones was reduced substantially.
4.4 Conclusions Numerous studies have shown the relative importance of aspen in the winter diet of ruffed grouse, especially in the northern tier of states, Canada, and Alaska. During the present study, trembling aspen was used extensively in all three winters. Big-tooth aspen was used heavily when grouse populations were high, substantially less when densities were intermediate, and was not used when populations were low. Total use of paper birch and black cherry also declined as grouse density declined at Cedar Creek, but not as markedly as that of big-tooth aspen. Of 7,341 budding observations in all three winters, 87% was in trembling aspen, 5.6% in big-tooth aspen, 3.2% in paper birch, 2.5% in black cherry, and 1.7% in six other species. Arboreal budding was concentrated during the snow cover period and peaked in late December or early January each year. The frequency of arboreal observations in December (the first month of snow cover) was five to seven times greater than during November (largely snow free). Grouse crop samples confirmed this trend and indicated 8.2% aspen in the diet in September-November compared with 30.2% during December-February. Heavy budding at highly preferred clones reduced the male flower-bud supply in 1971-72 by 80-90%. Overall consumption at all aspens used for feeding that year was estimated at 63.5% of the available male bud crop (Table 4.5). Consumption during 1972-73 and 1973-74 was sharply reduced and accounted for 25.7% and 37.6% of available male buds. Grouse were under considerable competition when the population was high in 1971-72. They responded by utilizing a wider variety of food resources, as would be predicted by most foragingstrategy models (Pyke et al. 1977). The number of displaying territorial males was substantially reduced in spring 1973 following the most severe winter during this study (Table 4.7). The fall flushing rate also continued to decline (Table 4.1). The spring decline continued in 1974 following a winter with generally favorable snow conditions. Consumption of male flower buds along the transect was substantially reduced in each winter, and no shortage occurred at even the most heavily used aspen clones (Table 4.7). These data suggest no direct relationship between the quantity of winter food and the density of territorial males the following spring (Table 4.7). Breeding densities declined throughout the study, but the decline was least following the winter of most-intense food competition in 1971-72 (Table 4.7). Once the decline began, the number of males was further reduced following a winter of favorable snow conditions. The fall population, based on flushing rates, apparently declined
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
153
Table 4.7. Summary of snow conditions, food competition and utilization, and population trends during three winters at Cedar Creek, Minnesota
a
Winter
Snow roosting conditions
Food competition
Aspen utilization (%)
Change in drumming males3
1971-72 1972-73 1973-74
best worst intermediate
high low intermediate
64 26 38
0.93 0.64 0.44
Proportional change; males in spring of year 2 -H males in spring of year 1.
immediately following both intense winter competition for food (1971-72) and severe winter roosting conditions (1972-73) (Table 4.1). Our data indicate that competition for aspen buds is more a function of snow cover than bird densities.
4.4.1 Foraging strategy Ruffed grouse at Cedar Creek exhibited an overall preference for buds and twigs present in the upper third of the canopy in aspen (55.3% of all observations) over those in the middle (38.5%) and bottom levels (6.3%). Intensity of use of the upper canopy varied between years and appeared related to grouse density. In 1971-72, when grouse density was high, the upper canopy was used 64.6% of the time during evening feeding, compared with 52.8% in 1972-73 (medium population level) and 43.5% in 1973-74 (low population). Use of other tree species also increased in 1971-72, and this was the only year when several highly preferred clones were stripped of all but a few male buds. Bud and twig sampling from 17 trembling aspen suggests that: (1) grouse budding in the top canopy are selecting male buds that are 27-47 % heavier than those at the bottom and 20-33 % heavier than those in the middle levels of the same tree; (2) twigs in the upper canopy produce a more concentrated bud resource than other canopy levels (i.e., more buds/twig); (3) terminal twig fragments, a common component found in winter crops, may be heavier in the upper canopy than those obtained from other canopy levels; and (4) although twig diameters near the terminal end are larger in the upper canopy, this apparently poses no barrier to the utilization of this food resource. The mean duration of morning and evening budding was 12.2 + 0.9 and 17.6 + 1.0 minutes, respectively. The overall mean feeding rate was 25.1 + 0.9 bites/minute. Each bite was assumed to represent one bud or twig consumed. Grouse in flocks (two or more birds) fed at significantly faster rates than those feeding alone. Grouse budded in an average of 3.0 trees per feeding session, and a single grouse required a minimum of 8.4 aspen trees to sustain it through a typical winter. Continuous observation of individual grouse during entire feeding sessions
154
R. A. HUEMPFNER AND J. R. TESTER
also suggests that they bud at a 35% faster rate (bites/minute) in the second 10minute interval of feeding than in the first. This period of accelerated feeding coincides with, and may be facilitated by, budding in the upper canopy. Because buds and twigs in the upper canopy are heavier and higher in total energy (Huff 1970, 1973), ruffed grouse can maximize both quantity and quality of food ingested by budding in the upper canopy of aspen in winter. Upper-canopy preference, however, declined significantly as the grouse population declined from 1971 to 1974. This change in feeding behavior implies that a reduction occurred both in the quantity of food consumed per unit of time spent feeding and in the total energy obtained among winters. Behavior at feeding clones also underwent changes as the population declined. Under high and medium grouse densities, a portion of the population fed in "shifts" during the evening period, appearing at clones earlier, budding, and leaving before the main feeding period began. This behavior was most pronounced in 1971-72 (high grouse densities) and was not observed in 1973-74 (low densities). Initially, these factors seem to disagree with optimal foraging theory and the basic assumption that animals should be either time minimizers or energy (food) maximizers and hence should forage in such a way that their net rate of energy (or food) intake is maximized (Pyke et al. 1977). Pyke et al. further state, however, that whether or not an animal falls into one of these categories depends on whether or not other factors, such as predation or nutritional requirements, operate at the same time and behave in a manner that imposes costs on a tactic of increasing the net rate of energy (or food) intake. Our findings suggest that there are, indeed, behavioral conflicts operating that prevent ruffed grouse from feeding exclusively in the upper canopy of aspens, where the quantity and quality of food are greatest. The upward movement of grouse within the canopy as budding progressed may be a response, in part, to the greater security provided by lower canopy levels during initiation of arboreal feeding each day. The short period of roosting in lower levels, which accompanied arrival at clones, may provide an opportunity to survey the surrounding cover for predators. Although we agree with Doerr et al. (1974) and Morse (1970) that predation per se during foraging is relatively unimportant, Huempfner et al. (unpubl. data) and Gullion (1970c) portray the overall winter mortality of ruffed grouse through predation as an important part of total annual mortality. Ruffed grouse must remain alert to predators at all times, a response that appears to persist during arboreal feeding. After an initial period of surveillance and feeding in the lower canopy, grouse may be reassured of their own safety and gradually move to the upper levels where the rate of budding and caloric intake can be maximized. Reductions in the size of arboreal feeding aggregations may also be implicated. Evening feeding aggregations averaged 4.4 grouse in 1971-72 when grouse numbers were high. This was reduced to averages of 2.8 and 2.9 grouse
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
155
the following winters, during times of medium and low grouse densities, respectively. The proportion of grouse feeding alone increased in 1972-73 and 1973-74. Smaller flocks may have reduced surveillance capabilities, thereby reducing time available for foraging and/or time spent in the upper portion of the canopy. Surveillance strategy related to foraging has been stressed by Morse (1970) and Powell (1974) for a number of avian species. The significant shift to later evening feeding by ruffed grouse during the current study also suggests that a basic shift occurred in their feeding strategy. Age ratio changes of ruffed grouse as the population declined and genetic shifts to a more wary phenotype during times of peak densities (Keith 1963, Chap. 3) or immediately following peaks (Marshall 1954) may also be involved. In addition, changes in the color phase and its implied relationship to food utilization, physiological stress, and ultimately survival may be related (Gullion 1970c). The nutritional quality of aspen buds and twigs within each canopy level, a variable untested by this study, also may have partly dictated changes. The winter budding strategy of ruffed grouse at Cedar Creek appears to be a trade-off between an optimal foraging strategy based on maximizing the quantity and quality of food consumed and behavioral traits that limit this strategy from being fully implemented. The basic arguments for an optimal foraging strategy as summarized by Pyke et al. (1977) are: "(1) Behavior in general, and foraging behavior in particular, show heritable variation; and this entails variation in the contribution to subsequent generations; (2) There is a range of possible foraging behaviors; and, (3) Natural selection will favor those individuals which contribute the most to subsequent generations. Hence, natural selection will result in a change with time of the average foraging behavior in the population, towards that foraging behavior in the range of possible behaviors which gives maximum fitness. It is assumed (usually implicitly) that the rate with which these changes occur is much greater than the rate with which the position of the maximum fitness behavior changes." We contend that all of the required elements necessary for a basic change in ruffed grouse foraging strategy to occur (e.g., varied food availability, competition, predation, a declining population and associated genetic shifts, and variable snow depth and quality) were present during this study. Physical factors (especially variations in snow cover) and ruffed grouse population levels probably determined the changes in budding intensity observed during the three winters. Behavioral factors, however, may have been responsible for the major changes in timing and canopy preference during winter budding.
4.5 Summary Arboreal budding by ruffed grouse at Cedar Creek, Minnesota, was concentrated during the snow-cover period, peaking in late December or early January 1971-72 through 1973-74. The frequency of arboreal observations in December
156
R. A. HUEMPFNER AND J. R. TESTER
(the first month of snow cover) was five to seven times greater than during November (largely snow free). Grouse crop samples confirmed this trend and indicated 8.2% aspen in the diet in September-November compared with 30.2% during December-February. Male trembling aspen buds and twigs were the predominant winter food. Of the 7,341 budding observations in three winters, 87% were in trembling aspen, 5.6% in big-tooth aspen, 3.2% in paper birch, 2.5% in black cherry, and 1.7% in six other species. Big-tooth aspen was used heavily at Cedar Creek only when grouse densities were high. Paper birch was common at Cedar Creek but was used sparingly. Scattered black cherry trees received extensive use compared with their low availability. Ruffed grouse preferred feeding in the top canopy level of aspen (55.3%), compared with middle (38.5%) and bottom (6.3%) levels. Intensity of use of the upper canopy varied among years and appeared to be related to grouse density. In 1971-72, when grouse density was high, the upper canopy was used 64.6% of the time during evening feeding, compared with 52.8% in 1972-73 (medium population level) and 43.5% in 1973-74 (low population). Use of other tree species also increased in 1971-72, and this was the only winter when several highly preferred clones were stripped of all but a few male buds. Behavior at feeding clones also underwent changes as the population declined. Under high and medium grouse densities, a portion of the population fed in "shifts" during the evening period, appearing at clones earlier, budding, and leaving before the main feeding period began. This behavior was most pronounced in 1971-72 (high grouse densities) and was not observed in 1973-74 (low densities). The mean duration of morning budding was 12.2 minutes; evening budding averaged 17.6 minutes. The overall mean feeding rate was 25.1 bites/minute. Each bite was assumed to represent one bud or twig consumed. Grouse in flocks (two or more birds) fed at significantly faster rates than those feeding alone. Grouse also fed faster as the feeding period continued. This coincided with and/or was a result of feeding progressively higher in the canopy as budding continued. Grouse budded in an average of 3.0 trees per feeding session, and a single grouse required a minimum of 8.4 aspen trees to sustain it through an average winter. We estimate that 64, 26, and 38% of the aspen buds in preferred clones were utilized in the winters of 1971-72, 1972-73, and 1973-74, respectively. Over the same period the number of drumming males declined from 30 (1971) to eight (1974). However, this decline appeared unrelated to bud utilization; the winter of maximum use (1971 -72; 64 %) was followed by the smallest change in number of drumming males (from 30 to 28, 7%) decline). Behavioral conflicts appeared to have prevented ruffed grouse from following a strict, optimal foraging strategy. This was especially true in the winter of 1973-74, a time of low grouse densities and reduced flock sizes. The main
WINTER FEEDING BEHAVIOR OF RUFFED GROUSE
157
conflict may be related to maintaining a keen awareness of potential predators. The winter budding strategy of ruffed grouse at Cedar Creek appears to be a trade-off between an optimal foraging strategy based on maximizing the quantity and quality of food consumed and behavioral traits that limit this strategy from being fully implemented.
5
Spatial Patterns, Movements, and Cover Selection by Sharp-tailed Grouse M. W. Gratson
5.1 Introduction The sharp-tailed grouse (Tympanuchus phasianellus) is one of three North American tetraonids that live in steppe habitats—grasslands, savannas, and sagebrush (Artemisia spp.) plains —that show communal advertising by males at leks, and that are the most mobile of grouse. Demographic aspects of the lekking grouse also set them apart from the ptarmigan (Lagopus spp.) and forest species, perhaps as a result of unique characteristics associated with the steppe. This chapter presents evidence of changing spacing behaviors, daily and seasonal movements, and cover-selection patterns of male and female sharp-tailed grouse from data gathered during a study of three populations in Wisconsin. These findings are particularly useful in integrating results from other recent research, which has dealt specifically with nesting and brood-rearing stages (Artmann 1970, Christenson 1970, Pepper 1972, Schiller 1973, Kohn 1976, Ramharter 1976) and male territoriality (Rippin & Boag 1974a,b, Kermott & Oring 1975, Sparling 1979, 1983, Moyles & Boag 1981, Kermott 1982), and in formulating some general hypotheses concerning the strategies that sharp-tailed grouse have developed to survive and compete in the variety of open brushlands of North America. Although some of the behaviors discussed cannot be considered strategies themselves (Wittenberger 198la), they are nevertheless ecologically important to individuals and to the dynamics of a population as a whole. Finally, management biologists may find the material helpful in dealing with the lekking, open-land grouse, and the sharp-tailed grouse in particular, as resources or native fauna requiring management guidelines. 158
SHARPTAIL SPACING, MOVEMENTS, AND COVER
159
5.2 Study areas This research was conducted from April 1977 to July 1979 in northwestern Wisconsin (46°N, 92°W). Three study areas were chosen that varied in size, habitat complexity, and grouse population densities. The Douglas County (DC), Namekagon Barrens (NB), and Crex Meadows (CM) Wildlife Areas are within the state's 3,900-km2 Northwestern Pine Barrens, a glacial outwash plain of sandy soils that is level to gently rolling, sparsely pitted, and has a history of repeated wildfires. Douglas County (T44N, R12W) (1,522 ha), described as an oak (Quercus spp.)-aspen (Populus spp.)-pine (Pinus spp.) savanna (Hamerstrom 1963), consisted of 33% conifer and deciduous, wooded uplands (individual plant stems > 10 cm dbh dominating); 19% upland shrub, shrub-tree, and shrub-grass cover types (shrubs predominating); 44% grass-forb and grass-shrub cover types (grasses predominating); 3% sedge-meadows and shrub-marshes, and 1% (15 0.5-ha) clover, in dispersed patches (Fig. 5.1). This area is 5 km southwest of Solon Springs. Lands within 3 km of DC are forests and wetlands to the east, dairy and grain farms interspersed with woodlands northward, and wetlands to the south and west. The sharp-tailed grouse population at DC, suggested by counts of displaying cocks, was stable or increasing slightly during the period of study (Fig. 5.1). At only this area was the density of males, at approximately 1.0/km2, comparable to those at study areas in Wisconsin and Upper Michigan during the 1940s and 50s, as reported by Grange (1948) and Ammann (1957). There were approximately half the number of males in spring during 1977-79 as on the same area during the years 1950-57 (Hamerstrom unpubl. data). At NB (T42N, R14W), the northern unit (1,849 ha) was selected as an intensive study area. This large block of relatively homogenous shrub prairie, consisted of 7% conifer and deciduous woods; 76% shrub, shrub-tree, and shrubgrass cover types; 7% grass-forb and grass-shrub cover types; 8% pine slash; 1 % sedge-meadows and shrub-marshes; and 1% (9 1-ha) buckwheat and clover, in patches scattered throughout (Fig. 5.2). Forests in each direction and shrubmarshes to the north surround NB, which is 19 km west of Minong and 22 km southwest of DC. Counts of displaying cocks suggested the grouse population at DC was relatively constant from 1977 to 1979 (Fig. 5.2). Only two leks were present, although in 1977 a temporary shift of one occurred in mid-April, lasting through the remainder of the spring. Crex Meadows (T39N, R18W), a 12,185-ha unit of wetlands interspersed with shrub prairie and wooded uplands, was composed of 38% conifer and deciduous woods; 12% shrub, shrub-tree, and shrub-grass communities; 2% grass-forb and grass-shrub cover types; 19% deep marsh and open water; 28%
160
M. W. GRATSON
Fig. 5.1. Sharp-tailed grouse study area in northwestern Wisconsin, Douglas County Wildlife Area (DC), showing the spacing and locations of leks and the number of males per lek, 1977-79. Leks connected to black dots indicate that shifts in their locations occurred to the indicated spots.
sedge-meadows and shrub-marshes; and 1 % corn, rye, oats, buckwheat, and hay annually cultivated by the Wisconsin Department of Natural Resources (Fig. 5.3). Three platform corn-feeders were maintained for wildlife within a central 121-ha refuge during winter and early spring. Lands surrounding CM include pine and oak forest uplands and river-bottom hardwoods. Dairy and grain farms occur to the south. Crex Meadows lies 2 km north of Grantsburg and 45 km southwest of the NB. Although the density of males at CM (x = 0.93/km2) was lower than at DC, the CM population was clearly increasing (Fig. 5.3); the number of advertising males steadily rose during the period 1971-79 (Ramharter 1976, Evrard, Toepfer
SHARPTAIL SPACING, MOVEMENTS, AND COVER
161
Fig. 5.2. Sharp-tailed grouse study area, Namekagon Barrens Wildlife Area (NB), 1977-79.
unpubl. data). A newly reintroduced population of prairie chickens (Tympanuchus cupido) was also present; the number of cocks ranged from 16 to 25 during the 3 years (Toepfer & Anderson unpubl. data). Jack pine (P. banksiana) is the dominant conifer in this region. Tree species in deciduous upland woods include oaks (Q. macrocarpa, ellipsoidalis), aspens (P. tremuloides, grandidentata), and less commonly, a birch (Betulapapyriferd). Shrub and shrub-grass communities contain small pine, oak, and aspen (< 10 cm dbh), and in addition willow (Salix discolor, humilis) and hazel (Corylus americana) as characteristic species. Bluestem grasses (Andropogon gerardi, scoparius), bluegrasses (Poa pratensis, compressa), junegrass (Koeleria
162
M. W. GRATSON
Fig. 5.3. Sharp-tailed grouse study area, Crex Meadows Wildlife Area (CM), 1977-79.
cristatd), sedges (Carex pensylvanica, umbellata), blueberry (Vaccinium angustifolium), sweet fern (Comptoniaperegrina), and various forbs are dominant in grass-forb and grass-shrub cover types. Sedge-meadows consist of Carex stricta, rugosa, and oligosperma, bluejoint reed grass (Calamagrostis canaden-
SHARPTAIL SPACING, MOVEMENTS, AND COVER
163
sis), and bulrush (Scirpus cyperinus). Shrub-marshes also contain leatherleaf (Chamaedaphne calyculatd), Labrador tea (Ledum groenlandicum), willow, and bog birch (B. pumild). Spirea (Spiraea alba) and goldenrods (Solidago spp.) in shrub-grass communities and bluejoint reed grass on mesic and wet-mesic sites in grass-shrub communities also occur at CM. Further details of vegetation on these study areas are available (Hamerstrom 1963, Vogl 1964, Beck & Vogl 1972, Gratson 1983).
5.3 Methods Approximately four thousand radio locations of 26 male and 13 female sharptailed grouse provided many of the data that this paper is based upon. Specific details of methodology are available elsewhere (Gratson 1983); in brief, grouse were captured during all seasons and age and sex were determined using standard techniques (Snyder 1935, Gower 1939, Ammann 1944, Henderson et al. 1967). Juveniles were considered to become yearlings on 1 April each year, and yearlings to become adults on 1 September. Birds were radio-tagged with a backmounted radio package (Dumke & Pils 1973) and located one to three times per day following a rotational sampling schedule among study areas. This included at least three to four sample periods (each 3-7 consecutive days) per month per area. Frequently birds on two or three study areas were sampled on the same days. Estimated location error was + 33 m, (3,421 m2) (n = 46 trials) for stationary birds using a vehicle-mounted receiving system. Observations of birds supplemented these data gathered using telemetry. On average, radio-tagged individuals were "flushed" twice monthly to determine their flocking tendencies. I estimated the size of home ranges using the minimum-convex-polygon method (Mohr 1947) for birds located at least 15 times per month. For the same birds, I assigned cover types to their locations for cover use-availability comparisons (Neu et al. 1974). To determine dispersion and spatial patterns of birds, I measured the distance from each radio location to: (1) the home display ground (where males advertised and for only those females captured at leks); (2) the nearest display ground, which may also be the home lek; and (3) the nest, for hens. I supplemented these with home-range overlap data from the same birds. To estimate movements, I measured the distance from: (1) each radio location to a randomly selected location of the previous day (day-to-day movement); (2) each night location (assigned to the date of the preceding daylight period) to the previous night's location (night-to-night movement); (3) each daylight location to the night location of the previous day; and (4) each night location to a daylight location of the same day. Daylight-to-night and night-to-daylight data (3 and 4 above) were subsequently combined (daylight-to-night/night-to-daylight movement).
164
M. W. GRATSON
5.4 Results Male sharp-tailed grouse started regularly advertising at leks during early to mid April and continued through May. They irregularly visited grounds in March and June, and I observed cocks on leks in all months except July and August. Evening display during late April and May was noted only at the largest lek (A, Fig. 5.3), though birds from smaller leks often moved to areas near (< 50 m) their advertising ground in the afternoon. The period when most hens first mated on leks was the last week in April through the first week of May. For example, at NB, where there were only eight to ten cocks in 1978, eight hens copulated at lek B (Fig. 5.2) on 29 April.
5.4.1 Spatial patterns and dispersion of the sexes Display ground spacing, estimated by the mean distance between neighboring leks, was similar among years (P > 0.05) but different among areas (P < 0.01) (Figs. 5.1-5.3). There was no measurably significant relationship between lek spacing and lek density (r = 0.29, 7 df, P > 0.05), spring lekking cock density (r = 0.28, 7 df, P > 0.05), or to the mean number of males per lek (r = 0.57, 7 df, P > 0.05). Using the overall mean, lek spacing in northwestern Wisconsin was similar to that for two study areas in Alberta (Evans 1961, Rippin & Boag 1974b) and one in North Dakota (Bernhoft 1969, Christenson 1970), but less than those I computed from Grange (1948), Ammann (1957), Pepper (1972), and those reported by Sisson (1976) for sharp-tailed grouse in Wisconsin, Michigan, Saskatchewan, and Nebraska, respectively. Some leks were temporary. Displaying males disappeared or shifted locations (Figs. 5.1-5.3). Shifts occurred in April, between spring and fall, and between fall and spring; thus, they took place suddenly by entire groups of birds and gradually, between breeding seasons. Evidence indicates that the establishment of one lek and the relocation of a second and third were associated with the spring locations of females. Males spent more time nearer to home and to the nearest lek from early to late spring (Fig. 5.4). During May, adult and yearling males fed and roosted an average of only 353 + 80 m from home leks. Adults generally continued to remain at similar distances through June, though some yearlings moved farther away. Leks were often near the edge of the areas used by individual cocks for feeding and roosting during most of the display period (Fig. 5.5). Locations of birds during early April, before the period of intense display, were less confined to areas immediately surrounding leks, or to areas to one side of leks. One yearling male at NB did not display at a lek although he remained near one and advertised (Fig. 5.5, lower left). Ranges of males did not overlap with those of males at neighboring leks, even where the distance between two grounds was less than half the mean; this was consistent with data from Alberta (Moyles 1977).
SHARPTAIL SPACING, MOVEMENTS, AND COVER
165
Fig. 5.4. Monthly mean distances (m) ( + SE) of 26 male and 13 female sharp-tailed grouse from their home (males) and nearest leks (both sexes), 1977-79. Males and females were spatially separated during the period when males advertised at leks and females were nesting (mostly May).
166
M. W. GRATSON
Females were closest to display grounds during the prelaying period (mainly April) and also later, in November (Fig. 5.4). During prelaying, hens traveled widely (Fig. 5.6), although for one adult this was mostly a directional move to her nesting range. Two other hens (not shown), captured at leks, showed lesspronounced mobility, though again one adult moved directionally 4.3 km from the lek to where she eventually nested. Ranges of hens overlapped with those of cocks during the prelaying period, not only because hens probably copulated at leks, but also because both sexes used areas away from leks and their ranges were relatively large. The establishment of one CM lek (H, Figs. 5.3,5.6) and the "satellite" (F) ana shift (I) of two others in 1978 may have been influenced by the locations of hens during the prelaying period (Fig. 5.6). Two cocks established a display ground within the March and early April ranges of three radio-tagged hens where there had been no lek the previous year. Similarly, a satellite lek formed south of display ground F by one radio-tagged yearling and at least one other untagged male, and the shift in lek I to the east, appeared related to the activities of radio-tagged and other hens in these areas. Whether hens mated at the new lek locations is unknown. During the nesting period (mostly May), unlike during the preincubation phase, hens were on average well away from cocks and leks (Fig. 5.4). Hens averaged 1,024 + 287 m from the nearest lek, much farther (P < 0.01) than males. Four of six nests (including a probable first nest destroyed by fire) were near the edges of laying-incubation ranges, though hens apparently fed both near and far from nests during incubation (Fig. 5.7). Nests averaged 95 + 33 m from laying-incubation range perimeters, 342 + 168 m from prelaying range perimeters, and 998 + 329 m from the nearest lek. These findings are supported by data from other regions (Bernhoft 1969, Christenson 1970, Pepper 1972, Schiller 1973, Kohn 1976, Ramharter 1976), though these earlier works lacked information on both hens and cocks. Some males moved away from their display grounds during the summer, whereas others did not. By August, males were often closer to display grounds other than their own (P < 0.01) (Figs. 5.4, 5.5—lower right). As a result, the ranges of males from neighboring leks overlapped, although the birds were not members of the same flocks. Why some males moved away from home is unclear. The mean distances of cocks from home leks in August were not clearly related to parameters of lek status, e.g., number of displaying cocks on home grounds (r = -0.18, 9 df, P > 0.05), percent change in numbers the following spring at home leks (r = - 0.02, 4 df, P > 0.05), and whether or not the lek changed location (P > 0.05). Nor did mean distances vary in relation to factors that may have reflected the status of cocks—e.g., age (P > 0.05) and April weight (r = - 0.52, 8 df, P > 0.05) —although adults tended to remain closer to home (Fig. 5.4). The type of habitat near leks also appeared unrelated to whether cocks
SHARPTAIL SPACING, MOVEMENTS, AND COVER
167
Fig. 5.5. Spatial patterns of males (home ranges) during the spring advertisement period with respect to each other and to lek locations, and the movement of one of two yearlings away from the lek during the summer (lower right). Letters identifying lek locations correspond to Figs. 5.1-5.3; study area and year are shown for each set of males.
168
M. W. GRATSON
Fig. 5.6. Movements of three females, which were in the same winter flock at CM, during the prelaying period when males were advertising at leks (Fig. 5.5). Also shown are the shifts in lek locations and the establishment of lek H in relation to female prelaying movements.
SHARPTAIL SPACING, MOVEMENTS, AND COVER
169
Fig. 5.7. Nests were often near the edge of the laying-incubation ranges of sharp-tailed grouse hens, especially adults.
170
M. W. GRATSON
moved away or not. Males that remained an average of 200-500 m from home in August increased their use of shrub-grass cover from May to August. Those that were 1,000-1,500 m and 1,900-2,800 m from home decreased their use of that cover. Cocks that stayed near home did not use shrub-marsh, whereas of those that moved, some increased and some decreased their use of it. Similarly, use of other cover types revealed no pattern bearing on whether cocks moved, or how far (Table 5.1). In two instances at two study areas, one of two cocks moved away from a home lek and the other remained near it (e.g., Fig. 5.5 — lower right). These movements away from leks and other birds may have represented intrinsic tendencies of individual cocks. Spatial separation of hens from males and display grounds continued through the early brood-rearing period but not throughout the summer. Spatial overlap did not occur because radio-tagged hens and broods moved close to leks — they did not (Fig. 5.4). Rather, some cocks moved away from home to areas used by brooding hens. Also, at DC unbanded hens with older broods were found at or near grass-forb and clover food patches also used as lek sites in the spring. Thus, separation of the sexes during the late brood-rearing period clearly did not occur. Radio-tagged hens averaged 1,062 + 187 m from the nearest lek for all summer months combined. Movements from nest sites to brood-rearing cover averaged 285 ± 165 m, 1,502 ± 1,269 m, and 2,094 + 1,542 m for June, July, and August, respectively; but variability of these estimates was high. Not all hens radiotracked during this period were included in these estimates of young-brood moveTable 5.1. Cover use by sharp-tailed grouse cocks in relation to mean distance from home lek in August % cover use by those dc? 200-500 m from home lek Cover type Grass-shrub Shrub-grass Shrub Decid. woods Cropland Sedgemeadows Shrub-marsh n birds n locations a
% cover use by those c?c? 1,000-1,500m from home lek
% cover use by those &J 1,900-2,800 m from home lek
May
Aug
May
Aug
May
Aug
32 55 3 0 10
27 65 6 0 2
29 38 14 11 8
22 21a 28a Oa 1
23 34 7 6 6
50a 17a
0 0
0 0
0 0
0 28a
10 14
14 10
2 65
4 183
Greater than 10% change from May to August in cover use.
0 2 7
3 127
SHARP!AIL SPACING, MOVEMENTS, AND COVER
171
ments; some were first captured after leaving nests with broods, and the locations of their nests are unknown. Fall display was first observed on 14 September 1977 and 22 September 1978, and the September ranges of most radio-tagged adult males included the home display ground within their perimeters. Some juvenile males were captured while displaying at leks in mid-October, but two juveniles captured in brood traps did not establish a territory before they died in midwinter. The October range of one bird did not encircle a lek, but the November range of this bird and the October and November ranges of the other male did. Range overlap patterns of cocks during October, when most were regularly displaying, were similar to April and May patterns in that ranges of cocks attending the same ground overlapped, but they differed in that the ranges of some adult males from neighboring leks also extensively overlapped. Males from some neighboring leks were in the same flocks by mid-November, when permanent snow occurred during 1977 and 1978. For all fall months combined, adult males remained closer to the nearest display ground than did juvenile males (669 ± 59 m versus 1,014 + 134 m, P = 0.01), and closer than females (669 ± 59m compared with 1,017 ± 8 2 m , P < 0.001) (Fig. 5.4). Spatial patterns of females are further presented in section 5.4.4. With the onset of permanent snow and throughout the winter, males and females moved farther away from home and nearest display grounds (Fig. 5.4). Cocks were similar distances from home as from nearest grounds (P > 0.05). Hens were farther from the nearest display ground than were cocks for all winter months combined (1,676 ± 154 m versus 1,230 ± 91 m, P = 0.01), but complete segregation did not occur. Radio-tagged hens and cocks were occasionally in the same winter flocks (see also 5.4.6). Nine of 13 monthly home ranges of males (excluding ranges of males whose home lek was unknown) from November through February did not encompass a home display ground, but only two of six did not in March, during snowmelt. Winter ranges of both cocks and hens often included cropland and night-roosting cover (e.g., conifer shrubs, conifer woods, and wetlands; see section 5.4.5) at their perimeters. This pattern was also evident at DC, where birds used only natural foods; there aspen and birch stands were at range boundaries. Less than 5% and often no overlap occurred between the monthly ranges of adjacent flocks. Both cocks and hens moved off DC and NB during the winter of little snow (1977-78, x = 18.0 + 0.1 cm, 130 days) but not during the following winter (1978-79, x = 39.3 ± 1.7 cm, 151 days). Some of the last to leave at NB in 1977-78 included a flock of nine, in which a radio-tagged adult male was present. Observations of radio-tagged and untagged birds indicated that most grouse of both sexes moved off DC and NB by late December to mid-January. They moved to open (cutover) conifer and deciduous woods and wetlands. Most observations were less than 2.2 km from these study area boundaries, but many birds may have moved farther. In contrast, all eight radio-tagged birds remained on CM during
172
M. W. GRATSON
the winter of 1977-78, and observations of untagged birds indicated that they used similar areas the following winter. Birds at DC and NB began to reappear on the study areas, in large numbers, in mid-March 1978, when snowmelt was under way.
5.4.2 Home range sizes To avoid biases associated with unequal time intervals I standardized estimates of home range size into seasonal and monthly periods. Seasonal ranges of cocks, constructed by overlapping appropriate monthly ranges to produce concave polygons, averaged 348 ± 11 ha (n = 2), 82 ± 11 ha (8), 388 ± 113 ha (4), and 400 + 66 ha (4) during spring (March, April, May; minimum of 45 locations per bird), summer (June, July, August), fall (September, October, November), and winter (December, January, February), respectively. The annual ranges (12 months, minimum of 180 locations per bird) of two cocks averaged 593 + 50 ha. Ranges of hens for the same periods were 604 + 111 ha (3), 156 + 95 ha (2), 556 + 175 ha (2), and 251 + 84 ha (2). No estimate of annual range size was obtained for females. Analysis by month indicated that ranges of males were largest during January and November (Fig. 5.8). November variability was high in home range size of cocks, and apparently unrelated to the age (P > 0.05), October weight (r = -0.48, 3 df, P > 0.05), and display status of individuals (known displayers or not, P > 0.05). It was also unrelated to mean flock size of individual cocks (r = 0.47, 7 df, P > 0.05) and study year (P > 0.05). Males that used only natural foods did not have November ranges different in size from those that also used agricultural crops (P > 0.05); nor were those of males using wetlands for night roosting different in size from those that roosted in upland sites (P > 0.05). However, a positive relationship (r = 0.88, 6 df, P < 0.01) between range size and mean distance that cocks were from home display grounds in December suggests that the variability may reflect either different movement patterns by cocks to find or relocate winter habitat or that winter habitat of some birds was simply farther from their leks than for other birds. Females ranged over the largest areas in April, the prelaying period, and during October when many juveniles were dispersing from brood ranges (Fig. 5.8). Two adult hens that deserted their broods in September ranged over areas larger than juvenile females, still in broods, during the same month, before the juveniles dispersed. This trend was reversed in November when adults reached fall and winter ranges and juveniles were dispersing. Monthly ranges of both cocks and hens during the summer were generally less than 65 ha, similar to estimates reported by others in Wisconsin (Ramharter 1976) and Minnesota (Artmann 1970, Schiller (1973). Laying-incubation ranges of hens averaged 14 + 5 ha (n = 6), disregarding the monthly time frame (see Fig. 5.7). No differences in home
SHARP!AIL SPACING, MOVEMENTS, AND COVER
173
Fig. 5.8. Monthly mean home-range sizes (ha) (+ SE) of male and female sharp-tailed grouse, 1977-79, and individual ranges in the fall (inset). Ranges of females were largest during the prelaying period, those of males just before the onset of permanent snow cover in November.
range size throughout the year were apparent among study areas, nor between ages for males. However, cocks that used only natural foods during winter months ranged over larger areas than those that also ate corn (288 + 35 ha, n = 8 ranges versus 1 3 9 ± 4 1 h a , n = 4 , f > < 0.05). Ranges of males that used buckwheat were intermediate in size (231 + 46 ha, n = 6). Monthly ranges of females during winter were not significantly different in size from those of adult males (149 + 31 ha versus 259 + 33 ha, P > 0.05). Late fall and winter ranges
174
M. W. GRATSON
of males were comparable to those reported by others (Moyles 1977, Wells, 1981, Nielsen & Yde 1982); few or no published data from other regions exist for females. Home range sizes were not well correlated with monthly mean flock sizes for either sex (18 correlations, each P > 0.05). Even for combined winter months, no relationship was apparent (cocks, r = -0.08, 15 df, P > 0.05; hens, r = -0.27, 6df, P > 0.05).
5.4.3 Daily movement patterns Day-to-day movements by cocks and hens declined in distance from late winter and early spring through the summer, and then increased through the fall (Fig. 5.9). Movements by both sexes were reduced in midwinter; males moved less in January than in December and February (P < 0.05). For hens, this reduced mobility continued through February. Late winter-early spring (March) and prelaying (April) movements by females were longer than those by males (P < 0.05). On average, males moved 262 + 12 m (n = 32 monthly means combining months and ages) during the summer. Movements by hens averaged 132 + 37 m during early brood rearing and 263 + 41 m to 276 + 102 m during late brood rearing periods. Day-to-day movements by adult and juvenile/yearling males were not significantly different during any month, but the lack of adequate samples of both age classes precluded a fair test of this null hypothesis for most months. Night-to-night (inter-night) movements by individual cocks and hens showed seasonal changes similar to day-to-day movements but were often longer (Fig. 5.9). Adult and yearling cocks averaged 313 + 27 m (n = 22 monthly means) during the summer. Distances were again longest during the NovemberDecember and March-April transition periods from fall-to-winter and winter-tospring conditions. In winter, birds tended to night-roost in different locations on successive nights even where the confounding variable of feeding sites was held constant (Fig. 5.10). The distances that birds moved from feeding and daylight roosting-loafmg areas to their night roosts, and the reverse, were often shorter than day-to-day and night-to-night movements (Fig. 5.9). Cocks averaged 231 + 12 m (n - 28 monthly means) during the summer, shorter than night-to-night movements (P < 0.01). During the same period, hens averaged 136 ± 37 m with young broods and 118 + 26 to 214 + 30 m with older broods. Movements by females were longer than those by cocks from October through January (December and January combined 760 ± 117 m compared with 446 ± 50 m, P = 0.01) (P < 0.01). During snowmelt the feeding-to-night-roosting and night-roosting-to-feeding movements of hens were shorter than their day-to-day movements (P < 0.05); and again the latter were shorter than their night-to-night moves (P < 0.05).
SHARPTAIL SPACING, MOVEMENTS, AND COVER
175
Fig. 5.9. Monthly mean ( + SE) day-to-day, night-to-night, and daylight-to-night/nightto-daylight movements (m) of sharp-tailed grouse.
5.4.4 Brood breakup and dispersal Some insight into the process of brood breakup and dispersal by individuals is available. An adult, four juvenile females, and a juvenile male from a brood that included at least two more juvenile females were radio-tagged. The juveniles were about 13 weeks old when captured in mid-September at CM. Patterns of breakup and dispersal are shown (Fig. 5.11). These birds traveled as a brood for a week until the adult and juvenile females moved west between 21 and 26 Sep-
176
M. W. GRATSON
Fig. 5.10. Females roosted in different locations on successive nights in winter even where they fed principally at a single food source (platform with corn) at CM, 1977-78.
tember; the male remained near the mid-September range site. The adult female deserted the remaining brood members on 26 September. She moved back to the earlier brood range and was observed with 17 birds on 27 September. Between 28 and 29 September she moved 2.4 km north of the juvenile females. Observed alone on 28 September and with two birds on 2 October, by 4 November this hen joined a flock of 15-25 birds (x = 22.1 + 4.1, n = 8 observations) and remained with a group throughout November in the same general area. Between 10 and 15 October, a juvenile female ( $ No. 1) moved away from the other three, north to where the adult was located. This bird and the adult hen were in the same flock four of 11 times I observed them during October and November. Another juvenile female ( 9 No. 2) moved southwest of the two remaining radio-tagged females between 16 and 18 October. This juvenile reached a maximum of 5.8 km from her late September brood range but was killed by a predator by 26 October. The juvenile male moved southeast between 17 and 20 October, eventually reaching a maximum of 2.1 km from the early brood range. This movement was to an area near a display ground, but the bird was not observed displaying. His mean flock size during November was 6.4 + 2.2 (n = 5 observations). Between 20 and 21 October another juvenile female ( 9 No. 3) moved 5.8 km southwest of the female
SHARPTAIL SPACING, MOVEMENTS, AND COVER
177
Fig. 5.11. Movements of five juveniles and the adult hen of one brood before, during, and after brood breakup at CM, 1977. Juvenile females are designated by number and are referred to in text.
178
M. W. GRATSON
that remained at the brood range and she was also killed by a predator. The last female to disperse ( $ No. 4) moved north between 22 and 23 October to where the adult and other juvenile female were located. She was observed on three occasions during November with the adult or sibling or both, and her mean November flock size was 20.8 ± 3.2 (n =4 observations). The full dispersal history was obtained for one juvenile ( $ No. 1). She nested the following summer 1.4 km from her mid-September range site used as a chick (see also Fig. 5.6). One other radio-tagged adult female deserted her brood in mid-September at DC and moved 3.6 km. A juvenile male, captured in a lily-pad trap with another juvenile at NB, also dispersed in October and early November, moving a maximum of 3.7 km from his brood range.
5.4.5 Cover selection Analysis of cover-type selection was initially conducted for each study area (Gratson 1983). Data were subsequently combined because of cover-use similarities among areas, with the exception of wetland communities. Wetlands were essentially unavailable within DC and NB. Because of this, references to the use of sedge-meadows and shrub-marshes are for CM birds only and are presented graphically as percentages independent of other cover values. Cover types are named for the dominant (> 50% coverage) vegetation. Cocks selected open cover types with shrubs for daylight feeding and roosting during spring and summer (Fig. 5.12). They used cropland in spring, often for lek locations, but rarely in summer. Night roosts in open cover types during the summer were in the grasses, sedges, and forbs away from or between shrubs. Males also night-roosted in wetlands at CM where shrub-marsh was "preferred" (i.e., its percent use was greater than its percent availability within home ranges, 95% family confidence interval, P < 0.05, after Neu et al. 1974). Spring cover use by hens during the preincubation period (30 + 10 days, mostly April) was similar to that by cocks (Fig. 5.12). Hens selected shrub-grass communities during both preincubation and incubation periods. Four of five nests were in shrub-grass cover, one nest was in shrub-tree vegetation (shrub-grass cover with 3-12 trees [> 10 cm dbh] per ha), and the nest of an unbanded hen was in grass-shrub vegetation. Hens nested in cover that had not been burned for at least 4 years. Over 80% of the off-nest feeding locations were in the same cover type as the nest. Early (here 1-14 days post hatch) and late (15 days post hatch through 31 August) use of habitat by hens and chicks shifted to vegetation that was shorter, sparser, and with a smaller percentage of shrubs than during preincubation and incubation periods. Observations of unbanded hens with broods in 1977 (n = 33) agree with the findings from radio-tagged hens; Artmann (1970), Schiller (1973) and Ramharter (1976) obtained results that generally concur with those presented here for nesting hens. Hamerstrom's (1963) data on brood-
SHARPTAIL SPACING, MOVEMENTS, AND COVER
179
Fig. 5.12. Cover use by sharp-tailed grouse in Wisconsin, 1977-79. Availability-use comparisons and differences in cover selection among study areas can be found in Gratson (1983).
180
M. W. GRATSON
rearing habitat are also similar to mine. Spring and summer data from other areas for cocks are generally lacking. Cocks decreased their use of shrub-grass communities from summer to fall and began using more woodland during daylight (Fig. 5.12). Night roosts were still in open cover, and cocks began using wetlands at CM more heavily. Males preferred (P < 0.05) grass-shrub cover at two of three study areas during daylight and shrub-marsh at CM during night. Hens continued to use open cover from late brood-rearing through dispersal periods until permanent snow cover arrived. Grass-shrub communities and cropland were preferred during daylight (P < 0.05), and hens selected grass-shrub and sedge-meadow cover types for night roosts. The majority of fall cropland use by both cocks and hens occurred in November. The greatest disparity in cover use among study areas occurred in winter and generally was directly related to differences in cover availability (see also Gratson 1983). Conifer woods (10-14 cm dbh) at DC and deciduous woods at CM were preferred by cocks (P < 0.05) (Fig. 5.12). Although aspen and birch stands were used mainly by feeding birds, open pine woods were used for both feeding and daylight roosting. At CM cocks and hens used shrub-marshes for feeding and daylight roosting, but sedge-meadows provided little in the way of food resources. Cropland was used proportionately more than its availability (P < 0.05) by cocks at NB, and by both sexes at CM, but was unavailable because of snow depth at DC. Cocks night-roosted in grass-shrub, shrub-grass, and shrub (< 10 cm dbh conifers) cover and conifer woods (10-14 cm dbh) at DC and NB, and in wetlands and grass-shrub communities at CM; each type of cover was used at least 20% of the time per study area. Hens used conifer woods (again 10-14 cm dbh) at NB and wetlands at CM. Use of wetlands by sharp-tailed grouse has also been reported by Moyles (1981) and Wells (1981). Grange (1948), Hamerstrom and Hamerstrom (1951), Ammann (1957), Moyles (1981), and Swenson (1985) similarly noted the shift to more shrub and woodland habitats for the winter. The period of snowmelt (March) presented difficulties for analysis because snow melted in patches and it was not always clear whether or not grouse used particular cover patches that were generally snowless. Cover use by both cocks and hens appeared most similar to that for winter (Fig. 5.12). Snow-roosting presented a special opportunity to study cover use. Snowburrowing occurred as soon as snow depths in the field reached 18 cm, first in wetlands and later in other cover types (Table 5.2). Burrows were used both during the night and during daylight; in midwinter, birds remained in their burrows through the night, left to feed in the morning, burrowed in another location after foraging, and often stayed there through the remainder of the daylight and through the night. Daylight burrowing appeared to occur later in winter 1911-IS, when there was less snow than in winter 1978-79. (Table 5.2).
SHARPTAIL SPACING, MOVEMENTS, AND COVER
181
Table 5.2. Characteristics of snow burrows and snow-burrowing behavior of sharp-tailed grouse in Wisconsin, 1977-79 Characteristic
Night
Daylight
Date first observed 1977-78 winter 1978-79 winter
6 Dec. 17 Nov.
7 Jan. 20 Nov.
Group size (x)
5.9 ± 1.5
5.4
± 1.2
Distance between burrows (x) (m)a
1.6 ± 0.4
2.1
± 0.8
Length of burrows (x) (m)
2.4 ± 0.5
1.4
± 0.7
Flush distance (x) (m) n groups a
57
11.6 ± 4.1 101
Distances to nearest neighbor.
Burrows were difficult to locate initially without the aid of radio telemetry. However, once located, birds often allowed a close approach. The mean flushing distance of birds in burrows during daylight was shorter than birds on top of the snow or in snow depressions with their heads above the snow (11.6 + 4.1 m versus55.3 ± 2.2m, P < 0.01). Because burrows were long, birds flushed some distance from their entrance holes (Table 5.2). On average, the distance between entrance and exit holes was 1 m shorter during the night than during daylight (P < 0.01). Finally, it appeared that snow-burrowing opportunities influenced the flocking tendencies of birds (see section 5.4.6).
5.4.6 Flocking Flocks were the basic social unit. Even in summer cocks were in groups of 2-4 birds (Fig. 5.13). Flock sizes of hens and cocks generally averaged fewer than 6 birds from April through October, and greater than this for the remainder of the year. From November through March, females tended to be in larger flocks than males, but radio-tagged hens and cocks were occasionally in the same winter groups (6/141 observations December-February). Thus, segregation appeared incomplete, and flock sizes for radio-tagged hens probably included some males. In early winter, when a sample of both juvenile and adult males was radio-tagged, mean flock size was larger for the younger age class (adult x = 4 . 0 + 1.5, n = 3 versus juvenile x = 10.5 ± 1.5, n = 5, P < 0.05; P = 0.01 for combined winter months). There appeared to be no difference between years during this
182
M. W. GRATSON
Fig. 5.13. Monthly mean ( + SE) sizes of sharp-tailed grouse in Wisconsin, 1977-79. June, July, and August sizes for females are broods.
period (1977 x = 8.1 ± 1.2, n = 4 males versus 1978 x = 8.0 ± 1.1, n = 4 males, P > 0.05). Flocks were largest during the fall and spring transition periods (Fig. 5.13). Because of this, I further investigated the flock-size-snow-depth relationship (Fig. 5.14). Spring data were excluded because snow melted in patches and confounded analysis. Generally, flocks were largest when snow was on the ground but was less than 18 cm, a depth beyond which birds could no longer obtain ground foods easily but at which they could snow-burrow. Thus, mean flock size decreased when there was a reduced availability and distribution of ground foods, and also when there was increased snow-burrowing opportunities. These changes were statistically significant for males from conditions of no snow to those with snow less than 18 cm (P < 0.05), and for females from conditions of snow less than 18 cm to those greater than this depth (P < 0.01).
SHARPTAIL SPACING, MOVEMENTS, AND COVER
183
Fig. 5.14. Mean flock sizes during 10-day intervals in relation to snow presence and depth. Flock size increased when birds could not snow burrow (snow < 18 cm).
184
M. W. GRATSON
5.5 Discussion My data are generally too few to estimate reliable productivity and mortality statistics associated with certain habitat-use patterns, movements, spacing tendencies or levels of gregariousness. Thus, I am unable to measure the association of variation in these behaviors with some presumed index of individual fitness. Below I discuss the possible significance of some of the major findings of this study, comment on their generality, and explore some new and old hypotheses that may explain the data.
5.5.1 Prelaying ranges and activities of females Prelaying ranges of sharp-tailed grouse hens were twice as large as their latewinter-early-spring (March) ranges. Females ranged over areas that were considerably larger than did males, and their day-to-day movements were significantly longer. This was not merely a dispersal by females from winter ranges directly to breeding habitat, but rather, for yearling hens at least, a searching movement. As a result ranges were more circular or oblong than linear. Preincubation or prelaying ranges of females, and especially yearlings, of other tetraonids are also relatively large (Robel et al. 1970a, Maxson 1974, Herzog & Boag 1978, Connelly 1982, Hannon et al. 1982, Chaps. 6, 14); thus the generality of these findings has been established. There are at least three hypotheses that are relevant to the questions of why ranges are so large and what females are investing in during this period that requires such wandering movements. First, increased nutritional and energetic demands of later egg-laying may necessitate a foraging strategy that requires feeding over large areas. Second, if mate choice is important to females, they may be sampling many males before copulating. For the lekking grouse this could include visits to many leks and, hence, require long movements. Finally, females may be investing in a search for appropriate nesting sites, where later they and their clutch must spend a month or more. Because optimum nesting sites are not ubiquitous in the patchy steppe environment, large prelaying ranges may be a consequence of a sampling strategy. This greater prelaying mobility of sharp-tailed grouse hens cannot be explained solely by different food and cover requirements of the sexes, nor by changing habitat conditions from March to April. Shrub-grass was the most frequently used cover by both sexes during this period (x = 24-27% of the time). Cocks more often used grass-shrub communities and cropland and less often sedge-meadows than hens, but both used other cover types with about equal frequency (i.e. < 10% difference). Hens increased their use of shrub-grass by 15% and decreased that of cropland by 9% from March to April, but these vegetationuse changes also seem too small to explain the difference in range size between months. Daylight-to-night and night-to-daylight movements suggest that hens moved similar distances between feeding and night-roosting cover during the
SHARPTAIL SPACING, MOVEMENTS, AND COVER
185
prelaying period as they did during March, and that the distance between feeding and roosting cover was similar for hens and cocks. Food availability generally increased from March to April as snow melted. One might therefore expect smaller ranges and shorter movements, but these did not occur. I can find no support in these data that prelaying dietary considerations were important in determining this mobility. However, without data on the spring diets of hens and the distribution of selected items in relation to egg and offspring "quality" and quantity as well as data on mobility, this hypothesis cannot be fully rejected. Were the large prelaying ranges of hens related to their visits to display grounds to possibly choose mates and copulate, or rather, to their search for suitable nest sites, or both? On the basis of telemetry data it was evident that one radio-tagged yearling visited three display grounds within a 4-day period in midApril; another visited two leks. Also, one hen visited a lek 6 days after she was located near (less than 110 m) the area of her eventual nest site (Fig. 5.6). Although it seems reasonable to suspect that the process of lek and mate selection by hens may vary according to their age and experience, as well as other factors, some female prairie chickens and sage grouse (Centrocercus urophasianus) also visit more than one lek (Dalke et al. 1963, Robel et al. 1970a, Wallestad & Pyrah 1974, Dunn & Braun 1985, Chap. 6). Recently however, Petersen (1980) and Dunn & Braun (1985) found that only a small percentage of the sage grouse females monitored visited two or more grounds. If and why this investment in lek choice by females is different for sharp-tailed grouse and how this investment varies with other factors are problems for further research. Here I can only conclude that of the two yearlings radio tracked during the prelaying period, each visited at least two leks and these movements contributed to their increased mobility. Hens used shrub-grass cover more than 50% of the time during the preincubation period, and during nesting and early brood-rearing. Moreover, they nested at distances from nearest leks that were comparable to those long before nesting. In other words, females were moving through nesting habitat during the monthlong period immediately preceding laying. This association between prelaying cover use and large home ranges and selection of nesting cover and space is consistent with a hypothesis that hens search large areas for suitable nest sites during the prelaying period, but is insufficient for distinguishing between this and a "male and lek choice" hypothesis. It may be that females are investing time and energy in selecting both suitable nest sites and males. In a similar fashion Herzog and Boag (1978), Hannon et al. (1982), and Nugent and Boag (1982) have noted that female spruce (Dendragapus canadensis) and blue grouse (D. obscurus) are more mobile during the prelaying period and that yearlings cover larger areas than adults. Large prelaying ranges of these yearling forest grouse have been suggested to result from their interactions with adults at potential nest sites, in high-density populations. However, high density
186
M. W. GRATSON
and the resulting interactions appear unnecessary preconditions to larger, relative size ranges in sharp-tailed grouse; during this study populations were low and large ranges also occurred. Determinants of both the absolute and relative sizes of prenesting ranges of adult and yearling females may prove to be significant factors influencing the productivity of individual hens, and require additional study. Estimates for sharp-tailed grouse, prairie chickens, and sage grouse are especially few (Chap. 14). Prelaying mobility of females may have further significance to an understanding of spacing patterns of males. Bradbury (1981), on mainly theoretical grounds, suggests that large prelaying ranges of females are necessary for the evolution of clustered advertising males and hence lek mating systems in general. This hypothesis predicts that under conditions of large and overlapping female ranges (densities of females held constant), males abandon a dispersed spatial pattern and instead congregate at crossroads of female ranges and invest in self- rather than resource advertisement (i.e., form leks). Clusters of males are postulated to be spaced, on average, according to the size of female ranges. Data from the present study, although few, show a poor quantitative fit to this model in that the diameters of hen preincubation ranges averaged twice the mean distance between leks (1,215 m), but closely resemble Bradbury's (Bradbury 1981, p. 152) steadystate qualitative function of the spatial relationship of hen ranges to specific display grounds (e.g., Fig. 5.6). Moreover, the shift and establishment of new lek locations associated with female ranges documented in this study provide some evidence that males indeed track females.
5.5.2 Separation of cocks and hens during nesting Hens appeared to nest and rear young broods as far from cocks and display grounds as possible while remaining on summer habitat. As noted, support for the generality of these results for sharp-tailed grouse is found in North Dakota (Bernhoft 1969, Christenson 1970, Kohn 1976), Minnesota (Schiller 1973) Nebraska (Sisson 1976), Saskatchewan (Pepper 1972), and Wisconsin (Ramharter 1976) data. What can we conclude from this? First, it appears clear that because the trend is to be away from males, females cannot be selecting nesting sites because of advantages associated with being particularly near leks. Thus, females may distribute themselves randomly with respect to leks, or, space away from leks. If females are spacing away from males, at least two hypotheses may explain this. The first hypothesis is that the separation of cocks and hens in Wisconsin primarily resulted from different food and cover requirements of the sexes. However, both used shrub-grass cover 50-52% of the time. Hens used 15% more shrub-tree cover and 9% less grass-shrub vegetation than did cocks, but six other cover types were used with the same frequency (i.e., < 3% difference). Comparable cover-use data for both sexes are unavailable for sharp-tailed grouse in other regions, but data gathered from sage grouse support these findings
SHARPTAIL SPACING, MOVEMENTS, AND COVER
187
(Wallestad & Pyrah 1974, Wallestad & Schladweiler 1974). Sage grouse cocks used sage with canopy coverage that was the same as that at nest sites of hens, 20-30%, but as during this study, nesting hens were far from leks whereas males remained relatively near their advertising locations. The second hypothesis is that the strategy of hens may be to avoid conspicuous and advertising cocks (and leks) and reduce the probability of predation on themselves and their nests (see also Brown 1964, Chaps. 14, 16). Although knowledge of the foraging tactics of many predators is lacking, reports of the minimum frequency with which crepuscular and diurnal raptors do visit leks (Berger et al. 1963, Hamerstrom et al. 1965, Sparling & Svedarsky 1978) suggest that advertising males do attract predators. Experimental research on the impact of advertising males in the nesting areas of females will be necessary to investigate this hypothesis (Wrangham 1980). It should prove worthwhile.
5.5.3 Hen desertion and dispersal of juveniles After brood members are fairly old but generally before broods break up, hens appear to abandon their offspring. In a proximate sense, hen desertion may result from decreased social bonds and the increased independence of offspring, as is thought to occur during brood breakup (Alway & Boag 1979). However, the adaptive value of this behavior, if any, is unclear. Hens may improve their fitness if desertion results in increased survival prospects for their offspring or increased longevity and future productivity for hens themselves. Desertion may represent a switch from investing in current offspring, which are now relatively independent, to the chance for future reproduction. One possibility is that membership of hens and daughters in the same winter flocks is somehow disadvantageous, so hens abandon them early. However, during this study two siblings later joined the winter flock in which the brood female was also a member. If this is not uncommon, it suggests that desertion may have more to do with the timing and duration of brood breakup and dispersal than with a separation of related individuals. One possibility is that desertion may result in earlier independence of juveniles and hence added experience before winter conditions occur. Another possibility is that there is a parent-offspring conflict (Trivers 1972), and although offspring would benefit from continued association with the hen, she decreases her probability of higher fitness by further association. The mechanisms by which this might occur are wholely unknown. During their first fall, juvenile male sharp-tailed grouse commonly begin efforts to invest in territory acquisition (see also Moyles & Boag 1981). By contrast, there was no evidence in the data on movements or ranges of dispersing females that would indicate that females were in some way investing in activities specifically associated with successful reproduction. The process of selecting particular leks and particular spots on leks for displaying by juvenile males is only partly known. Juveniles should also have the option of forming new leks. Al-
188
M. W. GRATSON
though Moyles & Boag (1981) have made an important contribution to the knowledge of factors operating once the decision to establish a territory at a specific lek has already been made, data on movements of juveniles before lek visitation are needed to more fully understand the tactics and important determining factors involved with lek choice. One juvenile male from the radio-tagged brood of this study moved to one of the leks nearest to his late-summer brood range. Caldwell (1976) noted similar distances in Manitoba using capture-recapture data. These dispersal movements are relatively short, and are comparable to mean interlek distances. Because lek size, especially the number of adult males, and the locations of females should be important determinants of where juveniles can enjoy the best chances for future success, these conditions may have been met at closest leks during this and Caldwell's (1976) study. If juvenile females do delay fitness decisions closely associated with reproduction until spring, their initial dispersal in autumn may be related primarily to winter survival strategies. Four juvenile females from one brood dispersed in this study in the fall, but the full history was obtained for only one. Three general phases were apparent: (1) initial dispersal to a flock for the winter; (2) movements on the winter range; and (3) spring dispersal to a nesting range. Flocking and movements on the winter range are further discussed below (section 5.5.4). Spring dispersal was discussed previously.
5.5.4 Daily movements as a strategy How might the general patterns of day-to-day, night-to-night, and daylight-tonight/night-to-daylight movements be explained adaptively? Consider first summer movements by cocks when visits to leks were relatively unimportant and there were few other processes such as dispersal, changing weather patterns, and reduced food availability to confound analysis. Also, vegetation was tall and dense; birds were difficult to locate visually, by myself and probably other predators. Why should cocks move from feeding and loafing areas to night roosts, inasmuch as cover used for these activities was the same or similar? And, why do cocks move between night roosts, not returning to the same general area each night? The vegetation used for night roosts is not a depletable resource. Also, why were night-roosting locations spaced consistently farther apart for individual cocks than the distance they traveled from feeding and loafing areas to night roosts? The hypothesis I suggest is that the strategy of males involves roosting far enough away from feeding and loafing areas so that foxes (Vulpes vulpes} and coyotes (Canis latrans), nocturnal predators that hunt by scent, will not find these birds by locating feeding and loafing areas. However, the increased vulnerability of birds moving long distances from daylight locations to night roosts, to diurnal and crepuscular predators, mostly sight-hunting raptors, may result in a counterselection for shorter daylight-to-night movements than are optimum to "beat" the
SHARP!AIL SPACING, MOVEMENTS, AND COVER
189
canids. The distance between successive night roosts of males is longer than these daylight-to-night movements because the movements are actually during daylight, and males can move only short distances to roosting cover but still maintain long distances between where they roost each night. To counter canids and great horned owls (Bubo virginianus], the strategy may be to not roost in the same spot two, three, or even four nights in succession. Raptors have been documented returning to sites of previous success for the following hunting bout, and I presume for the argument above that the tendency is similar for sites where detection but not success has occurred. The habit of grouse of generally moving to a new location each day and night may have resulted partly from these types of predation pressures. In winter, hens at CM maintained a night-roosting pattern of shifting locations from night to night. Because the shifts occurred in the absence of variable food locations, it can reasonably be hypothesized that this pattern represents a strategy to counter predator hunting tactics. Just as foraging predators can be expected to have some general "rules" for hunting certain prey types and sizes, densities, and prey patches, so too can we expect at least some development of behavioral counter-strategies, such as movement tactics, in prey.
5.5.5 Winter flocking, snow-burrowing, and the use of wetlands My results concerning flock-size dynamics do not appear to support the "information-centers" (Ward & Zahavi 1973, Krebs 1974) and "more efficient food finding" (Krebs et al. 1972) hypotheses from other works. I have no data in support of or against the various "vigilance-optimal foraging" models (e.g., Pulliam 1973, Powell 1974, Caraco et al. 1980). Tendencies to join flocks (increasing sizes) or leave flocks (decreasing sizes) generally were consistent with changes predicted by some antipredator models of changing conspicuousness and flocking (e.g. Hamilton 1971, Lazarus 1972, Vine 1973, Treisman 1975). Flock sizes of male and female sharp-tailed grouse did not increase with decreasing food availability and distribution. At snow depths greater than 18 cm flocks were generally smaller than when snow was present but less than 18 cm. Food availability and distribution, certainly the latter, decreased with snow depth, if at all; however, more birds did not join flocks when snow was deep. Rather, mean flock size decreased. Also, flock sizes were not correlated with home-range sizes, but the sizes of home ranges depended at least partly on food type and distribution (corn, buckwheat, natural foods only). Thus, mobility was related to changing food distribution and type but the tendency to join or leave flocks was not. Mean flock size of birds using corn was 8.3 + 0.4 (n = 3 radiotagged birds); that of birds using buckwheat, 6.1 + 1.3 (n = 6); and that of grouse using only natural foods, 8.6 + 1.5 (n = 8). These were not significant differences (P > 0.05). Flock sizes were related to snow-burrowing opportunities. At snow depths that
190
M. W. GRATSON
allowed grouse to snow-burrow, birds left flocks. When depths were reduced and they had to roost exposed on the snow, birds joined others, and mean flock sizes increased. These data appear consistent with the general hypothesis that grouse join flocks when they are most conspicuous, on top of the snow. When sharptailed grouse snow-burrowed, the benefits of increased vigilance owing to flocking should have been fewer, and as predicted by the general antipredator models, birds left flocks. Tactics to avoid predators may include not only those that avoid detection by the use of physical objects, such as snow at roosting sites and the use of time and movements (previous discussion 5.5.4), but also tactics to avoid habitats where predators hunt. Sharp-tailed grouse in northern Wisconsin may use open wetlands for roosting primarily because fewer predators are hunting in these habitats. The lack of perches in sedge-meadows and shrub-marshes should reduce use by great horned owls of those open cover types (Petersen 1979). Schofield (1960) reported that red fox used marshes less often than other cover types in winter and that they avoided marsh, compared with its availability. Pils and Martin (1978) showed that fox used marsh cover in winter but also that they preferred retired croplands, grasslands, and forbs more than marsh. These predators were probably hunting where their primary prey were located (Pils & Martin 1978, Peterson 1979, Todd et al. 1981). I have investigated the relationship between sharp-tailed grouse use of wetlands and alternate prey abundance by examining the abundance of mammal prey in cover preferred by sharp-tailed grouse for night-roosting and in cover less often used. Of 2,400 trap-nights, using snap-traps in grass-shrub and shrub-grass cover on these three study areas, an average of 66.1 small mammals per 1,000 trap-nights (excluding the Soricidae (Pils & Martin 1978, Peterson 1979) were captured (R. K. Anderson unpubl. data 1978, Beck & Vogl 1972) (700 trapnights). In comparison, of 700 trap-nights in sedge-meadows and shrub-marshes, the cover preferred by grouse, only 33.3 small mammals per 1,000 trap-nights were captured (R. K. Anderson unpubl. data 1978). Similarly, an average of 27.6 small mammals per 1,000 trap-nights were captured in a sedge-meadow 43 km east of DC (D. Kent unpubl. data 1980). Snowshoe hares (Lepus americanus), an important prey for many predators of northern grouse, mainly use dense, wooded, and tall-shrub cover types and avoid open lowlands (Pietz & Tester 1979, Wolff 1980, Buehler & Keith 1982). Rusch et al. (1972) stated that "hares were usually absent in habitats where sharp-tailed grouse were found." Perhaps this pattern is part of a predator-prey, coevolutionary relationship in northern Wisconsin. The greatest preference for lowlands was shown in winter when birds were snow-burrowing, which would negate the possible thermal advantages of roosting in sedge-meadows and shrub-marshes when compared with other cover types. Similarly, although shrub-marshes are certainly used by feeding birds, the per-
SHARPTAIL SPACING, MOVEMENTS, AND COVER
191
sistent use of sedge-meadows, where there are no shrubs, by hens and the nonfeeding daylight use of wetlands by cocks suggest their selection for reasons other than foraging.
5.6 Summary Habitat selection, mobility, and spacing patterns of sharp-tailed grouse were investigated in northwestern Wisconsin from 1977 through 1979, principally by radio-tracking 26 males and 13 females. Monthly home ranges of hens were largest (464 + 145 ha) when they were visiting leks and potential nest sites during the preincubation period. These large ranges were probably unrelated, or only relatively weakly related, to the nutritional demands of laying eggs. Further, because densities of birds in the populations studied were low, the increased mobility of hens during the prelaying period was probably not a result of competition among hens for high-quality nest sites. Instead, it is proposed that large, prelaying ranges of females manifest a strategy for sampling males and thus mate choice, and/or, for sampling potential nest sites. Hens nested and reared young broods away from cocks and leks by using habitats between display grounds. Spatial separation apparently did not occur because of different habitat requirements of the sexes. It is hypothesized that this possible spacing, away from advertising males by nesting females, is a reproductive strategy to reduce detection by predators. Cocks and hens used grass-forb, grass-shrub, and shrub-grass cover types during the summer. Home ranges were less than 65 ha. Spatial separation of cocks and hens was not maintained throughout the summer. Some cocks moved away from home leks to areas used by brooding hens, and hens with older broods used areas near leks and males. The daily summer patterns of day-to-day, night-tonight, and daylight-to-night movements by cocks are hypothesized to represent an antipredator strategy to reduce detection by both raptors and canids. Dispersal by juveniles occurred in mid to late October after adult hens deserted their broods. The mean in home-range size of hens in October was 421 + 114 ha. Juvenile hens stopped dispersing when they joined flocks of 20-25 birds. Juvenile cocks stopped dispersing when they established territories at display grounds or stayed in areas near leks. Both cocks and hens were farthest (1,482 ± 242m and 1,806 ± 207 m, respectively) from the nearest display ground in February. Monthly mean home ranges of hens in winter (149 + 31 ha) were not significantly different in size from those of adult males (212 + 36 ha), but were smaller than those of juvenile males (259 + 33 ha). Cocks that used only natural foods in winter ranged over larger areas than those that also used corn (288 + 35 ha and 139 + 41 ha, respectively), but the tendency to join or leave flocks was apparently unrelated to food availability, distribution, and type. Although the winter ranges of hens were
192
M. W. GRATSON
smaller than those of cocks, hens moved farther (760 + 117 m versus 446 + 50 m) from their daylight feeding and loafing areas to night-roosting cover than did cocks. Further, hens maintained a pattern of moving to new roosting locations on successive nights even where they primarily fed at a single food source, suggesting the hypothesis that this night-to-night movement was a strategy to avoid nocturnal predators. Cocks and hens preferred to snow-burrow in open sedgemeadows and shrub-marshes, where there were fewer alternate prey (mice) of owls, hawks, and canids. Finally, sharp-tailed grouse joined fall and winter flocks when the depth of snow was less than 18 cm and they could not snow-burrow, and left flocks when snow was greater than 18cm and burrowing was a viable strategy. Flock sizes decreased when the availability of ground foods decreased, contrary to the predictions of hypotheses that suggest birds join flocks to obtain information on the whereabouts of food resources.
6
Reproductive Ecology of Female Greater Prairie Chickens in Minnesota W. D. Svedarsky
6.1 Introduction The greater prairie chicken (Tympanuchus cupido pinnatus) has declined in recent years, particularly in the eastern portion of its range (Christisen 1969). Intensified farming, overgrazing, and woody plant succession are the major factors diminishing the grassland habitat upon which prairie chickens depend. These factors probably exert the greatest effect on populations by limiting reproduction. Hamerstrom et al. (1957) and Kirsch (1974) noted that, of the various seasonal habitat needs, nesting and brood-rearing habitat appear to be the universal limiting factor for prairie chickens. Spring and summer ecology of females and broods is not well understood. Hamerstrom and Hamerstrom (1973; p. 28), in summarizing results of a 22-year study in Wisconsin, indicated that "we know nothing of movements during the summer.". Bowman and Robel (1977) noted that few data were available for greater prairie chicken broods. Although Toepfer (1973) in Wisconsin, Ammann (1957) in Michigan, and Janson (1955) in South Dakota collected nesting and brood data, their findings may not be directly applicable to the northern extreme of the range due to differences in climate and vegetation. Thus, Farmes and Maertens (1973) identified the need for more detailed knowledge of nesting and brood habitat in Minnesota in order to more appropriately manage grasslands for prairie chickens. The major goal of this study was to gain information for developing management recommendations, and to better understand life-history strategies of female greater prairie chickens. Radio-tagged females were monitored from early spring 193
194
W. D. SVEDARSKY
to late summer in 1975, 1976, and 1977 to document reproductive behavior, movement patterns, mortality factors, and habitat use. 6.2 Study area The Kertsonville Study Area (Fig. 6.1), approximately 18,900 ha, is in northwestern Minnesota about 25 km southeast of Crookston. Intensive field work was carried out in the center of the study area on the 582-ha Pembina Trail Preserve (hereafter, the "Preserve") established by The Nature Conservancy in 1974. The study area is in the continental forest-prairie transition zone where trembling aspen (Populus tremuloides) clones and willow (Salix spp.) clumps are interspersed throughout tallgrass prairie. Dominant prairie grasses are big bluestem (Andropogon gerardi), little bluestem (Andropogon scoparius), indian grass
Fig. 6.1. Location of the Kertsonville Study Area in Minnesota and the Pembina Trail Preserve.
GREATER PRAIRIE CHICKENS IN MINNESOTA
195
(Sorghastrum nutans), and switchgrass (Panicum virgatum). In the absence of disturbance, particularly fire, the prairie subclimax tends to be replaced by forest vegetation (Ewing 1924). Physiographic factors are described in more detail by Svedarsky (1979). Jorgenson (1977) estimated the extent of different habitat types in the Kertsonville Study Area in 1975 as follows: cropland, 36.5%; grass, 49.5% (mostly heavily grazed); trees and brush, 10.4%; slough, 1.5%; and miscellaneous, 2.1% (highways, farmsteads). Habitats were subjectively classified in 1975 on that part of the area intensively studied (Fig. 6.2). Some habitat types off the Preserve were modified during the course of the study. Habitats included in Fig. 6.2 and their percent occurrence are as follows: (1) Bluestem (41%): Areas moderately drained and dominated by warmseason prairie grasses, notably big and little bluestem, indian grass, and switchgrass with an intermixing of prairie forbs. Scattered willow clumps and small aspen clones are also present. (2) Sedge (4%): Lowland areas with standing or slowly flowing water in the spring and after heavy summer rains. Areas are dominated by graceful sedge (Carex praegracilis), northern reed grass (Calamagrostis inexpansa), Baltic rush (Juncus balticus), and prairie cord grass (Spartina pectinata). Willow is usually present but with less than 25% ground coverage. (3) Cropland (15%); Well-drained areas, spring-seeded to wheat, barley, sunflowers or corn, and generally fall plowed. (4) Alfalfa (8%): Well-drained, level areas dominated by alfalfa with considerable bluegrass (Poa spp.). (5) Brome (2%): Generally, well-drained areas planted to smooth brome (Bromus inermis) with intermixed alfalfa and bluegrass. (6) Sweet clover (6%): Abandoned cropland located on a dry ridge and dominated by sweet clover (Melilotus spp.), quackgrass (Agropyron repens), some bluegrass, and scattered patches and plants of sage (Artemisia spp.). (7) Redtop (1 %): Located primarily in poorly drained portions of abandoned cropland and dominated by redtop (Agrostis stolonifera). (8) Bulrush (2%): Marsh dominated by bulrush (Scirpus acutus} and cattail (Typha latifolia) with standing water present most of the summer.
196
W. D. SVEDARSKY
(9) Willow (14%): Grass and sedge lowlands with greater than 25 % coverage of brush composed of willow, bog birch (Betula pumild), and small aspen (< 10 cm dbh).
Fig. 6.2. Habitat types on and adjacent to the Pembina Trail Preserve, 1975-77.
GREATER PRAIRIE CHICKENS IN MINNESOTA
197
(10) Aspen (7%): Forested areas dominated mostly by pole-sized trembling aspen and generally lacking in understory except for occasional red osier dogwood (Cornus stoloniferd) and willow. Fallen trees are common in older stands.
6.3 Methods Counts of males on leks were conducted from 1 April to 30 May 1974-1978, using the procedures of Kirsch (1956). Females were trapped with a cannon net on one lek, and with a long-handled dip net on nests and roosts. Radio telemetry procedures were essentially those of Dumke and Pils (1973) with variations described by Svedarsky (1979). Nests were found by locating radio-tagged females on nests and through the use of a cable-chain drag (Higgins et al. 1977). Plant species composition at nest sites was measured with a 10-point frame (National Academy of Sciences 1962). Scientific names of plants follow Gleason and Cronquist (1963). Vegetation density was determined with a density pole (Robel et al. 1970b). Litter was measured at sampling points around the nest site, and a "brush index" was calculated by counting the number of brush clumps (mostly willow and aspen) within a 50 - m radius of the nest. Distances to the nearest travel way (vehicle tracks, roads, pasture edges) and to short cover (< 10 cm) were also measured. Radio telemetry data were grouped into five major periods related to reproductive chronology. These periods were: (1) preincubation—from 2 days after attachment of the radio to the onset of incubation; (2) incubation—from the onset of incubation to the hatching of a brood or to nest destruction; (3) early brood— from hatching until a brood age of 2 weeks; (4) late brood —from a brood age of 2 weeks until loss of the brood or the end of monitoring; and (5) postbrood —after the loss of a brood or nest (if no renesting attempt occurred) until the end of monitoring. For home-range analysis the preincubation period was further separated into the prelaying and laying periods, and the brood and postbrood periods into 2-week intervals. The same separation was used for movement analysis, except the brood period was separated into weekly intervals. Movement indexes in meters per hour were calculated from location data in which at least 6 hours but less than 36 hours had elapsed between consecutive locations. Movement per habitat type was determined when two consecutive locations occurred within the same habitat. A movement index for incubating females was determined from distances between feeding sites and the nest. In addition, the sum of distances between consecutive brood locations by weekly intervals was calculated and termed "minimum movements." Home ranges were determined when at least nine locations per female occurred within a reproductive period. The "modified minimum area" method (Harvey & Barbour 1965) was used with one variation. In this method, if two outer
198
W. D. SVEDARSKY
locations are a greater distance apart than one-fourth the range length, they are not directly connected; instead, the boundary line is drawn from one of these points to the next outermost point that is no more distant than one-fourth the range length. Location points falling in excess of one-fourth the range length from any other point are excluded from the main home range but are connected to the nearest point by a straight line. Harvey and Barbour considered these lines to be 1 ft wide, whereas I considered them to be 10 m wide. This method has been used by Haas (1974) for spruce grouse (Dendragapus canadensis) and by Porter (1977) for wild turkeys (Meleagris gallopavo) and closely approximates a method used by Weigand (1977) for Hungarian partridge (Perdixperdix). Although this method has not been applied to prairie grouse, I believe it more closely delineates the area actually used than the "minimum area" method (Mohr 1947) used for prairie chickens by Robel et al, (1970b) and Bowman and Robel (1977); especially for broods. In this study, broods occasionally made back-and-forth movements in a small area (< 16 ha), then made long, unidirectional movements (> 2 km) and began intensively utilizing a smaller area again. To connect all the outermost points according to the Mohr method would grossly overestimate the area used in these situations. Habitat use was evaluated according to the major reproductive period, habitat type, and land use condition of that habitat. In addition, locations were classified according to the daily periods suggested by Pepper (1972). These were: morning (sunrise to 1030 hours, CST), midday (1030 to 1630), evening (1630 to sunset), and night (sunset to sunrise). Night locations were taken between one-half hour after sunset and one-half hour before sunrise to ensure that birds were roosting. A variety of methods has been used to delineate "available" habitat in order to evaluate preferential use. Robel et al. (1970b), studying prairie chickens, and Maxson (1978), studying ruffed grouse (Bonasa umbellus), considered all the area within a defined study area to be available habitat. This probably includes some area to which individual birds have never been exposed and thus may overestimate available habitat. At the other extreme, Haas (1974) considered the estimated home range of spruce grouse to delineate available habitat. This method is probably biased toward preferred habitat, assuming the animal has experienced the surrounding habitat. I consider the methods of determining available habitat in the following description to be an acceptable compromise between these cited methods. Habitat preference was evaluated using a method modified slightly from Maxson (1978). The mean percentage of female locations that occurred in a habitat and land use type was calculated along with the mean percentage of that habitat considered available. Available habitat was calculated separately for each female. In this method the use of percentages gives each female equal weighting, although unequal numbers of locations were recorded per bird. A measure of habitat preference or avoidance is obtained by subtracting the mean percentage of avail-
GREATER PRAIRIE CHICKENS IN MINNESOTA
199
able habitat from the mean percentage of locations falling within that habitat and land use type. A positive value indicates preference and a negative value avoidance. For example, if the mean percentage of locations during the preincubation period of four females that were in grazed aspen was 10% and the mean percentage available was 2 %, that habitat and land use type would have a preference rating of +8.0. If no locations had occurred in this type when the same quantity was available, an avoidance rating of -2.0 would be assigned. Only females for which a minimum often locations were recorded during a period were included. Habitat considered available to females during the major reproductive periods was determined as follows: (1) Preincubation: The area contained by a circle having as its diameter a line connecting the two locations of a hen farthest apart during this period (Fig. 6.3A). This corresponds to the "activity area" used by Schiller (1973). In cases of renesting, when the female moved to a new nest site so that the activity areas of consecutive nests were nonoverlapping, the new nest was considered a different sample with respect to associated habitat use and preference evaluation. (2) Early brood: The area contained by a circle using the nest site as a locus and a radius equal to the mean of the maximum straight-line distance that broods of all radio-tagged hens had moved away from the nest by 2 weeks (Fig. 6.3B). If a ditch that contained at least 20 cm of water or a strip of trees bisected the circle, it was considered a barrier to young broods, and only that habitat on the brood side of the barrier was considered available. In instances where the movements of a brood exceeded this radius, an overlapping circle was drawn similar to the preincubation period with the nest site at one end of the diameter and the most-distant brood location at the other. Available habitat was then the total area outlined by both circles. (3) Late brood: For broods from age 2 to 4 weeks the same radius was used as for the early brood period. For broods older than 4 weeks, a similar circle was drawn using as a radius the mean distance that three radio-tagged brood hens had moved by a brood age of 4 weeks. Ditches and trees were not considered barriers to broods during this period, and, if all locations did not fall within the reference circle, an additional circle was drawn as in the early brood period. (4) Postbrood: The area within a circle drawn as for the preincubation period. Habitat preference was not evaluated for postbrood females, which made long-range movements and showed little affinity for a particular area.
200
W. D. SVEDARSKY
Fig. 6.3. Habitat considered available to female prairie chickens during the preincubation and postbrood periods (A) and during the early brood period (B).
Nesting-habitat preference was evaluated by considering available habitat to be that within a circle having as a radius the mean distance between nest sites and the lek of capture or the closest lek. This distance was used as a mobility index for nesting females. The nest site rather than the lek was used as a circle locus because this circle included most of the prelaying locations of the hens and was considered to be the area with which females were familiar.
6.4 Results During this study, 1,113 locations were recorded for 21 female prairie chickens that were radio-tagged and monitored during spring and summer for an average
GREATER PRAIRIE CHICKENS IN MINNESOTA
201
of 57.1 days. The average number of locations per female was 53.0. Two females were monitored for two consecutive reproductive seasons. Sixteen females were cannon-netted on the Pembina lek (Fig. 6.4), and five were nest-trapped between incubation day 11 and 15. Ten recaptures were made on roost sites to inspect the condition of birds or to remove the radio at the end of the monitoring period.
6.4.1 Breeding chronology The peak of female attendance on leks was about 12 April, and the peak of copulation was about 20 April. After the copulation peak, visits by females to leks gradually declined until few or none came. Two to 3 weeks after the copulation peak, visits to leks and copulations resumed but numbers were more sporadic than earlier. Late copulations probably represented renesting activity, because welldeveloped brood patches were noted on females banded 20 May 1975, 5 May 1976, and 13 May 1976. Cannon-netting and radio-tagging were carried out near the end of the first copulation period in 1975 and during the peak in 1976, and probably delayed some copulations. In 1977, all trapping occurred before 13
Fig. 6.4. Location of prairie chicken nests and leks on the Pembina Trail Preserve, by year.
202
W. D. SVEDARSKY
April. All the observations from blinds occurred during the morning display period except three afternoon sessions. Fewer females came to display grounds in the afternoon than during the morning of the same day. Nevertheless, on 10 May 1975 a successful copulation was observed in the afternoon.
6.4.2 Nesting chronology Four radio-tagged females, whose nest were located and back-dated, commenced egg-laying an average of 3.8 days (range 1-5) after they were observed to copulate. The rate of egg-laying observed for five females was 1.0 egg per day. Assuming this laying rate, two females (E-75 and B-75) were present on the lek 3 and 2 days, respectively, after they had initiated egg-laying. One of these (E-75) was observed to feed on corn placed on the lek; this may have been the stimulus prompting her return. On the other hand, three eggs in her clutch of 15 were infertile and may have been laid before successful copulation. The other female (B-75) was not observed from a blind, and her behavior on the lek is unknown. Another (E-76) was trapped on the lek the day before the calculated onset of her egglaying. It is unlikely that she began laying the day after trapping because birds typically showed reduced activity for 1-2 days while adjusting to the transmitter. Thus she probably also had established a nest and commenced laying prior to visiting the lek. While on the lek she fed on corn and acted receptive to males but was not observed to copulate. Her nest was destroyed by a predator before egg fertility could be determined. Most females visited nests to lay eggs between 0800 and 1400 hours. Once incubation began, females were on nests constantly except for brief feeding periods in the morning and afternoon. The average length of the incubation period was 25.2 days. Nests hatching in late June and early July tended to hatch about 2 days faster than those hatching in early June. Hatching apparently occurred about 1 day before the broods left the nest. A nest was discovered at 1100 hours on 20 May with one pipped egg. The hen was still present at 1700 on 21 May, but she left the nest the next day.
6.4.3 Clutch size and fertility Females that initiated nests early in the year averaged 14.6 eggs per clutch and had larger clutches than those nesting later (Table 6.1). Because of the uncertainty in distinguishing between first attempts and renests, nests were grouped by periods for clutch-size evaluation. Only clutches being incubated were used in these calculations. Fertility was determined from completed clutches where presumably no partial nest predation had taken place. Of 246 eggs, 226 were fertile (91.9%). Infertile eggs occurred in six of 19 (31.6%) reference nests with a mean of 3.3 (range
GREATER PRAIRIE CHICKENS IN MINNESOTA
203
Table 6.1. Prairie chicken clutch sizes (means) for periods when first egg was laid, Kertsonville Study Area, northwestern Minnesota, 1975-77
Period I II III IV V
(15-30 April) (1-15 May) (16-31 May) (1-15 June) (after 16 June)
Mean + SE
n
Probable nesting attempt
0.43 0.45 0.43 1.98
9 6 9 2 1
1st 1st and 2nd 2nd 2nd and 3rd 3rd
14.6 13.0 12.8 9.0 5.0
± + ± ±
1-6) infertile eggs occurring per nest. No relationship between egg fertility and nest chronology, nest characteristics, or female age was noted.
6.4.4 Nesting success and renesting A nest was considered successful when at least one egg hatched. Nest failures directly influenced by man (2 cases of researcher-induced abandonment and 3 cases of grass fires) were excluded from the overall success calculation so as to more closely approximate the extent of natural nest losses. Nesting success for the 3 years was as follows: in 1975, 8 of 10 nests (80%) hatched; in 1976, 9 of 14(64.3%); and in 1977, 3of7 (42.9%), for an overall nesting success of 62.4%. Striped skunks (Mephitis mephitis) were the most frequent nest predator (Table 6.2). Red foxes (Vulpes vulpes) destroyed three nests and, in two instances, preyed upon the nesting female. The female apparently escaped after capture in the third instance as evidenced by feathers found at the nest. When nests were destroyed, females generally renested (Table 6.3). Of six radio-tagged females known to have lost nests, five renested; one female renested twice. The hen that did not renest had lost her nest about incubation day 20. Females that lost nests during the laying period renested promptly and laid eggs in a new nest within 3 or 4 days. The first nest of female C-76 was destroyed around incubation day 4, and 9 days later she began laying in a second nest. This nest was destroyed around incubation day 2, and only 4 days later she began laying in a third nest. Her clutch size per nesting attempt was 13, 11, and 11. Female F-77 started laying in a second nest 12 days after her first nest was destroyed around incubation day 16. Her clutch size per nesting attempt decreased dramatically from 13 to seven. Both adult and yearling females renested. It was uncertain whether females copulated between nesting attempts, but two renesting females were recorded on or very near leks. Renests were an average of 760 m from the previous nest except that of female F-77, who moved 5.6 km to renest after she was nearly depredated on her first nest.
W. D. SVEDARSKY
204
Table 6.2 Causes of prairie chicken nest failures, Kertsonville Study Area, northwestern Minnesota, 1975-77 n Nest (%)
Cause Striped skunk (Mephitis mephitis) Red fox (Vulpes vulpes) Fire Hen killed off nest during egg laying or incubation by raptor Researcher-influenced abandonment Raccoon (Procyon lotor) Franklin's ground squirrel (Citellus frankini)
4 (25) 3a (19) 3 (19) 2 (12) 2 (12) 1 (6) 1
Total a
(6)
16
Two hens killed while on nest.
Table 6.3. Prairie chicken renesting data, 1975-77
Female
A-75 F-76 E-76 C-76 C-76 F-77 Mean a b
Date of nest destruction or abandonment 26 23 22 10 29 3
May May May May May June
1975 1976 1976 1976 1976 1977
Time interval between nests (days) ?a 3 4 9 4 12 6.4
Evidence of female visiting Completed clutch size Distance booming between nests of consecutive ground (m) between nests nests 200 600 1,060 1,430 510 5,600
? ? 9+ 13 11 13
? ? 13 11 11 7
No No Yes Yes No No
760b
Second nest depredated before egg count taken. The extreme value of female F-77 was excluded in calculating this mean.
6.4.5 Nest placement Thirty-six nests were located during the study period; 22 were nests of radiotagged females, and 14 were located through the use of a cable-chain drag. Three additional nests were found incidentally by personnel, not in this study, one in 1973 and two in 1978. Nests were not always located closest to the lek of copulation (Table 6.4). One
GREATER PRAIRIE CHICKENS IN MINNESOTA
205
female (E-76) copulated on the Pembina lek in 1977 when it had 19 males and was 2.5 km from her nest (Fig. 6.4). Two other leks were located 1.2 and 1.4 km from her nest and had six and nine males, respectively. These were positioned somewhat between her nest and the Pembina lek, suggesting strong fidelity to the lek of prior experience; she was trapped and probably copulated on the Pembina lek in 1976. At least four females trapped on the Pembina lek returned there for copulation. The remainder of the radio-tagged females either copulated on another lek or on the Pembina lek when observers were not present. One radiotagged female visited another lek in addition to the Pembina lek, but was not observed to copulate. Females radio-tagged on the Pembina lek nested on or within 700 m of the Preserve. Nests were concentrated on the westerly portion of the tract; a 106-ha area contained 21 of 39 nests observed from 1973 to 1978 (Fig. 6.4). Thus, 54% of the nests were on approximately 15% of the Preserve. Two concurrently active nests discovered in 1978 were only 7.9m apart. It is interesting that no nests were found close to the Pembina lek even though 16.2 ha of undisturbed alfalfa and bluegrass were located 400 m to the northeast. Also, 40 ha of undisturbed prairie and over 100 ha of lightly grazed prairie were located within 1.0 km to the north and northeast. Although this area was not searched for nests, it received no use
Table 6.4. Distances (m) between prairie chicken nests and leks Relationship of reference lek to nesting female
a
Mean ± SE
No. of nests
Copulaton site preceding a nest Yearling Adult Total
1,221 2,235.3 ± 269 1,981.8 + 317
1 3 4
Capture site preceding a nest Yearling Adult Total
1,642.8 ± 306 1,727.5 + 147 1,708.7 ± 129
4 14 18
Closest lek to nest Yearling Adult Total3
859.0 ± 148 1,283.4 ± 109 1,044.6 + 73
6 16 37
Closest lek to nest Successful nests Unsuccessful nests
1,171.4 + 103 974.7 ± 149
20 10
Total includes additional birds of unknown age.
206
W. D. SVEDARSKY
by radio-tagged birds. This may have been because it was separated from the Preserve by a well-traveled road and railroad tracks. Two females, which were radio-tagged in two consecutive seasons, nested the second year close to their successful nests of the previous year, demonstrating nest-site tenacity. After the loss of her first nest in 1976, female E-76 moved 950 m to successfully renest along a roadside. After copulating on the Pembina lek in 1977, she nested only 4.6 m from her successful nest of 1976. Female H-76 was trapped on a nest in 1976 and was followed throughout the summer until the transmitter was removed in the fall. She was trapped on the Pembina lek in 1977 and later radio-tagged when her nest was discovered by cable-chain dragging. Her 1977 nest was located 29.8 m from her successful nest of the preceding year.
6.4.6 Nest characteristics and success Sixteen of 36 (44%) nests were found in bluestem habitat. Eight of 36 (22%) were found in brome, with lesser proportions in other habitats. Nests were mostly (89%) in habitats that were undisturbed for one or more growing seasons, but two renests established in recently burned areas illustrated the range in acceptable nesting habitats. Nesting success was higher in nonnative habitats (85.7%) than in native habitats (53.3%), but significant differences were noted only with respect to the brush index; nonnative habitats had fewer brush clumps (Table 6.5). A total of 73 plant species occurred at nests in native habitats compared with 55 species in nonnative habitats. When comparing successful and unsuccessful nests (Table 6.6), significant differences were noted for the brush index (P < 0.01) and litter depth (P < 0.05). These data suggest that more nest predation occurred in undisturbed habitats with greater amounts of litter and more clumps of brush. Such areas were possibly more attractive to mammalian predators, and thereby increased the likelihood of nest discovery. Canopy coverage and vegetation density did not significantly affect nest success (Table 6.6). Therefore, nest concealment and impediments to predator travel appeared less important than other factors in accounting for differences in nest predation.
6.4.7 Brood mortality factors Eleven radio-tagged females hatched 116 chicks during the study. Only two chicks in the brood of one hen (C-75) were known to be alive at the end of August (Table 6.7). The average length of time that radio-tagged females had broods was 23.9 days, excluding two hens (B-75 and E-76) because of transmitter-influenced mortality and loss of radio signal, respectively. It was generally difficult to locate broods in dense cover to confirm their existence or estimate number of chicks. This was especially true for young broods because females would commonly flush some distance from the concealed brood
207
GREATER PRAIRIE CHICKENS IN MINNESOTA
Table 6.5. Comparison of nesting success and structural characteristics of prairie chicken nests in nonnative and native habitats Nonnative habitat (success = 85.7%, n = 14) mean + SE
Characteristic Visual obstruction reading (dm) at :
Canopy coverage (%) Litter depth (cm) Brush index 3 a b
100% 50% 0%
2.2 2.9 4.7 53. 6 8..6 2 .3
Native habitat (success = 53.3%, n = 15) mean + SE
0..1 0..1 0..5 7.5 ± 0.9 b ± 11. l
2..0 2..5 4..4 58.7 9.5 27.1
0.2 0.3 0.3 ± 10 .1 ± 1 .5 1. 0
± ±
±
Number of brush clumps within 50 m of nest. Significantly different, Mest, P < 0.05.
when approached by the investigator. These periodic disturbances possibly contributed to mortality because they appeared to stimulate movement. Documenting brood mortality was also difficult; however, two occasions were recorded. A 10day-old brood of at least five chicks was killed by a fox on the roost site, and a 30-day-old chick of one female (C-75) was killed by a female harrier (Circus cyaneus) while the brood was feeding in a mowed alfalfa field (Svedarsky 1980). Because young galliform chicks may be adversely affected by wetness and associated chilling, hatching dates were examined relative to precipitation and ternTable 6.6. Comparison of characteristics between succussful and unsuccessful prairie chicken nests
Characteristic Visual obstruction reading (dm) at: 100% 50% 0% Canopy coverage (%) Litter depth (cm) Brush index3 Distance to travelway (m) Distance to short cover (m) Number of days nest initiated after earliest nest per year a b c
Number of brush clumps within 50 m of nest. Significantly different, r-test, P < 0.05. Significantly different, Mest, P < 0.01.
Successful nests (n = 20) mean + SE
2.2 2.7 4..4 51..0 7,,8 4..6 70.0 106 .4
± 0.2 ± 0.,2 ± 0..2 ± 8.,3 ± 1,, 1 ± 2 ,1 ± 19,,5 ± 27 .3
15 .8 ±
3,.0
Unsuccessful nests ( = 9) mean ± SE
2.,0 2.6 4.,7 67.8 11. 9 38..6 51.,6 97., 1
± ± ± ± + ± ± ±
12,.4 ±
0,, 1 0.2 0,,8 5. 9 0.,7" 17,,3C 15.,4 25,.0 4,.0
208
W. D. SVEDARSKY
Table 6.7. Longevity and mortality factors of radio-tagged praire chicken broods
Female (age)
Hatching date
No. of chicks
Estimated longevity (days)
E-75
22 June 1975
10
31
Extensive early movements through undisturbed cover, 5.0 cm precipitation during first 4 posthatching days, harness entanglement of brood female
D-75 (ad)
2 July 1975
5
17
Extensive early movements through areas with 10-20 cm of standing water
C-75 (ad)
28 June 1975
15
51 + a
B-75
1 July 1975
5
3b
Transmitter-influenced mortality of hen
1 July 1976
13
?a
Unknown owing to loss of radio contact after 3 days
H-76 (ad)
26 June 1976
14
21
Extensive early movements through undisturbed habitat containing 10-20 cm of standing water
G-76 (ad)
8 June 1976
10
10
Extensive early movements, 7.5 cm precipitation following hatching, depredated on roost by fox
C-76 (ad)
7 July 1976
11
28
Extensive early movements, 3.8 cm precipitation during first 3 posthatching days, researcher disturbance
E-76
3 June 1977
13
22
Extensive early movements, and twice crossed a ditch with 10 cm of running water
H-76 (ad)
16 June 1977
7
25
Unknown
D-77 (ad)
7 June 1977
13
9
10.5 (1.06)
23.9 (4.3)
(yr)
(yr) E-76
(yr)
(yr)
Mean (± SE) a b
Possible or observed mortality factors
Extensive early movements, 5.2 cm precipitation during first 3 posthatching days
Suspected predation
Minimum value as 2 chicks were with the female at the end of the monitoring period. Not use in calculation of mean longevity.
GREATER PRAIRIE CHICKENS IN MINNESOTA
209
perature patterns—especially for the primary hatching period of 25 May-14 June (Fig. 6.5). Of the 3 years, 1976 was the warmest, and was relatively dry during this period. Regionally, 1976 was the second-driest spring on record in Minnesota. That year, the majority of broods hatched before significant rains began in mid-June. A 25% increase in the number of displaying males in the Kertsonville Study Area was noted in spring 1977 and may have reflected favorable weather conditions for young broods in 1976. However, chick numbers in unmarked broods, another indicator of brood survival conditions, were lower (Jc = 4.6, n = 5) in 1976 than in 1975 and 1977 (jc = 6.2, n = 4 and jc = 6.4, n = 6, respectively), but this may have been the result of a small sample. Radiotagging was carried out in late May 1975 and in late April and May 1976. This, coupled with the period of adjustment to radio packages, resulted in the broods of six radio-tagged hens hatching later during periods of heavy precipitation (Table 6.7). This delay may have partly accounted for the high mortality of radiotagged broods. Precipitation, temperature, and movements in relation to habitat structure were felt to be key determinants of chick mortality. Movements in turn were probably influenced by disturbance and habitat quality. Female G-76 had at least five chicks 10 days after hatching even though 7.5 cm of precipitation had occurred in that period. However, most of this occurred after the brood was 5 days old, and habitat use consisted mainly of a well-drained ridge with sparse cover interspersed with shrubby lowlands. This habitat combination provided "drying off' areas following rainfall and protection during storms. The area surrounding nests of some radio-tagged females was interlaced with shallow swales that contained 10-20 cm of water for up to 2 weeks following heavy rains. Radio-tagged brood females were known to cross these areas with chicks less than 1 week old. Although no drowned chicks were found, wetting of chicks surely occurred, and some members of young broods may have become separated from the hen and subsequently died. In 3 years, I observed an average of 5.8 chicks (SE = 0.78, n = 15) in unmarked broods 2-6 weeks old that were observed under conditions where it was likely the entire brood was counted. Single chicks observed without a female were not included in this estimate. Based on an average clutch size of 10.9, this represents a brood survival rate of 53.2% for chicks averaging about 4 weeks of age.
6.4.8 Female mortality factors Of 21 females that were radio-tagged during spring or early summer, nine were alive at the end of summer, an estimated survival rate of 42.9% from 1 May through August. This converts to a summer mortality rate of 19.0% per month or an annual mortality rate of 228.4%, if the mortality rate was constant per month. Evidence that summer mortality was disproportionately greater than dur-
210
W. D. SVEDARSKY
Fig. 6.5. Chronology of prairie chicken hatching and occurrence of precipitation on the Pembina Trail Preserve 1975-77. Insets include the numbers of displaying males in the Kertsonville Study Area, the mean numbers of chicks per brood (2-6 weeks old), and the mean temperatures for the primary hatching period (25 May-14 June).
GREATER PRAIRIE CHICKENS IN MINNESOTA
211
ing other periods of the year was that four of five radio-tagged females alive at the end of the summer in 1975 and 1976 were observed the following springs, for a fall-spring survival rate of 80%. The summer survival rate of42.9% is probably lower than normal due to the presence of radio packages which, in two instances (B-75 and E-75), were known to disadvantage females and possibly increased the likelihood of predation for others. The 80% fall-spring survival estimate may be high relative to the population because four of the five spring survivors were initially marked as adults. Although differences were not significant (P > 0.05), the mean length of time that adults were monitored was 63.4 days compared with 44.6 days for yearlings. Thus, the summer survival of adults may be greater than that of yearlings. Ten radio-tagged females were lost to predators. Three were taken by mammals, presumably fox, before nests were located. One female was killed during the egg-laying period by a great horned owl (Bubo virginianus) and another while off the nest feeding during incubation by an unidentified raptor. Two females were killed on their nests during incubation. One carcass was located near a fox den 2.0 km from the nest, and signs at the other nest also suggested fox predation. One female was depredated on a roost site with 10-day-old chicks. Her remains and the chewed transmitter were recovered from a fox den 1.0 km from the predation site. The brood of one hen (D-77) may also have been preyed upon, as she made a long sporadic movement after brood loss. Two females were depredated during the postbrood period, one by a mammal and the other by a raptor. Raptor kills were found 0, 30, and 50 m from trees taller than 8 m; these may have served as ambush sites. Nine often instances of predation occurred within 43 days after radio-tagging. This may suggest that females had not adjusted well to the radio package and, because of excessive preening or some other factor related to the presence of the transmitter, were more susceptible to predators. However, I believe that at least three occurrences (two females on nests and one with a roosting brood) of predation were not related to radio-tagging. Quantitative estimates of fox populations were not made in the study area, but at least two dens (possibly the same family) were known to be active each year. These were on or within 1.6 km of the Preserve. Fourteen incidental fox sightings were made during the 1976 study period, and three foxes were seen from a car on 2 January 1977 in 2 hours of driving in the vicinity of the Preserve. These observations and the fact that foxes were protected from trapping on adjoining cattle ranches to the north and east of the Preserve suggested that the fox population was moderately high in the area.
6.4.9 Movements Movements tended to be greatest during the period when females were visiting leks and establishing nests (Fig. 6.6). All radio-tagged females restricted their
212
W. D. SVEDARSKY
Fig. 6.6. Movement indexes of female prairie chickens by reproductive period.
prelaying movements to the vicinity of the Preserve except two yearlings that moved steadily away from the Preserve after tagging. One was last located 8 days after tagging, 5.5 km from the lek of capture. Radio contact was lost with the other, 6 days after tagging when she was 2.8 km from the lek where captured. Her remains were later recovered from a fox den 4.7 km from the same display ground. Movements decreased once egg-laying commenced and were later restricted to feeding locations during incubation, a mean of 396.2 m from the nest (Fig. 6.7). Distances to feeding locations did not vary according to the stage of incubation, but the number of observations was low (18) and feeding near the nest may have been undetected due to limitations in telemetry accuracy. Young broods had somewhat greater movement indexes and minimum movements than older broods, although differences were not significant (P > 0.05) (Figs. 6.6, 6.8). It is noteworthy that although young chicks were smaller than older chicks and likely encountered more difficulty in moving through ground cover, they traveled comparable distances. The largest "minimum movement" for a weekly period (4,075 m) was for a 1-week-old brood (Fig. 6.8). Movements of young broods (1-14 days old) were evaluated to determine the effect of habitat
GREATER PRAIRIE CHICKENS IN MINNESOTA
213
Fig. 6.7. Frequency distribution of feeding-location distances from the nest during incubation.
and land use (Table 6.8). Movement differences were not significant (P > 0.05) but were generally less in disturbed habitats compared to monotypes (cropland and brome) and undisturbed habitats. One hen (F-77) made a 6.4-km move after being nearly depredated on a nest. After this, she was "night-lighted" and netted on an assumed roost that was, instead, a second nest. She immedately left the area after being trapped, and 52 days later was located 12.3 km from the second nest site. Encounters with predators
Fig. 6.8. Minimum movements (m) of prairie chicken broods per week of age. Differences not significant (P > 0.05) between age groups.
214
W. D. SVEDARSKY
Table 6.8. Movement indexes (m/hr) of 1- to 14-day-old prairie chicken broods per habitat and land use type Movement index mean + SEa
Habitat type Burned brome Undisturbed bluestem Undisturbed sweet clover Cropland Burned willow Undisturbed brome Undisturbed alfalfa Hayed alfalfa Burned bluestem Grazed bluestem a
34.8 30.8 28.7 21.9 19.3 16.8 16.3 14.3 12.5 6.8
± 9.7 ± 14.8 ± 13.4 ± 13.3 ± 7.8 ± 4.1 ± 9.0 ± 1.1 ± 2.4 ± 1.8
n 11 6
5 1 8 6 4 10 17 13
Differences not significant at the 5% level by Scheffe's (1953) s-test.
clearly stimulated long movements by this female. Another long movement occurred when one female (D-77) moved 2.7 km in no more than 2 days after losing her brood. In 6 days she had moved 6.0 km away from the brood area, but then began moving back and was eventually located very close to the last brood location. Sixty-nine days after brood loss she was trapped on a roost 4.8 km away from the brood area but in a different direction from the previous long movement. No other female made a comparable movement after brood loss and, consequently, postbrood movement data of D-77 was not used in the determination of means. A predator encounter may have accounted for the loss of the brood and the ensuing long movement. Movements of broodless females were not significantly different from females with broods (Fig. 6.6). Furthermore, movements of broodless females did not appear to change with time following brood loss.
6.4.10 Home range Modified, minimum home ranges were determined for reproductive periods when at least nine (x = 15.7) locations per female were recorded. Home ranges were largest during the prelaying period, progressively decreased during the brood period, and tended to increase slightly during the postbrood period (Table 6.9). The cumulative home ranges for two females (C-76 and E-76), monitored from prelaying into the postbrood period, were 163.1 and 104.4 ha, respectively. Both females lost broods during the third week. Another female (D-77) also was monitored from prelaying to the postbrood period and lost her brood during the second week. Her postbrood home range during the first 2 weeks was 347.2 ha, markedly greater than the 31.5-ha mean of five other hens for the same period, and was not used in calculating the mean. After the second week following brood
GREATER PRAIRIE CHICKENS IN MINNESOTA
215
loss, this hen (D-77) moved into a ranch where access was limited, and location data were not collected until the female was trapped at the end of the summer for radio removal. Her cumulative home range was 503 ha through the second postbrood week, three times as large as that of any other female. The cumulative home range for the only female (C-75) followed for a full season and having chicks at the end of the summer was 82.6 ha. Most of the laying home range was included within the prelaying range (Table 6.9). To a lesser extent, brood home ranges during the first 2 weeks were included within the laying range. Older broods tended to move into new areas more than young broods. Once females lost broods, they generally remained within the home range of the preceding brood period, at least for the first 2 weeks. Later in the postbrood period females began to move more, and into new areas as they joined other birds in loose flocks. Four females (F-76, E-76, D-77, and F-77) were all in the company of other prairie grouse when trapped on roosts in late August or early September for transmitter removal. When females renested, the overlap in laying home range per nesting attempt was less than 5% in all four instances where sufficient locations (> 9) were recorded for home-range estimation. This suggests a strong avoidance of those areas where nest predation had occurred. No difference was observed in homerange size per nesting attempt. In 1975 and 1976 and to a lesser extent in 1977, there was overlap in the preincubation (prelaying and laying) home ranges of some radio-tagged females (Figs. 6.9, 6.10, 6.11). These areas of overlap were usually in habitats used for feeding Table 6.9 Home ranges of female prairie chickens per reproductiove period and percent of home range not included in the preceding period % of new area
Home range (ha)
a
Period
Mean + SEa
Prelaying
82.0 ± 15.6 A
Laying
31.4 ±
5.8B
Brood: 0-2 Weeks 2-4 Weeks 4-6 Weeks
18.3 + 10.9 ± 9.7
Postbrood: 0-2 Weeks 2-4 Weeks 4-6 Weeks
Mean + SE
n females
11
37.0 ± 14.4
3
3.9 B 1.9 B
9 3 1
64.9 + 7.9 79.5 ± 16.9 91.8
4 3 1
31.5 ± 11.9 B 14.2 ± 6.4 B 20.0 ± 0.6 B
5 5 2
47.4 ± 22.9 71.0 ± 31.8 82.3 ± 7.5
4 5 2
n females 5
Means followed by the same uppercase letter are not significantly different at the 5% level using Scheffe's s-test.
216
W. D. SVEDARSKY
such as small-grain and hayed alfalfa fields. Occasionally, unmarked prairie chickens were flushed in these fields and were probably feeding there as well. By contrast, in all 3 years, there was little overlap in the home ranges of radio-tagged broods (Figs. 6.9, 6.10, 6.11). This may have resulted from nest placement in 1976 and 1977; however, three nests were relatively close together in 1975 yet the broods used different areas. There were no apparent differences between adult and yearling females in these spatial relationships. Certain home-range comparisons were made for two females (E-76 and H-76), which were monitored during 1976 and 1977. Most of the prelaying and laying home ranges associated with E-76's successful nest of 1976 overlapped with those
Fig. 6.9. Spatial relationships of preincubation and brood home ranges of female prairie chickens, and to nest sites and leks in 1975.
GREATER PRAIRIE CHICKENS IN MINNESOTA
217
of the same periods in 1977 (Fig. 6.12). Radio contact was lost early in the brood period of 1976, but initially, brood ranges were identical between years. This could be partially explained by her nests of 1976 and 1977 being very close together along a road ditch containing water. In 1977, about half the brood range
Fig. 6.10. Spatial relationships of preincubation and brood home ranges of female prairie chickens, and to nest sites and leks in 1976. Home ranges associated with renests are shown if a shift occurred.
218
W. D. SVEDARSKY
overlapped the prelaying and laying home ranges, but the postbrood range did not overlap with any previously monitored home range. Female H-76 was nest-trapped both years and showed no overlap in brood ranges although nests were close together (Fig. 6.13). This could have resulted
Fig. 6.11. Spatial relationships of preincubation and brood home ranges of female prairie chickens to nest sites and leks in 1977.
GREATER PRAIRIE CHICKENS IN MINNESOTA
219
from land use conditions—she took her brood to recently burned prairie each year. Most of the 1977 brood and postbrood home ranges were within the 1976 postbrood range. The brood and postbrood ranges used in 1976 were about twice the size used in the same periods in 1977.
6.4.11 Habitat use and preference The primary consideration in selecting the Preserve for intensive study was that a variety of habitat types and land uses occurred within a relatively small area (10.24 km2). Efforts were made to diversify land uses on the Preserve by coordinating haying leases and conducting prescribed burning. Thus, by providing a
Fig. 6.12. Home range of a female prairie chicken (E-76) during the prelaying and laying periods of 1976 and the prelaying, laying, and brood periods of 1977.
220
W. D. SVEDARSKY
Fig. 6.13. Home range of a female prairie chicken (H-76) during the brood and postbrood periods of 1976 and 1977.
broad spectrum of habitat conditions to females, their use and preference for cover during each reproductive period could be evaluated. Preincubation Undisturbed bluestem was the most frequently used habitat during the preincubation period (Fig. 6.14), although the preference index (Fig. 6.15) suggested it was used in relation to availability. This use was associated with nest establishment and egg-laying. Morning and afternoon feeding locations were mostly in cropland and hayed alfalfa. A higher preference index was shown for alfalfa than other cropland due to differences in availability, but cropland was the most frequently used habitat in this period. Ten of 17 radio-tagged females were located using cropland at least
GREATER PRAIRIE CHICKENS IN MINNESOTA
221
Fig. 6.14. Major habitats used by female prairie chickens per reproductive period and time of day. Includes habitats with at least five of the locations per reference period. Land use prefixes: UD = undisturbed for 1 year or more, HA = hayed, GR = grazed, BU — burned the current spring or fall of preceding year.
once. Grazed aspen was preferred during the midday period, presumably for shade and to feed on dandelions (Taraxacum spp.), which emerge early in these areas. Sedge habitat accounted for 63% of the roost locations. Incubation and nesting Off-nest feeding locations during incubation were within the area used during the preincubation period, but a narrower range of habitats was used. Of 18 locations, nine were in burned or grazed bluestem; six in cropland; and one each in hayed alfalfa, roadside and undisturbed sedge— suggesting a strong preference for disturbed habitats. Feeding in the immediate vicinity of the nest may have occurred, but it was not documented. Habitat use and preference were evaluated for 36 nests located during 1975, 1976, and 1977. Nests were located a mean of 1,290 ± 97.2 m from the lek of capture, or the closest lek in instances of nest-trapped females. This distance was used as a reference radius to delineate available habitat around the nest for preference evaluation. Brome and redtop were preferred for nesting; bulrush, aspen, and cropland were avoided (Fig. 6.16). Other habitat types were essentially used relative to
222
W. D. SVEDARSKY
Fig. 6.15. Preference or avoidance of habitats by female prairie chickens per reproductive period. Values based on at least three (x = 5.4) females having a habitat type available during a given reproductive period.
GREATER PRAIRIE CHICKENS IN MINNESOTA
HABITAT
223
TYPES
Fig. 6.16. Nest-site selection for habitat type in 36 prairie chicken nests. Brome was used more than expected, P < 0.01.
their availability except roadside, which was used in 2 consecutive years by the same female. Characteristics of nests in brome and redtop were compared with all other nests in an effort to determine what characteristics nesting females might be selecting for (Table 6.10). Both redtop and brome had stems bearing leaves for 25-30 cm of their height, which persisted after frost. This resulted in nesting cover that was as much vertically oriented as horizontally and accounted for brome nest sites having a significantly higher, 100% visual-obstruction reading than other nests (x = 2.7 dm versus Jc = 1.8 dm respectively, P < 0.01). The leaves of most other grasses arose near the base of the plant and tended to become Table 6.10. Comparison of prairie chicken nest-site characteristics in brome, redtop, and all other habitats used Habitat used for nesting
Characteristic Height (dm) where visual obstruction was: 100% 50% 0% Canopy coverage (%) Litter depth (cm) Brush index 0 Distance to short cover (m) a b 0
Brome (n = 8) mean ± SE
2.7 3.5 4.9 75.0 9.6 6.9
a ± 0.3 b 0.3 ± + 0.2 + 7.5 b + 2.0 + 4.4
90.4 + 41.4
Redtop (n = 3) mean + SE
1 .8 2..4 5 .2 21 .7 5 .3 3 .7
± 0.,2 ± 0..2 ± 0..9 b ± 14.,2 + 3..4 + 2..0
93 .3 + 33 .3
Means significantly different at 1 % level using Scheffe's .y-test. Means significantly different at 5% level using Scheffe's s-test. Number of brush clumps within 50 m of nest.
All others (n = 25) mean+ SE
1.8 2.4 4.4 58.4 9.4 24.4
+ O.l a b ± O.l ± 0.3 ± 6.4 ± 0.9 + 7.8
121.8 ± 23.6
224
W. D. SVEDARSKY
flattened by winter snows. Litter (horizontally oriented, residual vegetation near the ground) was apparently not important in nest site selection, although brome nests did have the greatest litter depth (x = 9.6 cm) of any nest category. Two successful prairie chicken renests were established in recently burned areas (no litter). Another successful prairie grouse nest was found in a burned area and was likely a renest also, as it contained no breast feathers. In 1978, two nests were located in brome burned the previous year, in which there was no litter. Nesting preference was also evaluated with respect to land use (Fig. 6.17). Undisturbed habitats were preferred, and heavily grazed and hayed habitats were avoided. Apparently, recently burned habitats did not greatly deter nesting because use of these was comparable to their availability. Hens used a variety of topography for nesting. This ranged from a hummock in a sedge lowland surrounded by 20 cm of standing water to a dry ridge top with sparse vegetation. Although the range of acceptable nesting conditions was broad, undisturbed sites with dense residual cover near ground level were preferred, with other characteristics appearing less important in nest site selection. Early brood period Habitat use was recorded for nine young broods (Fig. 6.14), which had moved an average of 983.2 + 171.7m away from the nest by 14 days after hatching. This distance was the reference radius used in delineating the circle of available habitat for calculation of preference ratings (Fig. 6.3). After hatching, broods generally moved directly from undisturbed cover surrounding nests to habitats that had been disturbed. Young-brood locations in undisturbed habitats resulted, in part, from nest placement rather than habitat
Fig. 6.17. Nest-site selection for land use type in 36 prairie chicken nests. Land use undisturbed (> 2 years) was used more than expected, P < 0.05.
GREATER PRAIRIE CHICKENS IN MINNESOTA
225
selection. Undisturbed bluestem contained 7.4% of the total locations but had a preference rating of —6.7. In contrast, burned and grazed bluestem contained 31.2% of the total brood locations, and both showed preference ratings of +8.6. Disturbance apparently did not affect the use of alfalfa; both undisturbed and hayed areas were preferred. Willow was generally not used unless it had been disturbed, and then most locations were utilized during the midday periods, suggesting its use for shade. Cropland had a preference rating of - 13.8 and was clearly avoided by eight young broods. Late brood period The circle of available habitat used for young broods was also used for older broods, from 2 to 4 weeks of age. The maximum distance that three broods had moved by 4 weeks (x = 1,580 + 528.4 m) was used to delineate available habitat for broods older than 4 weeks. Late-brood habitat use was similar to that of young broods in favoring disturbed habitats (Fig. 6.14). Alfalfa was used most and was highly preferred (Fig. 6.14). Undisturbed bluestem was clearly avoided by four broods but was used after it had been grazed or burned. Disturbed willows were used mainly during midday and evening periods, as in the case of young broods. Cropland, with a preference rating of - 16.7, was avoided by five broods. Postbrood period Postbrood habitat use and preference were similar to that of the preincubation period (Figs. 6.14 and 6.15). Bluestem was the major habitat used, and grazed bluestem was preferred (+12.9). Cropland accounted for 15.3 % of the total locations; this use occurred mainly in the latter part of the summer as females began using harvested small-grain fields. The preference rating of - 1.2 likely underestimated the importance of cropland during late summer and early fall because home ranges were generally shifted to these areas and thus changed their availability. Willow and aspen habitats were used more when disturbed and during midday and evening periods. As during the preincubation period, most roost locations were in sedge habitats. Overall, 70.9% of postbrood locations and 58.0% of the preincubation locations were in disturbed habitats.
6.5 Discussion This was a management-oriented study of the reproductive ecology of female prairie chickens. This portion of the life cycle is generally considered to be the key factor in determining the status of prairie chicken populations. The study provided life-history data of the species at the northern edge of its present range. Below, these are compared with findings from other parts of the range.
6.5.1 Courtship and breeding The breeding chronology at the northern edge of the range in Minnesota is about the same as in central Wisconsin and the Lower Peninsula of Michigan, but about 1 week later than near the southern edge of the range.
226
W. D. SVEDARSKY
Female visits to leks peaked about 12 April, 6 days earlier than the average peak day for central Wisconsin (Hamerstrom & Hamerstrom 1973). Early springs in 1976 and 1977 may partly account for earlier attendance in this study area, located 400 km north of the Wisconsin area. Copulation peaks occurred about 8 days after female attendance peaks, in contrast to an average of 3.2 days in Wisconsin. The timing of copulation peaks was similar, however, occurring about 20 April. Ammann (1957, p. 150) reported that 22 April was the "height of courtship activity" in the Lower Peninsula of Michigan and that 2 May was for the Upper Peninsula. Arthaud (1968, p. 81) noted that the first 2 weeks in April were the peak of the "breeding season" in southwestern Missouri. Robel (1970) noted that the first copulation peak occurred between 21 and 30 April in northeastern Kansas but observed some copulations before 10 April. In this study the earliest copulation was observed on 14 April. I observed a small, second copulation peak in two springs, which occurred 20-30 days after the first peak as similarly reported in Kansas and Wisconsin.
6.5.2. Preincubation movements Movements and home ranges during April and early May tended to be greater than for other periods, agreeing with the findings of Robel et al. (1970b). Yearling females appeared to move more than adults; they apparently had not developed as strong an attachment to a particular area as influenced by nesting and/or breeding experience. Both of the radio-tagged females that left the immediate area of the Preserve in the spring were yearlings. Of five radio-tagged females alive at the end of the monitored seasons, four returned to the same lek where they were trapped the preceding year. In 1977, three females were radio-tagged early enough in the season to provide reliable, preincubation home-range estimates. These had a mean diameter of 2.2 km. Bradbury (1981), in a discussion of the relationship of hen mobility to lek spacing and evolution, predicted that interlek distances should be the sum of one female home-range diameter (here recorded as 2.2 km) plus twice the distance the lek can be detected (here conservatively estimated at 3.2 km). He might thus predict, using the above samples, an average of 5.4 km between leks. This is more than three times the 1.6 km mean distance between three leks associated with the Preserve in 1977. It is also more than 2 km longer than the shortest distance between neighboring leks throughout the Kertsonville Study Area, in 1977 (3.1 km). Intensive land use has dramatically reduced the nesting value of most intervening lands between preserved habitats in this study area, and leks were closely associated with preserved habitats (Svedarsky 1979). Thus, the distance between preserved habitats was a key factor in determining interlek distances, and 3.1 km probably represents a maximum. It is unknown whether a vast expanse of grassland, as was presumably present when prairie chickens evolved,
GREATER PRAIRIE CHICKENS IN MINNESOTA
227
would lead to a greater density of leks with about 3-km interlek distances, or a similar density to that at Kertsonville, with leks simply spread farther apart. A second prediction by Bradbury is that most females should visit a single lek in a given season. I recorded two females visiting at least two leks. Such interlek movements of females are infrequently reported, not only because few females are marked, but also because it is uncommon for adjacent leks to be observed at the same time, Interlek movements of females, however, may be a common and regular feature of greater prairie chicken biology, especially where leks are close together, where population density is great, and following nest failure. Regardless of interlek movements, females may develop an attachment to a particular lek, not only during a given season, but in subsequent years as well. Of 15 females trapped on the Pembina lek, at least seven revisited the lek; five were observed to copulate the same spring; and at least three revisited the lek the following spring.
6.5.3 Preincubation habitat use Habitats used by females in spring and early summer are particularly important because they must satisfy physiological demands associated with egg-laying and incubation. One female (C-76) laid at least 35 eggs in three nests. Assuming an average egg weight of 24 g (Johnsgard 1973), this amounted to 84% of her body weight and necessitated the intake of food resources high in protein and calcium. Examination of a dropping during the third nesting attempt indicated that she had fed upon June beetles (Phyllophaga sp.), as well as plant material. When available, insects may be an important protein source in the diet of egg-laying prairie chickens, as for some waterfowl during the same period (Swanson & Meyer 1973). Over 58% of preincubation locations and 94% of feeding locations during incubation were in disturbed habitats. In these areas, plant growth generally commenced earlier than in undisturbed habitats having a mulch layer. Drobney and Sparrowe (1977) felt that the accessibility to new growth in grazed prairies was a key factor contributing to their increased use in spring by prairie chickens in Missouri. Bendell (1974) noted that a number of disturbances (fire, mowing, plowing, removal of litter) may increase the nutrient content of plants, especially protein and phosphorus, and this also may have been a factor in females utilizing disturbed habitats. Cropland was used by more females than any other habitat type, because the birds apparently fed upon waste grain and shoots of newly planted grain. Korschgen (1962) reported that corn was the leading food of prairie chickens throughout the year in Missouri, and that the greatest seasonal use occurred during March and April. Korschgen also noted important summer use of Korean lespedeza (Lespedeza stipulaced) and soybean leaves. Alfalfa, another legume, was an important food item in this study, with 11.3% of preincubation
228
W. D. SVEDARSKY
locations in hayed areas having new leaf growth. Hayed alfalfa and grazed aspen stands had the highest preference ratings of all preincubation habitats. Roadsides, also preferred, apparently were used for loafing and to obtain grit. Undisturbed habitats were used for roosting, activities associated with nesting (including nest site selection), and loafing. Lowland areas, usually dominated by sedges, were selected as roost sites during the preincubation period and throughout the summer, even though cover was sometimes short following disturbance by grazing or fire. Consequently, low topography and wetness may be as important as heavy cover in determining roosting areas. Hamerstrom et al. (1957, p. 15) referred to habitats of reed canary grass (Phalaris arundinaced) and "coarse sedges" as providing excellent roosting cover. Ammann (1957, p. 61) noted that "marshes and bogs are often sought as roosting cover, particularly by prairie chickens, although these types may not serve any other purpose" and that "they show a preference for the lowland types if the water level is not so high as to prevent their finding dry spots." Presumably, these low areas represent habitats that are the most secure from predators. Foxes may avoid such areas because of the wetness and would have difficulty ambushing a roosting bird in the dense cover commonly found there.
6.5.4 Nesting Egg-laying in first nests usually began the last week of April with some visits to leks by females after laying had begun. Why these visits were made is unknown, but recopulations between nesting attempts have been documented (Robel et al. 1970a) and probably occurred at least twice in this study. Clutch size decreased in nests initiated later in the year, similar to reports by Baker (1953) and Robel (1970) for Kansas. Robel noted a mean clutch size of 13.8 for six nests begun between 15 and 30 April, whereas I recorded a mean of 14.6 for nine nests in the same period. The mean clutch size of 29 nests observed throughout the season in this study was 12.7. In Illinois, a mean of 12.3 for 12 nests was reported (Yeatter 1943); in Kansas, a mean of 11.6 for 19 nests (Robel 1970); in Wisconsin, a mean of 12.0 for 66 nests (Hamerstrom 1939); and in Michigan, a mean of 11.4 for 13 nests (Ammann 1957). Thus, clutches at the northern edge of the range in Minnesota appear larger than in other portions of the range. The incubation period of 25.2 days was longer than the 23-24-day period reported for Missouri by Schwartz (1945), for Wisconsin by Gross (1930) and for Kansas by Silvy (1968). However, Arthaud (1968) reported an incubation period of 25 days in southwestern Missouri, as did McEwen et al. (1969) for incubator-hatched eggs. The variation in reported incubation periods may be a result of temperature differences as reported for mallards (Anas platyrhynchos) (Hess 1972). Like Hess, I observed the incubation period to decrease progressively from early nests to late nests.
GREATER PRAIRIE CHICKENS IN MINNESOTA
229
The 91.9% egg fertility was somewhat lower than other reports. In Wisconsin, 98.0% was reported (Hamerstrom 1939); in Missouri, 100% (Schwartz 1945); in Kansas, 100% (Silvy 1968); and in Illinois, 93.0% (Yeatter 1943). The overall nesting success of 62.4% was higher than the 50% generally reported for other parts of the range. Westemeier and Vance (1975) noted that nesting success must average 50% for populations of prairie chickens to maintain their numbers in Illinois. Census results for the Kertsonville Study Area indicated a relatively stable population from 1974 through 1978, except for a 25 % increase in 1977 (Svedarsky 1979). Hence, mortality factors other than nest predation apparently have a greater influence in Minnesota than in Illinois. The date of nest initiation did not significantly affect nest predation, although successful nests were initiated an average of 3.4 days later than unsuccessful nests. Chesness et al. (1968) found overall predation of pheasant (Phasianus colchicus) nests in Minnesota to be greatest early in the season, but noted that the portion lost to mammals (the major nest predators in this study) steadily increased with the season. In Kansas, Robel (1970) and Baker (1953) found that early nests were more successful than later ones, as did Horkel et al. (1978) in Texas, who suggested that this could be a result of delayed development of a search image by predators. I found unsuccessful nests to have significantly greater brush indexes and litter depths than successful ones. Litter accumulations may indirectly attract nest predators. Foxes feed on meadow voles (Microtus pennsylvanicus), and Tester and Marshall (1962) found vole populations positively associated with increasing litter. Also, a considerable amount of skunk feeding sign was observed in habitats with litter accumulations. Tester and Marshall noted that Orthoptera (a major food item of skunks) was most abundant where light to moderate amounts of litter were present. However, in dummy nest studies in Kansas, Bo wen (1971) found skunks were the primary nest predators in spring-burned sites where no litter was present. Successful nests were located somewhat farther (70.0 m) from travelways (vehicle tracks, roads, pasture edges) than unsuccessful nests (51.6 m), and this may have affected predator detection. Kirsch (1969) found that foxes use vehicle trails in idle cover and that nesting success is diminished by this enhanced predator access. I found that females were persistent renesters after nest destruction or abandonment; at least one female (C-76) established three nests. This agrees with findings of Robel (1970). I did not observe a reduction in home-range size per nesting attempt, contrary to the 55% reduction per attempt in Kansas reported by Bowen (1971). Females always moved to new areas when renesting; one female (F-77) moved 5.6 km to renest after escaping from a fox, which temporarily caught her on the nest. Schiller (1973) reported that a sharp-tailed grouse (Tympanuchus phasianellus) hen renested 20.0 km from her first nest, which was apparently destroyed by a thirteen-lined ground squirrel (Citellus tridecemlineatus).
230
W. D. SVEDARSKY
Two females returned to nest within a few meters of their successful nests of the previous year. This demonstrated "nest-site tenacity" for prairie grouse, for which no published account could be found in the literature. A preference was shown for nesting in habitats that were undisturbed for 2 or more years. Kirsch (1974) emphasized the importance of residual cover for nesting throughout the range of prairie chickens. In this study, females selected nesting habitats with dense cover close to the ground (20-30 cm in height) and generally not tall (40 cm average height). Smooth brome and redtop habitats provided attractive nesting conditions, and habitats with tall ( > 1 m) vegetation appeared to be avoided. Westemeier (1972), after evaluating 242 prairie chicken nests in Illinois, concluded that redtop is attractive as nesting habitat after seed-harvesting has reduced the cover to 25-35 cm in height. This permits easy visibility to a standing prairie chicken yet provides ample concealment for nesting. He found no nests in undisturbed stands of native grasses that had developed a rank, impenetrable layer of residual cover. In Oklahoma, Jones (1963, p.772) found nine nests in "taller (x = 45 cm) and heavier cover than was usual for the tallgrass community," but tallgrasses would generally not attain as great a height in Oklahoma as in Illinois. Species composition of nesting habitats was not important in itself so long as proper density requirements were met, agreeing with Hamerstrom et al. (1957) and Kirsch (1974). Habitats undisturbed for over 4 years did not occur on the study area and hence long-term, idle grasslands were not evaluated as to prairie chicken nesting use.
6.5.5 Brood movements, habitat use, and disturbance Early reports indicated that broods probably stayed in the vicinity of the nest for the first few weeks (Schwartz 1945, Hamerstrom & Hamerstrom 1949). In this study, some broods less than 1 week old moved extensively. One brood (E-76) moved at least 3.8 km in the first 6 days after leaving the nest, and a 1-week-old brood (C-75) moved 1.4 km in 28 hours. Other studies of radio-tagged prairie chicken broods have also noted extensive early movements. Silvy (1968) observed a brood to move 3.2 km from the nest site in 6 days. Viers (1967) observed a brood movement of nearly 3.2 km in 7 days after leaving the nest. Large early movements have also been reported for other grouse species. Barrett (1970) recorded a movement of 396.2 m in 4.5 hours by a 1-day-old ruffed grouse brood. Schiller (1973) reported that a sharp-tailed grouse brood moved 960 m the first 2 days after hatching. Sonerud (1985) notes the following hypotheses to explain these extensive early movements in grouse: (1) dispersal, (2) a means of maximizing chicks' feeding rate on insects, and (3) an antipredator defense strategy. My observations tend to support the latter two hypotheses, which I believe are of equal importance in brood ecology of prairie chickens in Minnesota. Brood habitats must provide the proper combination of insect quantities, ease
GREATER PRAIRIE CHICKENS IN MINNESOTA
231
of brood mobility, physical protection from weather, and concealment from predators. Although differences were not significant, the smallest movements in this study were recorded in grazed and burned bluestem, and hayed alfalfa. Greatest movements were recorded in burned brome and undisturbed bluestem and sweet clover. These differing movements may have reflected differences in insect abundance, but no quantitative insect data were collected. Erikstad (1985) found that willow ptarmigan (Lagopus I. lagopus) broods moved less in insect-rich habitats and grew faster when home ranges were smaller. Southwood and Cross (1969) estimated that a partridge (Perdix perdix) chick would have to move over five times as far to obtain insect food in a herbicide-treated barley field as in downland (natural grassland), based on insect abundance. Subjectively, habitats that contained alfalfa and those recently burned appeared to have more insects than other types during this study. Burning has been shown to significantly increase numbers of Hemiptera and Homoptera on a Missouri prairie (Cancelado & Yonke 1970), Orthoptera and other herbivorous insects on a Mississippi right-of-way (Hurst 1970), and certain families of Coleoptera, Diptera, and Homoptera on a Minnesota prairie (Van Amburg et al. 1981). In all instances, broods moved directly from the nest site (usually in undisturbed cover) to areas disturbed by burning, grazing, or haying. Disturbed habitats accounted for 68.8% of early brood locations and 78.0% for the late brood period. Ten of the 15 habitat and land use types that received positive preference ratings had been disturbed (Fig. 6.15). This indicates that preferred brooding areas were not the same as preferred nesting areas. The term "nestbrood habitat" (Hamerstrom et al. 1957, Kirsch 1974, Drobney & Sparrowe 1977) suggests that a given habitat provides for both activities. The term has limitations in northwestern Minnesota, however, unless the land use condition is specified. For example, bluestem might be generally considered "nest-brood habitat," but specifically, undisturbed bluestem was nesting habitat, and grazed or burned bluestem was brood habitat. Disturbed habitats were also found to be preferred brooding areas in Missouri (Skinner 1977), Wisconsin (Toepfer 1973), and Oklahoma (Jones 1963). In Illinois, Yeatter (1963, p.755) reported that farmers mowing fields of clover and mixed hay "not infrequently" encountered young broods. In Missouri, Arthaud (1968, p. 85) reported that "men haying on the Taberville Prairie often mowed several hundred acres without seeing any prairie chickens (broods)." In that area broods were observed either on prairie edges or "within 50 yards of some type of rough cover such as fencerows, briar patches or rock outcrops," suggesting that expanses of undisturbed prairie vegetation were not used by broods. However, Arthaud did not indicate the land use condition before haying. Schiller (1973, p. 144) found disturbance to be a "vital factor" in opening up areas for sharp-tailed grouse broods in northwestern Minnesota. Kessler (1977) found fallow rice fields adjacent to nesting sanctuaries to be preferred brood-rearing cover for Attwater's prairie chicken (Tympanuchus cupido attwateri) in Texas. Kessler also recommended that efforts be directed at
232
W. D. SVEDARSKY
providing brood habitat within sanctuary boundaries, because intensifying land use on adjacent private land reduced its value to broods. During 1977, private, legume-grass hay fields and grazed prairies, adjacent to the Preserve and previously used by broods, were converted to row crops and subsequently received no use by broods. Interviews with neighboring farmers in 1975 indicated that they "always saw three or four broods" while mowing alfalfa hay. In 1976, no broods were observed by these individuals while mowing the same fields; however, about 285 ha of spring-burned bluestem were present on the Preserve along with 32.4 ha of regrowing alfalfa that had been hayed during 1975. Presumably, the disturbed habitats on the Preserve in 1976 attracted broods and reduced the use of adjacent alfalfa hayfields. Movements did not appear reduced in habitats where an accumulation of residual vegetation was present, thus the ease of mobility apparently was not a direct factor. Disturbed habitats (facilitating easier movement) were preferred, however, and extensive movements noted in undisturbed habitats may have represented searching for habitats with better overall brood values, including enhanced mobility. Very short cover apparently prompted increased movements. One female (E-76) moved her 1-2-day-old brood into a 15-20-cm small-grain field and then moved 800 m in no more than 7 hours. Willow and aspen habitats were used more by broods during midday, presumably for shade. Also, because brushy habitats were used more after being disturbed, the improved ease of movement may have been important as well. Pepper (1972, p. 31) found that sharp-tailed grouse broods preferred "heavier cover" at midday and "heavier-than-normal" cover during very hot weather. Sonerud believed the antipredator advantages of long movements to be their most important benefit to young grouse broods in Scandinavia. He reasoned that those broods immediately leaving the site of prey capture would be less likely preyed upon once more than would broods staying. In this study, taking direct locations of broods probably increased their subsequent movements. It was frequently necessary to approach within 100 m to determine the habitat type, particularly in transition areas. In such instances, the brood female was probably aware of my presence. This may have represented almost as much of a "predator encounter" as an actual flushing and was similar to Sonerud's "site of prey capture." Artmann (1970) noted that sharp-tailed grouse brood females made long movements after being captured during the brood period. He felt these long movements were definitely stimulated by the capturing. Long movements from the nest soon after hatching could also serve to reduce mammalian predation. Nests probably produce a considerable amount of scent from the hatched eggs and droppings.
6.5.6 Brood mortality Brood mortality of radio-tagged females in this study was very high; only two chicks in one radio-tagged brood were known to be alive at the end of the sum-
GREATER PRAIRIE CHICKENS IN MINNESOTA
233
mer. Radio-tagged broods hatched somewhat later than unmarked broods in 1975 and 1976 as a result of late-season trapping and renesting. Consequently, they were exposed to greater amounts of precipitation than earlier broods and probably experienced greater mortality. Also, the number of chicks in unmarked broods that hatched earlier in 1975 and 1976 tended to be greater than later broods (both radio-tagged and unmarked). If early-hatched broods generally contributed more to the population in Minnesota, as they have been reported to do in Texas (Lehmann 1941) and Kansas (Baker 1953), this could be partly explained by early clutches being larger (Table 6.1) and the seasonal precipitation pattern. In northwestern Minnesota, 32.2% of the 52.5 cm of annual precipitation occurs during the primary hatching months of June and July (Severson 1982). Early nests would hatch between 24 and 30 May, when the probability that a given day (at Crookston, Minnesota) will receive at least 1.27 cm of precipitation is 0.061 (Feyerhern et al. 1966). This increases to 0.072 (the highest for the year) in the period 28 June-11 July, when late nests would hatch. Therefore, in the long term, early nests in northwest Minnesota would tend to avoid more precipitation than later ones. The prairie chicken range in Minnesota is located at the eastern edge of the prairie biome, where precipitation is more than sufficient for the vegetation growth needed to provide preferred nesting cover height of about 25 cm. This is especially true on most of the preserved grassland habitats, which tend to be poorly drained, for two reasons. Their wetness often was a factor in landowners wanting to sell them, and it also made the tracts qualify for acquisition through state and federal wetland programs (Nielsen 1973). This moisture relationship contrasts, however, with dryer areas farther west where early precipitation is needed for the growth of adequate concealment cover for nests and broods. Much of the upland habitat there is grazed in late fall thus reducing residual cover the following spring and placing a greater importance on regrowth cover. Martinson (1963) felt that low spring and early summer precipitation reduced pheasant and sharp-tailed grouse productivity in western North Dakota. Precipitation can also affect broods if it reduces the time available for foraging and the activity levels of insects, particularly if accompanied by low temperatures. Jorgensen and Blix (1985) observed willow ptarmigan chicks to have low growth and high mortality when food availability and temperatures were reduced. Chicks were unable to compensate for increased energy expenditures and reduced food availability by increasing food intake. If prairie chickens in Minnesota respond similarly, cool temperatures, long movements, and low insects during the early brood period would result in chick mortality resulting from starvation. Precipitation effects are also related to habitat structure and topography (as it affects drainage). Brood habitats should provide physical protection from heavy rain yet contain openings for drying off and warming up. In Saskatchewan, Pepper (1972, p. 31) noted that sharp-tailed grouse broods used "lighter than usual cover" after heavy rains or on cool sunny days.
234
W. D. SVEDARSKY
Open expanses of relatively short cover may also increase the vulnerability of broods to predation. Both documented instances of chick predation (1 chick killed by a female harrier and a brood killed by a fox) occurred in large (> 32 ha) expanses of short (< 20 cm) cover. Schiller (1973) observed two instances of sharptailed grouse chick predation by raptors in northwestern Minnesota that occurred in abandoned hay fields, which presumably had shorter cover.
6.5.7 Female mortality Radio-tagged females experienced a summer (1 May-31 August) mortality rate during this study of 57.1 %, which is high compared with an average annual mortality rate of 56.0% calculated for female prairie chickens in Wisconsin (Hamerstrom & Hamerstrom 1973). Predation was responsible for most of the summer mortality observed in Minnesota. Vulnerability of females to predation probably increases in late spring and summer owing to: (1) the dissolution of winter flocks and the attendant reduction in predator detection potential; (2) nest establishment and incubation activities; (3) the need to actively feed while off the nest during incubation; (4) greater attraction of predators during brooding as a result of increased scent, movement, and sound; and (5) female distraction behavior during defense of brood. Evidence that female ground-nesting birds are susceptible to mammalian predation can be found; 79.5% of the mallards found at fox dens in North Dakota were females (Johnson & Sargeant 1977). Johnson and Sargeant (1977) concluded that fox predation was sufficient to cause an unbalanced sex ratio in favor of males in spite of hunting mortality being selective for males. The 1974 fall sex ratio for hunter-bagged, adult sharp-tailed grouse in northwestern Minnesota was 178 males per 100 females (Berg 1975). Although this ratio may be influenced by "lek hunting" (presumably biasing it toward males), it may also be partly caused by predation of nesting hens, and a similar ratio may exist for prairie chickens. Evidence indicated that seven (33.3%) radio-tagged females in this study were taken by mammals (mostly fox) and three (14.3 %) by avian predators. Schiller (1973) reported that at least nine of 26 radio-tagged, sharp-tailed grouse females in northwestern Minnesota were depredated during the reproductive season, with canid and avian predators being about equally responsible. In Wisconsin, Dumke and Pils (1973) noted that the nesting period of pheasants was a time of accelerated predation, primarily by red fox. They also found that 70.3% of the mammalian predation occurred during a 24-hour period in which precipitation occurred, which enhanced scenting conditions. Three of five birds killed by mammals, for which predation dates were known during this study, were also associated with rainfall. Females may be less alert for avian predators while actively feeding, particularly when away from the nest during incubation. One female (C-77) was killed
GREATER PRAIRIE CHICKENS IN MINNESOTA
235
while presumably feeding in an alfalfa field during the incubation period. Schiller (1973) reported that a sharp-tailed grouse female was killed by a raptor while feeding in a small-grain field. The remains of all three females taken by raptors in this study were found close to trees. Because trees were sparse on the study area, particularly that portion used by prairie chickens, this suggests that the association between trees and raptor kills was not a chance occurrence and that trees may have served as ambush sites. The overlap in female preincubation home ranges (Figs. 6.9, 6.10 and 6.11) and several observations of radio-tagged females in the company of another bird (presumably a female) during this period may indicate an adaptation for enhanced predator surveillance, as suggested by Wittenberger (1978). Disturbed habitats, preferred at this time, provided reduced concealment, and presumably a group of birds would be safer than if they were solitary. Dumke and Pils (1979) recorded a similar overlapping of early-spring home ranges of female pheasants. The association of broodless females with other birds in late fall would likewise provide an increased potential for predator detection as they began frequenting more-open habitats, such as harvested grain fields. Brooding prairie chicken females were not depredated more than broodless females, contrary to the findings of Maxson (1978) for ruffed grouse. Radio-tagged females may have been more susceptible to predation than unmarked females. No remains of unmarked females were found at five fox dens, three of which were partly excavated to recover transmitters. Radio-tagged females probably made up between 30 and 40% of the total female population in the immediate area based on lek observations and nest searching. It is doubtful whether the population associated with the Preserve could have maintained stable numbers during the study if mortality rates recorded for radio-tagged females and their broods were consistent throughout the population. I believe, however, that the overall effect of radio-tagging did not result in birds exhibiting atypical behavior concerning habitat selection, general movement patterns, and reproductive activities. That one bird (E-76) was carrying a radio package for 15 months indicated that at least some individuals adjust to its presence. Dumke and Pils (1973) used the percentage of radio-tagged female pheasants that hatched broods (50.0% compared with 52.8% for birds not carrying radio packages) to indicate the impact of radio-tagging on behavior. I used the same type of radio package and recorded 52.4% of the radio-tagged hens to hatch broods, but the proportion of unmarked hens hatching broods was unknown.
6.5.8 Broodless-female movements and habitat use Broodless females were observed to have somewhat greater home ranges than females with broods, but they showed similar movement indexes. No published prairie chicken data were located in which brood and broodless females have been
236
W. D. SVEDARSKY
compared, but Maxson (1978) reported that ruffed-grouse brood females used larger ranges and moved into new habitats more frequently then broodless females. The interseasonal movement pattern of female prairie chickens in Minnesota, of extensive spring mobility decreasing to smaller movements and home ranges later in the summer, is consistent with findings of Robel et al. (1970a) in Kansas. In late August and September, I observed movements and home ranges of broodless females to increase slightly as they began to associate with other birds in small flocks, feeding in harvested small-grain fields. Robel et al. (1970a) also noted that flocks of broodless females and males used grain fields in Kansas during summer and fall. As in the preincubation and incubation periods, most of the postbrood locations were in disturbed habitats; grazed bluestem was the most important type. The average home range of two females, followed from prelaying through the broodless period, was 133.8 ha, compared with a home range of 82.6 ha for a female that had a brood at the end of the summer. The average for three females was 116.7 ha and could be considered the minimum size of summer home ranges of females on the study area. As a general reference for management, this could also be used as the minimum size of nesting and brood-rearing sanctuaries, but would vary depending upon habitat quality, the nature of adjacent land uses, and the distance to other sanctuaries.
6.5.9 Population density and possible limiting factors The density of males on the Kertsonville Study Area was approximately 1 male/259.0 ha (640 ac) or 1 male/25.1 ha of preserved habitat. A 3.2-km2 area, which included the westerly portion of the Preserve and contained most of the locations of radio-tagged females, maintained a density of 6.3 males/259.0 ha for the study period. This was the highest density in the Kertsonville Study Area and was thereby judged as the "best" habitat area. Hamerstrom et al. (1957) reported the following male densities in the "best" prairie chicken areas of selected states: in Wisconsin, 28.0-29.5/259.0 ha; in Missouri, 34.0/259.0 ha; and in Kansas, 38.8/259.0 ha. Westemeier (1971a) reported a density of nearly 100 males in a 259.0 ha area in Illinois, one of the highest densities recorded throughout the prairie chicken range. Thus, the density in the best portion of the Minnesota Study Area is substantially lower than in the best portions of other states, suggesting that the present density is less than the potential carrying capacity. Nesting cover was not considered the critical limiting factor on the Preserve. The preferred cover (brome) contained an average density of 0.07 nest/ha compared with Westemeier's (1972) reported nest densities of up to 3.1 nests/ha in certain cover types in Illinois. Further, nest success was generally higher than reported for other areas. Therefore, reduced reproduction is thought to be caused by the following factors operating in combination: (1) inadequate brood-rearing habitat, (2) predation, and (3) precipitation and associated temperature conditions
GREATER PRAIRIE CHICKENS IN MINNESOTA
237
during the early brood period. Brood habitat must be within reasonable proximity to nesting cover so as to reduce energy-consuming, initial movements. They must provide insect food, physical protection from the sun and storms, concealment from predators yet be open enough to facilitate movement at ground level, and loafing and dusting openings. Predators can increase mortality by killing chicks and brood females, by scattering broods, and by stimulating movements at unfavorable times (cool and rainy) through detrimental habitats (thick litter, standing water). The likelihood is greatest that a mammalian predator will disturb a brood when scenting conditions are best—during the night or during cool, wet periods when brood vulnerability to chilling is also greatest. It is interesting to consider the relationship of the number of prairie chickens (total cocks) in Polk County, Minnesota, to December fox-fur prices (presumably inversely related, to some degree, to fox numbers) (Fig. 6.18), and this should merit attention in the future. Acquired lands supporting prairie chickens in Minnesota are fast becoming "habitat islands" in an intensively farmed landscape, with the predominant land use being row-crop agriculture (small grains, sunflowers, corn). The surrounding private lands presently serve as display sites and attractive feeding areas for females and males (Jorgenson 1977) during spring, early summer, and fall. However, they are of no value for nesting and roosting, and, except for alfalfa hay fields, are used little for brood-rearing. Furthermore, Westemeier and Vance
Fig. 6.18. Relationship between December fox-fur prices (presumably inversely related to fox numbers) and the counts of male prairie chickens on leks in Polk County two springs later.
238
W. D. SVEDARSKY
(1975) suggest that high prey populations supported on such habitat islands may attract significant numbers of predators. They attributed the recent decline of prairie chickens in Illinois to accelerated nest predation. This study demonstrated that a variety of habitats and land uses are used by female prairie chickens during spring and summer for various reproductive activities. To accommodate prairie chicken requirements, Hamerstrom et al. (1957, p. 59) recommended that a "scatter-pattern" of small (16.2 ha) tracts of "nestbrood" cover be acquired throughout private land in the Wisconsin prairiechicken range. This concept of "ecological patterning" has been the central theme of the management effort there and in Illinois. Land ownership within the prairie chicken range in Minnesota is such that few small tracts (< 32.4 ha) have become available for acquisition. Consequently, the average size of tracts owned by the U. S. Fish and Wildlife Service, the Minnesota Department of Natural Resources, and The Nature Conservancy within the range is approximately 132.3 ha, with several tracts considerably larger. Because current land use on private land fulfills a limited amount of prairie-chicken habitat requirements and because the availability of small tracts on which nesting and brooding cover could be developed is also limited, I suggest that the "scatter-pattern" concept be applied to these larger tracts. For example, a 64.8-ha (160 ac) tract separated into four management units (16.2 ha) and disturbed on a 4-year rotation should provide nesting and brooding habitats within reasonable proximity. A compromise must be made, however, between the need to maintain large, undisturbed, nestingcover plots and recently disturbed, "patchy" cover for broods.
6.6 Summary This study examined the reproductive ecology of female greater prairie chickens at the northern edge of their range. Home ranges were greatest in late spring when females were visiting leks and establishing nests. They were smallest during nesting, increased in the early brood period, and then decreased in the late brood and postbrood periods. Breeding commenced about 1 week later than in more southerly parts of the range. There was evidence of lek fidelity in consecutive years and within a year by renesting females. Nest-site fidelity was also shown by successful females radiotagged for 2 consecutive years. A strong tendency for renesting was observed. Undisturbed habitats were preferred for nesting; species composition was not important so long as cover was dense at ground level, and about 25 cm high. Undisturbed cover in lowlands was preferred for roosting; otherwise, prairie chickens preferred habitats disturbed by burning, grazing, haying, or cropping. This was especially true for broods, which in all cases moved directly to disturbed areas from nests that were generally in undisturbed cover. Clutch size and nesting success observed here were generally higher than in
GREATER PRAIRIE CHICKENS IN MINNESOTA
239
other studies, but, in spite of this, population density was lower. I attribute this largely to brood mortality factors. Mortality of radio-tagged broods was considered greater than that of unmarked broods, but both marked and unmarked broods tended to have fewer chicks than had been reported in other studies. The poor drainage of the study area and the early summer peaks in precipitation were believed to contribute to poor brood survival. Exposure to precipitation and cool temperatures were affected by hatching dates, characteristics of brood habitats, and extent of movements as influenced by disturbance and proximity of nest sites to brood cover. Climatic factors may affect chick survival directly or indirectly by affecting insect populations. These relationships and the role of brood-habitat composition and structure in moderating climatic factors are in need of further study. Predation was a significant mortality factor for females and, perhaps, for broods as well. The extent to which intensive land use on adjacent lands concentrated predation pressure on the sanctuary was unknown, but has been reported to be significant in other parts of the range. Because nest predation was indirectly related to litter buildup and brush growth, the potential exists to reduce this effect through management by burning, haying, and in some cases, by grazing. Other than for feeding and courtship, private lands surrounding the sanctuary provided a minimum of the varying habitat needs of different stages of the prairiechicken reproductive cycle. This emphasizes the importance of a rotation management system to provide a mosaic of habitat conditions within sanctuary boundaries.
7
Mate Choice by Female Sage Grouse /. E. Hartzler and D. A. Jenni
7.1 Introduction The communal displays of male sage grouse (Centrocercus urophasianus) are by themselves spectacular enough to attract and intrigue biologists. Moreover, it is especially interesting that all females attending a particular lek mate with only a few of the many males available to them. This strong sexual selection shown by hens attracts the attention of sociobiologists and evolutionary biologists. Many efforts have attempted to understand the evolution of leks (Wittenberger 1978, Bradbury 1981), or to discover the basis of mate selection (Wiley 1973b, Gibson & Bradbury 1985). This paper addresses some proximal questions; we are especially interested in discovering what happens within the lek, the mechanisms of mate selection by hens, and the clues females use in mate selection. Each male defends a small territory on the arena (hereafter used to refer to the place where the lek forms) where he displays. Males near the center of the lek (hereafter used to refer to the aggregation of birds on the arena) are more likely to copulate than those immediately peripheral to them. Yearling males establish territories around the edges of the lek and rarely breed. This spatial segregation of males by age and the restriction of copulation to a few central males led Scott (1942) to a set of descriptive terms (master, guard, and peripheral cocks) that implied a hierarchical rather than spatial relationship among males. In fact, Scott described the social organization of males in the lek as a dominance hierarchy with the most dominant males becoming master cocks; less-dominant cocks, guards; and least-dominant cocks (yearlings), peripheral cocks. Unfortunately This chapter was written by D. A. Jenni and is based on J. E. Hartzler's (1972) dissertation and additional data provided by him. 240
MATE CHOICE BY FEMALE SAGE GROUSE
241
the criteria he used for classifying cocks was their location on the arena rather than their behavior. Because many subsequent workers followed Scott's convenient nomenclature, cause and effect became interminably confused. Although sage grouse are important game birds, they attracted relatively little attention until Patterson's (1952) monograph. Additional data on sage grouse leks began appearing about a decade later (Dalke et al. 1960, 1963, Eng 1963). Researchers then began comparing the behavior of different grouse species (Lumsden 1968, Hjorth 1970), and, more recently, a series of papers by Wiley (1973a,b, 1974, 1978) has focused attention on male sage grouse behavior in an attempt to understand how the lek functions. The variety of mating systems in grouse has attracted the attention of sociobiologists (e.g., Wittenberger 1978). Most recently, following the recognition of lek behavior in other taxa, there has been increased interest in trying to unravel the evolutionary history of lek behavior in a broader sense (Bradbury 1981). We initiated this study before we were aware of Wiley's research, and fortunately we approached the problem differently. Because of this, our results are complimentary. At the outset, it seemed inconsistent that male sage grouse could express a dominance hierarchy on the arena at the same time that they occupied mutually exclusive territories. One objective was to resolve this inconsistency. We also wanted to learn whether territories were established in exactly the same locations each year, whether each male returned to the same territory in subsequent years, whether the same males were breeding year after year, and whether breeding success increased with age. Finally, we wanted to learn how or by what criteria females selected their copulation partners. In the lek setting female choice is blatantly conspicuous, with each male tied to a territory and females free to wander through the lek and select a male from which each solicits a single copulation.
7.2 Study area and methods Observations were made at the Ford's Creek lek, 11 km north of Grass Range, Fergus County, central Montana (47°08'15"N, 108°53'05"W; R22E, T16N, section 24, from 19 March to 28 May, 1969; 25 February to 28 May 1970; and 7 March to 20 May 1971. Altogether, 172 mornings and 95 evenings were spent observing lek behavior of sage grouse. The same lek had been studied earlier (1965, 1966, and 1968) by, respectively, Lumsden (1968), Hjorth (1970), and Wiley (1973a). The area in which cocks consistently gathered formed a rough oval approximately 100 m wide and 250 m long. Within this large area was a smaller one, about 50 m wide and 90 m long, where hens consistently gathered and in which most of the older cocks were located. This smaller area was watched intensively. Grassy areas alternated with gravelly washouts that were 0.1 m to 0.8 m lower
242
J. E. HARTZLER AND D. A. JENNI
in elevation. Sagebrush (Artemisia tridentata) bushes up to 0.6 m high were unevenly scattered on the arena. In 1969 and 1970 males were captured by night-lighting and were permanently marked with metal bands to which colored, plastic leg boots were attached. Cocks near our blinds were individually marked by clipping the tips of their rectrices by shooting through them with a .22 caliber rifle. Because cocks usually fanned their tails at the lek, clipped rectrices permitted fast, accurate, individual identification of all males on the central part of the arena. Cocks showed little disturbance during the shooting, rarely interrupting their ongoing behavior. Capture by night-lighting also had minimal long-lasting effects. Six males that had been previously tail-clipped were captured at night, and all reappeared at their accustomed sites the next morning. Late in 1969 and 1970, five males that had been marked by tail-clipping were captured with loop traps, similar to the bal-chatri traps used by falconers. These were also permanently banded. Although this trapping occurred late in the season, when males were abandoning the lek, all but one of the five returned to his accustomed site the following morning. In 1969, display rates were collected by recording the struts of three to five adjacent males while hens were present. Fighting data were collected by locating the position of each fight on a daily map for the same central males and noting whether or not wing-beating occurred in each fight. In 1970 and 1971, we watched adjacent males in groups of three to five birds and tape-recorded the following data for each cock: location and duration of each fight, number of wingbeats for each fight, number of strut displays, and distance to the nearest hen and changes in distance. These data were obtained for 10-minute blocks on all central cocks beginning as soon as we could identify individual cocks in the morning, usually 45 minutes before sunrise, until lek activity slowed in late morning. We attempted to gather 10 minutes of data on each central each morning. Data were transcribed daily onto standard forms. An 8-m grid of wooden stakes 60 cm high was placed on the central portion of the arena. Daily data maps were prepared that showed this grid pattern and obvious landscape features. In 1969, movements of central cocks were drawn on the maps. In 1970 and 1971, locations of individual cocks were recorded at 15minute intervals every morning. Each year, the location and identity of copulating cocks were recorded for all observed copulations. An attempted copulation was considered successful if the cock mounted the hen and lowered his tail over her cloaca, and if the hen performed a vigorous postcopulatory display. A certain amount of disturbance is impossible to avoid in a field study, but we took strong measures to minimize it. The 1.5-m high, sage green tents we used as blinds were erected before most lek activity began in early March. The observer entered the blind in the evening, stayed there overnight, and did not leave until after the grouse had departed in the morning.
MATE CHOICE BY FEMALE SAGE GROUSE
243
7.3 Territorial establishment and mating success 7.3.1. Fidelity of males to arena and territories Adult male sage grouse began going to the arena in early March of each of the 3 years. Their numbers increased through the peak of hen attendance and began to decline in late April to early May. Yearling cocks began to attend the lek at about the time hen numbers peaked in early to mid April. Yearling cock numbers peaked after hen numbers began to decline (Jenni & Hartzler 1978). At the beginning of the breeding season each cock restricted his movements to a limited part of the arena and returned to the same site on subsequent mornings. These areas of use became defended territories when additional cocks began attending the lek and were located on areas adjacent to those occupied by earlier cocks. Males displayed toward and fought with neighboring cocks as soon as they became available. In addition to interacting with neighbors, territory holders charged any other male intruder. Although they usually retreated, intruders sometimes persisted. The rare occurrences of territory takeovers are described later in this paper. Once they began attending the arena, most cocks showed strong territory fidelity, at least through the peak of hen attendance. Toward the end of the lek season, when the number of visiting hens fell to few or none, some males became less faithful in returning to the lek (Jenni & Hartzler 1978). Some that continued to attend wandered about the lek establishing temporary territories at unoccupied sites. Nearly all cocks responded to nearby hen clusters by displaying on the side of their territory closest to the hens. This often resulted in slight shifts in territory boundaries toward the hens but not in the relocation of entire territories. Significant shifts in territory boundaries occurred when a territory owner that had been either a successful breeder or had been located close to a successful breeder died or for some other reason failed to return to the lek. Sometimes, however, unoccupied sites near hen clusters were taken over by males from other parts of the lek and no boundary shifts occurred. Cocks banded as yearlings were just as likely to return to the arena the following year as were cocks originally banded as 2 years old or older. The return rate for both groups was 29%, for both years combined (Table 7.1). There was, however, a significant difference in the return rate between years. Of the 20 cocks banded in 1969, 50% returned in 1970. Of 31 banded cocks present on the lek in 1970 (ten returnees from 1969 plus 21 more banded in 1971), only 16% returned in 1971 (P < 0.05, X2 = 4.72, df = 1). One cock banded as an adult and therefore at least 2 years old in 1969 returned to the lek in both 1970 and 1971. Another adult cock that returned in 1970 was found freshly dead on the lek at the beginning of the 1971 season. There was a strong year-to-year territory fidelity. Most permanently banded
244
J. E. HARTZLER AND D. A. JENNI
Table 7.1. Numbers of male sage grouse permanently banded in 1969 and 1970 at the Ford's Creek lek in central Montana and numbers returning to the lek in 1970 and 1971 No. banded Age of grouse Yearlings 2-year-olds 2 + years old 3 + years old Total
1969
12 8
20
1970
5 16
21
No. available to return
Total
1970
1971
Total
Actual returns 1970
1971
Total
No. %
No. %
No. %
5(42)
0(0)
5(29)
26a
17 34a
31a
51a
17 24
41
12 -
20
5
-
-
-
5(63)
5(19)
10 (29)
10(50)
5(16)
15(29)
Sixteen banded in 1970 plus ten banded in 1969 that returned in 1970 (five as 2-year-olds and five as 3-year-old or older cocks).
cocks 2 or more years old that returned to the arena in subsequent years returned to the same territories they had occupied the year before. However, there was significant recruitment of adult cocks into the central area of the lek each spring. In 1969, approximately 20 adult cocks attended the lek, and in 1970 and 1971, 48 and 50 attended. The new cocks did not have territory sites to return to but were influenced in selection of sites by the distribution of returning cocks that settled on their former territories. In 1970, a majority of the adult cocks established territories around the tip of a low ridge about 35 m west of where cocks were concentrated in 1969. In 1971, most older cocks established territories about midway between the 1969 and 1970 centers. In both 1970 and 1971 permanently marked cocks that had been highly successful during the preceding breeding seasons failed to return to the lek and their former territories were available to other males. In 1970, cock G, 1969's most successful cock, failed to return to the lek and was presumed dead. His four nearest, surviving neighbors of 1969 all returned to approximately their same territories (Fig. 7.1). Although three shifted their boundaries slightly to include parts of the unoccupied territory, none exerted any intensive effort to occupy G's old territory and most of it remained unoccupied in 1970. Cock A was again highly successful, but H was not successful until he abandoned his old territory and established a new one farther west in the lek. The other two did not copulate in 1970. Thus what had been an important mating center in 1969 did not function as an important center in 1970 when the most successful of the 1969 cocks failed to return. In 1971 none of the four permanently marked, successful cocks of 1970
MATE CHOICE BY FEMALE SAGE GROUSE
245
Fig. 7.1. Map showing locations of territories on and near the 1969 mating center. When cock G failed to return in 1970, his returning neighbors made small adjustments in their boundaries, but most of G's territory remained unoccupied.
returned, but four of the 23 unsuccessful ones did (Fig. 7.2). One of the four (2R) established a 1971 territory approximately 24 m from his 1970 territory at a location where he was about as far from any mating center as he had been in 1970. Two of the four (P and 13) established territories that included much of the area they had occupied in 1970, but made boundary shifts to include parts of the 1970 territories held by successful neighboring cocks that failed to return (Fig. 7.2). One cock (13L) made a minor boundary shift to include part of a 1971 center, but otherwise occupied almost exactly the same site in 1971 and again failed to breed. Thus, of eight returning cocks permanently marked and at least 2 years old the previous season, none simply occupied a previous year's mating center even though the former holders of those sites failed to return. One of them made a large move but failed to obtain copulations. The other seven made relatively small shifts in territory boundaries; three of them were unsuccessful both years, and of the four that were successful the previous year, two were successful again their second year. Again, there is no evidence that returning cocks regularly attempted to occupy previously successful territories even when they were no longer occupied.
246
J. E. HARTZLER AND D. A. JENNI
Fig. 7.2. When the most successful cocks of 1970 failed to return to the lek in 1971, adult neighbors made small adjustments in their territory boundaries, but none simply took over the unoccupied territories.
During the three years, 14 established cocks "vacated" their territories; 11 of these 14 vacancies provided opportunities for peripheral neighbors to move closer to active mating centers. Three of the 11 territories were taken over by neighbors that had been encroaching on the territories previously. Two cocks shifted their boundaries to include parts of the adjacent vacancies and spent most of their time in the newly acquired areas. Four cocks shifted boundaries to include vacated areas, but made little use of the new areas. In one case a vacancy remained empty for 2 weeks before a new cock occupied it. In the final instance a new cock occupied the territory the day it was vacated, and may have driven the former owner out.
7.3.2. Fidelity and return of hens Hens first began to attend the lek in mid-March each year. One to a few hens attended each morning and wandered over the arena seemingly oblivious to displaying males. No day-to-day pattern to their wanderings was apparent. These hens
MATE CHOICE BY FEMALE SAGE GROUSE
247
moved through all the central territories. In early April the number of hens coming to the lek increased significantly (Jenni & Hartzler 1978). They gathered in clusters and remained together throughout their stay at the lek; they also left in groups. Hens arriving in small groups joined hens already present. Hens arriving on foot moved, as they fed, across the arena to join already existing clusters. Almost all copulations occurred in these compact clusters. Once the copulation period began hens tended to cluster on one or two territories each morning, but not always on the same territories, especially early in the copulation period (Table 7.2). This lack of territory fidelity by hens was most conspicuous in the 1971 season when they first clustered in the territory of cock 13; then 13 and 5; then 13 only; then 6 and 5; then 6 and 3; then 6 and 5R; then 5R, 3 and 5; then only 6; then 6, 5R, and 90; then 6 and 5R; and finally only 5R. Table 7.2. Dates on which hens clustered on individual cock's territories. (The maximum number of hens counted in the cluster is followed by the number of copulations performed by that cock during that morning.)
1969 Date
1970
1971
Ga
J
A
Su
7
B
_ 15-? 18-? 23-3 30-1 20-7 22-9 27-8 20-1 15-5 -
_ _
_ _
_ _
_ _
_ _
Mar 78 ivial Z.O
29 30 Apr 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 20 21 22 23 24 May 2 a
13
6
5R
90
3
_ -
-
5
18 0
-
_
_ _ _ -
6-2 _
17-2 _ _ _ 18-0 30-2 49-29 20-8 15-6 32-5 _ 18-3 32-21 7-1 18-11 _ _ _ _ _ _ _ _ _ _ _ _ 45-8 - 8-1 50-10 55-14 24-17 35-33 _ _ _ _ _ 10-4 _ 14-12 17-6 14-2 12-2 _ 6-1 -
Letters identify specific cocks.
10-0 23-0 _ -
_ _ -
_ 14-2 52-12 33-13 _ 15-5 25-12 30-8 20-5 30-8 _ 32-25 15-9 6-0 14-6 9-2 8-5 9-5 14-5 11-6 _ _ _ _ 6-2 -
10-0 10-0
_ -
5-3
_ _
6-0 8-1 15-2 7-1 10-5 40-0 _ _ _ -
-
5-4 5-2 _ _ — -
248
J. E. HARTZLER AND D. A. JENNI
7.3.3 Nonrandom mating of cocks In all 3 years a small proportion of the older cocks performed most of the copulations. In 1969, cocks G, F, H, and J —20% of the 2-year-old and older cocksperformed 80% of all copulations. In 1970 one of the approximately 48 adult cocks copulated 169 times, performing 68% of all copulations that year. In 1971, 6 cocks, 12% of the total, performed 85% of the copulations (Fig. 7.3). Each year one or two cocks were clearly more successful than the others. In 1969 cock G performed 41 % of all copulations, in 1970 cock Su performed 68%, and in 1971 cocks 6 and 5R performed 33% and 31% of the copulations, respectively. Of 515 copulations observed, 378 (73%) occurred in hen clusters. Nonrandom mating was a direct consequence of hens clustering, during the peak of the copulation period, around one or a few of the many cocks available to them. Each year a few isolated copulations occurred outside the hen clusters. Although these isolated copulations were performed by central cocks, they were not necessarily neighbors of the most successful males. In 1969, however, three of the five neighbors of the most successful cock also bred frequently. In the other 2 years neighbors of the most successful cocks did not appear to gain any special advantage by being adjacent to a mating center. In 1970, the two most successful cocks had nine neighbors, only one of which copulated frequently. In 1971, again only one of the nine neighbors of the two most successful cocks copulated frequently. Thus, in 1 of 3 years neighbors of a highly successful cock were more likely to be successful than other cocks; in 2 of 3 years the proportion of neighbors that became successful (0.11) was not different from the probability (0.13) that any adult cock on the lek would be successful. The situation in 1969 appeared to be unusual. The most successful cock, G, became obviously tired during the peak of hen attendance and failed to copulate with all the hens that indicated readiness. Shortly after the peak of hen attendance, but while many hens still attended, he suddenly quit displaying and fighting and simply stood on his territory with his wings and tail folded. The hens drifted off to his neighbors. After 2 days he resumed displaying, hens again came to him, and he copulated. Thus only in 1969 did the greater success of the neighbors of the most successful cock appear to be influenced by his inability to maintain the heavy schedule of displaying, fighting, and copulating.
7.3.4 Fidelity of cocks compared to mating success During the 3 years, six of the 14 cocks that vacated their territories did so after fighting with other cocks. Four incidents occurred well after the peak of breeding for that year, and one occurred before many hens attended. In the remaining instance cock 9L drove a successful cock from his territory during the peak of hen attendance in 1971. When hens began attending in large numbers, cock 9L abandoned his territory next to two cocks that did not become successful breeders that
MATE CHOICE BY FEMALE SAGE GROUSE
6 5R 5 3L 90
249
3 12 OTHERS
INDIVIDUAL COCKS Fig. 7.3. Copulations were not evenly distributed among the cocks in any one year. The numbers of times each cock copulated is indicated on top of each bar.
250
J. E. HARTZLER AND D. A. JENNI
year. Three days later 9L appeared adjacent to the territory of cock 3, who had copulated nine times in the preceding 6 days and who had approximately 40 hens clustered on his territory that morning. After a long, violent fight between 9L and 3, cock 3 sat quietly. He never returned to the lek, and cock 9L occupied 3's territory for the rest of the season. The day after the fight a few hens visited the territory briefly, but they never clustered there subsequent to the fight and 9L did not copulate. Thus for 9L, at least, forcefully taking over a mating center did not result in breeding success. Successful cocks were no more likely to return in subsequent years than were other 2-year-old or older cocks that were not successful. Five of 16 (31 %) permanently marked cocks that obtained one or more copulations returned the following year, and five of 15 (33%) permanently marked, adult nonbreeders returned the following year (Table 7.3). Of the five successful cocks that returned, two were again successful. Of the five unsuccessful cocks that returned, two were successful the next year. Thus, successful cocks that returned a second year were no more likely to breed successfully than were previously unsuccessful returnees; 2 of 5 returnees in each category bred the second year. Yearlings were as likely as adult males to return the next year. Four of 11 cocks banded as yearlings returned as 2-year-olds, compared with nine of 31 older cocks. Although a higher proportion of returning 2-year-olds (0.75) obTable 7.3. Ages and breeding success of permanently banded cocks that attended the Ford's Creek lek for at least 2 years between 1969 and 1971
1971
1970
1969 Individual cock
Age
No. cops.
Age
No. cops.
Age
No. cops
Banded as adults P A H J F 1, 3L IB 2R 5R
2+ 2+ 2+ 2+ 2+ -
>1 6 11 9 12 -
3+ 3+ 3+ 3+ 3+ 2+ 2+ 2+ 2+
0 18 7 0 0 0 0 0 0
4+ 4+ 3+ 3+ 3+ 3+
0 0 0 2 63
Banded as yearlings 1 2 b D PO
1 1 1 1 1
2 0 9 1 0
—
—
0 0 0 0 0
2 2 2 2 2
MATE CHOICE BY FEMALE SAGE GROUSE
251
tained copulations than older cocks (0.44), the difference was not statistically significant.
7.4 Hen selection of mating centers or cocks The two most conspicuous differences between cocks that mate frequently and others are: (1) the presence of hen clusters in the successful territory and (2) the copulatory behavior of the successful males. It is not obvious why hens cluster on so few of the territories and copulate with so few males. Hens arriving late in the season could simply be joining the existing hen clusters. However, this "success breeds success," or reinforcement, hypothesis does not answer the critical question of what initially attracts the hens to particular males.
7.4.1 Hen clusters and mating centers In preference to mating with neighboring males who were immediately available, hens in clusters often waited for opportunities to crouch in front of the preferred male, who was sometimes too exhausted to mount. Even though neighboring cocks that had not copulated that morning were available only a few meters distant, hens persisted in copulating with males that had been and were likely to continue to be the most successful. In 1970, cock Su copulated with as many as 33 hens in a single morning, and in 1971, cock 5R copulated with 25 in one morning. Although all cocks appeared to spend most of their time on the arena either displaying or fighting, there were obvious individual differences in behavior. The most important difference was that only one or two copulated frequently, some copulated occasionally, and most never copulated. In 1969, ten of 20 cocks 2 years old or older never copulated. The amount of time spent copulating was relatively small; cock Su performed 169 of the copulations in less than 0.2% of his time. All the observed copulations occurred in the central part of the lek in an area approximately 90 m long and 25 m wide. However, the mating centers, where copulations were concentrated, did not always occur at exactly the same location within this central area. The primary mating center in 1969 was not a mating center in 1970 (Fig. 7.1). Although we know some copulations occurred in the western part of the central area in 1969, data on copulation locations for the entire central area are available for only 1970 and 1971 (Fig. 7.4). One of four 1970 mating centers (20 copulations) became a mating center again in 1971 (12 copulations). The other three 1970 mating centers, though reoccupied by adult cocks in 1971, failed to develop into mating centers again. One of the 1971 mating centers overlapped partly with the major mating center of 1969. The three other 1971 mating centers were at sites that did not function as centers in 1969 or 1970, even though they had been occupied by adult cocks in those years. If hens were mating with cocks on the basis of where the cocks were, mating centers should
252
J. E. HARTZLER AND D. A. JENNI
7.4. Locations of copulations on the central part of the lek, 1969-71. The area mapped in 1970 and 1971 is 60 x 30 m. Locations of copulations were recorded only for the eastern part of the central area in 1969.
MATE CHOICE BY FEMALE SAGE GROUSE
253
have occurred on the same sites in subsequent years. If they were using some other basis for selecting mating centers, location should be of no more than secondary importance, and mating centers should have occasionally recurred on the same sites, simply as a consequence of random placement on the relatively small, central area of the lek.
7.4.2. Strutting frequency compared to mating success Cocks that copulated more often also appeared to display and/or fight more often. These cocks had hens near them more often, and the differences in their activity levels could have been a consequence of hen proximity rather than a reflection of intrinsic differences between cocks. We investigated this possibility. Because temporal properties of the male strut display are remarkably stereotyped (Wiley 1973b), it is unlikely that females could select males on the basis of temporal variation. However, it is possible that females could respond to qualitative characteristics of displays that we have not been able to detect. It is also possible that hens select males on the basis of some physical characteristics. Cocks do differ in rates at which they strut. Number of struts per uninterrupted, 2-minute time intervals were measured repeatedly for central cocks in 1969. To minimize seasonal or situational influence, the rates of breeders and nonbreeders were measured simultaneously. All 2-minute intervals interrupted by fighting, copulating, or freezing in response to predators were excluded from analysis. Successful cocks strutted more often than unsuccessful cocks during 288 of 341, 2-minute intervals in 1969 (P < 0.001. X 2 = 225, deleting 23 tied scores). Although as a group successful cocks strutted at faster rates than did unsuccessful ones, strutting rates did not correlate with number of copulations (rs — 0.4, n — 4, Spearman Rank correlation). Early in the spring of 1970 and 1971, before most hens began attending the lek, there were no differences in the number of struts per 10-minute periods by cocks that later copulated five or more times and those that copulated zero to four times (P > 0.10 both years, Mann-Whitney U tests, Table 7.4). The number of struts performed during 10-minute blocks when many hens attended, however, was significantly greater in both 1970 and 1971 for cocks that copulated five or more times than for cocks that copulated zero to four times (P < 0.05 both years, Mann-Whitney U tests, Table 7.4). It was also clear from these data that cocks strutted much more frequently when many hens were attending the lek than they did earlier in the season; 25 of 26 cocks increased and only one decreased his rate of strutting. The increase in strutting rate when hens began attending in large numbers was 75% in 1970 and 177% in 1971 by those cocks for which data for both periods were available. Cocks that copulated more often than others typically had hen clusters in their territories more frequently than others. To determine whether the higher strut
254
J. E. HARTZLER AND D. A. JENNI
Table 7.4. Mean strutting rate (struts per 10-minute interval) of cocks prior to the appearance of large numbers of hens, and while many hens attended, 1970-713 1970, before
Cock 9 7* 3 11* 4 12 6 1 P 13
Strutting rateb
33.3 28.2 26.0 25.6 24.0 24.0 20.0 14.0 13.6 13.6
1971, before
Cock
Strutting rate
3L* 2R 2 13 6* 9R 8 5* 1,3L 8 3* 39 56 P 9L 5R* 79 7 8L 1/2
25.4 24.2 24.0 22.8 22.8 21.5 19.9 19.0 18.9 18.2 17.7 16.9 15.0 14.9 14.2 13.8 13.1 12.9 12.1 12.3
1970, present
Cock
Strutting rate
Su* 9L* +9R H* +P 7* + 3R A* + 13 12 13* +9 11* + 8L 1 +D +6 4
55.4 55.3 54.7 49.3 47.7 47.6 46.0 43.8 43.4 43.2 41.5 40.7 38.6 37.5 34.5 29.9 27.9 21.3
1971, present
Cock
Strutting rate
+ 13 +95 3+ 6* 3L* 5R* 2R +8 90* 5* + 79 + 8R +9L + 8L +P 1 +9R 1,3L 1/2
62.2 60.6 60.3 58.2 57.0 56.0 55.5 54.1 51.7 51.7 51.2 50.9 48.2 47.8 46.2 42.3 41.3 40.8 38.3
a
Prior to 99 = 19-27 March 1970 and 7-28 March 1971; while many 99 present = 29 March-6 May 1970 and 3-18 April 1971. b The rank of an individual cock corresponds with the order of his strutting rate; e.g., in 1970, prior to many hens visiting, cock 11 averaged 25.6 struts/10-minute interval and was ranked fourth of ten males. * Cocks that mated five or more times.
rate of these successful cocks was in response to hen clusters or was independent of them, we measured strut rates in 1970 and 1971 for individual cocks when hens were near (< 3 m) and when hens were at medium distance (4-18m) (Table 7.5, 7.6). Both classes of males (those that copulated five or more times and those that copulated fewer than five times) strutted much faster when hens were within 3 m than when they were 4 to 18m distant. The more successful males strutted at faster rates when hens were near than did the other males, but the differences were less obvious when hens were more distant. Thus, the difference in strut rate between the two classes of cocks was not simply a consequence of having hens near them more often; rather, when hens were near, the most successful cocks as a group tended to increase their strutting rate more than other cocks.
MATE CHOICE BY FEMALE SAGE GROUSE
255
Table 7.5. Mean strutting frequencies (average number of struts per minute) of cocks that mated frequently (copulated at least five times) and mated infrequently (less than five copulations) when the nearest hen was within 3 m and when the nearest hen was 4-18 m away, 1970
9-20 Apr
2-8 Apr
Individual cock
21 Apr-22 May
< 3m
4-18 m
< 3m
4-18 m
< 3m
4-18 m
5.2 5.9 6.0 6.1 5.5 6.6 6.2 -
3.8 3.3 4.8 4.2 4.0 4.1 7.0 4.5 -
6.9 6.5 5.6 6.1 6.7 6.7 6.8 7.0 7.1 6.9 6.7 6.7
3.8 3.8 5.3 5.1 7.1 4.0 6.4 6.3 6.2 6.2
6.9 6.3 6.7 6.3 6.2 6.8 7.3 6.3 6.5 6.7 6.7
4.6 5.6 5.0 6.4 4.0 3.4 5.6 6.3 5.8 4.9 5.6 5.6
Averages for cocks that copulated 5 + times 6.1 5.2 6.8
5.8
6.6
5.3
Averages for cocks that copulated 0-4 times 5.8 4.0 6.6
5.0
6.6
5.1
P D 9 13 12 9R 11* 7* A* 9L* Su* B* H*
* Cocks that mated five or more times.
7.4.3 Agonistic behavior compared to mating success Agonistic behavior of central cocks was another conspicuous variable that hens might have used in selecting mates. Females could have used either the relative amount or the relative conspicuousness of fighting to evaluate males. We measured agonistic behavior by counting the number of fights in 10-minute periods and the total time spent fighting in 10-minute periods. We measured conspicuousness of agonistic behavior by counting the number of wing beats in 10-minute blocks, and by calculating the percentage of fights that included wing-beating. For the 1969 data, only the percentage of fights that included wing-beating were compared, because of biases in the way in which the data were gathered. All four parameters were compared for the 1970 and 1971 data. Cocks that later obtained five or more copulations did not score significantly higher than the other cocks on any of the individual fighting parameters during the period before many hens attended in any year (P > 0.075, Mann-Whitney U tests, data in Hartzler 1972). During the period when many hens attended in 1970, however, the most successful cocks scored higher than other cocks for
J. E. HARTZLER AND D. A. JENNI
256
Table 7.6. Mean strutting frequencies (average number of struts per minute of cocks) that mated frequently (copulated at least five times) and mated infrequently (less than five copulations) when the nearest hen was within 3 m and when the nearest hen was 4-18 m away, 1971 3-8 Apr
Individual cock
21 Apr- 15 May
9-20 Apr
< 3m
4-18 m
< 3m
4-18 m
5.7 5.7 5.8 5.2 4.4 6.0 5.6 5.3 6.8 6.1 4.1 6.5 5.9 6.5 6.3 6.1
_ 4.8 4.5 5.2 6.4 5.0 5.9 4.0 2.5 3.6 5.0 5.4
3.4 6.5 5.0 5.7 6.6 6.3 5.8 5.6 6.4 7.2 6.2 5.8 6.5 7.3 6.6 6.6 6.8 6.7
_ -
< 3m
4-18 m
4.3 4.5 7.0 4.3 5.7 6.7 6.4 5.0 6.1 6.0 4.4 2.8 2.9
6.8 6.4 7.2 7.3 5.9 7.1 6.1 6.4 7.1 7.4 6.7 6.7 6.5
6.6 4.0 5.5 4.7 2.9 1.8 4.6 6.1 6.3 6.3 6.2 6.3 3.8 4.6 4.4 6.1
Averages for cocks that copulated 5+ times 6.3 4.7 6.7
4.0
6.9
5.0
Averages for cocks that copulated 0-4 times 5.5 4.8 5.9
5.6
6.6
5.0
9R 8 1,3L 1 P 95 8R 8L 2R 13 79 1/2 9L 3* 5R* 90* 6* 3L* 5*
-
* Cocks that mated five or more times.
three of the four agonistic parameters (time spent fighting per 10-minute period, number of wing-beatings per 10-minute period, and percentage of fights including wing-beating, P < 0.05, Mann-Whitney Utests, U = 11, 14, and 13, Table 7.7). In 1971, differences were not significant (Hartzler 1972). We also calculated an overall fighting rank by combining the ranks for each parameter for each male. Again we compared the most successful cocks with the others. Before many hens attended, these two groups of males did not differ in either 1970 or 1971. During the time many hens attended in 1970, the most successful cocks had higher overall fighting ranks than the others (P < 0.025 MannWhitney t/test, U = 9), but the two classes did not differ in 1971 (Table 7.8). There appeared to be a general trend for the most successful cocks to have
MATE CHOICE BY FEMALE SAGE GROUSE
257
Table 7.7. Rank of cocks by fighting parameters while many hens attended, 28 March-6 May 1970 Fighting
No./lO min.
Wing-beating
Time/10 min.
No./lO min.
% of fights with wing-beating
Rank
Cock
Ave.
Cock
Ave.
Cock
Ave.
Cock
1
1 B* 11* 6 7* 9R 8 P 13 9 12 D Su* A* 3R 9L*
2.5 2.4 2.4 2.2 2.1 1.9 1.6 1.5 1.4 1.4 1.2 1.1 0.8 0.8 0.6 0.4
7* 9L* 11* Su* 6 B* 8 P 9 9R 3R 12 13 1 D A*
2.1 2.0 2.0 1.9 1.7 1.6 1.6 1.6 1.5 1.4 1.3 1.2 1.2 1.2 1.1 0.6
Su* 9L* 8 9 B* 7* 3R 11* 13 9R P 12 A* D 6 1
2.1 1.7 1.5 1.2 1.2 1.2 1.0 1.0 1.0 0.9 0.6 0.6 0.4 0.1 0.0 0.0
Su* A* 9L* P B* 3R 9R 9 8 7* 13 12 11* 1 6 D
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
% 46 43 41 37 33 33 29 28 26 23 22 20 17 14 12 09
* Cocks that mated five or more times.
higher fighting ranks or higher strut-display rates, but results were not consistent. It is possible that females were reacting to overall male activity. That is, females might have responded to how much males fought and displayed. To approximate total activity we combined the strutting and total fighting ranks to derive a total activity rank. We simply added the two ranks, although arguments could be made that the activities were not equally important to females or that their relative importance might have changed during the season. Early in the season, before many hens attended, this total activity score was significantly higher for cocks that eventually obtained five or more copulations than for the other cocks in 1970, but not in 1971 (P < 0.04 and > 0.05 Mann-Whitney U tests, U = 1, and 22). However, during the period when many hens attended, the total activity ranks for the most successful cocks were significantly higher than for the other cocks in both 1970 and 1971 (P = 0.025 and 0.05, Mann-Whitney U tests, U = 10, and 17, Tables 7.9 and 7.10). It was, of course, during the period of peak hen attendance that hens had the opportunity to compare the cocks' activity patterns. Attempted copulations were interrupted when neighbor cocks attacked in all 3 years; 17% were interrupted in 1969, 7% in 1970, and 17% in 1971. Sometimes
258
J. E. HARTZLER AND D. A. JENNI
Table 7.8. Total rank of fighting, obtained by averaging the four fighting parameters measured during the time when many hens attended 28 March-6 May 1970 Rank
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Cock
B* Su* 7* 9L* 8 11* 9 P 9R 6 3R 13 1 A* 12 D
3-18 April 1971 Rank
Cock
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
6* 3L* 9R 3* 8 9 1/2 5R* 8L 79 13 90* 2R 8R 1 95 5* P
* Cocks that mated five or more times.
the attack was too late and the copulation was complete by the time the mounted cock was hit. Attacks were successful in preventing successful copulation 75 % of the time in 1969, 56% in 1970, and 78% in 1971. Cocks that were interrupted most frequently were those that performed many matings and had many close, older neighbors. In 1969, cocks G, H, and F performed the most matings and were interrupted the most often (Table 7.11). Cocks Su and 6 both mated and were interrupted most frequently in 1970 and 1971, respectively (Table 7.11). Cocks 5R and 5 in 1971 and cock A in 1970 each performed many matings but were not interrupted (Table 7.11). Each of these males had only one older neighbor whereas the three cocks (G, Su, 6) that were interrupted most frequently had three or more close, older neighbors with which they frequently fought.
7.5 Discussion While on the lek, adult sage grouse cocks maintain exclusive territories and display and fight with their neighbors for much of their time; only rarely do they have an opportunity to interact with intruding males other than neighbors. When
MATE CHOICE BY FEMALE SAGE GROUSE
259
Table 7.9. Rank of cocks by total activity before (19-27 March) and while (28 March-6 May) many hens attended Ford's Creek lek, 1970 19-27 March 1970
28 March-6 May 1970
Rank
Cock
Rank
Cock
1 2 3 4 5 6 7 8 9 10
7* 9 11* P 1 3 13 12 6 4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Su* 9L* 7* B* 9R P 3R 8 11* 9 13 A* 12 6 1 D
Total matings 159 6 12 9
5
18
* Cocks that mated five or more times.
they leave the arena in the morning, however, they often leave in one or more groups, and it appears that group members often spend the day together. During the day they presumably have access to one another and could establish a dominance hierarchy in the way that black grouse (Tetrao tetrix) do in the Netherlands (Kruijt et al. 1972). Even if they do, however, we have never seen such a hierarchy expressed on the arena among the lekking cocks, nor can we find any descriptions of such behavior in the literature. This does not preclude the fact that one male can dominate another on a territory boundary or area of overlap. The notion that cocks are arranged hierarchically on the lek is based on Scott's (1942) unfortunate assumption that success in obtaining copulations is correlated with rank.
7.5.1 The site hypothesis for mating success Most copulations occur with hens that are in hen clusters. It is clearly advantageous to the few cocks that do the majority of the mating to have the hens cluster in their territories. The advantage of clustering to the hens, however, is not obvious. Before the hens cluster in the spring, when only a few of them visit the lek, they wander through the arena in what appears to be a random way of passing
260
J. E. HARTZLER AND D. A. JENNI
Table 7.10. Rank of cocks by total activity before (19-27 March) and while (28 March-6 May) many hens attended Ford's Creek lek, 1971 19-27 March 1971
28 March-6 May 1971
Rank
Cock
Rank
Cock
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
3L* 2R 6* 13 9R P 5* 8R 1,3L 8 5R* 9L 1/2 39 3* 79 7 8L 56
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
6* 3* 3L* 13 8 5R* 95 2R 9R 90* 79 8L 1,3L 1/2 8R 5* 1 P
Total matings 67 9 11
63
11
13
* Cocks that mated five or more times
into and out of most of the territories. Later, when the hens cluster and begin to copulate they typically cluster on more than one territory before they begin returning to the same one or two territories morning after morning. At the Ford's Creek lek in 1971 the hens clustered again at one of the 1970 mating centers, but they did not cluster at the other three mating centers of 1970. One of the other four mating centers of 1971 did overlap partly with the major mating center of 1969. Thus, on the Ford's Creek lek in 1969, 1970, and 1971, mating centers did not develop at the same locations any more frequently than one might expect on the basis of random placement on the central area of the lek. These observations are inconsistent with Wiley's (1973a) conclusion that mating centers occur on the same sites in subsequent years. Wiley's evidence came from Scott (1942), who studied six strutting grounds for a total of 12 "nights" in 1940. In 1941 he studied the largest of these for 15 nights and wrote, "Since the four principal mating spots had (in the present incidence) the same general position on the strutting ground as in the preceding year. . . . " (p. 484). His second piece of evidence is Patterson's (1952) report that "Several of these prospective 'mating spots' would be located on a large strutting ground, and the same spots
MATE CHOICE BY FEMALE SAGE GROUSE
261
Table 7.11. Interrupted matings at Ford's Creek lek
Interrupted cock
No. of attempted matings
Attempted interruptions in which mating was successful
Attempted interruptions in which mating was Total unsuccessful i nterruptions
1969 G F H J I A
37 16 15 9 7 6
2 by F 1 by G, 1 by I -
2 by 4 by 3 by 1 by -
H, 1 by F G G, 1 by I
5 by 9L, 1 by 9R, 1 by 9R & 9L 1 by B
1 by 9R, 1 by 9L
G
2 4 6 0 1 0
1970 Su
178
A 7 H
16 13 9
9L 11
7 5
-
10 0
1 by P 1 by B, 1 by 1 by 13 1 by Su 1 by Yl
1 B & 13,
4 1 1
1971 6 5R 5 3L 90 1/2 9R
90 63 13 16 11 7 1
4 by 3L
\ by 6 1 by 3L 2 by 79, 1 by 2R
22 by 3L, 1 by 8 5 by 6 2 by 79, 1 by 2R 1 by 6
27 0 0 6 1 5 1
would be utilized by groups of hens during successive seasons" (p. 158). Wiley's own observations are similar. On two leks that he studied, "mating centres occupied similar positions in two years. The Dry Sandy Lek had four mating centres in 1969. When I visited the lek on 30 April of the preceding year, two of those mating centres were in at least roughly the same positions. Since very few females were seen on that date late in the season, I had to use my judgment on the fact that territorial males cluster more densely around mating centres, information that only approximately specifies the position of a mating center. . . . The mating centre . . . on the Muddy Springs Lek in 1967 was located in precisely the same spot in 1968. Its location, determined on 3 May 1968, was immediately apparent, since the packs of females that had gathered there had trampled the grass to its roots, as they had done in the preceding year" (Wiley 1973b, p. 103). These reports show that mating centers occur in the same general part of the arena each
262
J. E. HARTZLER AND D. A. JENNI
year but do not, in our view, constitute any real proof that mating centers occur in precisely the same location each year. Scott, for example, is no more specific than "same general position"; Patterson wrote about "potential mating spots"; and Wiley relied on trampled grass and male clustering. We feel that these observations do not constitute sufficiently hard proof to justify the conclusion that mating centers are precisely fixed from year to year. Our interpretation of Scott, Patterson, and Wiley in the preceding paragraph differs significantly from Wiley's 1978 interpretation. Wiley wrote that, "the mating center, a small inner area about 10 meters in diameter where most of the copulations occur, is found in the same place year after year. Scott, Patterson, and I have all observed this consistency of location, although in small leks or leks disturbed by human activities the location of the mating centers seems to be less consistent" (1978, p. 114-16). Because it is impossible to study a lek without causing a minimum of disturbance, the last caveat allows one to explain any inconsistency of location as disturbance-induced. This passage does suggest that the difference between our interpretations might be exaggerated by differences in what we mean by "in the same place." Wiley (ibid.) describes the mating center as "about 10 meters in diameter," or about three times as large as the territories of his most successful cocks. We believe, however, that Wiley really does mean precisely the same spot, and these slight differences in square meters are trivial. If females are coming to specific locations on the lek and mating with the male that occupies that site, regardless of other characteristics of that male, we would expect fierce competition among males for those sites. Once during the 3 years of this study we saw a male defeat the owner of a territory where hens were clustering, a mating center in Wiley's sense, and take over his territory during the peak of hen attendance. Females had been clustering there, and a few came back the next day, but none copulated and no females visited that territory subsequently. We saw no serious attempts to take over the other nine successful territories that we observed. When hens clustered on a territory the neighboring males often intruded on that territory when they moved toward the hens. These intrusions led to fights between the cocks but these fights did not appear to differ significantly from other boundary fights. Even if cocks do not fight more with cocks whose territories attract hen clusters after the hen clusters appear than earlier in the season, we would expect to see returning cocks attempt to occupy those areas at the beginning of breeding activities the following spring if the site is critical in attracting hens. Instead, returning cocks show a high degree of site fidelity. Even when the holder of a territory who copulated many times the previous year fails to return, his old neighbors typically reoccupy their old territories. They make only minor changes in boundaries to include a small portion of the formerly successful but now vacant territory. The net result is that what was a successful territory in one year is sometimes subdivided into parts of several territories the next year and fails to attract
MATE CHOICE BY FEMALE SAGE GROUSE
263
hens. Wiley (1978) also reported that when vacancies occurred on the lek that allowed neighbors to move closer to a mating center, there was no conspicuous, aggressive competition among the neighbors for the area. Neighbors of the most successful cocks are not the only cocks that know the location of the mating centers. The arena is open and the birds have good visibility. Cocks that are two or more territories away from a mating center move toward the hen clusters as much as their neighbors will tolerate. Yet when birds return the next spring there is no more competition for last year's successful territorial sites than for any other site on the lek. Vacancies on the lek, regardless of where they occur, often become occupied by adult cocks that are being recruited to the central part of the lek. We found no evidence for the centripetal movement of the cocks toward a central mating center that Wiley proposed (1978). Not only did we find strong site fidelity rather than centripetal movement, but also we found hens clustering at different locations each year. Because yearling cocks are typically on the periphery of the lek and adult cocks occupy more central territories, there is an initial trend toward centripetal movement. Yearling cocks, however, do not attend the lek until well into the season and their numbers do not peak until after the peak of hen attendance (Jenni & Hartzler 1978). By this time some of the older cocks are no longer attending the lek regularly. During the last few weeks of lekking, territorial fidelity, in general, tends to break down and more adult cocks abandon the lek. The remaining cocks, of both age categories, move through the lek and often establish temporary territories at central, formerly occupied sites. Thus, adult cocks that are recruited into the central part of the lek each spring could be two-year-olds that are returning to sites they had explored as yearlings late the previous spring. One cock that, as a neighbor of the most successful cock in 1969, copulated 11 times, returned to its territory the next year. When hens failed to cluster there in 1970 he leapfrogged across several central territories, which is contrary to the site hypothesis. He then took a previously unoccupied site near where the hens were clustering and copulated seven times that year. If the site hypothesis were valid, and hens simply mated with any cock occupying a traditional territory, those always successful territories would be limited resources for which males should compete. We have already discussed the surprising absence of more intense competition at the beginning of the lek season for territories that had been successful the previous year than for any other territory. One possible explanation for a lack of conspicuous aggression is the generally accepted belief that cocks on the lek have a dominance hierarchy. Such a hierarchy would allow for a priority of access to limited resources. It is true that some males have a priority of access to females, but that access comes about because females come to specific males, not because they come to specific territories regardless of which males hold them. We disagree with the conclusion that male sage grouse "are ranked in a dominance hierarchy in accordance with the distance
264
J. E. HARTZLER AND D. A. JENNI
of their territories from the mating center" (Wiley 1978, p. 117). Wiley's "polarized territorially" is based on the assumptions that mating centers are permanently fixed in space and that a centripetal progression occurs toward a geographically fixed mating center. We have already discussed that, except for the yearlings on the lek periphery, with older cocks more central, we have no evidence for centripetal movement. Two-year-old cocks were scattered throughout the center of the lek in the second and third year of this study; they were not located between the yearlings and the old cocks. Territoriality is a form of social dominance (Kaufmann 1983) and its effectiveness clearly depends on the ability of the territorial holder to dominate conspecifics that intrude. This does not lead to a dominance hierarchy, however. The only example of a species in which individuals simultaneously exhibit both territoriality and a dominance hierarchy is the polarized territoriality as described by Wiley. Because this appears to be an invalid hypothesis, we are left with no examples of both social systems existing simultaneously in a single group of conspecifics. In the black grouse, territoriality is limited to the lek early in the morning, and a form of space-related dominance is found at other times of the day (Kruijt et al. 1972). Female sage grouse cluster at different locations during and between seasons; these locations are not fixed resources that males can compete for. To correlate dominance with priority of access without some independent measure of dominance/subordinance behavioral patterns tells us nothing about dominance relationships among the cocks.
7.5.2 Male behavior and mating success If females do not go to a specific site to copulate with the cock that holds that site because he is the alpha in a dominance hierarchy of males at that lek, how do females choose males? Clearly the males do not choose females. The possibilities seem limited to individual differences among the males in morphology, or in some behavioral characteristic(s), or in some combination of morphology and behavior. Except that yearling cocks have conspicuously different plumage than older cocks, no one has described morphological differences among sage grouse cocks, let alone correlated any differences with cocks' success in obtaining copulations. Presumably, of the possibilities, only behavioral differences remain tenable. Wiley's (1973b) demonstration that the temporal properties of the strut display are highly stereotyped suggested that there was little likelihood that females could be discriminating between males on the basis of these displays. However, Gibson and Bradbury (1985) have discovered a significant relationship between mating status and an acoustic component of the strut display in male sage grouse. This discovery raises the distinct possibility that there may be other behavioral differences between cocks. The most conspicuous factors associated with the most successful males are:
MATE CHOICE BY FEMALE SAGE GROUSE
265
presence of hen clusters in their territories, copulations, and postcopulatory displays of hens. Although these clues are all available to hens arriving late in the season, they are insufficient to explain why hens began going to particular males in the first place. They are also insufficient to explain why hens in clusters do not necessarily return to the same male territory on subsequent mornings. Two conspicuous factors for potentially distinguishing among the adult cocks were the differences in their fighting and display behavior. Early in the season, before many hens attended, there were no differences between the most successful cocks and the less successful ones in fighting (four parameters) or strutting rates. Once hens began to attend the lek in large numbers, however, the more successful males, in comparison with the less successful ones, fought at faster rates and more intensely in 1 of 3 of the strut display years. During the period when many hens attended, the most successful cocks also performed more strut displays per unit of time in each of the 3 years. This was an important finding: there was a behavioral difference between the most and least successful cocks other than the differences in success, and this difference was available for assessment by the hens. Similar results have been reported from a much smaller lek in California (Gibson & Bradbury 1985). Because cocks react to close approach by hens by strutting more frequently, Wiley (1978) proposed that the faster strutting rate of the central cocks, compared with the peripheral cocks, was a consequence of having hen clusters in the central part of the lek. Wiley compared strut rates and discounted the differences. However, the ways in which he gathered and reported data were very different from ours. Wiley compared central cocks as a class with peripheral cocks as a class; we compared cocks that copulated five or more times with cocks that copulated zero to four times. We made our comparisons among central cocks only. Another important difference was that we compared strut rates when hens were within 3 m of the observed cock and when the nearest hen was more distant, 4 to 18m. Wiley compared three classes of cocks: those with hens in their territories, those with hens in adjacent territories but not in theirs, and those with no hens in either their or any of their neighbors' territories. To gather data in either of Wiley's two latter categories, we would have had to measure strutting rates of our most successful cocks —in fact, probably for all of our central cocks, when there were no hens on the lek. We could have done this only early in the season before many hens attended, late in the season when hens no longer attended the lek, or late in the morning after the hens left and lek activity was at its low for the day. We think such a comparison would be invalid and we are confident that our data reflect real differences among the central cocks on the Ford's Creek lek in 1970 and 71: the most successful cocks strutted faster when hens were close to them than did the less successful cocks. Display rate has also been found to be significantly related to mating status by Gibson and Bradbury (1985). We propose that the apparently aimless, random wandering of hens across the
266
J. E. HARTZLER AND D. A. JENNI
lek early in the season, as well as their frequent wanderings during the peak of hen attendance, allows the hens to compare the way the different cocks respond to their approach. Although there is a strong tendency for hens to follow one another on the lek (Bradbury & Gibson 1983), not all hens are followers. Some females express mate choices more independently (both Wiley's 1978 data and ours). In our study 42 cocks obtained one or more copulations (and two of these cocks copulated in two different years). A total of 55 copulations were performed by cocks that obtained from 1 to 4 copulations each. In our fascination with the cocks that copulate the most, we tend to ignore these others, but they are important because they do show that some hens express mate choices independent of the hen clusters (Bradbury & Gibson 1983). At leks with fewer hens, the hens may show even more independence of choice. Of 23 hens that mated at a small California lek, three cocks were chosen by three hens each, four were chosen by two hens each, and six were chosen by one hen each (Gibson & Bradbury 1985). We did not find, however, a correlation between strutting rate and success, suggesting that hens may use other criteria in addition to strutting rate in choosing which cocks they will mate with. All cocks, regardless of their success in obtaining copulations, strut faster when they have hens near them. Perhaps the hens are making some overall assessment of the males. Fighting and strutting could both be important. When fighting and strutting ranks of the cocks were combined into a single activity rank, the most successful cocks had significantly higher ranks than did the less successful cocks. There is no reason to presuppose that female sage grouse are cuing on any single variable nor only on those measured by grouse biologists. Many of the potentially important variables remain unmeasured, and the potential interactive effects remain untested. Cocks often attempted to disrupt the copulations of their neighbors. In our study, attempts were made to disrupt 12% of the attempted copulations, and 71 % of these attempts were successful, resulting in the successful disruption of 9% of all attempted copulations. Most disruption attempts were made against the most successful cocks, which, of course, provided the most targets for their neighbors. The most successful cocks seldom retaliated. Their neighbors had fewer hens that they attempted to copulate with, and they were busy with hen clusters in their territories. Disruption does not appear to gain copulations for the disrupting cock; we never saw a hen go into the territory of a cock that had disrupted her mating and mate with him. The hen typically mated with her first choice. Foster (1983) proposed that disruption would be advantageous to the disrupter if he could, on average, decrease his age of first reproduction or obtain more copulations. Neither condition appears to apply to sage grouse (Wiley 1973a, this study). Foster also proposed that selection should operate to minimize disruptions at leks through: (1) the evolution of strict dominance hierarchies among lek males, but as already discussed, dominance hierarchies do not occur on sage grouse leks, or (2) by increasing the separation of lek males, which may result in the formation
MATE CHOICE BY FEMALE SAGE GROUSE
267
of exploded leks. The function of disruptions on classic, compact leks, such as sage grouse leks, is not at all clear. They do not appear to have any serious consequences on average, for either the disrupter or disruptee, but are expensive acts in terms of energy expenditure.
7.5.3 Conclusions Clearly the site hypothesis does not appear valid. We reject the hypothesis on the basis of our observations. The literature includes no conclusive proof that hens select mates on the basis of their occupying a traditionally successful territory. Nevertheless, mating is restricted to the central part of the lek, and because space is limited there, mating centers could recur in precisely the same site by happenstance. It is also possible that hens are attracted to the most "conspicuous" cocks, and that cocks may become conspicuous by displaying more frequently than do other cocks when hens are close (Gibson & Bradbury 1985 and this study), or by displaying in qualitatively different ways (Gibson & Bradbury 1985). On leks of only three or four cocks (Gates 1985), the most central cock has more neighbors and hens closer to him for him to react to with higher display/fighting rates than do the more peripheral cocks. The centripetal hypothesis of how cocks progress toward a mating center is considerably weakened by the rejection of the site hypothesis. This hypothesis assumes that it takes several years before a cock becomes successful. However, one of our two-year-old cocks obtained nine copulations, and at a new lek with 11 hens in attendance a two-year-old cock was the only one seen copulating (Gates 1985). Our data do not support the notion that cocks become more successful as they get older. Successful cocks do not seem to be any more successful in subsequent years than any centrally located cock. In our study the two most successful cocks did not return the year after they were so successful. Until someone does a longitudinal study of a permanently gridded arena and demonstrates that hens come to precisely the same spot and mate with the same male that occupies that spot, regardless of his behavior, the site hypothesis must be rejected.
7.6 Summary During a 3-year study of sage grouse lek behavior on the Ford's Creek lek in central Montana, we permanently marked 41 cocks. Except for recent work by Bradbury and his colleagues we know of no other long-term studies of sage grouse behavior on a single lek in which cocks were individually marked. It is not surprising that our findings are inconsistent with previously available reports that were based on short-term studies. Unfortunately, since the pioneering work of Scott (1942), researchers have been influenced by his interpretation that sage
268
J. E. HARTZLER AND D. A. JENNI
grouse cocks are arranged hierarchically on the lek, and this false notion has become firmly established in the literature. In a trivial sense Scott was correct (some cocks breed and others do not), but this has been misinterpreted as a dominance hierarchy. It is not. Male sage grouse exhibit classic avian territoriality on the lek. The limited resource they are defending is a place where they can display and copulate, if they are able to attract hens that will copulate with them. Neither we nor anyone else reports seeing the kinds of dominant/subordinant interactions among cocks that would allow one to rank them. All rankings appear to be based on success at obtaining copulations. Adult cocks that return to the lek in subsequent years show a remarkable degree of territorial fidelity. Almost all return to their previous year's territory regardless of whether they were successful there or not. A few cocks establish territories at new locations in the central part of the lek, but not necessarily closer to the sites where many copulations took place the previous year; and a few may change locations during a single season, usually moving to a site closer to hen clusters. Territorial boundaries do not remain the same from year to year. Almost all returning cocks made at least minor shifts in their territorial boundaries, sometimes toward sites where many copulations occurred the previous year, but not always. Only 29% of the permanently marked cocks returned to the central part of the lek the next year. The new recruits occupied space between the returnees and established territory boundaries without regard for history except as limited by the returnees. In this study neither older cocks nor returning successful cocks were any more likely to become successful than any adult cock on the lek. Yearling cocks apparently never or rarely breed, but we have established that two-year-olds do. We saw no more competition among males for sites where many copulations occurred the previous year than for any other sites in the central part of the lek. More important, in subsequent years, hens cluster and copulate at sites in the central part of the lek that are no more similar in location to those of previous years than would be suggested by random selection of sites. Even within a single season hens cluster on several different territories. There are three possible reasons why hens might sometimes cluster at precisely the same sites in subsequent years: (1) random coincidence; (2) the central area of a very small lek might include only one territory; or (3) there could be leks where cocks on particular sites are so much more conspicuous than other cocks, regardless of differences in their strut rates, that they always attract hens. Sage grouse leks in general do not have traditionally located "mating centers" for which cocks compete and toward which they centripetally progress year by year. Instead, our observations show that hens may use behavioral cues to select mates. Not only do the most successful males strut faster than other males during the period when many hens are attending the lek, they also strut faster than other
MATE CHOICE BY FEMALE SAGE GROUSE
269
males when both have hens close to them. The most successful males are also conspicuous because they have hen clusters in their territories, they copulate frequently, and hens perform conspicuous postcopulatory displays. Hens tend to follow one another on the lek and to join already existing clusters. The initial choice of where to cluster must be made early in the season when hens visit the territories of most of the central males and have more than adequate opportunity to assess the males' response to close approach. Even during the peak of hen attendance a few hens wander throughout the lek during, but especially toward the end of, morning attendance. Some hens do make choices independent of hen clusters. Hens that do not mate with the most successful cocks distribute their copulations widely among the other males, presumably making independent choices. Perhaps some hens choose and others follow, and others choose but are not followed.
8
Behavior of White-tailed Ptarmigan during the Breeding Season R. K. Schmidt
8.1 Introduction The white-tailed ptarmigan (Lagopus leucurus) is the smallest grouse and the only species of ptarmigan whose range is not circumpolar. White-tailed ptarmigan occupy alpine regions of the Rocky Mountain Cordillera, from Alaska and Yukon Territory south to Colorado and parts of New Mexico, with isolated populations extending into the Cascade Ranges of the Pacific Northwest. In Alaska and northwestern Canada they coexist with rock and willow ptarmigan (L. mutus and L. lagopus, respectively), although these latter species generally favor less-exposed tundra and tundra-edge habitats found at lower elevations (Weeden 1959a,b). Over the past 15 years, a good deal of research has been carried out on the ecology, movements, and population characteristics of white-tailed ptarmigan (Braun 1969, Braun& Rogers 1971, Braun& Schmidt 1971, May &Braun 1972, Hoffman & Braun 1975, May 1975, Herzog 1977b, Giesen & Braun 1979a, Giesen et al. 1980). Published accounts of ptarmigan behavior have been limited largely to wintering flocks (Braun & Schmidt 1971, Braun et al. 1976) and to activities of pairs during laying and incubation (Giesen & Braun 1979a,b, Giesen et al. 1980). To date, the most comprehensive work on the breeding behavior of this species has been done by Choate (1960, 1963a,b). Others such as Bradbury (1915), Jensen and Ryder (1965), Lewis (1904), and Weeden (1959a), have contributed information dealing with certain general aspects of white-tailed ptarmigan behavior. The present study was initiated in March 1967. Field work was carried out 270
BEHAVIOR OF WHITE-TAILED PTARMIGAN
271
from mid-April to 31 August 1967 and was continued from 26 March to 12 September 1968. Intensive investigations were made in Rocky Mountain National Park, Colorado, during late spring and summer; late winter and early spring studies were conducted at Guanella Pass, Clear Creek County, Colorado, in 1968. Originally, I had intended to study only those behavior patterns relating to courtship and territorial activity. Later, the study was expanded to include behavior of nesting females, broods, and postbreeding male flocks.
8.2 Study areas Most field work was done at two study areas along Trail Ridge Road in Rocky Mountain National Park (Fig. 8.1). The Sundance Mountain site (2.9 km2) was the main area of research from April to mid-July; late summer studies (mid-July to early September) were carried out at Toll Memorial and supplemented by visits to Fall River and Iceberg Lake, areas also in Rocky Mountain National Park (RMNP). In April 19681 studied ptarmigan at Guanella Pass, located 9.7 km west of Georgetown, Colorado. White-tailed ptarmigan territories in Colorado generally lie along the ecotone separating true alpine tundra and subalpine forest. In this and other respects, the Sundance Mountain (SM) area is typical of ptarmigan breeding ranges. The area consists of a basin or benchland at 3,508 to 3,692 m above sea level bounded on the northwest by Sundance Mountain (3,815 m) and on the southeast by Tombstone Ridge (3,600 m). To the east, the terrain dips sharply down a headwater drainage and into a large cirque. The lower portion of SM (3,569 m) is a gently sloping terrace, much of it wet and marshy in late spring and summer, containing stands of willow (Salix spp.) and sedges (Carex spp.), termed willow-sedge hummock stands by Marr (1961). Clumps of stunted Engelmann spruce (Picea engelmannii) grow throughout this basin. In winter, areas containing willow and spruce tend to accumulate snow, whereas the drier uplands often remain snow-free. Hillsides west of the SM basin are terraced and have slopes of 30% to 35 %. Strong westerly winds in winter leave the leading edge of benches bare and deposit snow on concave leeward slopes. The major snow-accumulation areas (nivation depressions) on these east-facing slopes contain rock fields (with stones of 15 to 61 cm in diameter) and support snow-tolerant plants such as hair grass (Deschampsia spp.) and buttercup (Ranunculus spp.). The drier benches are not as stony and support cushion plants and lush vegetation such as avens (Geum rossii) and bistort (Polygonum spp.). Large rock outcrops are conspicuous along the uppermost ridges; rock rivers and broad talus slopes occupy the upper portions of East Sundance Mountain. The Toll Memorial area (TM) at 3,692 to 3,754 m is relatively flat to gently
272
R. K. SCHMIDT
Fig. 8.1. Sundance Mountain (A) and Toll Memorial (B) ptarmigan study areas, Rocky Mountain National Park, Colorado.
sloping, with rock outcrops running northeast-southeast along the ridgetop. Boulder fields and areas of patterned ground are scattered throughout the TM unit. Vegetation is primarily dry tundra (sedges, cushion plants, Kobresia spp., and Geum rossii), although snow-tolerant plants occur in nivation depressions, most of which are located on north- and east-facing slopes. Dominant vegetation at both the SM and TM study areas is described in some detail by Braun (1969).
BEHAVIOR OF WHITE-TAILED PTARMIGAN
273
8.3 Methods I located ptarmigan by searching on foot and scanning surrounding terrain with binoculars. In 1968,1 used tape-recorded ptarmigan calls to elicit vocal and flight responses. This technique proved effective in locating broods and territorial pairs (Braun et al. 1973). Ptarmigan were captured using a modified telescoping pole 4.9 to 6.8 m long with nooses of 60-lb (for adults) or 40-lb (for chicks) leader wire attached to the end. Birds were banded according to a banding program designed by Braun (1969). Adults and chicks older than 2 weeks were banded on the right leg with numbered, aluminum bands (size 8, National Band and Tag Company, Newport, Ky.). Adults were banded on both legs with plastic wrap-around bandettes (sizes 5,6, and 7) with black numerals. Blue (B) bandettes were used in 1966, yellow (Y) in 1967, and red (R) in 1968. Green (G) and white (W) bands were used as replacements. In 1968, chicks younger than 2 weeks were marked with plain, expandable, plastic leg bands (size 3). Each brood in a given area received a different color. Most behavioral data were collected by watching ptarmigan activities and taking notes. Displays were also recorded on 16-mm and super-8-mm movie film. Sound spectrographic analyses of taped calls recorded at 2.95 cm per second were made on a model 661 A Sonagraph (Kay Electric Company, Pine Brook, N.J.) using both wide- and narrow-band Sonagrams. Day-to-day locations of known territorial males on SM were marked with stakes and plotted on aerial photos (scale approximately 1:7920) from which distance measurements were made. All (12) but two nests were located by watching occupied territories in early morning and late evening during the incubation period. Females observed off their nests were followed back to the nest. Two other nests were found by radiotracking hens fitted with transmitters according to methods employed by Marshall and Kupa (1963) and Brander (1965).
8.4 Early territorial behavior From early November through most of April, white-tailed ptarmigan live primarily in loosely organized, segregated flocks. In Colorado, males winter close to treeline, and females move to lower elevations, usually to drainage basins and stream valleys where willows are abundant (Braun 1969, Braun & Schmidt 1971). From 23 March to 9 May 1967, and again from 20 April to 1 May 1968, 92% of the known birds at SM were males in flocks of 2 to 30 birds. A similar separation of sexes was observed by Braun and Schmidt (1971) at Guanella Pass, where flocks of 30 to 60 birds, mostly females, were common. Exhibitions of territorial behavior before May 1967 and 1968 were brief and sporadic. On 6 April 1968 at Guanella pass, two males associated with a large flock of females performed "territorial" flights and calls from 0515 to 0630. One
274
R. K. SCHMIDT
male also tried to court several hens, an act not repeated during subsequent observations. Territorial displays continued sporadically from 7 to 19 April during early morning (0500-0700) and evening (1900-1930). Because no birds were banded, I was unable to determine whether the same birds participated each time or whether the action occurred on established territories. At SM, I regularly observed groups of two to eight males from 23 March to 8 May 1967. During midday, these groups moved about the basin, feeding and loafing together and showing little sign of intolerance toward each other. Toward dusk, males tended to become aggressive, gave territorial calls, and chased one another; however, such activity was short-lived and did not take place each evening. Similar events were recorded from 12 April to 1 May 1968 for a group of about 15 males. Although displays were suppressed during daylight hours, four males —B21, B36, G32, and Y7 —were on their territories during 12 of 13 sightings. I also noted that when disputes arose, it was the territory holders of previous seasons that initiated display or attacked other birds. Frequency of agonistic behavior increased at SM from 24 April to 1 May 1968, although males remained associated in flocks during the day. I observed four prolonged bouts of aggressive behavior during that period: three in the evening (1915-1935 on 29 and 30 April, 1917-1940 on 1 May) and one in early morning (0510-0600 on 30 April). The evening displays took place in near darkness and always followed a period of intense, gregarious feeding. In each instance, the participants later claimed territories in the SM area; no females were seen during any of the displays.
8.5 Territories and territorial maintenance 8.5.1 Establishment and defense The location and extent of ptarmigan territories were determined through direct observation of individual males and pairs. Territory boundaries were identified by marking sites where neighboring males had fought or engaged in face-to-face confrontations, events which I termed "border disputes." Where there was no adjacent male to contest, the extent of territory was defined by connecting the outermost points from which the resident male performed displays of territorial advertisement. Intermediate display locations were also plotted to establish patterns of use within the defended area. Once established, each territory was maintained as the exclusive domain of the resident male or pair. White-tailed ptarmigan establish territories by means of conspicuous aerial and ground displays amplified by loud, penetrating calls. Vigorous defense of territories coincided with the arrival of female flocks to the SM area on 10 and 11 May 1967, and 2 to 4 May 1968. Principal displays associated with territorial establishment and maintenance are as follows:
BEHAVIOR OF WHITE-TAILED PTARMIGAN
275
(1) Male flight-scream display. The scream consists of four distinct syllables (Fig. 8.2) preceded by rapid, stuttered clucks. Each display lasts about 1.3 seconds, with the accent on the third and fourth syllables. Screams are audible at 0.8 to 1.6 km. The number of calls varies with length of flight, which is always direct and lacks the vertical component characteristic of willow and rock ptarmigan displays (see Weeden 1959a, MacDonald 1970). Periods of quiet at launch and between screams help frame the calls and pinpoint the bird's position and flight path, t (2) Ground challenges (Fig. 8.2) are a mixture of staccato clucks and screams usually given from prominent locations such as rocks or snow-free knolls and ridges. (3) In addition to scream-flight and ground-challenge displays, males assume aggressive threat postures (Fig. 8.3) in response to calls and flights of other ptarmigan. When moving cross-territory, threatening males run in stiff, strutting steps, with their plumage sleeked and body low to the ground. When stationary, the male stands upright and often waves his head rhythmically from side to side. Whether moving or standing still, threatening males always have their scarlet eyecombs fully exposed and utter low, evenly spaced "buc" notes. (4) A previously unreported female flight-scream display (Fig. 8.4) that resembles the male flight scream was recorded six times from five different hens during the spring courtship period, 1968. In each case, the female gave this 3-4 syllable call as she accompanied a male who also performed the flightscream display. I also observed a broody female perform flight-screams in response to chick distress cries, further indicating that females, sufficiently stimulated, are capable of expressing a type of agonistic behavior more commonly associated with males. Even the postures of females before and after flight displays resembled those of threatening males. Unfortunately, I was unable to determine whether the spring displays were directed at other ptarmigan or were simply triggered by intense territorial activity. The mechanisms by which territories were maintained involved three basic categories of display —challenges, border disputes, and fights and aerial chases. In challenging, there was no physical contact or direct confrontation among participants; males claimed or advertised sites by moving from place to place and calling repeatedly. Border disputes were challenge displays carried out at close quarters (with participants 1.5 to 10 m apart), but involved no prolonged invasion of a neighboring territory. Disputes were noisy and kinetic and lasted anywhere from 15 to 300 seconds, although few continued for longer than 60 seconds. Fights and aerial chases took place when one male invaded another's territory. When attacking, males took on a sleek, bulletlike appearance and rushed at their opponents while giving a harsh, chatter call (Fig. 8.5). In most instances, pursued birds evaded attack by flying a short distance, whereupon the sequence was
276
R. K. SCHMIDT
Fig. 8.2. Top: Wide-band sonagrams of a male, white-tailed ptarmigan flight scream (four syllables). Bottom: Sonagrams of a male ground scream and cluck calls.
repeated until the intruder had been driven from the territory. During such episodes, neither male uttered flight calls, and ground screams were rare. I documented only one instance in which an invading male (B36) annexed a small portion of a resident male's (B24) territory.
BEHAVIOR OF WHITE-TAILED PTARMIGAN
277
Fig. 8.3. Typical threat attitude of a defensive, territorial male white-tailed ptarmigan.
In general, fights and border disputes were more common in 1967 than in 1968. During 22 days of observation (from 11 May to 8 June 1967) I witnessed 17 disputes on the SM area, versus 8 encounters in 35 observation days (2 May to 24 June) in 1968. I attributed most of this difference to the loss, in 1967, of one of the dominant territorial males of 1966 (Braun, pers. comm.), thereby allowing two subdominant males (both subadults in 1966) to establish themselves on territories after numerous conflicts along a contested border zone. In 1968, territories remained essentially the same as in 1967, indicating there was less need to redefine positions through direct confrontation.
Fig. 8.4. Wide-band sonagram of a female flight scream (three syllables).
278
R. K. SCHMIDT
Fig. 8.5. Wide-band sonagram of the male's "chatter" call.
Conflicts between territorial and nonterritorial males had little in common with between-territory disputes. Nonterritorial males always assumed submissive, skulking attitudes; once spotted, they were promptly attacked and driven from the territory, usually after a long, silent aerial chase. Like other grouse, white-tailed ptarmigan performed territorial activities in early morning and late evening. Most morning displays occurred during pair formation (11 to 19 May 1967; 2 to 30 May 1968) when flocks of hens were present. At these times, males repeatedly challenged one another, fought over hens, and disputed territorial boundaries. I witnessed such behavior as early as 0355 and as late as 1000. On eight of 11 mornings in 1968, territorial activity began between 0435 and 0520, in semidarkness, and ended between 0600 and 0800; similar patterns were recorded in 1967. Once pairs were established, morning displays became sporadic and less intense. Evening display began shortly after sunset (Fig. 8.6) and continued for 15 to 30 minutes. Displays lasting 30 to 55 minutes were recorded during incubation when males accompanied hens that had left the nest to feed. Territorial behavior usually ceased during the middle hours of the day, except when fog or snow squalls moved into the SM basin. Such conditions restricted visibility, which may have triggered (or prolonged) daytime display. One advantage of crepuscular display is that it reduces the risk of predation by diurnal predators while avoiding the most effective foraging period of nocturnal predators. Given their noisy, conspicuous mode of display, white-tailed ptarmigan make exceptionally good targets, especially for avian predators, and would seem to benefit by being most active during the darkest periods of twilight. The utility of low-visibility display also makes sense with respect to daily terri-
BEHAVIOR OF WHITE-TAILED PTARMIGAN
279
Fig. 8.6. Estimated time of occurrence and duration of evening territorial display by male ptarmigan, Rocky Mountain National Park, Colorado.
tory maintenance. During midday, from about 1000 to 1830, territorial pairs were relatively sedentary and inconspicuous. Nevertheless, there was solid evidence that males, when with females, monitored the whereabouts of other males on a more or less continuous basis. Even when resting or feeding, males remained watchful and stationed themselves on a prominent site such as rocks or tussocks where they could see (and be seen by) other pairs. That ptarmigan are able to detect other ptarmigan on the ground at considerable distances was demonstrated in the reactions of territorial males to nonter-
280
R. K. SCHMIDT
ritorial intruders. On several occasions, males spotted skulking intruders up to 100 m away, even though these "floaters" were submissive and crept about slowly, as if deliberately to escape detection. Presumably, males would also be able to keep track of neighboring pairs without resorting to overt advertisement that might attract predators. Under conditions of reduced visibility, however, contact with other birds would be interrupted and the integrity of the territory placed in jeopardy. Under these circumstances, the best way to reestablish contact and maintain territorial position would be to increase the level of activity and thus become more visible. As already described, displays of white-tailed ptarmigan have a high visual and aural resolution well suited for late evening/early morning conditions in mountainous terrain. Movements are predominantly straight-line and wideranging. With straight-line flights, the dark slope of the territory makes an effective backdrop against which the male's white wings are clearly visible, even in very dim light. Movements on the ground are much less obvious, although flashes of the male's white primaries and underparts can be seen at distances of 100 m or more. Taken together, the air and ground displays, punctuated by screams, clucks, and chatters, provide an extraordinarily graphic representation of each pair's position and the extent of their territorial holdings before roosting. A similar sequence of behavior in early morning darkness "resets" each pair's position until improved light conditions eliminate the need for overt display. Fog and snow squalls restricted visibility to a much greater degree than was noted at twilight. At such times, neighboring males could not see each other except at close quarters (within 100 m or so) no matter how actively they displayed. The result was a state of confusion in which males approached and overflew shared boundaries, thereby precipitating a round of fights, aerial chases, and border disputes that seldom occurred during twilight displays.
8.5.2 Size, occupancy, and distribution The status, distribution, and pattern of occupancy of seven territories at SM were studied intensively during the 2 years of investigation (Fig. 8.7). Observations of birds banded in RMNP from 1966 to 1968 showed that many males returned to the same territory in successive springs (Table 8.1). The 1966 records and most sightings from areas other than SM were provided by Braun (pers. comm.). Males such as B9, B23, B4, B34, B14, and B19, which failed to reoccupy their territories in 1968 and which were not seen again on their traditional summering areas, were presumed dead. Choate (1960) reported that males in Glacier National Park, Montana, occupied territories for only 10 to 40 days (average, 20 days). Territories at SM were occupied and defended for about 2.5 months —the period of occupancy coinciding roughly with that of the pair bond, which lasted from early May until midJuly. Thus, each territory provided the living requirements for the male and
BEHAVIOR OF WHITE-TAILED PTARMIGAN
281
Fig. 8.7. Locations of breeding territories at Sundance Mountain, 1967 and 1968.
282
R. K. SCHMIDT
female and nesting cover for the hen. In most instances, the pair abandoned the territory shortly after hatch, although a few broods stayed on territory for several weeks. White-tailed ptarmigan at SM defended large territories, ranging from 6.5 to 9.3 ha in 1967, and from 9.3 to 19 ha in 1968 (Table 8.2). Indeed, the defended areas computed in this study area were larger than those determined for other nonlek species of North American grouse, including other ptarmigan (Weeden 1959b, Boag 1966, Ellison 1971, Herzog 1977b, McNicholl 1978, Hannon 1982). Although it appears that territory size increased in 1968, the discrepancy more likely reflects my greater familiarity with the study area that year, based on more observations of territorial activity. The areas computed for 1967, then, probably underestimate actual territory size. Territories at SM and in other parts of Colorado did not include all alpine habitats, but occurred only in areas containing willow, the chief food source of ptarmigan in winter and early spring (May & Braun 1972). Because snow cover is present on breeding areas during the period of territory establishment, it is necessary that willows be exposed and therefore available. Thus, it is not surprisTable 8.1. Territories occupied by the same white-tailed ptarmigan males in 2 or more successive years, Rocky Mountain National Park, Colorado Territory and <$S identification no. B25 B9 B24 B36 B23 G32 Y7 B4 Y41 B3 B12 Bll B32 B17 B35 B34 Y39 B14 B75 B19 B8 a
Seasons occupied
1966
1967
1968
Location3
X X
X X X X X X X X X X X X X X X X X X X X X
X
SM SM SM SM SM SM SM SM TR TR TR TR TM TM TM TM FR FR FR FR FR
X
X X X X X X X X X X X X X
X X X X X X X X X X X X X X
SM, Sundance Mountain; TR, Tombstone Ridge; TM, Toll Memorial; FR, Fall River.
BEHAVIOR OF WHITE-TAILED PTARMIGAN
283
Table 8.2. Estimated size (ha) of territories at Sundance Mountain, 1967 and 1968
1967 Territory and dd identification no.
Area defended
Area of home base
Area defended
Area of home base
B24 B32 B36 B23 G16 Y7 R12 R43 B9
8.5 6.5 7.3 9.3 —
2.8 2.0 1.3 2.6 —
7.3
2.0
13.4 9.3 17.0 19.0 15.0 13.4 13.8 -
3.8 2.1 4.2 6.4 3.5 (13.4)a 3.3 -
7.8
2.1
14.4
3.9
Average a
1968
Distinct home-base area not identified; figure excluded from average.
ing that all territories examined during the study were clustered around clumps of exposed willow; areas that supported little or no exposed willow either were uninhabited or were occupied by nonterritorial males. The solid boundaries shown in Figure 8.7 represent the outermost limits that males visited when challenging other territorial males. Most of the time, males and their mates stayed within much smaller, "maximum use" areas (see Table 8.2 for estimates of size) that supplied feeding and loafing cover and served as a departure point from which to challenge other males. The maximum use areas also contained areas of exposed willow, important in the early stages of territory establishment and pair formation. To estimate the fidelity of males to maximum use areas, I recorded the number of times B24 and B36 were observed on and off these areas during the breeding season (locations of males during territorial disputes were not included). From 1 May to 19 July 1968, B24 was on his home-base area on 39 of 41 occasions (95%); B36 occupied his area 22 of 27 times (81 %) from 1 May to 1 July 1968. Other males at SM showed a similar attachment to particular parts of their territories; however, I did not compute frequency of occupation on a per-sighting basis. One male (R12) apparently used most of his territory uniformly, possibly because this area lacked vegetative and topographic variety and was hemmed in by the territories of B24, B36, G32, and Y7. Territories and maximum use areas were not fixed entities, but varied in size and shape as the breeding season progressed. Through most of May, when snow covered all but the most exposed sites, males kept to lower slopes and benchlands where willows were plentiful and visited windblown uplands primarily for display. As snowfields receded in June and food plants favored by ptarmigan
284
R. K. SCHMIDT
emerged along upper slopes, most territorial males (B23, B24, B36, B9, G16, and G36) gradually moved uphill and seldom revisited the low benchlands, which had turned marshy after snowmelt. From mid-June to mid-July, pairs inhabited high, rock-strewn slopes that contained remnants of large snowfields. They fed and loafed along snow lines where fresh, lush vegetation was emerging, and returned to nivation areas even after snowfields had disappeared. Pronounced uphill movement was not observed on the territories of Y7, R43, and R12. Unlike 1957 resident B23, R43 in 1968 never abandoned the lower reaches of this territory even though his maximum use area was situated well upslope. Presumably, territorial activity north of Trail Ridge Road (see Fig. 8.7), and particularly that of R12, drew R43 to low-lying areas well into summer. Male R12 had no opportunity for uphill movement, which was blocked by the territories of B24 and B36 to exactly the same area as his predecessor, B23. Another new male in 1968 (R12) occupied what seemed to be a marginal territory (not occupied in 1967) that lacked access to well-drained upslope sites that for most pairs provided nesting and summer foraging habitat. It is perhaps significant that R12 failed to attract a female and therefore did not breed in 1968, despite his aggressive behavior. Finally, returning residents B24, B36, and G16 occupied essentially the same areas in 1968 as in 1967. Minor shifts in boundary positions, particularly in the territories of G16 and R43, apparently occurred in response to territorial activities of other males in the SM basin. The overall pattern of territorial occupancy with respect to chronology of events and spacing of territories was observed elsewhere in Colorado (Braun, pers. comm.). There was little indication that the relatively large size of whitetail territories reflected low breeding densities. Investigations conducted from 1966 to 1968 (Braun, pers. comm.) showed that breeding populations in RMNP remained at relatively constant, high-density levels in all 3 years. If anything, it appeared that male breeding densities approached saturation levels at RMNP study areas. In 1967 and 1968, an undetermined number of unmarked males (presumably juveniles) appeared periodically on breeding ranges. Some of these males (e.g., R43) appropriated territories recently vacated; others (e.g., R12) established marginal territories but failed to attract mates. Still others, probably the majority, were "floaters," who inhabited but did not defend fringe areas. The presence of marginal territories and nonterritorial "floaters" on breeding ranges suggests a surplus of males vying for breeding space. While nonmated males were observed throughout the breeding period, unpaired females were never confirmed after late May, when the pairing process was considered complete. May (1975) at Niwot Ridge, Colorado, also reported an absence of unmated females, as did Braun (1969) for other study areas in RMNP. The significance of territoriality and the probable reasons for its evolution in avian populations have been reviewed by Brown (1964). He noted that species
BEHAVIOR OF WHITE-TAILED PTARMIGAN
285
that tend to occupy large territories probably do so (1) as a result of competition for a limited food source or (2) to provide an opportunity to breed (as determined by ownership of a suitable area in terms of feeding and nesting habitat). Most evidence favoring the first explanation is based on studies of species in which the young are dependent on the territory for their food supply. Since ptarmigan chicks are typically raised outside the territory (Schmidt 1969, Herzog 1977b), the food-for-young interpretation is not applicable. The second hypothesis, however, seems to adequately define the basic functions of territoriality in whitetailed ptarmigan. As noted, ptarmigan territories in most cases have two distinct parts: a central area for feeding and loafing, and a perimeter area that is used primarily for territorial advertisement. The inner portion—the maximum use area—supplies the physical needs of the pair; the outer portion provides a buffer zone that ensures the pair's exclusive rights to prime foraging and resting habitat. But perhaps more important, the perimeter area contains the nest site, which lies 100 to 300 m from principal foraging areas, as determined from nest-territory observations in Colorado (Schmidt 1969, Giesen & Braun 1979b). The separation of nest from the maximum use area is advantageous because it limits cues that might attract predators to the vicinity of the nest.
8.6 Pair bonds 8.6.1 Establishment White-tailed ptarmigan establish pair bonds that typically last for more than one breeding season (Table 8.3). Data indicate that hens, like males, "home" to territories they occupied in past seasons. The chronology of pair formation and reestablishment in 1967 and 1968 was as follows. Male ptarmigan were on breeding ranges and had begun to defend territories by the time the first female flocks arrived in early May. With hens present, males soon had well-defined, vigorously defended territories across which they courted and pursued incoming females. Unlike males, hens did not stay in one place but wandered among territories, apparently in response to male territorial activity. I noted that females tended to go to territories where a male was calling and performing flight displays; their arrival invariably elicited a courting response from the resident male (see Schmidt 1969), which in turn seemed to stimulate territorial display among neighboring males and precipitate renewed movement of females. Sexual/aggressive displays toward hens in the form of courtship strutting and chasing occurred from 11 to 20 May 1967 and from 2 to 9 May 1968. The period of pair formation was considered complete by 21 May 1967 and 31 May 1968, when lone females were no longer seen. Male courtship displays appeared to have little influence on initial pair formation or reestablishment of previous-season pairings. I never observed males
286
R. K. SCHMIDT
Table 8.3. Territories occupied by the same ptarmigan pair in 2 or more successive years
a b
Male
Female
1966
1967
B23 B9 G32 B24 B3 Y41 B17 B19 B8 B14
B94 BIO B10b G52 B59 G81 G57 B20 B66 B15
X X
X X X X X X X X X X
X X X X X
1968
X X X X X X
Location3 SM SM SM SM TR TR TM FR FR FR
SM, Sundance Mountain; TR, Tombstone Ridge; TM, Toll Memorial; FR, Fall River. BIO paired with B9 in 1966 and part of 1967, but B9 was replaced by G32 after May 1967.
courting their mates of past years, although these same males frequently chased and courted other hens that came into their territories. Re-pairing birds seemed familiar to each other, as if they had remained together between breeding seasons. The effect of courtship activity by previously unpaired and widowed males was equally ambiguous. Male B36, for example, courted at least eight females from 11 to 24 May 1967 before obtaining a mate. From 2 to 30 May 1968, B36 accompanied at least ten females but failed to pair that year. During both seasons, hens remained on B36's territory for several days. While they remained (from 2 to 7 days), their behavior appeared indistinguishable from that of paired females. These and similar observations at other study areas indicate that actions of the male are not crucial in deciding pairings. At the same time, I was unable to confirm any clear pattern of overt, aggressive competition among females. Hens accompanied males during challenge displays and border disputes, but did not take an active part. Evidence of aggression was limited to female flight screams and small signs of intolerance (pecking, short chases) among "rival" hens similar to that observed in winter flocks. The principal strategy governing pair formation was not readily apparent in the pairing histories recorded at RMNP study areas. The fact that most pair bonds were maintained through successive years was marked by exceptions that seemed to follow no consistent pattern. For example, female Y81 returned to her 1967 territory and mate (Y7) on 2 May 1968, but left after 9 May to pair elsewhere. Another hen (G15) reappeared on her former breeding area where she accompanied unfamiliar male B36 from 11 to 19 May 1967 and from 2 to 10 May 1968 before leaving the study area. Still other females exhibited a strong, persevering attraction to location, whether last year's mate was in residence or not. Thus, hen BIO paired with male G32 on her former territory after B9, her mate in 1966, suddenly disappeared early in 1967. In 1968, males G32 and Y7 occupied differ-
BEHAVIOR OF WHITE-TAILED PTARMIGAN
287
ent parts of B9's original territory, which BIO seemed to ignore as she moved between the two males without ever abandoning the previous, familiar territory. Taken together, these variations indicate that pair formation involves something other than mate recognition or simple philopatry to previous breeding sites. In the few instances cited here, female attraction to a prior breeding area appears stronger than fidelity to a previous mate, to the extent that attachment to mate may be of marginal importance in deciding the permanence of pair bonds. Still, the question remains: Why do some females return to familiar territories (and mates in some cases) only to abandon them, when the majority of hens exhibit mateterritory fidelity that spans several breeding seasons? First, it seems unlikely that "quality" of territory or resident male per se is responsible. There was no indication that conditions on the rejected territories deteriorated between the 1967 and 1968 breeding seasons, since other pairs successfully established on territories that adjoined and were vegetatively similar to those abandoned. Nor was it likely that some males became less fit as mates, since most of them later paired and successfully bred with other hens. Another possibility is that human interference forced females off territories, but this too is improbable, given that many pairs withstood repeated capture and periods of close observation without noticeable detriment to the pair bond. A third explanation is that female fidelity was influenced by last season's breeding experience which, because white-tailed ptarmigan nest but seldom raise broods on the territory, is synonymous with nesting success. My own observations are too limited to test the hypothesis that successful hens are more likely to return to (and nest on) the same territory in successive years than unsuccessful hens, although indirect evidence lends some support. Braun (1969) has noted, for instance, that successful hens appear more likely to come back to the same brood area than unsuccessful hens (he is careful to point out, however, that successful and unsuccessful birds do not have the same probability of being observed). My observations at SM and TM also showed that several females with and without broods in 1968 returned to areas where they had raised broods in 1967, and that three of the females who re-paired on the same territory in 1968 were known to have nested successfully the year before. Clearly, some females do home to breeding and brood ranges where they were previously successful. If past performance, success or failure, does influence the current year's pairing, it must also account for these major sources of failure: (1) failure that stems from breeding incompetence in the male, (2) failure that reflects an inadequate territory, (3) failure caused by physiological or behavioral incompetence of the female, and (4) failure caused by events more or less independent of the pair, e.g., a visit by a predator not signaled by activity of the male of female, or accidental loss such as through trampling or exposure to adverse weather. A pairing strategy based on past (and hence expected future) success makes sense in the case of incompetent males or males with inadequate territories. In both instances, the
288
R. K. SCHMIDT
male would be "penalized" by being deserted by his former (unsuccessful) female, forced to try for a new mate, and perhaps deprived of the opportunity to breed at all for one or more seasons. At the same time, the female (who is assumed to be competent) would increase her chances for improved fitness by moving. An incompetent or inexperienced female, although she is deprived also of a known breeding/nesting area, is not prevented from attempting to breed in the future by going elsewhere. In both scenarios, then, movement by an unsuccessful female improves the opportunities for successfully nesting that might otherwise be blocked if philopatry to site or mate were absolute. Nesting failures not simplistically related to the quality of the territory or the mate would also, under a success-driven pairing strategy, result in dissolution of the pair bond, even if that pair had bred successfully before. This would appear to be a "foolhardy" strategy because it would penalize the pair bond without offering some distinct selective advantage. Such a conclusion holds true only if one assumes that the so-called chance event that destroys the nest is indeed independent of site and behavior of the pair, and this may not always be true. A nest site may be framed by topography and ground features such that it is more likely to be found by predators or trampled than a nest hidden in krumholtz or tucked behind a large rock; a nest too exposed may be subject to lethal chilling or be more visible to avian predators. Such accidental losses in many respects are not independent of the pair's territorial quality or behavior. Certainly, the female that selects a nest site that is detectable by predators or is too exposed to weather is to a significant extent responsible for that nest's failure. Thus by seeking a different territory and mate in the year following an accidental nest loss, the likelihood of repeated failure might be reduced. Similarly, a male that defended a territory that lacked suitable nesting cover would have fewer opportunities to improve his fitness than other, more successful males.
8.6.2 Maintenance Once the pair bond was established, the male accompanied the female until incubation began. Daily activities generally were synchronized, with both birds feeding and resting at the same times. During feeding periods, males uttered continuous, irregularly spaced "assurance clucks" at a rate of 50 to 80 per minute, the frequency of clucks increasing with distance from the female. Paired males also gave low "churring" calls (Fig. 8.8) when the hen approached within 0.3 to 1.0 m. Churring was accompanied by head-bowing and prominent display of eyecombs, such as accompanied courtship amd precopulatory displays, and caused the female to move away. The threat postures and calls associated with territory maintenance were also used to defend the female. Defense of the hen was most evident when I disturbed the pair, causing the male to place himself between me and the female while as-
BEHAVIOR OF WHITE-TAILED PTARMIGAN
289
suming aggressive attitudes. If the female flushed, the male invariably followed and resumed display. Even when I captured a female for banding, the male would remain nearby (within 1.5 to 3.0 m), clucking loudly and assuming threat postures. In 1968, pair bonds were temporarily disrupted several times by violent weather. During blizzards on 6 and 7 May, 13 and 16 May, and 24 to 27 May, females abandoned their mates and moved (presumably) to lower elevations. At the same time, males gathered in small flocks as they had in early spring and did not consistently defend or even remain on territory. What territorial activity there was, was halfhearted and strictly crepuscular. With the return of good weather, hens reoccupied their former positions and males again maintained exclusive territories.
8.6.3. Nesting period During the laying period, females accompanied their mates most of the day. Because only two nests were located at this time, I obtained little information on the behavior of the pair in relation to the nest site. Both hens visited their nests between 1130 and 1640 (based on four observations). While laying, females were strongly attached to the nest and had to be pushed off to examine the clutch. Hens so disturbed usually moved 0.3 to 3 m away, and some performed distraction displays. Males sometimes accompanied hens to the nest and remained nearby while the nest was occupied; at other times, males remained alone until after the egg had been laid. With the onset of incubation in late June 1967 and 1968, the pair's activity changed dramatically. Males restricted their movements to a small portion of the
Fig. 8.8. Wide-band sonagram of the male's "churring" call.
290
R. K. SCHMIDT
territory containing rock outcrops and snow-accumulation areas that usually overlooked, but were not always close to, the nest site. When disturbed, unattended males remained sedentary or showed alarm, but seldom reacted in a defensive manner. In this respect, they were indistinguishable from nonbreeding birds and males in postbreeding flocks. All nests found during the study period were on-territory, usually near the territorial boundary. Incubating hens normally left the nest once a day to feed, although some were off the nest twice in 24-hour periods. Females usually flew from the nest to a feeding area early in the evening (1900-2010), thereby attracting her mate and initiating a flurry of territorial activity. While with the hen, males resumed their defensive roles and occasionally revived border disputes with adjacent males. Hens for the most part ignored these antics and continued feeding. They favored snow-accumulation areas and consistently returned to the same place each night. Distances separating feeding areas and nest sites ranged from about 25 m to over 300 m, with most nests located 50 to 100 m away. Hens fed avidly for 15 to 30 minutes before returning to the nest. Although birds sometimes walked back to the nest during the incubation period, most flew at least partway; however, I never saw a female fly directly to or from the nest site. It is worth noting that visibility was extremely poor when females made their feeding forays, and birds on the ground, including males, were extremely difficult to see. As described earlier with respect to territorial display, this is not the case with flying birds, whose white wings are startlingly conspicuous even in very dim light. Thus, one must assume that the female's flight, which attracts her mate's attention, is also likely to attract the attention of crepuscular or nocturnal predators (presumably mammalian predators, because the principal avian hunters, prairie falcons [Falco mexicanus} and common ravens [Corvus corax], rarely forage at night). This risk is compensated by the appearance of the conspicuously displaying male who in effect becomes both guard and decoy for the now inconspicuous, feeding female. The decoy value of the male is improved if the female remains inconspicuous (i.e., not flying) whenever she is in the vicinity of the nest, thereby reducing the opportunities for predators to key-in to the nest site. It is significant, I think, that although males accompany hens on postfeeding flights, they usually stay near the landing site, leaving the hen to walk back to the nest alone.
8.7 The brood period 8.7.1 Dispersal from breeding areas Nine of 12 nests studied in RMNP in 1967 and 1968 produced broods. Data on seven of them suggested that newly hatched ptarmigan abandoned the nest site soon after the last egg hatched. Six broods left the nest early in the morning (or
BEHAVIOR OF WHITE-TAILED PTARMIGAN
291
late the previous evening), probably before 1900. One hen apparently led her brood from the nest between 1100 and 1500. Movements following hatch varied among broods. On the first day, most broods remained close to the nest and chicks showed little inclination to forage far from the female. In each of three broods observed over several hours in 1968, the female strayed only 10 to 100 m from the nest the first day. By day two, broods began to venture farther, although, again, there was considerable variability among families. Some broods (G57, R40, R36) remained on or near the breeding territory for several days or weeks; others left the nest and territory immediately. Female G52, for example, moved her brood 300 m in 2 days and in 4 days was 1.45 km from her territory. Long-term brood movements with respect to nest and territory location followed much the same pattern and variability as the early movements. Three broods at Fall River ranged from 350 to 1200 m from their nests and tended to move farther away as the season progressed. One brood at TM was seen three times between 8 August and 12 September 1968, never more than 500 m from the nest, but never inside the hen's territory. Braun (pers. comm.) noted that brood female B31 remained in the territory throughout summer 1966. The following 2 years, B31 bred and apparently nested 1.1 km from her 1966 territory but used that area for brood-rearing all 3 years.
8.7.2. Brood ranges After dispersing from breeding sites, many broods moved to areas that in effect served as communal "brood ranges" through mid- and late summer. Brood ranges in RMNP included the TM area, where most of the observations to follow were recorded, Iceberg Lake, and the upper slopes of SM. About 70 ha of the 130-ha study area at TM were used consistently by 8 broods between 18 July and 31 August 1967, and by ten broods between 1 August and 12 September 1968 (Fig. 8.9). On this and other brood ranges, females most often selected rock habitat (with rocks of 20 to 60 cm in diameter constituting 50% or more of the ground cover) that contained lush tundra vegetation and usually one or more snow-accumulation area(s). Brood ranges, like territories, typically were used by the same females for more than 1 year. At TM, three hens returned with broods 2 years in a row, and one (G57) raised broods there for 3 successive years. Similar fidelity was observed at SM, where two hens (B31 and B91) reared broods in 1966 and 1968. Several females that were broodless in 1968 returned to the areas they had occupied when successful. Thus, hens Y87, G58, and B73 were observed at TM with chicks in 1967 but returned alone in 1968. These brood ranges also attracted successful year-old hens that apparently had had no previous attachment to those sites. Preference for particular habitat types (high, rocky slopes containing nivation
292
R. K. SCHMIDT
Fig. 8.9. Ranges occupied by six broods at Toll Memorial, 1968.
areas) and the tendency of females to return to previously used sites, probably accounts for the phenomenon of communal brood ranges. There was little indication that broods occupied communal ranges because females were in some way attracted to each other. At TM, I often observed two or more broods foraging 15 to 30 m apart; however, such encounters appeared unintentional, evidently the result of brood-paths intersecting, and did not give rise to lasting associations between particular broods. Distances traveled by broods at TM in a 24-hour period were relatively small. During 6 hours of observation (0730 to 1330), one brood remained in a 60 by 40 m area and followed the same elliptical route every 3 hours until disturbed by a human intruder. From 1330 to 1730 this brood wandered 115 m to the east then circled back to within 70 m of its original position. The total distance covered in 10 hours did not exceed 750 m and included no more than 1.2 ha. Similar localized movements were recorded for three other broods, each of which covered only 280 to 500 m over a three-hour period. Some marked broods favored some parts of the brood range more than other parts. Day-to-day locations of six broods observed from 1 August to 12 September 1968 were plotted on aerial photos, which revealed a broad overlapping of individual brood ranges (Fig. 8.9), particularly among broods G37, G57, Y85, and G35. Two of these broods (G37 and G57) frequented the same areas at about the same time each day. Both favored a gentle, northwest-facing slope during early morning and late afternoon, and more precipitous, snow-accumulation areas through midday.
BEHAVIOR OF WHITE-TAILED PTARMIGAN
293
Although ptarmigan broods on communal ranges normally functioned as independent family units, hens occasionally responded to the distress calls of unfamiliar chicks and thus created opportunities for brood mixing. Brood mixing was most obvious in August when I observed the same hen with a variable number of chicks on successive days. Some mixing occurred as a result of predation, as when two females adopted the young of female Y88 (killed by an avian predator), or because of human interference. My unintentional disturbance of broods Y94 and Y96 at SM, for instance, was responsible for the transfer of one chick to Y96's brood. Similar incidents were recorded at other locations, and probably for other reasons, as evidenced by chick exchanges. But, for the most part, there was surprisingly little shuffling of chicks, given the large number of broods present on most communal ranges. The term communal brood range used here refers to a sharing of habitat, as opposed to an aggregation dictated by social instinct; young broods never gathered in compact flocks. By remaining independent, broods remained relatively inconspicuous and were seldom in potentially competitive situations with other ptarmigan. Such tenuous associations would tend to discourage the sort of predator "trap" that more-compact gatherings of animals usually invite, particularly when the gathering place is traditional. One obvious benefit of brood-range sharing is that it ensures better survival among chicks, since lost or orphaned chicks are likely to be adopted by other broods. In a sense, brood-range sharing also enhances or justifies protective behaviour by the female (see 8.7.3) whose investment (i.e., her brood) is underwritten by her brood-hen neighbors. By late August and in September, individual broods began to form "gangs." Such associations were temporary but probably resulted in a good deal of brood reorganization. In large gang broods observed at TM on 11 September and at SM on 6 September 1968, it was impossible to differentiate the individual families. By late September and early October, broods had largely dispersed from brood ranges without their brood hens.
8.7.3. Brood maintenance and defense In very young broods, chicks were brooded at least half the time and seldom ventured more than 1.5 to 3.0 m from the female. At this stage, chicks spent little time feeding and seemed to peck at surrounding vegetation without actually taking much food. Females were extremely vigilant and would not leave the area even when chased. Protective behavior of females varied with age of brood and the context of disturbance. (For instance, hens behaved much more aggressively when disturbance was accompanied by chick distress cries). Hens with broods of 2 weeks or younger often "attacked" a human intruder in response to distress calls. Attacking females sometimes flew at and struck the investigator; more often, they charged
294
R. K. SCHMIDT
on the ground, with wings slightly drooped and spread (emphasizing the white primaries and carpal patches), head and tail raised, and the pinkish eyecombs exposed (Fig. 8.10). Attack displays were accompanied by loud hissing. Females with concealed or noncalling chicks usually "feigned injury" by skittering erratically for 3 to 15 m with wings dragging and head held close to the ground. If not pursued, the hen would stop, still in a crablike crouch (Fig. 8.10), and emit soft "ticking" and "tucking" sounds. Feigning hens also showed carpal patches and eyecombs, and if closely approached would nervously peck toward the ground or make head-jerking movements. Some females, when closely approached, temporarily abandoned their crouch and flew or ran a short distance ahead. During these displays, brood hens ventured no more than a few dozen meters from the concealed chicks. The most common calls associated with feigning were "craaow" and "brrrt" calls. These harsh, sharply defined sounds were heard often during the brood period and apparently served to signal the chicks to freeze. Females with broods older than 2 weeks usually responded to distress cries by "alert-calling," in which they uttered shrill, rapid clucks from a prominent, upright position (Fig. 8.10). If peeping continued, hens moved toward the cries, occasionally pausing on rocks to call. Alert-calling was often punctuated with headjerks and tail flicks. All the protective displays described shared certain features (the exposure of carpal patches and eyecombs, and the lowered wing and head positions of attack and injury feigning; the jerky head movements of feigning and alert-calling; the aggressive, upright attitudes of alert-calling and attack), and all were intended to draw attention away from the brood. On several occasions, encounters with broody females elicited the entire distractive repertoire in rapid, jumbled succession, with the female seeming to attack and draw off the intruder at the same time. These heightened displays were noted during banding operations when chick distress calls were intense and continuous. As chicks reached flight age (about 10 days and older), reactions of females became more subdued. My movements through a brood capable of flight rarely caused the female to display in a distractive manner. Instead, the hen would walk slowly away, clucking softly to her chicks as she went. Defensive displays were uncommon unless distress calls were given. By September and October, brood hens frequently failed to respond to distress cries, although some females gave alert-calls into late October and early November (Braun pers. comm.). Alertcalls were also recorded in spring when males were courting hens on territories. As with the female flight scream, heard on rare occasions during breeding and brood-rearing seasons, alert-calling appears to be a behavior capable of being expressed under extreme levels of stimulation, even when the context of expression appears inappropriate.
BEHAVIOR OF WHITE-TAILED PTARMIGAN
295
Fig. 8.10. Top: Typical "attack" posture of broody female white-tailed ptarmigan. Middle: Female "injury feigning." Bottom: The "alert-calling" posture of a brood female.
296
R. K. SCHMIDT
8.8 Male postbreeding flocks Activities of male flocks of TM were studied from 21 July to 30 August 1967, and from 14 July to 12 September 1968. Locations mentioned in the following discussion refer to Figure 8.9. Breeding males abandoned their territories and began gathering on summer ranges during and shortly after the peak of hatch, which took place from 18 to 22 July 1967 and from 14 to 20 July 1968. In 1967 the first male groups were observed at TM on 21 July; flocks in 1968 began to appear on 14 July. Male flocks frequented the same general areas as broods and, like broods, preferred rocky habitat associated with snow-accumulation areas.
8.8.1 Movements Male flocks tended to use different portions of the TM area as the season progressed. In July they favored areas of lush tundra bordering the upper edges of recently exposed, snow-accumulation areas. As these sites began to dry out in early August, males moved downslope to fresh snow-accumulation areas that still supported newly emerged plants. This trend in seeking recently exposed nivation areas continued through August and into September when flocks shifted to slopes lying east and south of the Quarry (see Fig. 8.9). Although the same males inhabited the TM areas more or less continuously, some moved between concentration areas. Thus, males B24 and R29 were seen at both TM and SM areas in summer 1968, whereas G74 and RW90 wandered between TM and Iceberg Lake. The distance between TM and SM areas is 2.1 km; Iceberg Lake and TM are about 3.2 km apart.
8.8.2 Flock structure Flocks at TM contained from two to 30 males and averaged ten to 15. Other concentration areas supported fewer birds in smaller flocks of five to ten. Broodless females appeared singly or in small groups (of two to five birds) among male flocks. Unsuccessful hens were less often associated with large, male flocks and tended to form all-female groups. It also appeared that fewer hens were present at TM late in the year, indicating that broodless females are more inclined to wander. Braun (pers. comm.) has noted that unsuccessful hens will move considerable distances (up to 20 km) during summer; similar extensive movements are uncommon among males. Male flocks were not strongly cohesive; individuals readily shifted between groups and subflocks, as evidenced by observations of marked birds. Flocks that remained intact during the day disintegrated toward evening, when growing darkness (and the presence of females) triggered brief bouts of display normally associated with territorial defense. Sexual displays, courtship strutting, and bowing
BEHAVIOR OF WHITE-TAILED PTARMIGAN
297
(Schmidt 1969), directed at broodless hens sometimes caused fights among males, disrupting flock structure. There appeared to be a form of dominance hierarchy among the males in a given flock. Males that had failed to obtain mates during the breeding season seemed to dominate males that had bred successfully. At TM, R28 (a subadult) consistently threatened and chased other birds but was never successfully intimidated himself. During the breeding season, this male never paired, although he defended a small territory. Braun (pers. comm.) noted similar dominance by a nonbreeding subadult at Crown Mountain/Crown Point, Colorado. While the two observations of subadult dominance reported here cannot be generalized to ptarmigan populations in general, it does suggest that breeding season aggressiveness incompletely expressed may carry over into later periods.
8.8.3 Late-summer association with breeding areas Most males moved to "neutral" concentration areas after nests hatched, some remained on their breeding ranges in close association with their territories until late summer. Such attachment was observed in four males (G36, G97, G67, and G75) at Fall River, RMNP, in 1968, and among three birds at SM in 1966 (Braun, pers. comm.). In all instances, males stayed on breeding territories and continued to perform defensive displays only when females were present. These males accompanied broods (or unsuccessful hens) and exhibited defensive and sexual behavior typical of the pair-bond period. Female alert-calls played from a tape recorder elicited vocal challenges and flight screams from two males at Fall River; continued playback triggered precopulatory bowing accompanied by full eyecomb display. At Fall River, males sometimes responded to chick distress calls and would approach in various threatening attitudes, but without attacking or attempting to distract the investigator. Defensive reactions of males to distress cries always accompanied female distraction/chick-protection displays, suggesting that movements of the hen, not the chick calls, accounted for the male's response. The immediacy of response, however, may also indicate that males are capable of associating a distress call with female presence. The result is a sort of rudimentary brood-rearing activity seldom seen in white-tailed ptarmigan. It is interesting that brood (hen) defense reactions were not observed on male concentration areas such as TM where brood hen and chick calls were common, daily events.
8.9 Summary Studies of the breeding behavior of white-tailed ptarmigan were conducted at several locations in Colorado from 1967 to 1968. In late winter and early spring, ptarmigan formed sex-specific flocks that frequented areas containing willow.
298
R. K. SCHMIDT
Males first displayed conspicuous, agonistic behavior in mid-April; however, territories were not delineated or routinely defended before early May, when groups of hens visited breeding areas. Territories were occupied and defended from early May until mid-July. Most males returned to territories they had used in previous years and males alone participated in territorial defense, which was expressed mainly through calls and flight displays; fights and face-to-face border disputes were less frequent. Conspicuous, territorial display was restricted primarily to morning and evening twilight hours, usually from 0500 to 0600 and from 1900 to 2030, and to foggy or snowy conditions that restricted visibility. Most territories were occupied by adult males, although some juveniles also claimed territories. Estimates of maximum territory size ranged from 6.5 to 9.3 ha in 1967 (likely low estimates) and from 9.3 to 19 ha in 1968. Males generally used only a small portion of the total defended area for routine feeding and loafing. These maximum use areas contained from 1.4 to 6.4 ha. When females were present on breeding areas, resident males never allowed other males inside their territories; when hens were absent, such as occurred during severe blizzards in 1968, males defended their territories sporadically and tolerated the presence of neighboring males. Adult females, like adult males, tended to return to the same territories in successive springs and re-pair with their previous mates. Exceptions to this suggested that philopatry to a prior breeding area was more important than mate recognition, and that a female's choice of breeding site may be influenced by nesting success the preceding spring. The period of pair formation and reestablishment lasted from about 11 to 21 May 1967 and from 1 to 31 May 1968 (no unpaired females were observed after 1 June either year). Once paired, males constantly accompanied and defended their mates until incubation began. During incubation, males remained on the territory and accompanied the female only when she left the nest to feed, which normally occurred during deep, evening twilight. Pair bonds dissolved after eggs hatched, when most broods and males abandoned territories for their summer ranges. While some broods stayed on the territory and were occasionally accompanied by the male, most moved to communal brood ranges where they remained throughout the summer (mid-July to mid-September). Brood hens tended to return to the same ranges in successive summers. In Rocky Mountain National Park, brood ranges featured extensive rock areas interspersed with snowaccumulation areas and lush tundra vegetation. Shuffling of chicks among broods, following human disturbance and loss of the brood hen, was observed on communal ranges. It appeared that loose associations of broods on these traditional ranges served to ensure greater survival of lost or orphaned chicks. Why hens adopted the chicks of other hens is unknown.
BEHAVIOR OF WHITE-TAILED PTARMIGAN 299
Males, like broods, gathered in loosely organized flocks of five to 30 birds, and inhabited the same general areas as brood hens, although a few males remained on or near their breeding territories until mid-August. These latter males often accompanied broods and were observed behaving sexually and aggressively toward broody females; however, they did not participate directly in broodrearing or defense. A dominance hierarchy appeared to exist among summering males, with unsuccessful juvenile males seemingly dominating other males in two instances. Broodless hens also associated with these flocks and were sometimes courted.
9
Cyclic Population Changes and Some Related Events in Rock Ptarmigan in Iceland A. Gardarsson
9.1 Introduction The aim of this paper is to describe the main results of a population study of rock ptarmigan (Lagopus mutus) in Iceland, and to offer some observations that may improve our understanding of population cycles. Demographic events during 1963-70 are examined in relation to spacing behavior and food. Population work was initiated by the late Dr. Finnur Gudmundsson in 1963. My field work, centered on spacing and food, was carried out in close cooperation with him during 1965-69. The vertebrate herbivores native to Iceland are all birds, including several anatids and one gallinaceous bird, the rock ptarmigan. Like many other herbivores, Icelandic rock ptarmigan are subject to cyclic population changes, with a periodicity of approximately 10 years (Gudmundsson 1960). In comparison with other cyclic populations that have been studied, this one is unique in that no other cyclic vertebrate herbivores, such as microtines and lagomorphs, occur within its range.
9.2 Study areas Population studies and observations in summer were made on Hrisey, a low island, 767 ha, in the fjord Eyjafjordur, North Iceland (Fig. 9.1). The island is
300
ROCK PTARMIGAN IN ICELAND
301
Fig. 9.1. Map of Iceland showing study sites mentioned: (/) Hrisey, (2) Birningsstadir, (3) Kvisker, (4) Herdisarvik, (5) Heidmork, (6) Armannsfell, (7) Strutur, (8) Hvammur, (9) Fornihvammur, (10) Trollakirkja, ( / / ) Vididalsfjall, (12) Vatnsdalsfjall. Main icecaps outlined with dotted lines. Inset map of Hrisey with 20-m contours, village shaded, outlines of intensive study area.
302
A. GARDARSSON
about 3 km from the nearest mainland coast. Apart from a few old males, the ptarmigan breeding on the island are migratory and probably winter mainly on the nearby mainland, which is characterized by basalt mountains averaging about 1,000 m in height and divided by numerous valleys. Hrisey has a human population of about 300; the people live in a village on the south end of the island, and mainly fish for a living. Farming on the island has been decreasing in recent decades. Because of reduced sheep grazing, the vegetation at Hrisey was changing rapidly during this study. The vegetation of most of the island is an EmpetrumCalluna heath with many hummocks. Depressions in the heath are generally dominated by either Vaccinium spp. or grasses. Meadows are dominated by various grasses and sedges with much Polygonum viviparum. In winter, the Hrisey population of ptarmigan could not be followed and field work was necessarily somewhat opportunistic. Age and sex ratios were obtained mainly from birds shot by hunters in the open season, 15 October-22 December, over wide areas, but mostly in the western half of Iceland. Most direct observations in the winter were made in the area around Fornihvammur, West Iceland. This area of about 250 km2 covers part of the watershed between the west and north coasts of Iceland (Fig. 9.1). Most of it is a poorly drained basalt plateau, about 300 m in altitude. The highest peak in the area, Trollakirkja (alt. 1,001 m), is the easternmost mountain of a range that separates the west coast lowlands of Faxafloi and Breidafjordur. The climate is characterized by cool summers, much snow in the winter, and high precipitation. The vegetation in the Fornihvammur area can be divided into a high alpine zone (above approximately 500 m altitude), a low alpine zone (about 300-500 m), and lowlands (below about 300 m). The high alpine zone has open, snowbed vegetation and extensive boulder fields. In the upper, high alpine zone there are large boulder-strewn areas and vegetation is sparse. The lower, high alpine zone consists of extensive snowbeds and meadows with much Salix herbacea, a preferred food of ptarmigan. The low alpine zone in the Fornihvammur area is largely poorly drained and boggy with large patches of moss heath. Between the bog and the moss heath, there is a belt of meadow with much Salix herbacea. In the lowlands the climax vegetation up to about 250 m is birch scrub (Betula pubescens), of which only remnants are left. The extensive areas cleared of scrub and the lower slopes above the scrub line are mainly covered with mixed heath with much Empetrum.
9.3 Methods Density of ptarmigan was estimated by counting cocks about 25 May each year. At this time territories had become stabilized and the mostly white, territorial cocks stood out conspicuously against the largely snow-free heath. The number
ROCK PTARMIGAN IN ICELAND
303
of known casualties up to the time of census was added to the total of live cocks. On Hrisey the whole island of 161 ha was counted once per year. Observations during 1966-68 on a 100-ha intensive study area suggested that the efficiency of the annual census was close to 100%. Survival of chicks in their first month of life was estimated by flushing a sample of hens on a few days around 25 July, when the chicks were mostly 3 to 4 weeks old. Before this age the chicks of a brood tended to flush individually and often did not flush at all. Later on broods began to amalgamate and hens began to leave the chicks. Losses of young in August-September, before they left the breeding grounds, could be estimated from the numbers found dead in autumn. Rock ptarmigan are shot for the market and for sport during an open season 15 October to 22 December each year. Shot birds from several localities were examined each year and their age determined from the pigmentation of the 9th and 8th primaries (Bergerud et al. 1963, Weeden & Watson 1967). Accuracy of age determination was close to 100%. Segregation of the sexes in winter makes it impracticable to base the age ratio in early winter on young birds per old female (see Petrides 1949); therefore, I used the proportion of young in the total catch as the best estimate of the true age ratio in early winter. Hunters are likely to get relatively more young than old birds because (1) old birds often occur singly or in small groups, whereas large flocks frequently have a relatively high proportion of young birds, and (2) old birds are often wilder than birds of the year. This bias in hunting may not be constant from year to year. Annual changes at each locality, where samples from more than one year were available, showed the same trend as the overall age ratio, though there was considerable regional variation (Gardarsson 1971). Winter losses were estimated from return rates of the two age groups to Hrisey in 1964 through 1970, calculated from estimates of total numbers and age ratios. Age ratios of females were determined from birds taken on their nests. This was presumably an unbiased estimate because nearly all females found on nests were caught. Age ratios of males were from birds found dead in spring and, in 1966-68, individually recognized birds on the intensive study area. For the purpose of this study, return rates were assumed to equal survival rates, but it is obvious that differences in the tendency to disperse would influence return rates in a similar way as variation in survival. Methods of estimating available and ingested food, and chemical analyses are described by Gardarsson and Moss (1970) and Gardarsson (1971). Grazing on deciduous shrubs in winter was estimated by counting cut and intact shoots on a number of branches. Grazing on the male catkins of Betulapubescens was estimated by counting catkins on marked branches in early winter, before grazing began, and again in spring. Other grazing animals, mainly sheep, were a potential source of error in grazing estimates. For this reason I used, as much as possible, areas where domestic animals were excluded.
304
A. GARDARSSON
9.4 Results and discussion 9.4.1 Breeding density The density of ptarmigan on Hrisey varied between 7 and 34 males per km2 (Fig. 9.2). Numbers rose steadily by about 64% (54% to 73%) each year during 1963-66. This was followed by a rapid decline in 1966-68, the population decreasing by about 50% per year, and a slow decline of about 10% annually in 1968-70. During 1963-70 parallel fluctuations were found in three other study areas, where densities were always far lower than on Hrisey. At Birningsstadir in the Northeast, densities were 1-10 males per km2, at Kvisker in the Southeast they were 0.4-2.5, and in Heidmork in the Southwest 0.3-0.6 males per km2. Numbers of hens in spring were assumed equal to those of cocks. Direct counts of hens were not possible because of their secretive behavior. Instead, numbers were estimated from the ratio of marked (dyed primaries) to unmarked hens with broods in the peak year and 2 years of decline. This ratio was not significantly different from the number estimated on the basis of an equal sex ratio (Table 9.1). This result contrasts markedly with southern montane populations of Lagopus. Watson (1965) found a large excess of males in declining populations of Scottish rock ptarmigan, and Soikiro et al. (1969) reported an apparent 2:1 sex ratio in Japanese rock ptarmigan, where excess males held territories. In stable populations of white-tailed ptarmigan (Lagopus leucurus) an excess of males in spring was associated with a situation in which yearling males never held territories and 2-year-old males seldom did, whereas most yearling females laid eggs (Choate 1963a, Braun 1969). No doubt the balanced sex ratio in spring in Icelandic ptarmigan, and perhaps in other northern populations, contributes to their high natality rate and population turnover. The excess of males in the southern montane ptarmigans may be the result of an unbalanced predator situation, in particular the absence of gyrfalcons (Falco rusticolus).
Table 9.1. Number of breeding rock ptarmigan hens on Hrisey, Iceland Year
1966 1967 1968 a
Marked on nest
77 50 30
Sighted with broods, marked/total
Calculated (95% C.L.)
Expected3
46/138 33/91 19/37
231 (177-285) 138 (100-176) 58 (40-76)
248 124 65
May census of cocks minus known losses of hens.
No. of hens with broods
ROCK PTARMIGAN IN ICELAND
305
Fig. 9.2. Main demographic features of rock ptarmigan at Hrisey in 1963-70: density (males/km 2 ), % change in spring numbers between years, clutch size and brood size with 95% confidence limits, and age ratios (% old birds) in spring.
306
A. GARDARSSON
9.4.2 Adult mortality in summer Summer losses of adults (Fig .9.3) were highest in 1967 (38 %) and lowest in 1964 (6%), but in most years they averaged about 23%. The most important cause of mortality was predation by gyrfalcons which mainly took territorial cocks in spring and juveniles in autumn. During incubation and until chicks were about 4 weeks old (i.e., late July), the ptarmigan were extremely secretive and falcons took mainly other birds. The falcons that preyed on the Hrisey ptarmigan nested on the nearby mainland, at least 5 km away. In spring they visited the island at least daily, and possibly more than one aerie was supplied by Hrisey ptarmigan. In autumn, gyrfalcons often roosted on the island. Up to four falcons were seen at one time on Hrisey in August 1967 when predation was unusually high. The predator situation on Hrisey is not typical of Iceland as a whole, because there are no arctic foxes (Alopex lagopus) on the island. Foxes prey extensively on ptarmigan on the mainland and may prey selectively on incubating and brooding hens. On Hrisey the proportion of adult males in autumn was significantly lower than in the hunting season bag on the mainland (Table 9.2), and young and old birds in the bag had almost identical sex ratios. Predation on ptarmigan nests was practically nonexistent on the island, probably because nest predators, mainly ravens (Corvus corax), had available to them an abundance of more-accessible seabird nests, especially those of arctic terns (Sternaparadisaed), perhaps ten thousand nests, and common eiders (Somateria mollissima), about one thousand nests.
9.4.3 Production Mean clutch size varied little (10.4 to 11.7) during the study period, as did the proportion of young hatched (Fig. 9.2). Clutch size was significantly and inversely correlated with density (r = - 0.82, P < 0.05), but not with spring temperature (r = 0.06, N.S.), and by inference not with the state of the vegetation in spring (Gardarsson 1971). Table 9.2. Sex ratios of ptarmigan in autumn 1964 through 1969 %
(95% C.L.)
n
Open-season bag Old birds Young birds
47.0 47.7
(45.6-48.4) (46.9-48.5)
5,009 14,943
Hrisey autumn (calculated) Old birds
41.4
(38.6-44.2)
1,186
Ptarmigan
ROCK PTARMIGAN IN ICELAND
307
Fig. 9.3. Losses of rock ptarmigan each year, 1963-70, based on data from Hrisey and the hunting season, shown as k-values (Varley & Gradwell 1970).
308
A. GARDARSSON
The mean brood size (Fig. 9.2) was about nine in all years except two. In 1966 there was a significant drop in the brood size to 4.3 (X2 = 26.75, P < 0.001). This was caused by weather, which killed ptarmigan chicks over a large area in North Iceland during 23 to 24 July. Before the storm the mean brood size was almost ten (128 chicks/13 hens); thus, it seems unlikely that the chicks were any less viable in 1966 than in other years. In 1967, the first year of decrease, the mean brood was only 7.3, but this was not significantly lower than in the normal (i.e., excluding 1966) years combined (X2 = 1.42, P > 0.10). Losses of young in the period August-September were 5-14% except in 1967 when losses rose significantly to 36% (X2 = 407.9, P < 0.001). However, losses of young and adults on Hrisey were correlated (Fig. 9.4), both total summer losses (r - 0.85, P < 0.01) and autumn losses only (r = 0.93, P < 0.001), suggesting that a common environmental factor, rather than changes in chick viability, determined summer losses. The sex ratio of collected chicks 4-20 days old was 56% males (n = 91), not significantly different from an equal ratio (X2 = 1.10, P > 0.10). In a sample of 42 chicks about 20-25 days old, which died from exposure on Hrisey on 23-24 July 1966, 67% were males, significantly different from an equal ratio (X2 = 4.03, P < 0.05). Thus male chicks appeared more vulnerable to weather than did females. Different survival of the sexes in young birds is seldom reported. There are indications of such a difference in pheasant (Phasianus colchicus) where male and female chicks, respectively, are less tolerant to exposure (Lathan 1947). By October the proportion of males among fully grown juveniles in the bag was about 48% (i.e., similar to the adults; Table 9.2), but this may be a biased estimate because of the segregation of the sexes in winter (see section 9.4.9). At face value, the above proportions suggest that a sex ratio at hatching in favor of males changes during the chick stage, the males having a higher mortality rate than the females, and that this leads to an approximately equal sex ratio among full-grown juveniles.
9.4.4 Age ratios in the hunting season Age ratios in the combined yearly samples decreased significantly during the study period (Table 9.3). In 1964 and 1965 about 81 % of the catch consisted of juveniles, in 1966 about 76%, in 1967 and 1968 about 72%, and in 1969 only about 59%. The decreasing proportion of juveniles in the catch can mainly be attributed to increased mortality of young in August-October, since young-to-hen ratios in summer did not show a oneway trend (Fig. 9.2).
ROCK PTARMIGAN IN ICELAND
309
Fig. 9.4. Association between summer and autumn losses of adult and young ptarmigan on Hrisey.
310
A. GARDARSSON
Table 9.3. Proportion of young in the open seasons (15 October to 22 December), 1964 through 1969 Year
Young/Old
% young
(95% C.L.)
n
1964 1965 1966 1967 1968 1969
4.33 4.25 3.19 2.57 2.57 1.44
81 81 76 72 72 59
(79.6-82.8) (79.6-82.2) (74.9-77.3) (70.8-73.2) (70.3-73.7) (56.2-61.8)
2,338 3,808 5,049 5,046 2,799 1,226
Total
3.00
75
(74.4-75.6)
20,266
9.4.5 Winter losses Mortality in the winter, when birds could not be followed, accounted for the greatest part of the annual losses and determined the size of the breeding stock in the following spring. Although the apparent annual survival of old males was only 66% that of the females, the overwinter survival of the sexes was practically the same and averaged about 60% (Fig. 9.5). The difference was accounted for solely by the somewhat unusual predator situation on Hrisey. Low return rates of old birds (35-49%) occurred in 1964-65, during the increase, and during the crash winters 1966-67 and 1967-68. The return rate of yearlings was about 20% in the years of increase, but dropped during the crash to about 3% to 6% and stayed at this level during the low. Although decreased juvenile survival was without doubt the major factor contributing to the decline and subsequent low numbers, adult losses during 1966 through 1968 contributed significantly to the steep decline.
9.4.6 The territorial stage Many old males probably remain close to breeding grounds throughout winter. According to local information, scattered ptarmigan occur on Hrisey in most winters. In January 1968 I found a single male and a group of five apparently old males on the island, including one color-marked, local bird. On Hrisey, most males arrived in April 1967. In the first days of April only six or seven males were found in the intensive study area (Fig. 9.6). Two of these, both old males, were known individually. Both had their centers of activity in their territories of the previous year, but foraged widely (up to 1 km away) as did the others. Almost all other territory holders arrived simultaneously about 6 April, in warm, sunny weather with rapid snowmelt. The males established their home ranges immediately. Until about 23 April, boundary encounters occurred
ROCK PTARMIGAN IN ICELAND
311
Fig. 9.5. Apparent survival of male and female rock ptarmigan on Hrisey, based on birds found dead in summer and return rates in winter.
sporadically, the home ranges overlapped greatly (Fig. 9.6), and the males formed flocks during cold spells with snowfall. Warm, sunny days in April appeared to trigger territorial display in the males in 1966, but this was seldom seen in 1967. In mid-April 1966, when probably all the males had arrived, many territorial encounters were observed in the mornings. At the same time, other males, apparently mainly new arrivals, were in flocks elsewhere on the island. These flocks were joined during the afternoon by territory holders, probably because of better feeding in the localities occupied by flocks. During the establishment phase, males roosted either singly on their home ranges or in small groups. Territorial spacing was determined very early, and subsequent changes were insignificant. Latecomers generally did not establish themselves as territory owners.
312
A. GARDARSSON
Fig. 9.6. Territories of ptarmigan on a 100-ha study area at Hrisey in 1966-68.
ROCK PTARMIGAN IN ICELAND
313
Females arrived later and over a longer period than males. Most females arrived during the last days of April and in the first week of May, i.e., about 3-4 weeks later than most males, but some females may have arrived later. Several females were seen on 14 April 1966; the first females in 1967 were seen on 23 April. In late April all territorial males had established themselves. Boundaries remained more or less unchanged until nesting began (Fig. 9.6, early May). Each male held a piece of ground varying from a little over 1 ha (1966) to about 30 ha (1968). Maintenance activities, including roosting and feeding, were confined to the territories. Boundaries were rigorously observed except: (1) when chasing females, an activity especially noticeable in mid-May; (2) in response to difficult feeding conditions (fresh snow), when some flocking on favorable feeding grounds could still be observed; and (3) when avoiding gyrfalcons. Males escaping from gyrfalcons were sometimes displaced up to at least 2 km from their territories. Sometimes males moved gradually out of sight from a gyrfalcon perched nearby, taking cover in small groups, e.g., along a fence. This escape behavior was noticed throughout the territorial period. Cocks thus displaced temporarily behaved quite like submissive, nonterritorial cocks (see section 9.4.7). Cocks on their home territories in such situations also showed little aggression, but were quick to expel intruders when the falcon had gone. Before laying, females were apparently spaced out in widely overlapping home ranges, which in certain cases coincided roughly with territories of particular males. The typical situation, however, was probably that exemplified by female 29 in 1967 (Fig. 9.6), the home range of which partly overlapped territories of four, and later three, males. Observations in 1960, when density was very high, showed that territorial males typically associated with several different females during early May. Movements of these females could not be followed precisely because the methods of identifying them (color bands, plumage peculiarities) only sufficed to show whether or not the same female was with a particular male from day to day. Movements of females between male territories often resulted in encounters between males. In the middle of May, prolonged flights by females sexually chased by one or more males and passing a succession of displaying territorial males were frequently observed. The mechanism by which the females spaced themselves out appeared to be primarily inter-female aggression. However, females were difficult to observe and the data are fragmentary. Aggressive runs involving two females were observed a few times in May and once during the incubation period when two inattentive females met on a common feeding ground. Males were usually involved in female encounters. Probably most males had the opportunity to copulate with more than one female. Two and sometimes three females often associated at the same time with a male up to the time of nesting. This together with the evidence for independent
314
A. GARDARSSON
home ranges of males and females suggests that the mating system tends to the promiscuous and is often polygynous. Such a mating system of "open monogamy" with much opportunity for promiscuity would tend to maximize the reproductive output and would be advantageous to birds under heavy predation pressure. Probably the semipromiscuous system is a feature of relatively dense populations; in low-density areas this system may be indistinguishable from monogamy. During late May to early June while the females were laying, and therefore restricted to the vicinity of their nests, aerial chases became rare. Territory boundaries had by now become stabilized, but as in early May, slight day-to-day changes occurred. By this time the territories of individual males had often shifted, mainly as a result of removal of cocks by gyrfalcons (Table 9.4), but smaller changes were also observed (contrast early and late May 1966 and 1967 in Fig. 9.6) that were often related to the attachment of a male to a particular female. Thus a female spending much time off the nest with a male, but nesting just outside the border of his territory, often caused the male to take over that part of a neighboring territory, e.g., male 14 and male D in 1966 (Fig. 9.6). In one instance (male 12 in 1966) a physical change in the landscape caused a territory to shift. In early May this territory extended across a pond that was covered by ice and snow. After the thaw the male evidently had difficulties in defending all sides of the pond, and its territory shifted to the eastern bank where there was a nesting female. Table 9.4. Number of territorial males and fates of disappearing birds on a 100-ha study area on Hrisey (estimated number of yearlings in parentheses), 1966-68 Territorial males
1966
1967
1968
Estimated numbers in early May:
33 (24)
22 (9)
6 (2?)
Numbers of individually recognized males early May late June
19 (14) 14 (9)
21 (9) 10(4)
4 (1?) 4(1?)
Estimated numbers surviving until 1 July
24 (15)
11 (4)
6 (2?)
7 1 3
0 0 0
Fates of recognized males that disappeared Gyrfalcon Accident Unknown
2 1 2
ROCK PTARMIGAN IN ICELAND
315
At the onset of incubation (about 1 June) the males became secretive and they resumed their molt, which had been delayed since early May. At this time gyrfalcon predation became negligible, apparently because the falcons had difficulties finding ptarmigan and switched to other birds, such as shorebird chicks. Predation on ptarmigan began again in late July when the ptarmigan chicks were beginning to fly and were conspicuous. Territorial border encounters between males became much rarer during incubation, most observed encounters involving inattentive females that crossed boundaries on feeding trips. Females typically nested in good cover between hummocks on the dwarfshrub heath and foraged in food-rich habitats (heath depressions rich in Vaccinium or Salix herbacea, meadows with much Salix). They made short feeding forays (8-20 minutes long) about 4-6 times each 24 hours, usually to the nearest food patch. Dispersion of the nests (Fig. 9.6) was associated with the distribution of dense growth of the preferred food species. Feeding females were usually joined by a territorial male that showed watchful behavior and frequently displayed to the females. Often, however, females fed alone and were apparently not noticed by a male. The mean distance of a feeding trip from the nest was 131 m with a standard error of 14 m (n = 18); the longest distance recorded was about 270m. About the time of the hatch (late June to early July) the remaining territorial males often became more aggressive for a brief period (about 2 days), both toward other ptarmigan and toward whimbrels (Numenius phaeopus), which nested commonly (15-20 pairs) in the study area. It was not clear whether the numerous, interspecific encounters observed were started by the ptarmigan cocks or the whimbrels, which also had young chicks at this time. The males usually remained on or close to their territories for about 3 weeks after the chicks had hatched. Territorial defense was no longer observed, and two or three cocks from neighboring territories were often seen together on good feeding grounds. The males did not normally associate with females tending young chicks and only exceptionally showed distraction behavior when a brood was disturbed. Males often joined the creches in late July. In August most old cocks had left their territories and associated with the large flocks of young and females. Scattered males, single or in groups of up to five, were, however, observed until late August. Females with young chicks traveled widely and did not normally remain within the boundaries of territories. Broods were found up to 500 m from their nests in the first week and up to 1 km after 2 weeks. Coalescing of broods mainly took place about 3 weeks after hatching, but very young chicks were frequently seen mixed with other broods. One female that failed to hatch her clutch was seen with two young chicks from different broods on the day after she had deserted the nest. No aggression was observed between females tending broods.
316
A. GARDARSSON
9.4.7 Territorial behavior and changes in numbers Yearly numerical changes of ptarmigan on the intensive study area (Table 9.4) followed closely the cyclic changes on Hrisey as a whole. The mean initial size of territories (Table 9.5) increased from 2.7 ha in 1966 to 17.7 ha in 1968. Most territories increased in size during the territorial period, by about 60% on the average in 1966 and 117% in 1967. This resulted primarily from removal of territorial cocks by gyrfalcons and the subsequent enlargement of neighboring territories to fill the vacancies (Fig. 9.6). In 1968 no territorial male was killed in May or June on the study area and no territorial enlargements were observed. Vacant territories were almost always invaded by neighbors and not by "surplus" males, which were present in all 3 years. Thus, territoriality in this population was mainly a spacing mechanism of the males and did not determine numbers. Territorial cocks frequently invaded neighboring territories but were usually expelled immediately by the owners. Sometimes, however, such invasions lasted for longer periods, apparently because the invaders were not noticed and therefore not challenged. Because the removal rate of territorial males was high, such intrusions may enable territorial males to take over vacant territories quickly, and this was in fact observed. Males that did not hold regular territories were present in the study area every year and probably made up about 10% of the total. These were grouped according to status as: (1) submissive nonterritorial (several observations) (2) unsuccessful territorial (two instances) (3) aggressive nonterritorial (two instances) In May 1966 single submissive males were seen on ten occasions in the southcentral part of the study area; probably two or three such males were resident in
Table 9.5. Territory size (ha) for 3 years on a study area on Hrisey, 1966-68
1966
1967
1968
Initial size (early May) Mean + SE Range n
2.7 ± 0.26 1.3 - 5.1 (15)
4.8 ± 0.65 2.3 - 12.6 (14)
17.7 9.4 - 25.1 (4)
Final size (late June) Mean + SE Range n
4.3 ± 1.09 1.8 - 13.0 (10)
11.0 3.9 - 19.2 (4)
No change (4)
ROCK PTARMIGAN IN ICELAND
317
the area during this period. Circumstantial evidence, such as repeated observations of a submissive bird in the same spot, suggested that they were nonterritorial, and not birds that had been temporarily displaced by gyrfalcons. Sometime between 21 and 27 May 1966, male 7, which held a territory near the middle of the area frequented by the submissive males, was killed by a falcon. On 27 May the neighbor, male 4, was seen on this territory, but on the following morning a juvenile male (F) had taken the territory. Presumably this was one of the submissive males seen earlier in the area. Male F was seen repeatedly on this territory until 19 June. Its behavior never became as aggressive as that of typical territorial cocks. It was once seen defending its boundary against the neighboring 4, but males D and 5, which held territories on either side of F, often trespassed on its territory without any defense by F. A nest was found on the former territory of 7, but apparently male D had taken over the nest site in June (Fig. 9.6). The poorly expressed territorial behavior could have resulted from the absence of any nesting females on the territory (an intensive search failed to reveal a nest), or perhaps this male was somehow handicapped and unable to perform normal territorial activities. Male F molted almost completely into summer plumage by mid-June, about a fortnight earlier than other males in the study area. A submissive, nonterritorial male was seen on territory 13 on 25-26 April 1967. During 22-28 May up to three submissive males were seen in the same general area, mainly on the border of territories 26 and L. No other submissive males without territories were seen in 1967 on the study area, though such birds were present in May in other parts of the island. On 29 May 1967 an unmarked male attempted to defend a strip of ground some 150 x 20-30 m on the borders of four or five territories immediately south of the study area. It was repeatedly chased by three neighboring cocks, showed only weakly developed aggressive behavior, and was not seen again. In 1968 no submissive nonterritorial males were identified. Two nonterritorial, aggressive males were found. On 5 May 1967 a yearling male was caught in a mist-net on territory 16 while engaged in a fight with the owner. This male (banded) was taken again later the same day on territory 1, where it was crouching submissively. It was not seen again until 2 June when it attempted to defend a territory in the southeast corner of the study area, some 400 m southeast of the place where it was first caught. This attempt was apparently unsuccessful but the same male was again seen several times during early June, mostly in aggressive encounters, in an area of about 1 km 2 . On 25 May 1968 another aberrant yearling male (banded) was taken with snares. This male had lost most of the feathers on its throat, probably in fights, its belching was abnormally high pitched, and unlike any other male, it performed a sexual display to the mounted decoy cock as if the latter were a female. This male was seen several times until 15 June and ranged over 1 to 2 km2, including most of the study area, and was continually chased off by territorial cocks.
318
A. GARDARSSON
Both aggressive, nonterritorial males had home ranges centered close to the southeast corner of the study area. This part of the area was presumably unsuitable for nesting females (Fig. 9.6), and their absence, or the low density of preferred foods, may have reduced the chances of males establishing and retaining normal territories. However, both males may have been displaced to this lowquality area by territorial males. Many males that failed to establish normal territories may simply have arrived too late to do so. This seemed true for new males that appeared in the village on Hrisey as late as early May and set up transient territories and then disappeared. But in many cases the nonterritorial birds seemed physically abnormal, though the reasons for this were not determined. No nonbreeding females were observed, suggesting that "surplus" females were also rare, though such females may be difficult to detect.
9.4.8 An outline of behavior in flocks Flocking begins at the time of hatching, about 1 July, when territories break up, and lasts until April-May, when territorial spacing becomes prevalent again. During the first 3 weeks chicks are usually in discrete broods with their mothers, which spend much time brooding. Most maintenance and agonistic behavior patterns observed in full-grown ptarmigan appear during this period. Mother and chicks communicate largely by vocalizations. When not being brooded, small chicks spend most of their time feeding. In fair weather they come and go almost continuously, in wet and cold weather the whole brood usually feeds at once. Hens and broods are almost continuously on the move and tend to select patches that are rich in preferred foods, especially spikes of Polygonum viviparum and insects. Small chicks escape danger by crouching and creeping away into shelter. They begin to fly when about 10 days old. During the first month flying is used mainly for escape and the broods forage entirely on foot. Broods begin to aggregate when about 2 weeks old and often form flocks, or creches, of 2-4 broods. Flock formation is associated with the phenology and distribution of the food supply, mainly Polygonum growing densely in grassy habitats that may be relatively inaccessible to younger chicks. After the first 3 to 4 weeks chicks are no longer brooded regularly and they then roost in dense groups close to the hen. As the chicks grow bigger and more conspicuous, predation by gyrfalcons becomes important. The broods now tend to form relatively large flocks together with adults of both sexes and spend most of the day in good cover. On Hrisey many birds concentrated in an abandoned hay field with an abundance of preferred foods; elsewhere broods gather in lava fields and scrubs. After the age of about 5 weeks the chicks, now half-grown, roost in loose aggregates as do adults. At
ROCK PTARMIGAN IN ICELAND
319
Table 9.6. Age composition of rock ptarmigan in coastal mountains of North and West Iceland in autumn 1965-67 Period
% of juveniles
(«)
22-31 Aug 1-10 Sep 11-20 Sep 2 1-30 Sep 1-10 Oct
0 59 75 90 84
(11) (22) (12) (10) (70)
this stage the ptarmigan make regular movements, either on foot or flying, to roost on open heaths at dusk, returning to daytime cover in the early morning. Daily movements allow the birds to utilize foods at dusk in places where they would otherwise be exposed to gyrfalcon attacks. Roost movements, coincide with peaks in social activity and dust-bathing. Autumn flocks, numbering up to several hundred birds in a high year, remain on the breeding grounds until early September. In most parts of Iceland vertical migration to high alpine habitats takes place in autumn. Adults are the first to reach the high alpine, in late August, followed by the first juveniles after about 10-15 days (Table 9.6). In the volcanic zone, both in northeast and southwest Iceland, many ptarmigan remain until early winter in lava fields, which are often rich in food and provide excellent cover, whereas the mountains in these areas are usually poor in both. Other rocky habitats, such as stream gullies, may be used in autumn as daytime cover. Some ptarmigan, especially in parts of northern and northwestern Iceland, remain in lowland birch scrubs, where Vacdnium myrtillus, a preferred food, is important in the undergrowth. In certain cases, as on Hrisey, ptarmigan concentrate in villages and other built-up areas, where houses and other structures provide the cover and weeds are the main food. At this season, in September-October, the ptarmigan spend only the day in cover and make regular movements to roost in more open habitats. Movements of over 10 km have been recorded, and the birds probably make even longer movements to and from roosts. Long-distance migration also takes place during this season, mainly, it is presumed, at dusk or at night. Ptarmigan spend most of the day in good cover and feed at low rates until in the late afternoon when they move to good pastures, such as alpine meadows or snowbeds with a good growth of Salix herbacea, often close to boulder fields. At dusk or in total darkness they fly downhill to their roosting areas where additional foods, such as berries and Polygonum bulbils, become available. Ptarmigan that spend the day in lava fields or other lowland cover also make similar movements to roost on adjacent, open ground. Feeding is resumed at daybreak in the roosting area and the ptarmigan then fly back toward their daytime cover.
320
A. GARDARSSON
Snow falls frequently in the mountains in September and October, the high alpine generally becoming permanently covered in late October. At this time ptarmigan are molting into their white, winter plumage. Molting commences about 1 September and is usually over by 1 November. Energy and nutrient requirements must be relatively high during this season because of the simultaneous demands of molt and migration; at the same time, food is abundant compared with the rest of the winter. Before the ground becomes covered with snow, ptarmigan are vulnerable to gyrfalcon predation during the day. Flocks vary greatly in size, but undisturbed flocks in winter usually number from 5 to 30 individuals. Large temporary aggregations, to over one thousand birds, often form in early winter, always, apparently, in response to disturbance by predators such as ravens and gyrfalcons, which constantly follow ptarmigan flocks. When they are in good cover or well camouflaged by the background, the ptarmigan crouch or simply freeze in response to aerial predators. When surprised in the open where they do not match the background, or when in large conspicuous flocks, they fly to the nearest cover or to the tops of rocks or other tall objects. At such times scattered flocks will band together, but the aggregations break up soon after the danger has passed. Although snow conditions in early winter vary, the best Salix herbacea meadows are often covered with only thin snow and are therefore available to ptarmigan. As soon as snow falls, the ptarmigan decrease their diurnal movements and preferably roost in loose snow. They are no longer confined to boulder screes or other places where cover is good, because they are now mostly white and match the background of the snow cover, which is often thin and patchy. At the same time, the birds become more dispersed because a much greater area is available to them. In dense fog, white ptarmigan feed on open, snow-free sites during the day, again suggesting that aerial predators determine their diurnal movements in early winter. As snow accumulates at high altitudes and begins to cover lowland areas persistently, more ptarmigan move downhill and feed on exposed heaths or in scrublands. Winter weather and snow cover in much of the lowlands is variable. Ptarmigan take advantage of this by moving up to the hills during thaws to feed on Salix herbacea whenever snow conditions permit; uphill movements may also occur in spring before they move to breeding grounds. When feeding in winter, ptarmigan vary their foraging behavior with the amount and type of snow. When there is no snow, birds walk or creep along slowly, stopping for short periods in rich patches. When snow cover is thin, some scraping is usually necessary, but the birds also may select exposed patches where they do not need to scrape the snow. In soft snow, up to about 30 cm deep, and in thin, crusty snow the birds feed by scraping the snow away, mainly with their feet. Where the snow overlies a dense patch of preferred food the birds often feed for long periods in the same spot, but more often they move rapidly between sue-
ROCK PTARMIGAN IN ICELAND
321
cessive food patches. In thick snow the ptarmigan often feed on the tops or exposed sides of hummocks or on exposed ridges and slopes. In scrub they prefer to feed from the ground but may also feed by jumping up to get at low branches or by perching in the bushes, especially in late winter when preferred foods have become scarce. Diurnal activity is much affected by day length. In midwinter when the light period is only 6 hours, ptarmigan continue to feed at an increasing rate through the day. At other times they rest during midday —in the autumn in the daytime cover, later on in snow roosts close to the feeding patches. Daytime roosts are almost always in places offering a good view, e.g., high up, on steep slopes or snowdrifts. Such localities are seldom used for nighttime roosting, the birds then preferring fairly level ground, usually in the open and away from bushes. This difference in roosting appears to be an adaptation to predators, gyrfalcons during the day and arctic foxes at night. Times of resting during the day and also in the morning and evening are characterized by an upsurge in social interaction as well as snow bathing. Agonistic encounters and displays in flocks and broods differ little from such activities seen on territories, apart from a lack of permanence in space. They also have many similarities with true leks. The most commonly observed displays in flocks are wing-stretching, wing-flapping and flutter jumps; they often occur in an infective fashion when the birds become closely spaced, thus serving to increase individual distance. The most frequent aggressive encounters in flocks are, however, relatively unobtrusive displacements occurring both when the birds are feeding and when they are resting or snow bathing. Such encounters usually involve no more than approaches and avoidances, sometimes threat and, in rare instances, pecking. The winner of an encounter may gain a feeding patch or a place to rest. Such encounters are likely to lead to differential survival of aggressive and submissive birds. In adverse snow conditions submissive birds often initiate scrapes and are then chased off by more aggressive ones, which proceed to feed in the scraped patch. This process would lead to submissive birds expending more energy than dominant birds. In the autumn flocks, adults are clearly dominant over juveniles, and males seem generally more aggressive than females in winter flocks. In winter, adults tend to be in relatively small flocks, whereas large flocks sometimes comprise almost entirely juveniles, again suggesting that adults dominate juveniles.
9.4.9 Sex segregation in winter Males predominated in the kill in most of the northern half of Iceland (Fig. 9.7), reaching 80% and even more than 90% on some heaths in the North and Northeast. Females outnumbered males in the South and much of the West. The highest proportion of females (over 70%) was found on Mt. Strutur in West Iceland.
322
A. GARDARSSON
Fig. 9.7. Distribution of sexes in early winter, shown as proportion of males at each sampling locality. The overall pattern is indicated by means of shading: oblique cross-hatching over 70% males, oblique hatching 60-70%, unshaded 40-60%, vertical hatching 40% or less.
The most extensive and most densely populated breeding grounds of ptarmigan are in the Northeast (Gudmundsson 1960), where males predominated, hence the difference in the proportion of the sexes was caused by differential migration. Most females left the northern part of the country and migrated south, concentrating on the western, southwestern, and probably also the southeastern borders of the interior plateau. But migrating females from the North apparently did not reach coastal western and southwestern areas where the sex ratio was approximately equal. Banding returns (Gudmundsson 1960) have shown that ptarmigan originating in the Northeast migrate at least as far west as Hunavatnssysla on the north coast, and also to the southwestern and eastern coasts. There have also been two recoveries of ptarmigan banded in the Southeast and shot in the Southwest. Recoveries of birds banded on Hrisey were all in northern coastal areas within a radius of about 50 km from the island. In western Iceland, which was relatively well covered, males predominated in the northernmost localities, females in interior hills south and west of the watershed, and approximately equal ratios were found in southern coastal areas (Fig. 9.7). The Skagafjordur area west to and including Vatnsdalsfjall in
ROCK PTARMIGAN IN ICELAND
323
Hunavatnssysla was an exception to the general predominance of males in the North. Perhaps this area had low breeding densities and received females mainly from high-density areas farther east. There was apparently very little emigration from the cutoff Northwest Peninsula. The gradient between male and female areas was often steep. Thus there was a significant difference in the proportion of males on Vatnsdalsfjall (47%) and Vididalsfjall (61%), two mountains separated by a valley 10-15 km wide. Two samples from a plateau about 20 to 25 km south of these mountains contained 73 % and 84% males respectively. On mountains bordering the same plateau to the west and south, females predominated, reaching 72% at Strutur some 60 km south of Vididalsfjall. Sex ratios of adults and juveniles at any one locality were generally similar. Adults were often few and were usually included with the juveniles when analyzing the open-season sex ratio. However, important differences existed between the migration programs of adults and juveniles of each sex (Gardarsson 1971). Adult males were the least-migratory component and possibly stayed mainly within the general area of their breeding grounds. Juvenile males clearly went farther away than adult males, but not as far as females. Adult females appeared more migratory than juveniles. Most adult females left the northeastern areas, and the females found there were predominantly juveniles. Braun (1969) reported a similar situation for white-tailed ptarmigan in Colorado. Movements of ptarmigan in winter, as reflected in regional distribution, are best explained in terms of trade-offs between nutritional and social pressures. Because ownership seems to be decided immediately upon arrival on the territories in spring, there should be strong selection for males to remain as close as possible to their territories through the winter. Old males seem to do this, often at the expense of food quality. Juvenile males were more migratory than old males but less so than most females, suggesting that they compromised between proximity to the breeding area and quality of the wintering grounds. The winter distribution of females would tend to maximize food quality, because the areas where they concentrated are generally the richest in the preferred Salix herbacea. Females are presumably free to roam farther from the breeding grounds than are the males, because females' social spacing in spring is less rigid and because synchronous arrival may be less important to the females, which presumably became impregnated and lay regardless of social status.
9.4.10 Age-group segregation Age composition of ptarmigan flocks often differed markedly from the expected, or regional, composition, partly because adults tended to move ahead of juveniles. The age ratio on Hrisey began to increase in late August, suggesting that adult
324
A. GARDARSSON
females were leaving before the young. By late September the remaining flocks consisted almost exclusively of juveniles. Adults reached high-alpine habitats slightly ahead of the first juveniles (Table 9.6). The first young were seen in the high alpine on 7 September. By the middle of September the age ratios in the high alpine had become similar to those in midOctober, at the beginning of the open season. In the Fornihvammur study area, peak concentrations in the high alpine were reached in early October, provided the hills remained free of snow. Age ratios in the Fornihvammur area varied among altitudes during periods when ptarmigan were making movements in response to changing snow conditions (Table 9.7). In early winter the birds were mainly moving downhill as snow gradually made feeding impossible in the mountains. The first birds to move to the lower altitudes were adults and some juveniles, but many juveniles lagged behind in the high hills, often in larger flocks than farther down. If upward movement occurred again, the birds that lagged behind were also almost all juveniles. Table 9.7. Age ratios (% juveniles) of rock ptarmigan in winter in three altitude zones in the Fornihvammur area, West Iceland (arrows indicate likely direction of movements between zones, if any) Altitudinal zone High (> 400 m) % juv. («)
Mid (200-400 m) % juv. (n)
Low (< 200 m) % juv. (n)
1965-66 Sep-Oct Dec
82 (827) 94 (307)
91 ( 1 1 ) 62 (21) a
-
1966-67 Oct Nov-Dec Jan
78 (1076) 91 (281) -
77 (31)a 50 (10)
76(51) 94 (18) b
1967-68 Oct Nov-Dec Feb
82 (1040) -
61 (23)a 77(46) -
65 (158) 76 (86)
1968-69 Oct Feb
72 (189) -
76 (42) -
74 (55)
Date
-
a For difference between high and mid altitudes of all samples during downward movement, X 2 ± 16.98, P < 0.001. b For difference between altitudes in January 1967 (upward movement) P = 0.00457, Fisher exact probability test. Other differences not significant.
ROCK PTARMIGAN IN ICELAND
325
As a result of the differences in altitudinal movements, most adults probably enjoyed slightly better feeding conditions than most juveniles throughout the winter. This may have contributed to the different survival rates of adults and juveniles. 9.4.11 Food Ptarmigan feed mainly on willows (Salix spp.) and birches (Betula spp.) in winter, and on a variety of leaves and reproductive parts in summer (Fig. 9.8). In most of Iceland ptarmigan feed largely on the dwarf willow Salix herbacea in early winter. This species is especially abundant at high altitudes where the ptarmigan stay as long as conditions permit. Small quantities of evergreen leafy shoots, bulbils of Polygonum viviparum, and berries are also taken. All earlywinter foods are rich in energy, and S. herbacea is moreover the best source of protein available in winter. Grazing pressure on S. herbacea in alpine meadows near Fornihvammur, West Iceland, was studied in four winters (Fig. 9.9). Grazing on this species was affected by both ptarmigan population density and snow cover. In the mildest winter, 1965-66, about 25% of the S. herbacea shoots were eaten by ptarmigan; in winter 1966-67, about 10-15%; and in the two following winters, less than 10%. Grazing on shoots that were exposed in late winter was about twice as heavy as that on total Salix. These estimates suggest that grazing by ptarmigan may seriously deplete available S. herbacea, at least in some peak years.
Fig. 9.8. Generalized annual diet of Icelandic rock ptarmigan (% of aggregate volume), based on a total of about 1,200 crops. Winter data from Fornihvammur and Hvammur, West Iceland; summer data mainly from Northeast Iceland. In some areas Dryas, various dwarf shrubs, or willows are the main food in midwinter instead of Betula.
326
A. GARDARSSON
Fig. 9.9. Grazing by ptarmigan on Salix herbacea in winter on three study plots in hills near Fornihvammur, West Iceland, shown as proportion of number of shoots removed. Unfilled symbols indicate proportion grazed of total Salix, as measured in spring, after snowmelt. Filled symbols show proportion grazed of Salix that was available (exposed) in late winter.
As snow cover increases in alpine habitats in winter, most ptarmigan move to lower altitudes where there is considerable local variation in available foods. The most important food species in midwinter are shrubs, especially birch (Betula pubescens), dwarf birch (B. nana), and willows, mainly Salix phylicifolia, as well as Dryas octopetala and other prostrate dwarf shrubs on exposed sites. The food situation in mid and late winter is complicated, but food availability is largely determined by snow cover, though at times it is restricted by a combination of snow and the birds' grazing. When feeding on birch the ptarmigan preferred male catkins to woody shoots and buds, but during long spells of thick or crusty snow the birds exhausted the catkins and turned to feeding on shoots and buds. Birch catkins were heavily grazed (67-88 % removed) in all winters at the main study area, Hvammur in West Iceland (Fig. 9.10). Birch shoots were taken increasingly as catkins became exhausted; the heaviest utilization of birch shoots,
ROCK PTARMIGAN IN ICELAND
327
54%, was recorded in 1967-68. Heavy grazing of birch shoots was found over a wide area of northern Iceland in winter 1967-68 (Gardarsson 1971) and must occur frequently. In the lowlands of Southwest Iceland, where snows are typically much lighter, utilization of birch shoots was quite insignificant, though birch catkins (Fig. 9.10) and Salixphylicifolia shoots (Gardarsson 1971) were often heavily grazed. Generalizations about winter food summarized in Fig. 9.8 and in the preceding paragraphs provide an overall picture that on the individual level may be much affected by movements and by food selection. Many adults and some juveniles
Fig. 9.10. Grazing by ptarmigan on Betulapubescens in three study areas in West Iceland. Unfilled symbols show mean % of catkins removed over winter on transects of marked branches. Filled symbols show proportion of grazed shoot tips at Hvammur.
328
A. GARDARSSON
are probably able to maintain a diet much richer in relatively nutritious foods by responding promptly to changes in snow cover, which is the most important factor governing movements in winter. It is also suggested that the average diets of males and females may differ during the winter because of differences in the birds' distributions. Males often stay relatively close to breeding areas in winter, whereas females are commonly found in coastal mountains, often with a good supply of S. herbacea in early winter. In spring the females take more growing herbs and fewer berries than the males and therefore have a diet that is higher in protein and some minerals and lower in carbohydrates (Gardarsson & Moss 1970). Thus most females are likely to get somewhat more nutritious food in winter and spring than most males. This may have bearing upon reproduction, as Savory (1975) has shown experimentally that breeding performance of seasonally laying galliforms may depend on factors operating over a long period before they start laying. At all times of the year ptarmigan are highly selective in choosing their food (Gardarsson & Moss 1970, Gardarsson 1971). Selectivity coupled with heavy utilization as observed in winter implies that food is likely to affect the condition and life expectancy of ptarmigan. Direct proof of this may be difficult to obtain, however, and the fact that food can at certain times be limiting does not mean that it necessarily drives population cycles. In particular, no evidence of cyclic changes in plant nutrients, as would be postulated by the nutrient-recovery hypothesis (Pitelka & Schultz 1964), or in plant defense mechanisms as suggested by Haukioja and Hakala (1975). Food is often thought to limit populations by limiting production. A shortage of high-quality food may exist; in particular protein is proposed to be limiting to the very young (White 1978), often by way of maternal nutrition (e.g., Siivonen 1957, Moss et al, 1975), though food of growing young may also be important (Gardarsson 1979, Jorgensen & Blix 1985). It would appear that in Icelandic rock ptarmigan, variation in food quality does not cause variation in production. Clutch size was uniformly high and varied independently of spring temperature and by inference the quality of food plants on the breeding grounds. However, feeding in winter may also be important to females and their reproductive performance (see Savory 1975), making it difficult to measure relevant food availability. In comparison with many other studies (e.g., Jenkins et al. 1963, Watson 1965, Bergerud 1970a, Myrberget 1972), the production of ptarmigan young in Iceland is remarkably high and shows little variability. Perhaps this is because the chicks have available to them a highly nutritious, all-weather food, Polygonum bulbils, and do not depend on insects, which vary greatly in availability depending on the weather, as a source of protein (cf. Savory 1977).
ROCK PTARMIGAN IN ICELAND
329
9.5 Summary The study outlined in this paper was originally conceived as an inquiry into as many facets as possible of the life history of rock ptarmigan in Iceland, in the hope that this approach would lead to more refined ideas about the 10-year cycle of this population. The natural experiment provided by gyrfalcons preying on the territorial males in spring, and vacant territories being taken over by neighbors instead of floaters, suggests that the function of territorial behavior is to space out the birds on the breeding grounds but not to regulate numbers. Ptarmigan are migratory and their numbers are therefore not necessarily limited on the breeding grounds only. If behavior is important in regulating population, it may be social behavior in the winter flocks. If regulation occurs in winter, one would also expect widespread regional synchrony as observed in the 10year cycle in Iceland. In the population studied, production, including clutch size and chick survival in summer, was always high, and despite certain variation, could not be linked to the cyclic phenomenon. Clutch size was high but varied inversely with density. Chicks always survived well during the first month except in the peak year when poor survival was caused by weather; this poor survival was chronologically the beginning of the population crash. The key factor (Varley & Gradwell 1970) in the demographic changes is obviously winter losses (Fig. 9.3), in particular those of juveniles. First-winter losses can be partitioned into "early" (i.e., September-October) and "late" (November-April) losses. Late losses of juveniles, and presumably adults as well, caused the population decrease. Early losses increased during the population decrease and were responsible for continuing low numbers. Late losses of juveniles and heavy losses of adults during the decrease are best explained in classical terms as density-dependent. Increased early-winter losses during the low were caused by a delayed densitydependent factor. Only three such factors need be considered in view of present evidence: (1) increased predation along the lines suggested by Pearson (1966) and Keith (1974); (2) long-term changes in early-winter food, mainly Salix herbacea', and (3) intrinsic changes in the birds themselves, especially changes in aggressive behavior (see Chitty 1967, Krebs 1978a, Moss & Watson 1980). Clearly, we do not have enough information, after observing only one cyclic fluctuation, to decide between these possibilities. In conclusion, a direct density-dependent and a delayed density-dependent component can be identified in the decline and low phases of the 10-year cycle of rock ptarmigan in Iceland. More measurements of food and predation are needed before their roles are satisfactorily understood, whether or not intrinsic cyclic changes in the birds themselves can be demonstrated.
10
Winter Survival and Spring Breeding Strategies of Willow Ptarmigan D. H. Mossop
10.1 Introduction Willow and rock ptarmigan (Lagopus lagopus and L. mutus) in North America demonstrate classic, 9- to 11-year cyclic fluctuations in numbers (Bergerud 1970a, Weeden & Theberge 1972). I studied willow ptarmigan in Northern British Columbia from 1970 to 1974 to test some of the hypotheses that have been proposed to explain these regular fluctuations in numbers of breeding birds. In particular, I examined whether social behavior of willow ptarmigan in the winter or during the spring breeding period limited their numbers. Winter is the dominant season in the environment for these grouse that spend their complete annual cycles at high latitudes. Despite this, little is known about ptarmigan in winter. In fact, among all grouse, there is a paucity of comprehensive research in the interval between breeding seasons. Portions of winter ecology of ptarmigan have received some attention: Weeden (1964), Irving et al. (1967a,b), Pulliainen (1975b), and Hoglund (1980) have information on migration, possible sex segregation in winter, snow-roosting, and feeding rates. West (1968) has examined the physiology of wintering, captive ptarmigan. Many researchers have encountered other grouse species in winter and have provided a fair understanding of the habitats they utilize as well as some of the strategies they have adopted (Brander 1965, Godfrey 1970, King 1971, Chap. 5). Yet, with the possible exception of red grouse (Lagopus lagopus scoticus), which have been counted in fall and winter, and have been subjected to fall experiments (Jenkins et al. 1967), most of what we know about the behavior and ecology of grouse has come from summer studies. My first goal was to learn more about the behavioral adaptations of grouse to 330
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
331
survive the winter. I monitored ptarmigan movements, daily feeding and social behavior, changes in body condition and food supply, and the extent and causes of winter mortality. My second goal was to study the behavioral adaptations of ptarmigan when they abandon their winter survival strategy and space themselves in the spring for breeding. Here the emphasis was on the timing of breeding and whether all birds secured territories. Specific population-regulation theories addressed were: (1) that breeding numbers were determined by density-dependent mortality in winter owing to food shortages (Lack 1954, 1966); (2) that numbers changed as a result of densitydependent mortality in winter caused by predation (Lack 1954); and (3) that birds died because they failed to secure breeding territories in the fall and/or in the spring (Jenkins et al. 1967, and more recently Watson & Moss 1972, 1979, Watson 1985). A fourth hypothesis, proposed when my study began, was that annual mortality rates of adult ptarmigan were relatively constant, regardless of density (i.e., they were noncompensatory), and the force behind cyclic population changes was annual variations in breeding success, the survival of chicks (Bergerud 1970a,b).
10.2 The study area The study area was at Chilkat Pass in the extreme northwestern corner of British Columbia (Fig. 10.1). The weather in this wide, low pass in the Coastal Mountains is affected both by wet, coastal air from the Pacific and by dry, cold air from the interior of the Yukon Territory. This results in a luxuriant, alpine vegetation and a deep, winter snowpack dominated for long periods by dry, arctic air masses. Weeden (1959a,b) has given general descriptions of the vegetative associations in the area. Alpine, "high" tundra occurs above approximately 1,000 m ASL. This zone occurs in a relatively small area at the crest of the pass. By far, the most extensive vegetation zone lies inland from the crest. This subalpine zone is dominated by a luxuriant growth of shrubs that are interspersed between wet, tundra meadows, and is hereafter referred to as the shrub or shrub-tundra zone. The shrub zone extends from the top of the pass down the valley of the Nadahini River, and varies in width from 2 to 8 km (Fig. 10.1). At lower elevations the zone grades into a white spruce (Picea glaucd) zone on the interior slopes in the Yukon Territory. Willow ptarmigan were found at all times of the year in the shrub-tundra zone (Fig. 10.1). Immediately above the shrub-tundra zone, in an area of dwarf shrubs, were rock ptarmigan (see also Weeden 1959b). Common mammalian predators at Chilkat Pass were the red fox (Vulpes vulpes) (Jones & Theberge 1982), the short-tailed weasel (Mustela erminea), and the wolverine (Gulo gulo). Avian predators included gyrfalcons (Falco rusticolus) that nested on the adjacent mountains and hunted ptarmigan throughout the
332
D. H. MOSSOP
Fig. 10.1. The Highland study area was near the Nadahini River, close to the Chilkat Pass, and near the upper limit of the willow ptarmigan shrub habitat; it was used to study the breeding strategy of ptarmigan. The Lowland area was at Stanley Creek, where riparian willow was much taller, and was used as a reference area to study the winter survival strategy of ptarmigan.
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
333
year, and golden eagles (Aquila chrysaetos) that were summer residents only. Alternative prey for these raptors included rock ptarmigan, white-tailed ptarmigan (Lagopus leucurus), arctic ground squirrels (Citellus undulatus), and snowshoe hares (Lepus americanus), which occurred with increased frequency at lower elevations north of the pass. Two intensive study areas were selected—one at the Nadahini River at 1,000 m, just below the pass (Fig. 10.1), the other at the Stanley Creek area, at 750 m, 35 km north of the Nadahini area. The Nadahini area (hereafter referred to as the Highland area) was 2 km2 and its western perimeter bordered the upper limit of the shrub-tundra. Rock ptarmigan were immediately above this study area, which was used to investigate spring breeding behavior. The Stanley Creek area (hereafter the Lowland area) was 3.5 km2, and was used to research winter behavior. The lowland area supported riparian willow (Salix spp.) 3 to 8 m in height, typical of habitat chosen by wintering birds (Weeden 1964). Browsing by moose (Alces alces) and ptarmigan had altered the growth of many of the trees and shrubs to produce a dense life-form. 10.3 Methods My major field technique used throughout the winter was to record the behavior of birds while observing them from a heated blind. One blind was portable and transported by snowmobile. A second blind was built on a 7-m tower in the Lowland Area and overlooked the riparian canopy. Typically blinds were used from 2 hours before sundown, through the night, until 2 hours after sunrise. Ptarmigan sounds and behavior were recorded during 5-minute observation periods. This routine continued until no birds were seen or heard in 5 consecutive observation periods. I also recorded all predator activity. I counted the number of territorial males in May at the Highland Area on 194 ha in 6 years, 1970-1974 and 1976 (also 1977-79 and 1983). Both the Highland and the Lowland areas were searched at least once per week in the springs of 1973 and 1974. During these years nearly all males and females on territories at the Highland area were captured and banded, and their age, sex, and weight were recorded. I also searched for birds not on territories near the Highland Area and banded as many as I could capture. When I found birds snow-roosting, I recorded size of the flock, distances between birds, and distances between roosts and the nearest shrubs. The physical condition of ptarmigan was determined using measurements from approximately 456 banded birds and 421 collected birds. Most data were collected between November 1971 and June 1973. Weekly samples in the winter months averaged 16 birds. Four major measurements were: (1) total body weight minus crop weight, (2) amount of fat on the pectoral muscle, (3) amount of heart fat, and (4) sharpness of the keel interface with the pectoral muscle. The amount
334
D. H. MOSSOP
of fat on the pectoral muscle was measured by recording its width at the outer edge of the left muscle. Heart fat was similarly measured (width) along the aortic artery. Sharpness of the keel was given a subjective rating from 1 to 3 —class 1 birds had well-rounded breasts, and class 3 birds had deeply concave ones. I measured the availability of food throughout the study period, both on the breeding area and in the areas where birds wintered. Ptarmigan fed exclusively on browse. My technique was to follow the tracks of ptarmigan and count the unbrowsed (available) twigs on each shrub after birds visited them. Ten samples of ten twigs each were counted on each shrub. The proportion of browsed-tounbrowsed tips gave a relative measure of the food still available in the habitat. Most important, this avoided the problem of determining which shrubs were acceptable food for birds (see Moss 1972). It was also very much a minimum estimate of food availability. Obviously, my count was taken after at least one bird had already fed on each plant. Weather was monitored at four permanent sites that were systematically spaced from the top of the pass to the beginning of the boreal forest. At each station was a snow-depth gauge and a maximum/minimum thermometer. Sites were visited at least once each week, and snow depth, temperature, wind speed, and cloud cover were recorded. Also, a standard, government weather station was about midway across the shrub-tundra zone. Predators and their sign were recorded whenever encountered. I was interested in determining the density of the various predators at different times of the year, as well as gaining insight into their strategies and success at finding and killing ptarmigan. Mammalian predators left tracks, which I followed and which augmented direct observation. This was especially valuable in studying nighthunting activity of foxes. I accumulated a running tally of successful and unsuccessful attacks on ptarmigan. Experimental removals of ptarmigan were conducted to test the effectiveness of social competition in reducing the success of birds attempting to recruit into the breeding population. These were conducted at the Highland area, where two 1-km2 plots were used—one served as a control and a similar one as a removal area. All breeders were individually marked and careful counts were kept of all birds in flocks. Details of removal methodology are presented in section 10.7.4.
10.4 Numbers, breeding success, and mortality Density of male ptarmigan increased from 20 males/km2 in 1970 to a peak population of 57 males/km2 in 1971 (Fig. 10.2). The population then declined to a low of 16 males/km2 by 1976. Previous to this peak, Weeden (1959, 1966) had recorded a peak population of 77 territories/km2 in 1962 and a low population of 12 males/km2 in 1965. Another peak occurred in about 1979 when I counted 55 males/km2 in 1979. The population then declined to 20 males/km2 by the time
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
335
Fig. 10.2. Densities of territorial-male willow ptarmigan at Chilkat Pass. Counts for 1959 to 1965 are from Weeden (1959b, 1963, 1965b) and are at Mile Post 75. Counts from 1970 to 1976 (1975 missing) were made during this study at the Highland area (MP 70). Counts were also made 1977-79 and again in 1983. The population shows a cycle in numbers and is the highest density of willow ptarmigan reported for a noninsular population in North America.
I made another count in 1983. Thus, peaks occurred in 1962, 1971, and probably in 1979, or at intervals of 8 to 9 years, and the amplitude of the fluctuations was 400-500%. The ptarmigan population at Chilkat Pass during peak years was at higher densities than any other reported, noninsular population in North America (cf. Weeden 1963, Bergerud 1970a, Weeden & Theberge 1972). Production of young at Chilkat Pass showed great differences among years (Table 10.1). In all years some territorial males did not breed. I estimate that only 8% of the territories produced young in 1972. Chick survival from successful nests was significantly higher in 1970 and 1971 than in 1972 and 1973—5.8 chicks/brood versus 4.7 chicks/brood (P < 0.05) (Table 10.1). These findings are consistent with those of Bergerud (1970a) and Mercer (1967), who showed that breeding success was considerably reduced after a cyclic peak of willow ptarmigan in Newfoundland. Annual mortality rate of banded, established breeding adults between breeding seasons averaged 53% (Table 10.2). There was no significant difference between the mortality rates of 1972-73 and 1973-74, even though the population was
336
D. H. MOSSOP
Table 10.1. Population statistics from the Highland study area in spring and summer 1970-1974, Chilkat Pass, Bristish Columbia 1970
1971
1972
1973
1974
Territories/km 2
20
59
49
32
19
Territories without a female (%) a
18
20
29
34
15
Population parameter
Chicks/brood ± SE (sample size)
_
5.6 ± 0.35 6.1 ± 0.24 4.9 ± 0.60 4.5 ± 0.78 (21) (18) (7) (6)
% of territories producing young
52
15
8
13
low
Minimum number of chicks/km 2
72
48
14
18
low
60(15)
50 (22)
45 (24)
40 (10)
-
112
166
112
82
9
Number of juveniles in fall kill (%) b No. of fall birds/km2' a b c
Some males failed to acquire females in all years. This sample taken prior to migration outside study area. Fall population estimated by doubling territories/km 2 to account for females and adding the total to chicks/km 2 . This technique assumes an equal ratio of adult males and females and no oversummer mortality. Neither assumption is strictly valid but should be reasonable in all years. Table 10.2. Mortality rates of adults willow ptarmigan and a comparison of the proportion of males and females in the 1971-72 and 1973 cohorts. Chilkat Pass, British Columbia Adult mortality Birds banded
Mortality rate (%) ?-testa
Year 1972-73 1973-74 Total
42 31 73
16 10 26
50 68 58
63 40 54
t = 0.895 / = 1.566 / = 0.352
Cohort comparison
1971 1972 1973 1972 + 73 a
96 184 47 231
176 108 49 157
Mest of two percentages (Sokal & Rohlf 1969).
54 44 60 47
44 35 37 36
/ t / t
= = = =
0.033 1.484 2.269 2.258
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
337
higher in 1972 than in 1973 (Table 10.1). Young females were less abundant in the 1972 and 1973 cohorts than were young males. Breeding success was low in both 1972 and 1973 (Table 10.1), and females may have died at greater rates than males. Bergerud (1970a) thought that female young were less viable than male young during population declines, and Gardarsson (1971, Chap. 9) also showed that there were fewer female chicks than male chicks when the ptarmigan population in Iceland was declining.
10.5 The winter environment The three major factors that may result in overwinter mortality for grouse are deaths arising from the weather, food shortage, and predation. These environmental influences appear to be obvious problems that ptarmigan must cope with in their aparently awesome, arctic environment (cf. Salomonsen 1972).
10.5.1 Winter weather The weather at Chilkat Pass was, by arctic standards, variable but not particularly severe. The lowest temperature I recorded during the study was only -40°C, and average monthly lows were only - 15°C to -26°C at the Highland area. Ptarmigan winter successfully at far colder temperatures. In fact, birds migrated in the fall from the Highland to the Lowland areas, moving to colder temperatures (Fig. 10.3). Wind, as manifested by windchill and blowing snow, is probably the chief factor of winter severity. The windchill at Chilkat Pass was low enough that humans often found it difficult to function. However, in the microclimates that ptarmigan occupied, the birds virtually never exposed themselves to these extremes in wind. Even in the worst weather, ptarmigan were found with full crops and did not forgo feeding. I found no evidence that the formation of a surface crust prevented snow-roosting (cf. Hoglund 1980). Snow sufficiently loose to excavate was always available in the wind shadows of shrubs. The birds at Chilkat Pass were apparently exposed to relatively easy winter weather compared with some other populations in North America.
10.5.2 Winter food supply and physical condition of birds The food habits of willow ptarmigan during winter are fairly well known (West & Meng 1966, Weeden 1969, Moss 1973, Hoglund 1980). My collections and observations confirmed that for the entire period from early fall to late spring, the birds at Chilkat Pass used a relatively simple diet of willow and birch (Betula glandulosd) buds and twigs. These species are known to be excellent foods and are high in caloric content (West & Meng 1966, Moss & Hanssen 1980). Food supply is generally conceded
338
D. H. MOSSOP
Fig. 10.3. Ptarmigan moved downslope from 1,000-m elevation habitats to 750-m habitats late in the winter each year. This move took them to colder environments.
as the ultimate limiting factor of natural population growth (Lack 1954, 1966). One of the easiest ways to falsify the hypothesis of territorial self-regulation is to show that food is limiting during the winter (Watson & Moss 1970).
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
339
Food was not limiting at Chilkat Pass. When ptarmigan migrated in the fall and winter they left the Highland, which showed a larger proportion and total acreage of unbrowsed (available) food than was present in the Lowland area, to which they moved (Fig. 10.4). The snowpack was similar in both areas and ample food supplies remained above the snow throughout the winter in the Highlands. Similarly, Pedersen et al. (1983) and Braun and Schmidt (1971) watched ptarmigan leaving areas in winter in which ample food remained. I never recorded a foraging area denuded of twig tips, despite the fact that my measurements were a minimum indication of available food (Fig. 10.4). Nor was there evidence from behavior or from birds' crops that ptarmigan were under a
Fig. 10.4. When birds left the Highland area in February and March, they left an area with more food available than the area they moved to —the Lowland area. Food availability was measured in winters 1972-73 and 1973-74.
340
D. H. MOSSOP
food stress (cf. Irving et al. 1961 a). Willow ptarmigan are relatively agile at perching in shrubs to feed, availing to themselves a large food resource (Weeden 1959a). I never observed more than 20% of the birds in feeding flocks perched in shrubs; 80% were, even when food was scarce, not bothering to fly into the shrub crowns. The mean length of time that birds spent feeding outside their snow roosts was 71 minutes in the morning (n = 7) and 30 minutes in the evening (n — 27) in January-February. A proportion of that time was spent in nonfeeding behavior. This short period (7% of a 24-hour day) was sufficient for birds to fill their crops. Food actually became more available on the Chilkat Pass as winter passed and the snow level rose. Birds gained access to higher and higher levels of the shrub crown, where abundance and possibly nutritional content of buds and twigs were better. I recorded an increase in browse availability in food tallies immediately after blizzards. The physical condition of individual ptarmigan is an obvious gauge to measure food stress. Mean body weights of all sex and age groups steadily increased through the winter before the onset of breeding behavior (Fig. 10.5). The weights of six banded birds captured both in fall and later in winter or spring showed that five gained weight in the winter (Table 10.3). White-tailed ptarmigan in Colorado also gained weight in winter (May 1975), and willow ptarmigan at Hudson Bay maintained their weights throughout winter (Thomas 1982). It was also suggested that fat accumulated, and that birds' conditions improved or remained stable through winter (Table 10.4). Brittas and Marcstrom (1982) concluded that willow grouse do not store large energy reserves and suggested that this indicated they seldom undergo periods of malnutrition (cf. Myberget & Skar 1976). Food stress and deaths of ptarmigan as a result of the lack of food at Chilkat Pass appear highly unlikely.
10.5.3 Abundance of predators The important predators of ptarmigan at Chilkat Pass were the fox, gyrfalcon, and golden eagle (Table 10.5). The weasel and wolverine were never found hunting ptarmigan. Foxes were observed more often than any other predator (Fig. 10.6) and also were recorded hunting ptarmigan more often than any other predator. One fox was relatively tame, and I was able to follow him as he hunted. Serious hunting by foxes occurred only when ptarmigan were in snow roosts. Ptarmigan appeared safe from attack when they were on the snow surface. Upon the close approach of a fox, the reaction of birds above the snow was to utter the "duk" call and simply run out of the fox's way. On two occasions a fox rushed at a bird that was feeding. In each instance the ptarmigan dodged without flying and immediately began feeding on the same shrub. Foxes hunted birds in snow roosts during the day and night. They located
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
341
Fig. 10.5. Annual weight cycle for juveniles and adults. Birds generally gained weight through most of the winter when engaged in winter survival strategy, but males lost weight when they switched to breeding strategy between March and May. Brackets represent + 1 standard deviation.
342
D. H. MOSSOP
Table 10.3 Body weights of individual willow ptarmigan during winter 1972-73 (year of decline), Chilkat Pass, British Columbia Age
Sex
Period
Ad Ad
M M M M F F
11 Jan -> 7 Feb 10 Jan -+ 7 Feb 14 Nov ^25 Jan 3 Jan -» 16 May 16 Nov -> 14 Dec 29 Nov -» 11 May
ya
Y Ad Ad a
Winter weights (g)
525 550 460 450 470 490
-> -» -> -» -> -»
Direction change
+
540 540 470 590 480 525
+ + + +
Yearling
roosting birds by scent, up to 30 m away, then stalked and later rushed forward and plunged into the snow. I never saw a successful capture (n = 11). A local trapper caught 48 foxes in 1971, the peak ptarmigan year. His catch dropped when the ptarmigan declined between 1972 and 1974, suggesting that the number of foxes also declined. Snowshoe hares also peaked in 1971 and later declined. Foxes were seen carrying parts of hares and spent considerable time hunting them. Gyrfalcons were observed on most days in early and late winter when the weather permitted them to hunt (Fig. 10.6). Visibility conditions of approximately half the days at Chilkat in midwinter possibly prevented falcons from hunting. I estimate that six falcons spent part of the winter on the pass in 1970-74 (Fig. 10.7). Their success at killing ptarmigan once pursuit was serious was just over 50%. I also observed gyrfalcons obviously following foxes and pursuing the Table 10.4. Condition and body fat of willow ptarmigan in winter, 1970-74, Chilkat Pass, British Columbia Cocks
a b
Month
Condition 3
Sept. Oct. Nov. Dec. Jan. Feb. Mar.
2.0 2.4 2.0 2.1 1.2 1.8 1.2
2.4 2.0 2.7 3.3 3.3 3.3 3.0
Hens
Fat b
(n)
Condition
± ± ± ± ± ± ±
(12) (13) (28) (4) (15) (34) (26)
2.0 2.3 2.0 2.4 2.0 2.3 2.0
0.1 0.3 0.1 0.7 0.1 0.1 0.1
(n)
Fat 2.1 2.0 2.8 2.8 3.0 3.2 3.1
± ± ± ± ± ± ±
0.3 1.1 0.1 0.1 0.2 0.2 0.0
(6) (3) (14) (19) (5) (7) (6)
Condition ratings: 1 = excellent, 2 = medium, 3 = poor, using sharpness of keel (see Methods section 10.3). Heart and pectoral fat combined (mm) (see Methods).
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
343
Table 10.5. Number of winter observation periods (67 total) when specific predators or their sign were recorded at Chilkat Pass, British Columbia, 1970-71 to 1973-74 Predator Mammalian Fox (Vulpes vulpes) Wolverine (Gulo gulo) Wolf (Canis lupus) Weasel (Mustela erminea) Avian Gyrfalcon (Falco rusticolus) Peregrine (F. peregrinus) Goshawk (Accipiter gentilis) Golden Eagle (Aquila chrysaetos)
Fresh sign
Direct observation
67 58 4 58
52 2 5 0
0 0 0 0
36 1 5 18
ptarmigan that flushed from the willow patches when foxes entered them. Gyrfalcons occasionally followed me and on one occasion killed a ptarmigan I flushed. Golden eagles did not winter on the pass but were present in fall and spring (Fig. 10.6). They were effective in capturing ptarmigan. I observed five serious pursuits, and eagles captured ptarmigan in all 5 cases. This success was partly due to the fact that these eagles, unlike gyrfalcons, followed birds they were chasing into the shrubs that ptarmigan used as escape cover. I watched one eagle standing on the ground, among crouching ptarmigan it had just pursued into the
Fig. 10.6. Common predators at Chilkat Pass were gyrfalcons and colored foxes. Golden eagles migrated through the area in fall and spring.
344
D. H. MOSSOP
Fig. 10.7. Winter sightings of gyrfalcons from 1970 to 1974. Gyrfalcons were nesting at Chilkat Pass in the early 1970s, but since the end of the study have nearly disappeared as breeding birds.
shrubs. It simply grabbed a motionless bird. In the spring eagles could be expected to take two ptarmigan a day. It became obvious over the years that predation was the sole cause of death in winter; birds did not die from weather exposure or winter food shortages; they were killed by foxes, gyrfalcons, and eagles.
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
345
10.6 Survival strategies for the winter environment The problems for ptarmigan posed by weather, food supplies, and predators were countered by an array of tactics including the use of snow roosts, the adoption of a highly synchronous, crepuscular daily activity pattern, the formation of flocks, and migration.
10.6.1 Flock formation Willow ptarmigan began wandering from their spring breeding territories at various times, apparently depending on their success at breeding. Cocks unsuccessful in breeding were the first to wander and were immediately found in flocks as early as late June. The process continued through the summer until by mid-July only adults accompanying broods could be found dispersed over the breeding habitat. Flocks formed whenever the defense of territories was absent. Formation of flocks is the most obvious behavior associated with nonbreeding or "winter" activity. Some evidence indicates that flock formation is primarily an antipredator strategy. The mechanisms of avoiding predation by joining other birds have been explored theoretically and empirically by others (review by Edmunds 1974, Pulliam & Milikan 1982). Group formation may have the effect of diluting the probability that a particular individual will be singled out and killed. Increased vigilance is involved, but I think that the use flocks make of cover, especially snow as roosting cover, adds the more important value to flocking. This line of reasoning runs counter to some theoretical considerations of predator search (Krebs 1978b), but effective use of cover was missing in the prey species around which those ideas were built. Flock sizes varied and reflected the abundance of birds in general, rather than any functional relationship with a survival strategy. Flock sizes increased as each winter progressed (Tables 10.6 and 10.7), and the largest flocks were seen in the year of highest numbers (1971-72).
10.6.2 Snow-roosting Willow ptarmigan on Chilkat Pass began snow-roosting in fall as soon as sufficient snowpack was available and continued the behavior well into early spring. Birds dug roosts by scratching with their feet, and depending on snow conditions, could completely bury themselves in 10 to 30 seconds. Roost chambers were 20-50 cm from the entry hole and undisturbed exits were made by digging with the bill above the roost chamber. Snow-roosting behavior described by others compared favorably with my observations (Hoglund 1980). When viewed in association with other winter behavior, it appears that snowroosting is principally a predator-escape strategy. The idea that roosting primarily provides protection from weather is not supported by the facts that birds emerged only in morning and evening when temperatures were lowest and that
D. H. MOSSOP
346
Table 10.6. Bimonthly flock sizes at Chilkat Pass as the population increased in 1970-71 and declined in 1971-74 1970-71
Month
1971-72
1973-74
1972-73
5 ± 2 (4)
September
6 ± 9a (22) October
9 ± 36 14 + 49 47 ± 36 29 ± 33 23 ± 6 14 ± 4 66 ± 15 7± 5 14 ± 24
November December January February March April
Unweighed means a
(30) (26) (6)
(10) (6) (4) (4) (5) (4)
12 ± 11 (8) 15 ± 10 (3) 19 ± 40 (46) 26 ± 32 (21) 37 ± 30 (11) 39 ± 48 (19) 49 ± 68 (19) 45 ± 10 (35) 94 ± 53 (8) 114 ± 84 (16) 52 ± 20 (7)
36 + 58 (6) 36 + 62 (15) 63 + 71 (9) 65 ± 45 (10) 143 + 160 (9) 14 ± 13 (12) 24 ± 29 (13) 85 ± 141 (20) 16 ± 23 (10) 11 ± 20 (3) 48 ± 100 (16) 12 ± 8 (6)
46
46
23
15 ± 25 (10) 36 ± 58 (6) 60 ± 81 (15) 25 ± 18 (10) 6 ± 1 (12) 33 ± 39 (8) 33
Standard deviation.
roosting occurred on all days, even in fall and spring when temperatures would not warrant it. The use of vegetation cover by grouse is an advanced antipredator adaptation accentuated by cryptic coloration, as is well known. I view snowroosting as a sophistication of that tactic. In combination with flock formation it forms the most important survival strategy of willow ptarmigan in winter. The bonus of snow-roosting is that it has additional energy-conserving benefits that have been known for some time. The use of vegetative cover also has some of these advantages, and the progression from vegetation to snow appears a logical Table 10.7. Comparison of flock sizes between the Highland and Lowland study areas, winter 1972-73 No. of birds per flock x ± SE (n) Period Nov-Dec Jan Feb Mar Apr Total
Highland 68 28 12 16 8 22
± ± ± ± ±
23 (9) 9 (26) 3 (11) 8 (9) 2 (26) (81)
Lowland 76 126 183 54 157 119
± ± ± ± ±
35 (5) 45 (13) 48 (6) 38 (5) 71 (5) (34)
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
347
development of similar behavior. The physical link between vegetative cover and snow roosts is also maintained; snow roosts of willow ptarmigan are normally dug close to or in shrub cover (mean = 0.97 + 0.1 m, n = 101). This close association may be related to the vulnerability of birds on the surface while digging and emerging. In addition, birds roosting near shrubs gain advantage because predator search paths are partly blocked. Snow-roosting and flocking behaviors were associated. Roosting birds were normally in a flock and close to each other (mean nearest neighbor 1.7 meters; 2.1+0.2m,rc=101). Predators at Chilkat Pass in winter were primarily gyrfalcons and foxes. In snow roosts, ptarmigan were safe from falcons. The fox's hunting method seemed most dependent on its ability to accurately locate a snowroosting bird. In each instance the fox failed essentially for the same reason: the bird that was first flushed was not the bird being stalked (Fig. 10.8). The fox's stalk or rush always took it past a bird it was apparently unaware of. The synchronous and contagious flushing of ptarmigan from snow roosts may also startle foxes. During one observation a stalking fox stepped on a snow-roosting bird. The bird bursting from underfoot and the flushing of the rest of the flock visibly shook the fox, which made no attempt to grab at any of the birds.
Fig. 10.8. Typical unsuccessful attack by a fox of snow-roosting ptarmigan. Roosting birds flushed during the final rush of the fox, signaling all the birds to flight, including the one being stalked.
348
D. H. MOSSOP
10.6.3 Synchronized, crepuscular behavior One striking aspect of the winter behavior of ptarmigan was their short, strenuous bouts of synchronous feeding at first light and again at dark (Fig. 10.9). They were almost solely crepuscular. Active time averaged 52 minutes twice daily, or about 4% of the 24-hour period. The synchrony was reinforced at every bout by calls and close flock movements. Precise crepuscular feeding also occurs in Scandinavian birds (Hoglund 1980). Pedersen et al. (1983) noted that Scandinavian birds are more strictly crepuscular than red grouse in Scotland, where raptors are rare. At Chilkat Pass the crepuscular feeding schedule was effective against gyrfalcons. Another crepuscular activity I have termed "creeping." In the evening, birds left their snow roosts 30-60 minutes after sundown. At that time light was sufficient to discern the birds' shapes. At first the ptarmigan moved slowly for approximately 10 minutes; they paused several seconds between steps. This creeping left characteristic trails ranging from a meter to 50 m depending on the length of twilight (Fig. 10.10). About 50 to 90 minutes after sunset I could no longer discern the whiteness of the birds against the snow. Then birds simultaneously began to run from shrub to shrub while feeding. They also began to fly and call. The sound of these flocks, clamoring in willow patches, tearing browse and buds loose and flying between feeding sites, was audible for several hundred meters on still evenings. Braun and Schmidt (1971) have seen the same frantic crepuscular bouts of feeding in white-tailed ptarmigan. This behavioral sequence allows ptarmigan to avoid conspicuous motions until light intensity falls to the level at which the crypticness of birds remains preserved by darkness. Given the limited time of activity and feeding, combined with the energetic costs of emerging from roosts at the coldest times of day, the selective predator pressures to maintain this crepuscular habit must be strong.
10.6.4 Migration Willow ptarmigan are known to shift wintering locations, and some populations have been described as migratory (Weeden 1959b, Irving et al. 1967a,b, Moss 1972). At Chilkat Pass ptarmigan shifted north of the Highland breeding area as the winter progressed (Fig. 10.3). This took birds to lower elevations, lower temperatures, and generally to denser and taller shrub cover. Snowpack averaged only 9 cm shallower on the Lowland winter range than on the Highland breeding range. The longest movement by banded birds was 35 km. This winter shift of ptarmigan agrees with the view of Weeden (1959b), that birds moved toward more productive habitats; however, the productivity requisites sought by birds at Chilkat need not necessarily be food but rather more vigorous willow cover to escape predators. Timing of the movements can provide evidence of which requisites are in short
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
349
Fig. 10.9. During the winter birds moved out of their snow roosts and fed for brief intervals at first morning light and again as it became too dark for gyrfalcons to hunt.
350
D. H. MOSSOP
Fig. 10.10. When ptarmigan first appeared out of their snow roosts, they remained motionless for about 10 minutes, then they began "creeping," leaving behind a characteristic track. As it became darker, they started running and flying in voracious feeding bouts before moving to their night roosts.
supply. The major downward movement from the Highland to the Lowland area occurred in February and the major return migration occurred in April (Fig. 10.11). Males generally stayed closer to the breeding range in winter than did females (see section 10.6.5). Birds at Chilkat Pass showed no evidence of a classic migration as inferred by Irving et al. (1967b) for the fall migration of willow ptar-
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
351
Fig. 10.11. Birds left the Highland study are in late January and February when the snowpack exceeded 1 m and there was generally less than 1 m of willow shrub cover above the snow. The birds returned in the spring when snowpack still exceeded 1 m and shrub cover was still less than 1 m.
352
D. H. MOSSOP
migan through Anituvik Pass, Alaska. In one year, 1974, there was little egress at all from the Highland. Instead, my findings suggest that proximate factors in the winter environment act to trigger movements whenever a threshold is reached, and that this can occur at any time throughout the winter. When birds left the Highland area in February, adequate food remained available (Fig. 10.4). Birds actually moved to an area with more heavily browsed shrub resources. Further, they shared this food with more birds from many adjacent ranges. They did not make this move until late winter, after the worst weather had passed. Accumulation of the snowpack was the only weather parameter that I could causally relate to movement. As the accumulated snow reached about 1 m, leaving only 1 m of overhead shrub cover, birds began to leave upland sites (Figs. 10.11, 10.12) and move to tall patches of riparian willow cover at
Fig. 10.12. Top: Frequency of winter flocks that were recorded in the Highland area in shrub cover + 1 m in height. Bottom: Birds in the Lowland in the winter sought tall, dense shrub cover.
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
353
lower elevations (Fig. 10.13). These ptarmigan moved to select cover on their wintering areas that was well over their heads; this was much greater than that which they normally used on the breeding range (Figs. 10.12, 10.13). Cover to escape predators (both detection and once pursued, i.e., in the broad sense) is a sufficient explanation for this late-winter migration. When the snowpack had reached about 1 m on unprotected ground, drifting snow had inundated shrubs on the pass well into their crowns. This left ample food, in the form of shrub tips above the snow, but cover was lacking. Larger riparian shrubs at lower elevations were sought by the ptarmigan—this was the winter migration (Fig. 10.13). It was clear from the behavior of these birds that surface vegetation cover was essential for their escape from gyrfalcons when flushed from snow burrows.
Fig. 10.13. Birds left the Highland area in January-Feburary when approximately 50% of ptarmigan observatons were in willow cover that was less than 1 m in height, considered insufficient to hide birds. Ptarmigan left an area with ample food, and the increase in snowpack at the Highland in fact increased the availability of willow twigs and buds that birds could reach. Birds moved to the Lowland area where they fed on low growth-form willow shrubs that were immediately adjacent to the tall riparian willow used for escape cover.
354
D. H. MOSSOP
Winter movements or migration by Chilkat ptarmigan, I believe, are best interpreted as a process of "tuning" the basic winter, antipredator strategies of flocking and snow-roosting, by maintaining the essential ingredient of escape cover.
10.6.5 Sexual segregation Weeden (1964) found by sampling winter flocks of both rock and willow ptarmigan that hens and young traveled farther from breeding areas during the winter than did adult cocks. Sex and age ratios I encountered in wintering flocks showed a change in the proportion of flocks made up by females and yearlings as the winter progressed (Fig. 10.14). That hens were part of these flocks all winter, even at higher elevations, discounts the general notion that the sexes completely segregate. However, because the proportion of females was significantly less than the summer sex ratio, and because females seemed to be disappearing from the flocks all winter, it is plain that the sexes did tend to move differently. Moreover, the age ratio of the wintering flocks, although not in complete agreement with the idea
Fig. 10.14. As winter passed, males and females progressively segregated. Females moved farther downhill from the Highland study area to use more protective willow cover than that sought by males, particularly adults. Data from Highland area only.
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
355
Table 10.8. Age distribution in willow ptarmigan winter flocks, 1970-71 through 1973-74, Chilkat Pass, British Columbia % of birds captured (n) Age class
Highland
Lowland
/-value
Yearlings Adults Total
74 (151) 82 (165) 78 (316)
57 (42) 78 (94) 71 (136)
2.068 0.931 1.621
that yearlings dispersed with hens exclusively, indicated a tendency in that direction (Table 10.8).
10.6.6 Synthesis of winter survival strategies Predation has played the dominant role in shaping the strategy of willow ptarmigan. Solutions to adverse weather and food supplies are evident, but they are compromised by the overriding predator pressure. Birds often move away from adequate food resources and emerge from roosts and feed quickly at the coldest times of the day. Gyrfalcons played one of the most important roles in maintaining these behaviors. Colored foxes at the Chilkat Pass, hunting primarily by scent, added another dimension. The array of behaviors to counter these predators included: flock formation; snow-roosting; synchronized, crepuscular activity periods; and migration. Together these form a complex of behaviors—a strategy—that successfully counter gyrfalcons and foxes, and, at the same time are compatible with solutions to weather and food (Fig. 10.15).
10.7 Spring spacing behavior and its effect on the population The population-regulation model of Watson and Moss (1972, 1979) predicts that spacing behavior can regulate numbers by preventing some birds from breeding. In red grouse, this spacing occurs in the autumn when males return to territories. At Chilkat Pass, birds did not display fall territoriality, calling was infrequent (Fig. 10.16), and juveniles appeared too small to successfully compete with adults (Fig. 10.5). Later in the winter I did not observe any social behavior that might relate to spacing or formation of a social hierarchy. By February birds were largely absent from the breeding range and called infrequently except when feeding (Fig. 10.16). They remained inactive throughout the day, and at dawn and dusk engaged in vigorous bouts of feeding. Further, in winter birds were generally with strangers (Fig. 10.17), who later returned to their own breeding ranges. If social behavior with respect to space was causal in the regulation of ptarmigan numbers at Chilkat Pass, it would have had to occur in the spring when the birds competed for territories.
356
D. H. MOSSOP
Fig. 10.15. Winter strategy of ptarmigan to counter aerial, diurnal hunting gyrfalcons and diuranl-mocturnal, scent hunting foxes.
10.7.1 The early return Small numbers of birds were found in the vicinity of breeding territories all winter; the majority were cocks (Fig. 10.14). Sex and age composition of 135 birds captured on the breeding area in January and February was: 43% adult males, 37% juvenile males, 10% adult females, and 10% juvenile females. Birds began returning to the Highland breeding area in March, before testes enlargement (Fig. 10.18). These early returning birds were mostly males, and the majority were yearlings. The composition of 106 captured birds was: 26% adult males, 57% juvenile males, 2% adult females, and 15% juvenile females. This early return was a crepuscular activity. Movements typically began at
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
357
Fig. 10.16. Social calling per activity period throughout 3 years of population decline.
evening twilight. Single birds and small flocks were seen flying less than a meter above the snow surface in a goal-oriented, straight course up the pass. Mist nests set at right angles to the birds' line of flight began capturing them 1 hour after
358
D. H. MOSSOP
Fig. 10.17. During fall and winter, the proportion of banded poulation declined, suggesting ingress of birds to winter in the riparian cover of the Lowland study area.
sunset. By morning light, I could find only a small number of birds, although their calling was heard at early dawn (Fig. 10.16). Most birds I banded in the evening were downslope 3-8 km from the study area in the day, although a few were in the best cover on the study area. I have already presented evidence of the winter movements to effective antipredator cover. In late winter, with a deep snowpack, long clear days, and the onset of activity related to the breeding season, the necessity of finding cover would seem especially acute. Gardarsson (Chap. 9) has also noted that rock ptarmigan make long daily movements between cover and feeding areas in the spring. Both his and the populations at Chilkat Pass are hunted by the highly efficient gyrfalcon. Chilkat birds apparently expended extra energy flying to effective cover each day to maintain their winter survival strategy. I believe there must also have been a compensating benefit for the early return to the breeding range.
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
359
Fig. 10.18. Birds that wintered near the Highland area occupied territories before gonadal development had been completed (all years combined). Yearling and adult birds showed the same rate of increase of testes volume in the spring. Brackets include + 1 standard deviation.
This period of active movement between defended territories and effective cover could be an important time in the annual cycle of ptarmigan. Birds that were banded near the breeding study area in winter and during this early "invasion" period had a higher probability of occupying a territory on the study area than those arriving later (Table 10.9). Yearlings offer the best test of this idea because, in ptarmigan, only the location of their territory is in question. Evidence indicated
360
D. H. MOSSOP
Table 10.9. Proportion of male yearling ptarmigan banded on the Highland study area during winter and spring that were later found occupying territories, Chilkat Pass, British Columbia % of birds banded Winter (n)
1 Mar- 15 Apr (n)
16 Apr-31 May (n)
1971-72 1972-73 1973-74
80 (22) 67 (40)
42 (57) 27(11) 50 (6)
33 (21) 22 (9) 20 (10)
Total
71 (62)
41 (74)
20 (40)
Year
that adults surviving the winter and spring return to their breeding territory of previous years, or to one nearby (Chap. 8). This spring movement of mostly yearling males was my first evidence of social segregation relative to breeding. Yearlings that remained closest to the breeding range or returned first acquired an advantage in settling in the study area in preferred habitat. However, they risked returning to the breeding range while it was not as safe as those areas with taller and denser shrub cover at lower elevations.
10.7.2 Spring territoriality and social segregation The initiation of territorial behavior did not occur as a simultaneous populationwide event. Bouts of territorial defense began during clear days in March and intensified on warm, clear days as spring approached. In addition, territorial birds reverted to typical winter behavior during cold and snowstorm weather well into early summer. Significant differences occurred in the timing of territorial behavior between years. In spring 1971, a peak year of ptarmigan breeding density, territories were established about 2 weeks earlier than in the following 2 years. This early establishment was not obviously related to weather. The most severe spring weather occurred in spring 1971, and the snowpack increased throughout April. Yet in that year, by 20 April, 80% of the territories on my study area were occupied by a male. The following year, a relatively early spring when the snowpack declined in April, it was not until 10 May that 80% of the territories were occupied. Yet even in 1972, birds were calling much earlier than in 1973 and 1974 (Fig. 10.16). In 1973, the 80% occupancy mark occurred by 2 May. Events that were independent of the winter environment appeared to dominate the triggering of social competition, notably in the year of peak ptarmigan numbers. Change in population density (Fig. 10.2) over this period was the most obvious measurable parameter. The beginning of territoriality also did not occur simultaneously across the various sex and age groups. I recognized three stages of establishment. The first
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
361
was territorial defense by the few birds that either stayed on the breeding area or nearby all winter. In this stage few fights were seen and few aggressive calls were heard. These birds established breeding space while the gonads remained regressed. The second stage occurred when many of the males returned to the area in late April and early May. These were mostly yearling males. During this period I observed many interactions, including fights. Last, the hens began returning to the area, and social interactions increased in frequency. Adopting territorial behavior represented a major and obvious switch in behavioral tactics away from the survival strategy observed during winter. This changeover occurred when most of the environmental risks facing these willow ptarmigan in the winter were still present. The decision to switch behavior is an important one; weighed against the increased exposure to gyrfalcons, foxes, and at this time of year eagles, are the advantages of being early in breeding and first in social contests. The timing with which a particular individual makes the changeover should therefore reflect the varying importance of options available to that bird. The major return of males occurred in April. In the peak year, 1971, this return involved flocks of up to 200 males moving over the study area and adjacent breeding habitats. In subsequent years the invading male flock was smaller, until 1975 when these distinct flocks were hard to detect at all. A similar process occurred in early May when flocks of females arrived and moved over the breeding range. The segregation of the population into the group that had already acquired territories and those still in flocks became quite apparent by early May. As some males established territories, other males, mostly yearlings, segregated and moved to nonterritorial habitats—the taller and denser shrubs. An apparently similar sequence occurred among females. Some females joined males on territories and others, both yearlings and adults, frequented tall willows not contested. I have called these birds not on territories and traveling with other birds of the same sex the "waiting flock." It was a significant finding to observe birds on territories and, simultaneously, other birds in flocks at Chilkat Pass. The division of high-density, red grouse populations in Scotland into these two groups during typical years, and the fate of those birds in flocks, is the major dynamic parameter in the territorial expulsion theory of self-regulation (Jenkins et al. 1963, 1967, Watson 1985). I believe this flock occurred at my Highland study area because the mean density of birds at Chilkat Pass is extremely high and because I chose a study area at the top of the pass, just below rock ptarmigan habitat. There, the geographic difference between preferred and marginal habitats was in juxtaposition. It is reasonable that if birds were forced to wait for vacancies arising in preferred habitat they should do so where there is a sharp break in the habitat-preference continuum. I estimated the proportion of the total population in waiting flocks after 90% of the territorial males were established as: 40% in 1972, 20% in 1973, and less
362
D. H. MOSSOP
than 5% in 1974. Early in the season when the flocks were large they were predominantly males (mostly yearlings). Estimates of numbers in the waiting flock are: 190 males and 50 females in 1972; 70 males and 15 females in 1973; and 10 males and 5 females in 1974. In 1976 I could find no waiting flocks. The number of birds in the flock appeared to be dependent on the population density in a particular year. The process of segregation was clearly related to social behavior. As birds in the waiting flock wandered across the tundra, they were threatened and chased by territorial ptarmigan. My observations of newly territorial cocks appearing where a flock had passed support the conclusion that these wandering birds were searching for acceptable breeding habitat. By mid-May, movements of birds still in flocks were severely restricted because in most parts of the study area territorial boundaries abutted (Fig. 10.19). In the presence of territorial birds, flock members were submissive and generally silent. In lower elevations of the area, stands of taller shrub became the favored locations for these flocks. In 3 years, beginning with the peak year, I was able to predictably locate birds in these flocks in the taller shrubs after all the male territories on the study area were established. These flocks appeared to be a direct consequence of a severe selective process, occurring as a result of social competition at peak population density. The exact consequences of the contest are arguable, but the probability of at least banded males in the flocks later appearing on a territory in the preferred habitat on or near the study area decreased rapidly as spring progressed (see Table 10.9). The sex and age composition of birds localized on territories and those in the waiting flock constantly changed from April through May, as late birds arrived (mostly females) and others left the flock to localize on territories (Fig. 10.20). Based on the ages of captured birds, the sequence of territory establishment was: (1) first resident adult and yearling males established breeding territories; (2) latearriving adults (both females and males) settled; (3) yearling males left the flock and later appeared with territories; and finally (4) yearling and some adult females settled. Males were always slightly ahead of females, by perhaps a week; adult males settled before adult females; and yearling males established before yearling females (Fig. 10.20). Percentages of the total population that were composed of unpaired males on territories and paired males on territories confirmed the egress pattern observed in waiting flocks. In both 1972 and 1973 the number of single males and paired males increased during the last week of April and the first week of May (Fig. 10.21). Thus, at that time, yearling males were securing breeding space at the same time females (mostly adults) were pairing with already established males. In mid-May, the number of unpaired males declined as the number of pairs increased (Fig. 10.21). This implied that males were no longer establishing territories, It was shortly before this that I documented the disappearance of the last yearling males in the waiting flock. In early June, the number of pairs on the study
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
363
Fig. 10.19. Territories at the Highland control area from 1972 through 1974. The population was high in 1972 and territories abutted. Territory size increased in 1973 and 1974 as the population declined and there was more unoccupied space. Males with the smallest territories in each year failed to attract females. One male with a large territory in 1972 attracted two females. Other bigamous males are suspected.
364
D. H. MOSSOP
Fig. 10.20. Estimated number of birds in waiting flocks and age composition of females and males in the flock. As a rule, males left the flock for territories before females of their respective age group.
area declined and the number of single males increased (Fig. 10.21). Thus few females were localizing on territories, and many were disappearing and probably laying eggs. This nesting sequence agrees with the breeding schedule established by Hannon (1982) at Chilkat Pass from 1979 to 1981. The second week of June in 1972 and 1973 showed an increase in the number of pairs expressed as a percentage of the number of territories (Fig. 10.21); this occurred at the same time many hens were probably incubating. This increase closely followed the period
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
365
Fig. 10.21. Chronology of males establishing territories (singles) in 1972 and 1973, and the sequence of females joining males (pairs). A second peak of pairs occurred before midJune in both years, when some females probably joined males that were already paired but whose first female had begun incubating. Percentages of territories occupied by single males plus pairs equal the total percentage of territories occupied at one point in time.
366
D. H. MOSSOP
in June when I saw the last birds (females) in the waiting flock (Fig. 10.22). A tantalizing interpretation of this second peak in pairs is that the last females from the waiting flock may have been joining males that already had acquired a female, and that the first female was absent at the time, incubating. Cocks with two hens are reasonably common in ptarmigan (Watson 1965, Braun & Rogers 1971, Miller & Watson 1978, Hannon 1982). Hannon (1982) recorded that five of 57 hens (9%) shared a male in undisturbed habitats at Chilkat Pass from 1979 to 1981. The late increase in pairs I observed in 1972 and 1973 suggests bigamous males in both years, as in Hannon's study. An alternative hypothesis for the second peaks in the number of pairs is that many incubating hens lost or deserted their clutches. However, I had no evidence of inhospitable weather or an increase in the resighting of banded females paired earlier in a monogamous situation to support this. In all years I also searched for birds in the habitat around the study area, particularly in areas of suboptimum habitat on the mountainsides where rock ptarmigan
Fig. 10.22. Proportion of the population in the "waiting flock" declined between March and June 1972; the last females were seen in the first week of June. At times birds abandoned territories temporarily when snowstorms occurred.
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
367
bred (see Weeden 1965). During 1972, when waiting flocks were largest, eight territories were established by willow ptarmigan males late in May at the edge of and above the preferred shrub-tundra habitat; these males were in rock ptarmigan habitat (Fig. 10.23). All eight territories were occupied by yearling males. Only one male was observed with a hen. Significantly, four males were banded and had last been seen in the waiting flocks at lower altitudes. In later years, no willow ptarmigan were seen in these areas.
Fig. 10.23. Location of the waiting flock in all years that used willow habitat that was greater than 3 m in height, and the sites of eight territories of yearling males in 1972 located in rock ptarmigan habitat. Four of the males had been captured in the waiting flock in the taller willow cover and were banded; band numbers are shown. Note that the territories of willow ptarmigan above the shrub zone are larger than those of willow ptarmigan males without females below the shrub zone.
368
D. H. MOSSOP
Results suggest that the last disappearance of birds from the waiting flocks can be explained in part by: (1) males shifting to new habitats in suboptimum areas and (2) some females settling with males already paired with first females that were incubating. The separation of the Chilkat population into successful breeders and "doomed outcasts" was not as obvious as would be predicted by a simple hypothesis of social competition directly causing the eviction of surplus birds that were prevented from breeding.
10.7.3 Mortality and condition of birds in waiting flocks One significant finding was the observation that once formation of the waiting flock had occurred in spring, its disappearance was slow and gradual (Fig. 10.22). Any abrupt disappearance would suggest that birds were simply passing through to breed elsewhere. However, birds from these flocks were captured daily, and I concluded that many remained in the flock for long intervals in the early spring. Equally as important is that the flocks did in time disappear. The number of yearling cocks slowly dwindled to none about mid-May, and the hens were gone by the end of the first week of June. Once ptarmigan ceased to snow-roost and instead extended their daily activity period into the daylight because of the increasing social pressures occurring in spring, the success of the winter strategy appeared to collapse. The return of golden eagles and other avian predators occurred at that time. Gyrfalcons were well into their nesting behavior, and are known to depend heavily on ptarmigan in this early spring period (Barichello 1983). Golden eagles were partially successful with flocks. I located ptarmigan remains on or near the Highland area as the snow melted. These remains were known to be mostly from the late winter and early spring periods because very few ptarmigan actually wintered in the area, and because many were in spring plumage. Most remains were along the edges of the denser willow shrubs, where segregated flocks were also found (Fig. 10.23). Birds in the waiting flock and on territories died at a high frequency. Spring was clearly the season of greatest mortality for ptarmigan (cf. Weeden 1965, Mercer 1967). One interpretation of the effect of dominance in ptarmigan is that subdominants are found in less-favorable habitat, are stressed, and starve (Jenkins et al. 1963, 1967). Results of the present study do not support this view. Instead, territorial cocks lost considerable weight in the spring (Fig. 10.24). In contrast to the decline in weight of territorial males defending their spaces, males not contesting space in the waiting flock gained weight (Fig. 10.24). The hens did not show this weight difference; both females in the waiting flock and those on territories gained weight throughout the spring as follicles developed (Fig. 10.5). The mean size of the largest developing follicle for birds collected April-June was for hens from the waiting flock, 5.0 + 0.9 mm. (n = 14) and for females paired 4.2 + 0.5 mm. (n = 30). Females in the waiting flocks had large follicles even in
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
369
Fig. 10.24. Physical condition of territorial males declined in April and May, whereas males that remained in the waiting flock increased in weight. Brackets include 95% confidence limits.
the first week of June, 4.5 to 6.0 mm. (n = 2) versus 5.0 mm. for three females in pairs. Behavioral research conducted on red grouse suggests that birds not on territories in the spring are in poorer nutritional condition than territorial birds. Nonterritorial birds are thought to be expelled to a surplus flock because of inferior physical condition and nonaggressive behavior (Jenkins et al. 1963, 1967, Watson, 1985). My data suggest that differences in weight of "floater" and territorial red grouse may have occurred after social segregation. The floating, nonterritorial
370
D. H. MOSSOP
red grouse may have lost weight because they traveled in habitats with less heather (Calluna vulgaris} for food.
10.7.4 Test of the territorial self-regulation hypothesis If territorial establishment regulates population numbers by preventing some birds from breeding, and these birds die quickly, there should be replacement birds available if territorial birds are artificially removed. The replacement of territorial birds by floaters is the primary evidence for territorial regulation in ptarmigan (Watson & Jenkins 1968, Hannon 1983, Pedersen 1984). To test this hypothesis I removed territorial birds in 1973 and 1974. The differential migration of hens and cocks meant that by the time all territories were occupied by pairs, very few nonterritorial cocks remained in the vicinity. Removing hens and cocks had to be done as separate experiments. I removed cocks and a few hens from a general removal plot on 3-4 May 1973 and 30 April 1974, when occupancy of a control area by cocks approached 100%. The birds were removed at first morning light, and the area was rechecked for replacements 4 days later. A hen removal was conducted in 1 day in a second area on 23 May 1973 and 8 May 1974, at time when virtually all cocks in the control area had hens. I had banded about 30% of the birds in the waiting flock to document if they would replace the removed birds. In the general removal, conducted first, 32% of the males were replaced. There were no replacing adult cocks; all replacements were yearlings (Table 10.10). Only three of eight hens (38%) were replaced. In the hen removal about 2 weeks later, eight hens replaced the eight originally removed (Table 10.10). At the time of the general removal, there were few adult males in the waiting flocks. Five of the 19 (26%) replacing birds (4 males and 1 female) had been banded and came from the waiting flock. These results support the hypothesis that birds in the "waiting" flock were indeed waiting. But they do not show that occupancy per Table 10.10. Replacement of territorial birds removed in 1973 and 1974 by birds from the waiting flock, Chilkat Pass, British Columbia Type of removal General Removal Male Adults Male Yearlings Female Adults Female yearlings Hen Removal Adults Yearlings
No. of birds removed
No. of replacements
Age of replacements
12 13 6 2
8 3 -
All yearlings
5 3
8 -
2 adults, 1 yearling
5 adults, 3 yearlings
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
371
se was the only necessary condition for birds not localizing. Other factors must have been involved, because 22 of 41 (54%) vacancies were not filled. It was plain that hens and cocks approached the process of breeding in different manners. Hens, for example, may be immune to whatever factors limit cock replacement to only 32% (8 of 25); hens may simply be using males to locate prime nesting habitat. This would allow hens to specialize in nest-site decisions within a male's general area, as the male acted more in predator defense (Chap. 13). However, territorial competition did appear serious for yearling cocks. They had to arrive early, at considerable risk, and engage in obvious competition for space in breeding habitat.
10.7.5 Cost-benefit trade-off in tactics for breeding Males attempt to ensure breeding success by controlling requisites desired by females—nesting cover and space. Females seek nesting cover far removed from other nesting females, and they also seek vigilant males. If male vigilance and aggressiveness are correlated, a female that selected a male territory only on the basis of its large size would be assured of a vigilant male (Steen et al. 1985). My observations of territorial competition among males suggest that males used at least three tactics to control space and attract females, each with different benefits and costs. The first tactic was that of the residents. These males remained near the breeding range throughout the winter. This was risky because males that used it were conspicuous in poor winter cover and liable to predation. The benefits were that these males arrived first and may have had the best selection of female cover and space requisites. If a male can defend a territory twice the size of the average male, he has the possibility of breeding two hens (Hannon 1982, 1983). Another tactic appeared to be used by a group of males, mostly yearlings, that moved farther in winter from the breeding range. These males moved to tall shrub cover and remained less vulnerable than residents in winter. But they arrived on the breeding range after residents and thus had limited choices of female requisites. These males often had smaller territories than the first arrivals. If females are in short supply, these males should go without. A third tactic applies to males that also probably arrived very late, but remained in winter habitats the longest, and were presumably the least vulnerable in winter. On the breeding range they waited a long period for vacancies resulting from the death of territory owners. They benefited by maintaining weight and remaining in good physical condition. Some of these males may not have been prepared to occupy a small, henless territory even if one had become vacant. It is possible that these males could not have acquired a female by competing for a territory. At that time territories were already contiguous, and some males already present had inadequate space to attract a female. A further subdivision of
372
D. H. MOSSOP
the breeding range would have meant even smaller territories. In 1972 these remaining males moved to rock ptarmigan habitat, where I found them spaced far apart but in poor nesting cover (Fig. 10.23). One of these yearling males did attract a late-arriving female. Females also appeared to have trade-offs in spacing tactics. Early-returning, adult females were probably conspicuous to eagles and gyrfalcons against white snow (see Chap. 13). But by being first they could return to where they were successful last year or prospect for a new area and pick vigilant males with large territories and good nesting cover. Yearling females arriving later were more cryptic against dark substrates, but they probably had less choice among the territories remaining if they wished to remain monogamous and not share nesting space and male vigilance. Male territories below a certain variable-size threshold were not chosen (Fig. 10.19). A third option available to females in 1972 was to sacrifice cover for additional space and settle with a male in rock ptarmigan habitat. A final option for females became available when early-settling females were incubating, and thus not aggressive. A female could select a high-quality territory already chosen by a previous female. Nesting cover would be good, but her nest could be relatively close to the first female. Also, she could not count on the vigilance of the male in the brood season because bigamous males generally leave with the first brood that hatches (Hannon 1982). Late-nesting hens can, however, have a high nesting success (Parker 1981) and enjoy a high survival of chicks even without male vigilance (Hannon 1982). Chick rearing need not even occur on the nesting territory (Erikstad 1985).
10.8 Conclusions Chilkat Pass willow ptarmigan have shown strong 10-year cycles in spring breeding density. Ample evidence shows that increases in density occurred when the production of young was high and that the population declined when reproductive success was low. How the apparent sudden change in reproduction occurred at the peak of the cycle (Table 10.1) is generally unknown, but was unrelated to the survival of birds through the nonbreeding period and the social processes causing discriminative recruitment to the breeding population.
10.8.1 Winter environment and ptarmigan mortality The ptarmigan's strategies for winter survival are only slowly becoming clear. Willow ptarmigan on the Chilkat Pass spent most of the winter virtually secure. Their habitat included an abundance of food, which they needed access to for only short periods—choosing the security of dusk to feed. Ptarmigan of both sexes and all ages apparently passed winter in good physical condition; they maintained or
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
373
gained weight throughout the winter. I found little support for Salomonsen's (1972) view of ptarmigan struggling against a harsh arctic environment. Predation appeared to be the major problem molding ptarmigan wintersurvival tactics. A two-part behavioral strategy of flocking and snow-roosting was the basic winter system. Additional tactics of being highly crepuscular and synchronous while feeding, and migrating to tall and dense shrub cover, finetuned this strategy to changing winter conditions. Together they formed an effective, yet flexible system of defense against the simultaneous predation risk of raptors hunting by sight and mammals hunting by scent. The winter behavioral strategy began to diminish in effectiveness in late winter and early spring when ptarmigan began switching to the spring territorial distribution for breeding. Late spring was the only period in which ptarmigan body condition deteriorated. It was also the only time predators, notably golden eagles, apparently concentrated their attention on ptarmigan, especially those nonterritorial birds still in the waiting flocks. The ultimate factor in causing this mortality was changing social organization. The effect of mortality as a cause of changes in the subsequent breeding population can be discounted because: (1) it is not nearly of a magnitude that could remove a sufficient number of birds to explain the breeding density each spring; (2) it does not explain the reduced reproduction of birds that successfully took territories; and (3) many birds in the waiting flock do end up taking territories, even if in suboptimum habitat.
10.8.2 Social segregation of the population Jenkins and Watson (1962) found what they thought was the formation of an "outcast" group in the red grouse. Many aspects of the behavior of red grouse are comparable to those of willow ptarmigan at Chilkat Pass, but differences clearly exist and these become important points of comparison. In winter, red grouse in Scotland do not experience regular snow cover, and they do not migrate. This means that events can occur in the population during winter that are directly related to the quality and size of the breeding habitat. Competition and discrimination that decide which birds can and cannot go to various pieces of ground can occur in the fall and winter. The researchers found that by spring only birds making up the breeding population remained; those discriminated against socially had apparently disappeared. In contrast, the social behavior of willow ptarmigan at Chilkat Pass in winter appeared to relate best to survival. Flocking was essential. Males did tend to remain near the breeding range, and females sought better escape cover farther away, but there was no evidence of a winter hierarchy that might relate to population regulation. Competition in willow ptarmigan at Chilkat Pass did not occur until birds returned to the breeding range in late winter. Events that can occur over several months in the red grouse must happen over a few weeks, or even days, in willow ptarmigan. The red grouse researchers present few demographic data on their out
374
D. H. MOSSOP
cast group; they mention seeing a flock of about 40 on one occasion. I recorded a large number of birds not on territories in May when the rest of the population was strongly territorial. I observed this nonterritorial, waiting flock daily, counted them, banded them, classified them by sex and age, and measured their behavior. There is no support for the idea of a simple elimination of surpluses resulting in direct population control. Males in the waiting flocks maintained weight in the spring, whereas the males defending territories lost weight. This is in contrast to findings for red grouse, where the "eliminated" birds weighed less than the territorial ones and apparently starved (Jenkins et al. 1967). When I presented the waiting birds the opportunity to take optimum spaces, through artificial removal of the territorial birds, birds from the waiting flock did take them. In addition, at least some of these birds sought unoccupied spaces in suboptimum areas to breed. This evidence is compatible with the idea of strong selection in spacing strategies occurring at the peak of the cycle for a particular "type" of bird, which is analogous to its strategy in social competition (Chitty 1967, Bergerud 1970a, Moss et al. 1985). Further evidence of a process more complex than a simple press of numbers and an eliminated surplus was the differences I saw between the sexes. All males of the nonterritorial, waiting flocks, after all the territories of the study area were filled, were yearlings. Further, all replacement males were yearlings when I shot territory holders, and all males found in suboptimum, willow ptarmigan habitat were yearlings as well. Among hens, which returned later than cocks to breeding habitat, all collected in spring had developing follicles. Further, all females, regardless of social status, seemed to maintain weight. Few hens must have found it necessary to enter suboptimum habitats to breed. The nonbreeding, waiting flocks were mostly a male phenomenon. My data on sex ratios among flocks in the spring and on absolute numbers of hens suggest there were fewer yearling hens available as recruits than yearling males in 1972 and 1973, years of decline. Subsequent to the field work of this study, Hannon (1982, 1983, 1984) studied ptarmigan dynamics at Chilkat Pass. She concluded, on the basis of removal experiments in the spring, that females might prevent other females from recruiting to the population, thus regulating numbers. However, there was no evidence that females that filled vacancies might not have bred elsewhere, shifting to moreoptimum areas with respect to breeding options. Her experiments verified spacing behavior but not population regulation (Davies 1978). In addition, the stable population that occurred in her study area in 1980 and 1981 (Hannon 1982, 1983) would not have provided a good opportunity to test hypotheses of cyclic decline.
10.8.3 Population regulation The idea that predation during winter caused the disappearance of sufficient number of birds to explain changes in breeding density was not supported in this study. In 2,000 hours of observations of wintering flocks, I did not see the con-
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
375
centration of predators nor the killing that would be necessary to remove over half the ptarmigan. Through much of the late winter, foxes were territorial. Golden eagles, by my observations, were the most effective predator of flocks of ptarmigan. But they left the area completely in winter. Gyrfalcons also became rare in midwinter, probably because of poor hunting conditions. When they were present, I never observed more than two hunting a particular area, and my sightings were concentrated on three locations, suggesting territorial behavior. Gyrfalcons nest in late winter and are known to defend their territories in winter (Platt 1976, Barichello 1983). Predator switch-over is a possible mechanism for causing ptarmigan cycles. The decline in numbers of one-prey species as predator numbers increase may force predators to switch to other species, driving these latter populations down (Myrberget 1970a,b,c, Chap. 11). Most winter predators at Chilkat Pass do eat hares, and could conceivably switch from hares to ptarmigan if nearby hare populations crashed. The hare population did undergo a crash after 1971, but in almost complete synchrony with ptarmigan. Also, I saw no obvious influx of predators. I conclude that there was no support for a switch-over phenomena at Chilkat Pass. Density of ptarmigan was in a dramatic decline in 1971-74. A simple, functional relationship between available food and ptarmigan numbers would be reflected in increased density as food supply increased. Instead, at Chilkat Pass the opposite occurred; numbers of ptarmigan continued to decline for at least 2 more years in spite of a decreasing demand on winter food (Fig. 10.25). Further, birds gained weight in all winters (Fig. 10.26), even though the number of birds
Fig. 10.25. The ptarmigan population continued to decline in winter 1973-74, even though there were more unbrowsed willow in February and March of that winter than in February and March of the previous one.
376
D. H. MOSSOP
competing for food, food availability, and snow conditions varied among winters (Table 10.1, Fig. 10.25).
Fig. 10.26. Ptarmigan maintained or gained weight in all winters, even though there was a decline in the population from 1971 to 1974. The lowest gain in weight was 1970-71, when the population increased. Weight gain was independent of fall numbers.
SURVIVAL AND STRATEGIES OF WILLOW PTARMIGAN
CHICKS PER BROOD IN THE FALL
/ • • 0.388
377
TERRITORIES (%) PRODUCING YOUNG
MINIMUM CHICKS PER KM2 'N ™E FALL
JUVENILES (%) IN THE FALL POPULATION
r =0.992
/-=0.875
/•-0.9I6
ANNUAL CHANGES IN THE DENSITY OF COCKS BETWEEN YEARS Fig. 10.27. Correlation coefficients of proportional changes in the spring density of territorial males (year 2/year 1) with breeding success parameters in year 1.
The hypothesis that territorial behavior resulted in compensatory mortality can also be rejected. Annual mortality rates were not significantly different between 1972-73 and 1973-74, even though the fall population was much higher in 1972 than in 1973 (Table 10.1). My evidence suggests that all females alive in the spring took part in nesting, regardless of spring densities, and that changes in breeding numbers resulted from prior changes in breeding success (Fig. 10.27).
10.9 Summary Behavioral and population dynamics of ptarmigan at Chilkat Pass, British Columbia, were studied from 1970 to 1974, with an emphasis on winter survival and spring breeding strategies. The population reached a peak in 1971 of 57 territorial males/km2 and then declined to 12 males/km2 by 1976. Breeding success declined from 1970 to 1973, and these annual variations in productivity were positively associated with changes in density of territorial males in each following year. Adult mortality was an estimated 53% in 1972-73, and 61% in 1973-74. The population was hunted by gyrfalcons, colored foxes, and golden eagles. Winter tactics used by ptarmigan to minimize predation risk were: (1) forming flocks, (2) roosting in snow burrows, (3) movement or migration to taller escape cover, and (4) crepuscular and synchronized feeding periods that were brief in duration. In two winters of ptarmigan decline, food supplies were measured and found adequate. Birds maintained or gained weight, and were in good physical condition in winter, even when the population was declining. Social segregation was apparent during the period of competition for breeding space. Resident males, both adults and yearlings, took territories in March; they were joined by adult females and other yearling males in April and early May and by yearling and adult females later in May. In three springs—1972, 1973, and 1974—some birds were
378
D. H. MOSSOP
present in a waiting flock at the same time other birds were on territories. This waiting flock was largest in 1972 when the population was high, and declined as numbers decreased in 1973 and 1974. Males and females were removed from territories, but only 41 % were replaced. Replacement birds came from the waiting flock. In all years the number of birds in the waiting flock declined through spring, and the disappearance of the flock each spring coincided with the start of incubation. In 1972, some yearling males even established territories in rock ptarmigan habitat. It is argued that almost all males still alive obtained territories by the first of June, although many yearlings failed to secure a female. Data also suggest that all females ultimately mated and nested, even though some waited to select sites until June in 1972 and 1973, when breeding numbers were high.
11
Demography of an Island Population of Willow Ptarmigan in Northern Norway S. Myrberget
11.1 Introduction Many boreal and arctic bird and mammal populations show cyclic fluctuations, that is, they exhibit a significant tendency for variations in numbers to be repeated at intervals more regular than expected if occurring by chance (Keith 1963, Angelstam et al. 1985). Population cycles of willow ptarmigan (Lagopus lagopus} at 3- to 4-year intervals in Norway (Hagen 1952b), 6 years in Scotland (Williams 1974), and 10 years in Newfoundland (Bergerud 1971) demonstrate that the length of the interval is characteristic of the region and not of the species. Most Norwegian ptarmigan populations also fluctuate in synchrony with the 3- to 4year cycles of microtine rodents (Myrberget 1982a). The causes of cyclic fluctuations in the densities of tetraonid and microtine populations are thus far poorly understood (Stenseth 1978, Watson & Moss 1979). Whatever the causes, however, cyclic populations must show cyclic variations in reproduction, mortality, emigration, or immigration, or in a combination of these principal factors. A logical first step in the study of population cycles is to examine which demographic factors vary in a cyclic manner. This paper presents data obtained during the period 1960-80 for the willow ptarmigan population breeding on the small, coastal island of Tran0y in northern Norway. This population has shown a 3- to 4- year cycle that is in synchrony with the numbers of microtines there (Myrberget 1984a, b). The major purpose of this paper is to present evidence of which demographic parameters have varied in a cyclic manner. The relationship of the microtine cycle previously described by Myrberget (1982b) to these parameters is discussed, as well as the possible in379
380
S. MYRBERGET
fluence of other extrinsic factors on the cycles of ptarmigan. Also examined is whether variations may have occurred in the production of certain plants that might indicate a cycle in productivity of the ecosystem as a whole, and which in turn may cause the ptarmigan cycles (see Kalela 1962). Jenkins et al. (1963) and Bergerud (1970a), working on ptarmigan, have shown that some demographic parameters may be accompanied by changes in parental behavior; this paper therefore further includes data on breeding behavior of Tran0y birds.
11.2 The study area Tran0y (69°09'N-17°25'E) is located between the large island of Senja and the mainland (Fig. 11.1). It is 127 ha, with the highest point 32.5 m above sea level. The study area also includes three small islets nearby, with a combined area of 10 ha.
Fig. 11.1. Location of the Tranoy Study Area on the coast of northern Norway.
WILLOW PTARMIGAN IN NORTHERN NORWAY
381
Solid rocks here are mainly gneisses and granites (Myrberget 1975a). Because the island was below sea level during glacial periods, the soils in parts of the area are base-rich. Vegetation cover is patchy, and vegetation types are defined as 18% eutrophic, 7% intermediate, and 75% oligotrophic (Myrberget 1975a). The predominant tree is the mountain birch (Betula pubescens), usually less than 5 m in height. There is a 15-ha farm on the island. During 1960-65 it was mainly a dairy farm; from 1966 to 1975 it was not in use, and since 1976 has been maintaining sheep. The domestic animals have had rather notable though variable effects on the vegetation, as have fairly large numbers of moose (Alces alces) during the winters of 1966-75 and of mountain hares (Lepus timidus) during 1972-76. Most willow ptarmigan leave Tran0y around 1 September to winter on Senja (Myrberget 1975b). On Senja the shooting season lasts from 15 September to the end of February. Ptarmigan return to Tranoy in March and April, and territories of males are fully established by 1 May. Most or all surviving adults that had nested on Tran0y return there to breed, whereas first-time breeders may have hatched on Tran0y or on Senja (Myrberget unpubl. data). Egg-laying starts approximately 15 May, and most clutches hatch about 23 June (Myrberget 1972). The most common microtine rodent on Tranoy is the field vole (Microtus agrestis) (Myrberget 1982b). Characteristically, these rodents "crash," and the year in which they do so is defined as one that in May there are few field signs of live microtines, following a year with many signs. The year before a crash is termed the "peak" year. These rodent years are easily ranked in this fashion. Following the crash is a "postcrash" year and the year preceding the peak is a "prepeak" year. In 1959, 1962, 1966, 1970, and 1974 voles peaked. On nearby areas, including Senja, the vole population peaked in 1963 instead of 1962. All other peaks occurred in the same years on Tran0y and Senja (Myrberget 1982b). Although it was not evident from field sign on Tran0y, 1978 was a rodent peak year in many areas including parts of Senja (Myrberget 1982b), and is included among the peak years in this study. Highest vole numbers were usually evident in the autumns of peak years. In 1969, however, autumn vole numbers were probably higher than those in the following peak year. In 1969-70, but in none of the other years, large numbers of Norwegian lemmings (Lemmus lemmus) occurred on Tran0y.
11.3 Materials and methods 11.3.1 Collection of population data Population censuses were made in May and early June (breeding numbers assigned to 1 May) by charting the locations of territorial males, using the pointingdog method (Myrberget 1976a). Birds found dead were classified as having been
382
S. MYRBERGET
killed by predatory birds, "by accident," or from "disease." Whenever possible, the age of these (hereafter "yearling" or "old") was determined from the primaries (Bergerud et al. 1963), and the sex from feather color. A total of 604 nests were located, partly by using dogs. Nest cover was classified subjectively into five categories (from "very poor" to "excellent"). During the latter half of the incubation period, the distance between a human intruder (observer) and the nest was measured at the moment the incubating hen left the nest. Later, incubating hens were caught in nets, weighed, and their age determined (yearling or old). Nests were visited regularly during the last week of incubation to establish hatching dates and hatching results. Visits were made mostly at night, to avoid leading egg predators to nests. The starting date of egg-laying was calculated by assuming an interval of 1.1 days between each egg laid, and an incubation period of 21 days (Westerskov 1956, Jenkins et al. 1963). Newly hatched chicks were weighed collectively in each of 211 nests. With the aid of dogs and until about 25 July, broods were located and the numbers of chicks were counted. In total, about two thousand such observations were made. 11.3.2 Definition of parameters Parameters used in this paper are defined as follows. Male number. Actual census estimates plus birds found dead after 1 May. Female number. Refers to the number assigned to 15 May. Breeding age distribution. An equal age distribution for hens and cocks was assumed. Summer adult mortality. Number found dead from natural causes from May to July. Autumn adult number. Breeding numbers minus the summer mortality (as above), and minus a small summer loss caused by the study. Winter adult mortality. Autumn adult numbers minus the number of old birds recorded in May the following year. Nesting failure. Number of birds that left the area in May, before egg-laying. Egg-laying date index. Annual mean of egg-laying dates for those nests that started within a week of the medium date (defined as the medium date of the five successive dates when the highest number of hens started to lay). Clutch size. Annual mean numbers of eggs in nests where egg-laying started within a week of the medium date (as above) in order to exclude renests after the loss of initial clutches (Myrberget et al. 1985).
WILLOW PTARMIGAN IN NORTHERN NORWAY
383
Egg loss. Percentage of eggs that disappeared during incubation, multiplied by 1.5 (a factor based on data from 1974-78, Myrberget et al. 1982). This factor considers the losses during laying and incubation, as well as compensation of losses by renesting. Hatching failure. Percentage of eggs in the nests at hatching time that did not result in live chicks. Chick mortality. Percent loss between hatching time and an age of 3-4 weeks. The values for 1960-73 are based on the average number per brood at hatching and when 4 weeks old. Those for 1974-80 are the losses recorded for certain known broods. Juvenile number. Number of eggs laid (number of hens x clutch size) minus all losses of eggs and of chicks up to an age of 4 weeks. Juvenile rate. Percent juveniles in the population at the end of July. Based on autumn numbers and numbers of juveniles (as above). Juvenile winter loss. Number of juveniles minus the number of yearlings breeding in May of the following year. Population change index. Number of ptarmigan in any year expressed as a percentage of the numbers during the same season in the preceding year.
11.3.3 Estimates of plant production Female and male catkins were counted in late May and early June from 1964 to 1971 on 50 mountain birch trees, each about 2 m in height. Also, the width of annual rings for mountain birch was studied in six cross-section slices collected in 1971 and in ten slices from 1981. Three such samples each were also collected in 1971 from Pinus sylvestris, Picea abies, Populus tremula, Salix caprea, Sorbus aucuparia, and Alnus incana (18 total). Data on the berry production of bilberries (Vaccinium myrtillus) were obtained from reports collected in the autumn from Municipal Wildlife Boards in the county of Troms. Reports stating "high" production were given the value 2, "medium" 1, and "low" 0. Berry production indexes are the annual means of the above values for all localities. Data were also obtained from Steen (1978) on annual hay production in relation to the average value in the Tran0y municipality on Senja.
11.3.4 Adult behavior Parental behavior was studied during incubation and brood-rearing periods. The index during incubation refers to the percentage of incubating hens that flushed at distances of more than 0.67 m from the observers (1965-72). Brood-rearing
384
S. MYRBERGET
indexes are the annual means of the closest distance that the male or female parent, with chicks 15-30 days of age, was attracted by tape recordings of chick distress calls (1964-69). If the bird was only flushed and not seen afterward, the distance was recorded as 50 m. 11.4 General features of population variations 11.4.1 Breeding population The number of breeding birds averaged 101 (Table 11.1). The most striking feature in the numbers recorded is the decline in 1971 and the reduced density thereafter (Fig. 11.2). The mean for springs 1960-70 was 143 birds, and for the period 1972-80 it was 52 (Table 11.2). Table 11.1. Demographic parameters (see 11.3 Materials and methods) for willow ptarmigan on Tran0y, 1960-80, and their means and standard errors, range of values, and coefficients of variation (standard deviation as a percentage of the mean value) for a number of years («) Parameter
n
Breeding numbers Yearling hens (%) Adult summer mortality (%) Juvenile autumn numbers Juvenile rate (%) Total autumn numbers Juvenile winter loss (%) Adult winter mortality (%)
21 19 21 21 21 21 19 19
Mean + SE 101 42 2.1 152 58 252 68 46
± ± + + + + + ±
Range 27 18 0 9 28 43 42 29
12 3 0.6 22 3 31 3 3
-236 - 63 -9.2 -400 - 76 - 554 - 84 - 66
CV 54 31 124 67 24 56 18 24
Table 11.2. Demographic parameters for a number of years (n) within the periods 1960-70 and 1972-80. The significance of the differences as revealed by variance analyses (F-values) is given. 1972-80
1960-70 Parameters
n
Mean ± SE
n
Mean + SE F-value
Breeding numbers Yearling hens (%) Adult summer mortality (%) Juvenile autumn numbers Juvenile rate (%) Juvenile winter loss (%) Adult winter mortality (%) a
P < 0.001.
11 9 11 11 11 10 10
143 43 1.5 229 62 71 44
± ± ± ± ±
11 5 0.7 25 3 ± 3 ± 3
9 9 9 9 9 8 8
52 39 2.8 69 57 63 46
± 6 ± 4 ± 1 ± 11 ± 5 ± 5 ± 4
44.,7a 0..5 1..3 30..5 a 0..6 1,.6 0..2
WILLOW PTARMIGAN IN NORTHERN NORWAY
385
Fig. 11.2. Number of breeding ptarmigan on Tran0y 1960-80, number of juveniles in autumn and their percentage of total numbers, and percentage of yearlings among breeding hens (sample size above the curve). Vole densities are indicated at the bottom: black circles —high densities; open circles —increasing numbers; triangles —vole "peak" years in nearby areas, but not on Tran0y.
On average, willow ptarmigan numbers increased in rodent crash years and decreased in postcrash years (Table 11.3). One notable exception was in 1979, when numbers decreased during a vole crash. This occurred in only parts of the study area, however, and in others numbers increased slightly according to the previous trend (Myrberget 1983a).
386
S. MYRBERGET
Most cocks had a single mate, although in some years a few cocks had two mates and some cocks had no mates at all. No cyclic variation was found in spring sex composition (Myrberget 1984a). Of breeding females, on average 42% were yearlings (Table 11.1). Yearling rates were particularly low in vole postcrash years (Table 11.3), except in 1964 (Fig. 11.2).
11.4.2 Adult summer mortality The number of adults found dead in May through July accounted for, on average, 2.1% of breeding numbers (Table 11.1). Annual variations in mortality rate showed no significant relationship to the microtine cycle (Table 11.3). During the study 40 dead birds were found; 30% of these (12/40) were females. Two birds had flown against wires, and two were classified as "diseased." The rest (36) had been killed by predatory birds, mainly goshawks (Accipiter gentilis). The nearest goshawk nests were located 5-10 km away on Senja, and predation of ptarmigan on Traney may vary according to the extent to which breeding goshawks hunt in the area. However, in 1970 a female ptarmigan with a brood was killed by a nesting short-eared owl (Asio flammeus) (Myrberget & Aandahl 1976). In 1964, when predation was high (Fig. 11.3), the island was visited daily from May through June by goshawks, which were probably nesting some 10 km away (Myrberget 1970a). Goshawks also were often seen on Tranoy in 1975, another year of high predation. Table 11.3. Demographic factors in relation to vole cycles. Vole levels are divided into four phases: peak, crash, postcrash, and prepeak years. Variance analyses were made for the data for all groups compared with each other (F-values shown). Parameters Number of years3 Breeding pop. change Yearling hens (%) Adult summer mortality (%) Juvenile rate (%) Autumn pop. change Juvenile winter loss (%) Adult winter mortality (%) a
b
c d
Peak
Crash
Postcrash
Prepeak
F-value
5 101 45 3 67 100 66 43
6 117b 51 3 43d 65b 65 51
6 70C 29C 2 61 120 66 45
4 100 44 2 68 129 71 46
4.25b 4.00b 0.25 8.27° 1.29 0.96 0.42
No data exists for "breeding pop. change" in 1960 (a vole crash year); nor for "Yearling hens" in 1960 and 1961 (crash and postcrash years); nor for winter losses in 1960-61 and 1980-81 (crash and post crash years). 0.05 > P > 0.01. 0.01 > P > 0.001. P < 0.001.
WILLOW PTARMIGAN IN NORTHERN NORWAY
387
Fig. 11.3. Summer and winter mortality rates (%) of adults and winter loss (%) of juveniles (vole densities indicated as in Fig. 11.2).
11.4.3 Production and autumn numbers The mean juvenile rate (calculated % juveniles in the population at the end of July) was 58% (Table 11.1). No significant difference was found between the rates for 1960-70 and 1972-80 (Table 11.2). Rates were significantly lower in microtine crash years than in other years (Table 11.3). This held for all the crash years except 1960 and 1963 (Fig. 11.2). The annual number of juveniles averaged 152 (Table 11.1), with distinctly
388
S. MYRBERGET
lower numbers in 1972-80 than in 1960-70 (Table 11.2). The values peaked markedly in 1966 (microtine peak year), and in 1969 (prepeak year). For other years it cannot be said whether the ptarmigan peak occurred in rodent peak or prepeak years. The mean number of birds in autumn was 252 (Table 11.1). There were significantly fewer birds in autumn in microtine crash years (Table 11.3). Further, indexes of autumn population change and the juvenile rates for the same years were significantly related (r = 0.676, P = 0.01).
11.4.4 Winter losses The mean annual mortality rate of adults from August to April was 46%, and the loss of juveniles was 68 % (Table 11.1). No significant difference existed between the mean rates for 1960-70 and 1972-80 (Table 11.2). Also, no significant relationship was found between these winter losses and the phases of the vole cycle (Table 11.3); nor was there a significant correlation between autumn numbers and winter loss rates (for adults r = 0.060, for juveniles r = —0.333). The highest winter loss of adults (66%) was recorded for the winter of 1971-72 (Fig. 11.3). However, this value includes all birds that left Tran0y in May 1971 and did not return there to breed in 1972 (see 11.5.2 Nesting failure). The mean number of remains of birds that had died on Tran0y in the autumn constituted, on the average, less than 1 % of the autumn population. The value for winter losses also includes any birds found dead in the spring up to 1 May. These early spring losses represented on average 4.2% of the adult population (dead birds plus breeding numbers). The corresponding value for yearlings was 6.7%. Predatory birds, among them goshawks, were the main cause of mortality in early spring. The remaining winter losses occurred at some time after the birds left Tran0y in autumn, and before they returned in March and April. Except for those shot in autumn on Senja, about 5% of autumn numbers (Myrberget 1975b), the timing and causes of these losses are unknown.
11.4.5 Synthesis of changes in numbers The highest number of breeding ptarmigan recorded on Tranoy (236) corresponds to 85 pairs/km2. This exceeds the densities recorded for all other Lagopus populations studied in their natural habitats, although more than 100 territorial red grouse (L. lagopus scoticus) cocks/km2 have been found on managed heaths in Scotland (Watson & Moss 1980). Thus the present study, like most other studies on tetraonids (Watson & Moss 1979), has been conducted on a dense population. Spring and summer losses of adults and yearlings account for only a small
WILLOW PTARMIGAN IN NORTHERN NORWAY
389
proportion of the yearly loss of birds. The same has been found true for most other tetraonid species, e.g., for the rock ptarmigan (Lagopus mutus) (Weeden & Theberge 1972), though not for the spruce grouse (Dendragapus canadensis) (Herzog & Boag 1978). On average 58% of the birds in late autumns 1960-80 were juvenile, agreeing with the value of 59% estimated by Myrberget (1974a) from samples of shooting bags from many parts of Norway. However, in Newfoundland, juveniles accounted for as much as 70 % of the birds in bags of willow ptarmigan shot in October (Bergerud 1970a). Winter losses of juveniles were higher than those of adults, agreeing with returns of birds marked on Tran0y, on Senja, and in certain other Norwegian areas (Myrberget 1975b, 1976b), and data on white-tailed ptarmigan (Lagopus leucurus) (Choate 1963b, Braun & Rogers 1971). However, this does not apply to all Lagopus populations (Jenkins et al. 1963, Bergerud 1970a). As found in many other studies on tetraonids (e.g., Jenkins et al. 1964, Keith et al. 1977), predation seemed to be the main proximate cause of mortality of both yearlings and old birds on Tran0y. Changes in breeding numbers from year to year were usually preceded by corresponding variations in the juvenile rate (Fig. 11.4). If the extraordinary year, 1971, is excluded, the linear regression value is not significant (r = 0.538, P > 0.05). However, an equation of the third degree representing a curvilinear relationship fitted the distribution of the data. Using these data, an estimated juvenile rate of 60-65 % gave on average no change in the breeding numbers the following year. During the 1970s the percent change in breeding numbers never exceeded 110, whereas during the 1960s, this index on three occasions exceeded 130. A strong positive correlation was evident between autumn numbers and the number breeding the following spring. This relationship remained positive for both adults and juveniles even when data for the periods 1960-70 and 1971-80 were treated separately (not significant, however, for juveniles 1971-80, P = 0.07) (Fig. 11.5). These relationships suggest that a positive correlation should also exist between the net production (juvenile rate) and the percentage of yearling hens that bred in the following spring. This was true (Fig. 11.6). A similar relationship between autumn and spring total numbers has been found for most other Lagopus populations studied (Bergerud et al. 1985), except for red grouse in some habitats in Scotland (Jenkins et al. 1967). In summary, the following features characterize the 3- to 4-year ptarmigan cycle on Tran0y: (1) low net production in vole crash years, followed by a decrease in breeding numbers in post-crash years, at which time the proportion of yearlings among breeding hens was also low; (2) numbers of juveniles in autumn were highest in either vole prepeak or peak years; (3) high autumn numbers in such years were followed by high numbers of breeding ptarmigan in vole crash years. The time lags are consistent with both the strong positive relationship be-
390
S. MYRBERGET
Fig. 11.4. Juvenile rates (% of juveniles) in autumn in relation to the change in numbers of the breeding population the following years.
tween autumn and spring numbers, and the correlation between recruitment rates and subsequent changes in breeding numbers. No cyclic variations were apparent in winter losses. Variation in the juvenile rate represented the main demographic factor resulting in the cycles, as appeared true for other Norwegian Lagopus populations (Myrberget 1974a). Major questions raised by perusing these data are: (1) The 1971 decline in numbers resulted from an exceptionally poor production rate, but why did breeding numbers not subsequently increase to the same levels as those recorded in the
WILLOW PTARMIGAN IN NORTHERN NORWAY
391
Fig. 11.5. Autumn log numbers in relation to log numbers breeding the following spring.
392
S. MYRBERGET
Fig. 11.6. Relationships between juvenile rates (% of juveniles) in autumn and percentages of yearlings among hens breeding the following spring.
1960s? (2) Do the variations in juvenile rate and in autumn numbers drive the fluctuations in numbers the following spring, or are both the spring and autumn variations caused by the same factors? (3) At what stages of the ptarmigan reproduction cycle are the poor production rate observed in microtine crash years determined, and what causes the high losses in such years?
11.5 Annual variations in production An index of egg-laying date, clutch size, degree of nesting failure, egg loss, hatching failure, and chick mortality were measured. Losses caused by our study are not included, nor are the small, varying production losses owing to deaths of hens during the breeding period.
WILLOW PTARMIGAN IN NORTHERN NORWAY
393
11.5.1 Egg-laying date The mean index date of egg-laying was 23 May. Egg-laying was earliest in 1980 (index value 16 May) and latest in 1971 (2 June) (see Fig. 11.7). Egg-laying date indexes were significantly correlated with the date of snow disappearance at Troms0 in May (80 km NE) (r = 0.612,P = 0.003), but not with breeding numbers (r — - 0.055, P > 0.05) nor with weights of incubating hens (r = 0.091, P > 0.05).
11.5.2 Nesting failure Only in 1971, which was a vole crash year, was nesting failure observed. A census in early May indicated that about 70 pairs were present. By early June, only about 21 hens and 28 cocks remained; the rest had left the area. Two hens and three cocks were taken by predators in May. We found no deserted ptarmigan nests, nor a great number of shells of eggs preyed on, indicating that the birds that disappeared did not lay eggs on Tran0y. At least some hens that left Tran0y in May 1971 returned to breed in 1972. In 1971, 15 of the 21 hens were caught on nests or with broods. In 1972, 12 "old" hens were caught while incubating, but only one had been handled in 1971. Another four hens captured in 1972 had been recorded as breeding on Tran0y at least once during the period 1966-70. It is not known whether the birds that left Tran0y in 1971 bred elsewhere the same year, for example, on the nearby island of Senja. Probably they did not, because a parallel decline in breeding numbers was observed on Senja (casual observations, Olsvik & Olsen 1980). Small flocks of adult birds were seen on Senja in late June and early July 1971. Thus the recorded "nesting failure" probably reflects nonbreeding, a possible cause being the late thaw, the latest in fact for the study period.
11.5.3 Clutch size Average clutch size was 10.1 eggs, with a low coefficient of variation (Table 11.4), and there was no significant difference between the mean clutch for the periods 1960-70 and 1972-80 (Table 11.5). Observed clutch sizes may have been affected by the loss of some eggs before we found the nests, or we may have included some early, second clutches, which usually contain fewer eggs than do initial ones (Parker 1981). However, a significant correlation between mean and maximum clutch sizes in different years (Fig. 11.8) indicates that annual variation in recorded clutch size at least partly reflects a genuine variation in egg-laying ability. No significant relationship existed between clutch-size variation and the vole cycle (Table 11.6). There was a tendency, however, for the number of eggs per nest to be low in vole crash years. Small clutches (mean values less than 10.0
394
S. MYRBERGET
Fig. 11.7. Mean clutch size in different years. Also shown is the annual variation in some dependent variables: egg-laying date index; snow cover (i.e., numbers of days with at least 75% snow cover 1-20 May atTroms0); nest cover (i.e., percentage of nests in the two poorest of five cover classes); and hen weight (i.e., average weights of "old" hens during the last week of incubation). Sample sizes are given above some curves (vole densities as in Fig. 11.2).
WILLOW PTARMIGAN IN NORTHERN NORWAY
395
Table 11.4. Parameters of production (see 11.3 Materials and methods) for a number of years (n) between 1960 and 1980 Parameter
n
Mean + SE
Clutch size Egg loss (%) Hatching failure (%) Chick mortality (%)
21 21 21 21
10.1 28 5.6 50
+ ± ± ±
Range
0.2 4 0.6 3
8.09 0 25 -
12.2 86 11.2 80
CV 11 67 48 28
eggs) occurred in the crash years of 1967, 1971, 1975, and 1979, but not in 1960 and 1963 (Fig. 11.7). The mean exceeded 11.0 eggs in 1960, 1977, 1978, and 1980, each of which represents one of the four phases of the vole cycle. On average, clutch size was small in years with large breeding numbers (Table 11.7). This relationship was more evident when the study period was split (1960-70 taken alone, r= -0.677, P = 0.02; 1972-80, r = - 0.557, P < 0.05). Body weights of incubating, old hens were significantly heavier in years when mean clutch size was large than in other years (Table 11.7). However, this relationship is largely due to the particularly high values for both weight and clutch size in 1977, 1978, and 1980. Especially low breeding numbers were also present during these years (Fig. 11.2). For the period 1965-72, no significant correlation was found between body weight and clutch size (r = 0.236, P > 0.05). Clutch size and age distribution of breeding hens (% yearlings) were not correlated (Table 11.7). This is because, on average, the clutch sizes laid by old hens and by yearlings were similar (yearling hens 10.0 + 0.3 eggs (± 95 % confidence interval); old hens 9.7 ± 0.3 eggs; n = 344, P = 0.26). A strong tendency was found for an inverse relationship between clutch size and egg-laying date (Table 11.7). Three years deviated markedly from this pattern (Fig. 11.9). One was 1967, when clutch size fell far below the regression line, and the number of ptarmigan breeding was the highest for the study period. The other two years were 1977 and 1978, when clutch sizes were higher than expected considering the egg-laying date, and when breeding numbers were low. These exceptions to the general relationship between clutch size and eggTable 11.5. Production parameters for a number of years (n) within the periods 1960-70 and 1972-80 1960-70
1972-80
Parameter
n
Mean + SE
n
Mean + SE
Clutch size Egg loss (%) Hatching failure (%) Chick mortality (%)
11 11 11 11
9.8 24 5.6 47
9 9 9 9
10.6 34 5.3 55
+ 0.8 ± 17 ± 2.8 ± 13
+ 1.2 ± 20 ± 2.7 ± 15
F-value 2.8 1.3 0.1 1.7
396
S. MYRBERGET
Fig. 11.8. Relationships between annual clutch-size means and maximum clutch sizes observed in different years.
Table 11.6. Ptarmigan production parameters in relation to vole cycles Parameter
Peak
Crash
Postcrash
Prepeak
F-value
Clutch size Egg loss (%) Hatching failure (%) Chick mortality (%)
10.4 23 4 43
9.3 42a 6 58
10.2 22 6 54
10.7 21 7 41
1.86 1.83 0.87 2.25
a
0.05 > P > 0.01.
WILLOW PTARMIGAN IN NORTHERN NORWAY
397
Table 11.7. Correlation coefficients found by linear regressions between ptarmigan production variables and some independent parameters (see Fig. 11.7)
Parameter Breeding numbers Yearling hens (%) Egg-laying date Snow cover in May Poor nest cover Weight of "old" hens a b
Clutch size
Egg loss
Hatching failure
Chick mortality
Juvenile rate
-0.462a -0.185 - 0.406b -0.163 -0.306 0.631a
0.024 -0.035 0.104 -0.123 0.300 -0.165
0.067 0.136 0.421 C 0.430d 0.265 0.126
-0.215 -0.071 0.098 -0.300 0.204 0.062
0.046 -0.162
-o.sor 0.057 -0.5003 0.278
0.05 > P > 0.01. P = 6.8%.
c
P = 6.2%
d
P = 6.0%
laying date may thus be partly explained by a density-dependent effect on clutch size (see Fig. 11.10). If data for the year with the highest number of breeding ptarmigan (1967) and the 3 years with the lowest numbers (1978-80) are excluded, a significant correlation between clutch size and egg-laying date is indicated (r = 0.564, 0.05 > P > 0.01). Size of the breeding population, body weights of the hens, and egg-laying date are thus related to clutch size. Two of these are intercorrelated: breeding number and hen body weights (r = 0.545, 0.05 > P > 0.01). Egg loss may also have affected the observed clutch sizes to some extent. The size of the breeding population accounted for 67.5% of the annual variation in clutch size, as indicated by
Fig. 1 1 .9. Relationship between egg-laying date indexes and clutch size values for different years.
398
S. MYRBERGET
Fig. 11.10. Clutch size values in relation to numbers of breeding birds in different years. For 1971, the number of birds both before and after the nesting failure is given (the former value was used in the regression analysis).
a multiple regression analysis of the 13 years of adequate data. An additional 8.2% was explained by hen body weight, 6.6% by egg-laying date index, 5.3% by percentage of yearling hens, and 0.7% by rate of egg loss (total 88.3 %). When calculations are based on data for the 21 years, only 35.1 % of the variation in clutch size was explained by breeding number, an additional 19.3 % by egg-laying date, and 11.2% by egg loss (total 65.6%).
11.5.4 Loss of eggs Egg loss averaged 28%, and losses varied greatly from year to year (Table 11.4, Fig. 11.11). The most probable cause of egg loss was predation, but some eggs may have been eaten by the ptarmigan hens (Jenkins et al. 1963, Allen & Parker 1977). Egg predators on Tran0y are corvids, gulls, and stoats (Mustela ermined). In most years the principal egg predator was the hooded crow (Corvus corone cornix), especially territorial pairs (Myrberget et al. 1976, Erikstad et al. 1981). The number of hooded crows present ranged from one to four pairs; the variation probably resulted from changes in farming practices (Myrberget 1982c), and it was unrelated to the vole cycle. No correlation was found between annual number of breeding crows and egg loss rates (r = 0.099, P > 0.05).
WILLOW PTARMIGAN IN NORTHERN NORWAY
399
Fig. 11.11. Annual values for egg loss, hatching failure, and chick mortality. Figures give numbers of nests examined (vole densities as in Fig. 11.2).
On average, losses were high in vole crash years (Table 11.6). This was true for 1963, 1967, and 1979, but not for 1960 and 1971 (Fig. 11.11). Egg loss in 1975, a vole crash year, was average, but it may have been affected by an experiment that year in which egg predators were deliberately killed (Myrberget et al. 1976). The proximate causes of egg loss in the 3 vole crash years, when losses were particularly high, differ. In most years, either no stoats were present in the study area, or they were observed only occasionally. In 1967, many stoat-ptarmigan interactions were observed, and we found a large number of eggs that had been
400
S. MYRBERGET
stored by stoats. That year, eight of 61 ptarmigan nests observed each lost a few eggs, and for three nests the loss could be ascribed to stoat predation. Stoats may have taken some ptarmigan eggs in 1963 as well, but the main predators were a pair of nesting magpies (Pica pica). In 1979 the high loss of eggs was caused mainly by hooded crows, and probably also by ravens (Corvus corax) (Myrberget et al. 1981). One might predict egg predation would be higher in years with poor nest cover, which may occur in years with a late snowmelt. However, no significant relationship was found between May snow-cover indexes and those for nest cover (r = - 0.169, P > 0.05). There was a slight tendency for egg losses to be largest in years with relatively poor nest cover (Table 11.7). This was largely due to the poor nest cover in the year with the highest egg loss (1979) (Fig. 11.12). Size of the breeding population had no effect on rate of egg loss (Tables 11.5). Egg loss was high, however, in 1967, when the number of breeding ptarmigan
Fig. 11.12. Relationships between percentages of nests with poor nest cover and egg loss rates.
WILLOW PTARMIGAN IN NORTHERN NORWAY
401
was greatest (Fig. 11.10). That year many ptarmigan nests were on heaths, bogs, and meadows, and egg loss from nests in such habitats was higher than from nests in forests (Myrberget 1980). Egg predation thus varied from year to year, with particularly high losses of eggs in 3 crash years for voles. The presence of stoats in 1 year resulted in higher losses of eggs than usual. Nest-cover quality had little effect on egg-loss rate, but the distribution of the nesting habitats chosen may have had some effect on eggpredation rates.
11.5.5. Failure of eggs to hatch An average of 5.5 % of the eggs at hatching time did not hatch (Table 11.4). This value includes eggs destroyed by the hen, 0.9%; infertile eggs, 3.3%; dead embryos, 1.4%; and chicks found dead in nests, 0.5%. Mostly, hatching losses were one to three eggs in specific nests. In four clutches, two of which probably belonged to the same parents, all the eggs were addled. Two other clutches were flooded following rain, and in one nest placed on a bog most of the embryos died. Annual variation in hatching failure had no relationship to the vole cycle (Table 11.6, Fig. 11.11). Nor was any significant correlation found between hatching failure and the independent variables included in Table 11.7. There was a strong tendency, however, for hatching failure to be large in years with much snow remaining in May and delayed egg-laying.
11.5.6 Chick mortality About half the chicks disappeared before they were 4 weeks old (Table 11.4). This is a maximum value, because some live chicks were certainly overlooked. Some may have been killed by predators, by ravens for example (Erikstad 1979), but the main causes of chick mortality are unknown. Annual variation in chick mortality was not significantly correlated with the vole cycle, but there was a tendency for losses to be high in vole crash and postcrash years (Table 11.6). Losses were particularly high in the vole crash years of 1960, 1975, and 1979 (Fig. 11.11). No significant correlations were found between chick mortality rates and the independent parameters included in Table 11.7. Nor was there any significant difference between mean chick mortality for the two periods 1960-70 and 1972-80 (Table 11.5). For 7 years (1964-69 and 1975) Myrberget et al. (1977) found that the annual survival of chicks was significantly correlated with the mean temperature before hatching, that is, the period 1-15 June (r = 0.65, 0.01 > P > 0.001). No significant relationship was found between the chick survival rate and the mean ambient temperature during the 10 days following the average hatching date (r =
402
S. MYRBERGET
0.16, P > 0.05). Further, the annual mean weight of chicks at hatching and the recorded annual mortality rate were unrelated (r = 0.037, P > 0.05). For the same 7 years, a highly significant relation was found between the annual chick survival rate and chick body weight at an age of 2-3 weeks (r = 0.943, P < 0.001). I suggested that the annual variation in chick survival rate and chick body weight were dependent on the same causal factor, i.e., the quality and/or quantity of food available for newly hatched chicks (Myrberget 1981). The first days after hatching food consists of insects and reproductive parts of plants. Erikstad (1982) presented data from 6 years (1971 and 1974-78) on the size of insect populations on Tran0y during the first few weeks after the average hatching date of ptarmigan chicks. He found that a positive correlation existed between the annual index for the dry weights of insects and the mean body weight of chicks 6 days old. However, abundance of insects and chick mortality rates recorded in the same years were not significantly related (Fig. 11.13). Chick mortality was particularly high in 1979 (Fig. 11.11), and plenty of their preferred insect food (lepidopteran larvae) was available (Myrberget 1981). An abundance of insects is thus not the only factor that determines the chances of survival for ptarmigan chicks.
Fig. 11.13. Relationships between annual insect biomass (mg) caught in sweep nets (from Erikstad 1982) and chick mortality rates.
WILLOW PTARMIGAN IN NORTHERN NORWAY
403
11.5.7 Intercorrelation between productivity parameters No significant intercorrelation was found between the various demographic factors representing different phases of the breeding period (Table 11.8). Nesting failure data were excluded from the analyses because they represent only 1971. Nesting failure in other years was negligible. In 1971 the mean clutch size was small and the hatching failure rate was high. Also, egg-loss and chick-mortality rates in this year were lower than average. Similar analyses were made for the periods 1960-70 and 1971-80 separately (unpubl. data). In 1960-70, significant correlations were found between clutch sizes and egg losses (r = -0.749, P = 0.008) and between clutch sizes and hatching failure rates (r = -0.666, P = 0.025). Annual variation in autumn juvenile rate correlated best with variation in clutch size, but relationships between juvenile rate and both egg loss and chick mortality were also significant (Table 11.8). These relationships were good when the data were treated separately for the two periods (unpubl. data), although the correlation between the juvenile rate and chick mortality was significant only for the period 1960-70 (r = 0.811, P = 0.002). The relationship between clutch sizes and juvenile rates is mostly a result of particularly low values for both parameters in the vole crash years 1967, 1971, 1975, and 1979 (Fig. 11.14). Small clutch sizes in these years were not, however, the main cause of the reduced juvenile rates. Clutch sizes were on average 86% of the mean, whereas juvenile rates in these years were only 61 % of the study period mean. An estimated 52% of the variation in juvenile rate could be explained by variations in clutch sizes, another 19% by variations in chick mortality, 8% by egg loss, and 2% by hatching failure (total 81.4%, using multiple regression analysis). The remaining variation is largely due to nesting failure and loss of eggs and juveniles, as a result of our study (e.g., Myrberget 1983b).
Table 11.8. Correlation coefficients for linear regressions between production parameters
Parameter Chick mortality Hatching failure Egg loss Clutch size a
b c
0.05 > P > 0.01. 0.01 > P > 0.001. P < 0.001.
Egg loss
-0.358
Hatching failure
Chick mortality
Juvenile rate
-0.006 -0.263
0.037 0.312 -0.147
-0.5403 -0.286 -0.635b 0.718°
404
S. MYRBERGET
Fig. 11.14. Autumn juvenile rates (% of juveniles) in relation to clutch size in different years.
11.5.8 Comparison of ptarmigan production and plant growth None of the plant-growth parameters studied varied significantly in relation to the vole cycle on Tran0y (unpubl. data). Hay production by farmers, however, was significantly lower in vole crash years than in other years (0.05 > P > 0.01). Similarly, the annual increment of birch-ring widths was low in all the vole crash years after 1965 (Fig. 11.15). Peak catkin numbers occurred at 3-year intervals, i.e., in 1964, 1967, and 1970. However, the annual numbers of female and male catkins on birches were not significantly correlated (r = 0.468, P > 0.05). Hay production was best related to ptarmigan breeding success. It was significantly correlated with clutch size, hatching failure, and the juvenile rate (Table 11.9). Hay production and the nest cover index also were significantly correlated (r = 0.626, P = 0.022). Egg losses and annual birch-ring growth were inversely correlated (Table 11.9), largely because egg losses were particularly high and birch growth was particularly poor in 2 vole crash years (1967 and 1979) (Figs. 11.11, 11.15).
11.5.9 Hen behavior and production Attraction distances for females and males with broods were closely correlated (r = 0.939, P = 0.0005 (Fig. 11.16). For this reason, only data from hens will
WILLOW PTARMIGAN IN NORTHERN NORWAY
405
Fig. 11.15. Annual variations in plant production (see section 11.3 Materials and methods) (vole densities as in Fig. 11.2).
be referred to here. Attraction distances tended to be correlated with flushing distances recorded during the incubation period (r = 0.821, P = 0.089). For both parameters high individual scores are interpreted as suggesting low breeding effort. Values for the flushing distance during incubation varied significantly in
406
S. MYRBERGET
Table 11.9. Correlation coefficients for regressions between plant variables and ptarmigan production parameters
Variable Male catkins Female catkins Birch rings Blueberry index Hay production a b
Clutch size
Egg loss
Hatching failure
Chick mortality
Juvenile rate
0.089 -0.126 0.173 0.339 0.781b
-0.172 0.546 -0.434a -0.144 -0.416
-0.075 0.264 -0.146 0.058 -0.6043
0.686a 0.284 0.208 0.083 0.250
0.117 -0.213 0.261 0.073 0.721 b
0.05 > P > 0.01. 0.01 > P > 0.001.
relation to the vole cycle (Table 11.10, Fig. 11.16). Particularly long distances were recorded in vole crash years. The same tendency was found for attraction
Fig. 11.16. Annual variations in indexes of parental behavior during incubation and with broods (defined in section 11.3 Materials and methods).
WILLOW PTARMIGAN IN NORTHERN NORWAY
407
Table 11.10. Female behavior scores (mean values) in relation to vole cycle. Numbers of years in parentheses. Behavior
Peak
Crash
Postcrash
Prepeak
F-value
During incubation 3 At broodsb
14(3) 22(1)
47(4) 41 (1)
21 (5) 30(2)
16(4) 26(2)
6.8C 1.9
a b
c
Percentage of hens that flushed at distances longer than 0 . 6 m . Attraction distances in meters, i.e., the closest distance that hens approached an observer who played a chick-distress recording. 0.05 > P > 0.01.
distances for females with broods, but the data series included only a single vole crash year. In the years in which high values were recorded for flushing distance during incubation, egg-laying was later than usual (r = 0.578, P = 0.049) and hay production was low (r = 0.867, P = 0.026). Also in these years, clutch sizes and juvenile rates were low, and the hatching failure rate was high (Table 11.11). For old hens, but not for both age classes combined, a significant relationship was found between indexes for nest cover and flushing distance (r = 0.629, P = 0.045).
11.5.10 Synthesis of productivity Nesting failure, probably owing to nonbreeding, occurred once during the 1960-80 study period, in 1971, a marked year of decline in the ptarmigan population. Red grouse, too, have been known to leave their breeding area just before egg-laying during a decline (Watson & Moss 1980). The rate of hatching failure was always low, as found in other studies of Table 11.11. Correlation coefficients for female behavior indexes and ptarmigan production parameters Parameter Clutch size Egg loss Hatching failure Chick mortality Juvenile rate 3 b c
d
See Table 11.10 for explanation of index. 0.05 > P > 0.01. P = 5.3%. 0.01 > P > 0.001.
Incubating 3
At brood3
-0.598" 0.211 0.713d -0.163 -0.766d
-0.806C 0.748 0.857b 0.092 -0.774
408
S, MYRBERGET
Table 11.12. Variations of some summer production losses in rodent crash years Production losses Nonbreeding Clutch size Egg predation Chick mortality a
1960
1963
=a -
=
++
+++ +
1967
1971
1975
1979
=+ ++ +++ -
+++ +
= + + +++
= = +++ ++ +
-
= means that losses do no deviate more than 10% from the adverage loss; + indicates higher losses; — indicates lower losses. One symbol indicates deviation from mean of 11-20%; two, 21-50%; and three, more than 50%.
tetraonids (Bendell 1972, Voronin 1978). In fact, the observed hatching failure in some nests may have been because the hens were captured just before the eggs hatched. Clutch size was the only parameter of production that showed a significant correlation with breeding density. Of the different production parameters clutch size was best synchronized with the variation in juvenile rates. Variations in the juvenile rate were also significantly related to annual variations in egg losses (probably largely a result of predation) and in chick mortality (largely of unknown causes). Juvenile rates in autumn were markedly below average in most vole crash years. That the production parameter that contributed most to the high losses recorded in vole crash years varied from cycle to cycle is shown (Table 11.12). In 1 crash year nesting failure was high, in 3 crash years clutch size was particularly low, in 4 years egg-predation rates were high, and in 4 years chick mortality was high. On average, hay production was poor in vole crash years, and the parental behavior shown by ptarmigan hens could be characterized as "poor."
11.6 Hypotheses to explain changes in production 11.6.1 Clutch-size variation Of the different productivity parameters studied, the annual variation in clutch sizes gave the best correlation with the juvenile rate (r2 = 0.52). As suggested for many other bird species, clutch sizes of tetraonids may be affected by the quantity and/or quality of the hen's food before and during the egg-laying period (Lack 1966, Savory 1975, Beckerton & Middleton 1982). This egg-laying ability hypothesis may explain why clutch size tends to be low in years in which egglaying occurs late (Myrberget 1975c, Erikstad et al. 1986). In spring, there is a gradual shift in the diet of ptarmigan on Tran0y from woody plants to evergreen heath species (mainly bilberry stems) and finally to herbs (Blom 1980). Availability of the plants forming the ground layer to tetraonids, and their chemical composition, are affected by the date of snowmelt
WILLOW PTARMIGAN IN NORTHERN NORWAY
409
and by the weather in the spring (e.g., Siivonen 1957), factors that vary in a random manner. In arctic areas, however, cyclic variations in plant production have been observed, caused by herbivore-plant interactions (e.g., Pitelka 1958) or primary cycles of plant productivity (Kalela 1962). Herbivore-plant interactions may induce cyclic changes in the quantity of plant food available, and may be the result of overgrazing caused by high microtine populations (Pitelka 1958). Cyclic changes in plant quality may also occur, according to the nutrient-recovery hypothesis (Schultz 1964) or as a result of changes in the plant's chemical defense mechanism (Haukioja et al. 1983). The very poor production of ptarmigan on Tran0y in 1979, a year in which the effects of vole grazing there and on the nearby island of Senja were either absent or only slight, could not, however, have been caused by the effects of voles on either plant quantity or quality. Kalela (1962) considered the cyclic changes in the population levels of some herbivorous animals to be a result of cyclic changes in the production of their food plants. In arctic areas, a succession of favorable growth periods are required before a plant is able to flower and seed. Cycles in plant production may occur partly as a result of inherent rhythms in plant growth, and partly by the weather conditions. These cycles have been recorded for plants in the ground layer in some northern areas in Fennoscandia (e.g.. Tast & Kalela 1971, Laine & Henttonen 1983). The data suggest that the basic rhythm is largely unaffected by microtine grazing (Hansson 1982). The possible occurrence of these cycles in the foods of ptarmigan on Traney has not been studied, with the exception of the "mast year phenomenon" (Harper 1977) indicated in the production of catkins by mountain birch. However, no increase in the mean clutch size of ptarmigan was recorded in "mast years." No data exist that would indicate that the annual variation in ptarmigan clutch size (and juvenile rates) has been caused by primary cycles in the productivity of ptarmigan food plants. However, such a plant cycle may still have existed. Blom's (1980) results from a study of three successive springs on Tranoy (1976-78) do not support a hypothesis that the variation in the food supply of the hens is the main cause of the annual variation in ptarmigan production: In 1976, the diet of hens in May was of high quality (protein) but the clutch-size value was only average. In 1978, dietary quality was much poorer, but clutch sizes were by far the highest recorded during the 3 study years. Blom, however, did not study the diet in any of the years in which the clutch size values were extremely low, and his series includes the years 1977 and 1978, both years in which clutch size may have been partially density-dependent. His results, therefore, may not be representative for the entire study period. Furthermore, an important variable known to have caused differences in food quality during the period 1976-78 was the varying occurrence of protein-rich birch catkins, which in other years seemed to have had no or only little effect on the clutch sizes. It remains possible that some other aspects of the nutrient quality of the spring growth of the plants, other than
410
S. MYRBERGET
quantity or protein content, may be important for egg production, as Moss et al. (1971) found for red grouse. I conclude that the annual variation in clutch sizes of willow ptarmigan on Tran0y has affected the variation in the subsequent juvenile rate recorded in late summer. Breeding numbers may have affected clutch sizes in years in which extremely high, or low, breeding densities occurred. In other years, the phenology of the food plants in spring seems to be a major factor in determining clutch sizes. The results do not, however, explain why particularly low clutch sizes were generally recorded in vole crash years.
11.6.2 Predation On Tran0y, egg predation had a minor effect on the recorded clutch sizes. The main effect of summer predation on annual variations in ptarmigan production was the loss of many clutches during the incubation period, and probably also during egg-laying in some vole crash years. Annual variation in chick mortality could not be ascribed with any certainty to the effect of predation. Predation of adult hens in summer was too low to have had any great effect on juvenile production. High rates of egg predation in microtine crash years have also been observed in other Norwegian studies on the willow ptarmigan (e.g., Myrberget 1975c). Such annual variation in egg losses conforms with Hagen's (1952b) hypothesis, which suggests that high numbers of microtines may lead to an increase in the numbers of certain predators. When microtine populations begin to decline, many predators remain. These may then switch to alternative food items, such as bird eggs (Hagen 1952, Angelstam et al. 1984). This hypothesis involves a variation both in the number of predators and in their functional response to a varying food situation. On Tran0y, however, the numbers of egg-eating animals varied markedly only during one of the vole cycles, that is, in the 1967 vole crash, when a number of stoats arrived in the winter or spring and caused high losses of the eggs of ptarmigan and of other birds (Myrberget 1970b). This is consistent with Hagen's hypothesis. Also, during the 1963 vole crash, some of the relatively high egg loss was probably caused by invading stoats. In most years the main egg predators on Tranoy were corvids, and their numbers did not vary with the vole cycle. No data exist on the quantity or availability of the food of hooded crows in different years, nor on the ability of ptarmigan hens to withstand egg predation by crows. Ptarmigan nests may be more easily detected by predators in some years than in others. However, no significant relationship between egg-predation rates and nest cover was found. The data from 1967 indicate that predation rates may be influenced by differences in nest habitats from year to year, which in turn may be affected by the time of snowmelt (Myrberget 1976c), a random factor.
WILLOW PTARMIGAN IN NORTHERN NORWAY
411
Marcstrom and Hoglund (1980) remarked that the crash in some microtine cycles occurs in autumn or early winter. In such cases the "surplus" of predators may die or move away before the following spring, so that high predation of ptarmigan eggs cannot be expected. This argument may explain the low losses of eggs recorded on Tran0y during the 1971 vole crash, and perhaps in 1960 also (see Myrberget 1982b). I conclude that changes in the egg-predation rates have had a marked effect on the annual juvenile rate of willow ptarmigan on Tran0y. This seems to be true also for highland areas in Sweden (Marcstrom & Hoglund 1980), and for rock ptarmigan in Alaska (Weeden and Theberge 1972). However, the very high egg losses recorded during the 1979 vole crash, when no stoats were living on Tran0y, eludes understanding. With the exception of stoats in specific years, no other mammalian predators that eat both voles and birds (e.g., red foxes, Vulpes vulpes) occur on Tran0y (Myrberget et al. 1985). Therefore, in contrast to the situation in many other areas, no closer relationship is to be expected between vole cycles and the predation rates for ptarmigan eggs and chicks on Tran0y.
11.6.3 Environment of chicks Cold and rainy weather reduce the available time in which small ptarmigan chicks are able to feed, because they are dependent on the hen for warmth (Theberge & West 1973, Pedersen & Steen 1979, Erikstad & Spids0 1982, Erikstad & Andersen 1983, J0rgensen & Blix 1985). Cold weather may also reduce the availability of insects that chicks eat (Marcstrom & Hoglund 1980). But no relationship existed between the chick mortality rate and the air temperature just after hatching. This agrees with results obtained for other Lagopus populations (Jenkins et al. 1963, Watson 1965, Bergerud 1970a). In some areas, however, extremely unfavorable weather conditions, such as snow or sleet showers, have resulted in the death of many ptarmigan chicks (Gardarsson 1971, Marcstrom & Hoglund 1980). It has also been suggested that the weather immediately before hatching has a distinct effect on the phenological development of plants and insects, and thus on the food situation of newly hatched chicks, affecting their survival (Slagsvold 1975). On Tran0y, a relationship was found between the weather before hatching and the mortality rate of chicks, in agreement with Slagsvold's hypothesis. However, a close relationship between the amount of insects during the early stages of the life of the chicks and mortality rate was not apparent. In contrast, Erikstad (1985) observed that, in certain years, chicks hatched in poor habitats on Tran0y tended to move farther during the first few weeks of life, and to have a higher mortality rate and a slower rate of body growth than chicks hatched in more fertile habitats. This suggests that the chicks' food, of differing quality or quantity in different habitats, can affect growth rate and survival and that the chicks may
412
S. MYRBERGET
partly compensate for general poor food conditions by habitat selections (see also Steen & Unander 1985). At Troms0, annual variations in the survival of willow ptarmigan chicks kept in captivity were apparently caused by variations in plant phenology, but in such a way that more chicks died when the spring was early and warm, because of an early onset of the lignification of bilberry leaves (Hanssen & Ness 1982). In some years there may also have been a shortage of certain vitamins in the food available to chicks (Hanssen et al. 1979). The amount of flowering of some plants, and hence the quality of chicks' food, may also be affected by rhythms inherent in the plants themselves (Kalela 1962, Laine & Henttonen 1983). I conclude that the phenology of plants and insects in the spring may affect the quality of the food of ptarmigan chicks at Tran0y. The relationship between the spring weather, the chicks' food, and survival rate, however, would seem more complex than suggested by Slagsvold (1975).
11.6.4 Intrinsic phenotypic quality Siivonen (1957) has argued that the hens' diet before laying eggs affects the quality of the eggs, which in turn influences the vitality, and therefore the survival rate, of the chicks. Results of studies of red grouse in Scotland support this hypothesis (Jenkins etal. 1965, Moss etal. 1974, 1975), although conflicting evidence has been presented by Watson and Moss (1980). Annual variations in maternal quality may give significant correlations between clutch size, egg size, egg losses, chick mortality rate, and body weights and behavioral characteristics of hens (Jenkins et al. 1963, Moss et al. 1981). Moss and Watson (1982) found that a better correlation existed between chick survival rate and weight loss of eggs during incubation than with egg size. The results of the Traney study provide no support for the maternal quality hypothesis. This is consistent with the conclusions of Hanssen et al. (1982) based on studies of willow ptarmigan chicks kept in captivity at Troms0. No significant relationship was found between survival rates of chicks and clutch sizes, or with the maternal behavior indexes measured in Tran0y (see also Pedersen & Steen 1985). Similarly, no significant relationship existed between chick mortality and those factors indicative of the spring phenology (snow cover, egg-laying date, hen body weight, and hay production), and thus of the hens' food supply. During the period 1976-78, Blom (1980) found that chick survival was poorest in the year in which the greening of plants was earliest. No significant annual variation has been found in egg size (Myrberget 1977) or in weights of newly hatched chicks (unpubl. data).
11.6.5 Intrinsic genotypic quality Certain demographic characteristics of red grouse, such as clutch size, are partly determined by genotype (Watson & Moss 1982). Different genotypes may be fa-
WILLOW PTARMIGAN IN NORTHERN NORWAY
413
vored by selection when the population is increasing, decreasing, or remaining stable. In a cyclic population this may lead to a cyclic variation in characteristics of behavior and productivity (Chitty 1967, Smith et al. 1975, Charnov & Finerty 1980 Chap. 3, Chap. 10). An annual variation in the genetic composition of some Lagopus populations has been reported (Henderson 1977 Gyllensten 1984). On Tran0y, a correlation was found between clutch size and the juvenile rate. For 13 hens, clutch sizes in both 1966 and 1967 were known. The means were 10.3 eggs in 1966 and 7.9 eggs in 1967, values that lie close to those recorded for the entire population in the same years, that is, 10.1 and 8.0 eggs. Of these 13 hens, one had laid more eggs in 1967 than 1966, and all others had laid fewer eggs in 1967. The clutch size of one hen was greatly reduced in 1967—three eggs compared with nine in 1966, a fact perhaps best explained by partial predation in 1967. The data as a whole, however, indicate that phenotypic and not genotypic factors were the cause of the difference in clutch sizes observed in these 2 years. In no year was a significant difference between the clutch sizes of yearling and of older hens found (unpubl. data), a difference that might indicate that a variation exists in the egg-laying ability of birds reared in different years. Further, in no year were brood-size histograms bimodal (unpubl. data), a feature that may be consistent with Chitty's (1967) hypothesis (see Bergerud 1970a).
11.6.6 Synthesis I conclude that no evidence from the Tran0y data supports either of the two hypotheses that propose the annual variation in either phenotypic or genotypic quality of the population is the main determinant of the variations in the juvenile rate in the autumn. This does not mean, however, that these factors may not have had a minor influence on the annual variation recorded. Any hypothesis that adequately explains the annual variation in productivity of Tran0y willow ptarmigan must also explain why productivity was especially poor in vole crash years. None of the presently available hypotheses that fulfill this condition —e.g., neither the hypothesis of cyclic plant productivity nor the alternative prey hypothesis —would explain fully the observed variation in the juvenile rates in autumn. One major factor that determined productivity of ptarmigan appears to be plant phenology, which affected food of the hen before and during egg-laying, and probably also the quality of the plant food of chicks. However, plant phenology is largely determined by random fluctuations in weather conditions. I find it significant, however, that in some crash years plant (e.g., hay) production was poor, and the annual-ring increments of birch were also small. Hay production can hardly vary cyclically in the way Kalela's (1962) hypothesis implies. The above-mentioned findings indicate that in some vole crash years, weather conditions have been unfavorable for plant growth in general. But why should this occur more often in vole crash years than in other years?
414
S. MYRBERGET
I suggest the following explanation. Kalela's hypothesis is valid both for plantproduction cycles and for vole cycles. This means that vole peak years are most likely to occur in years in which weather conditions are favorable for plant production (early spring and warm summer), or in the second of 2 successive such years. Such a relationship between weather conditions and the occurrence of marked, vole peak years has in fact been reported (Kalela 1962, Myrberget 1974b). If weather conditions vary in a random number, the likelihood of favorable weather occurring in a year following a vole peak year (i.e., in a vole crash year) will then be slight for purely mathematical reasons, because in a series of random, fluctuating numbers, two to three high values seldom occur in succession (Cole 1954). My conclusion is that the poor productivity of ptarmigan in vole crash years occurs because the predator situation, the quality of the food supply, and the weather in the chick period are likely to be unfavorable in such years.
11.7 Hypotheses to explain changes in breeding numbers The hypotheses attempting to explain regulation and limitation of tetraonid breeding densities differ mainly in the emphasis placed on extrinsic versus intrinsic factors.
11.7.1 Winter food Winter food resources may be scarce, leading to a varying but density-dependent mortality rate (Lack 1966). On Tran0y, however, there was no correlation between winter mortality rate and autumn density, and the winter mortality rates were no higher in the 1960s than in the 1970s, although the density of adult birds differed considerably in these two periods. The data do not support a winter-food hypothesis. Most of the adult and juvenile birds died during the winter, before territories were taken in spring. This agrees with Gardarsson's (1971, Chap. 9) results from rock ptarmigan in Iceland, where most of the mortality was caused by gyrfalcons (Falco rusticolus) and in the absence of territorial spacing. However, Gardarrson suggests that in certain years these cyclic ptarmigan could reach higher densities than the limit set by late-winter resources available. If this is true also for the Tran0y population, we would expect winter mortality to have been particularly high in the 1966-67 winter immediately preceding the extremely high breeding densities reached in spring 1967. But this was not the case. In spring 1967, however, the most-preferred food, Salix spp., was markedly overbrowsed (unpubl. data), and many birds placed their nests in seldom-used habitats where egg predation was very high (section 11.5.4). Thus we cannot exclude the assumption that the resources set an absolute limit on the breeding density, a limit seldom reached in the study area.
WILLOW PTARMIGAN IN NORTHERN NORWAY
415
11.7.2 Predation of adults Predation may keep the number below the limit set by food resources, and varying annual pressure of predation may lead to variations in the breeding numbers of tetraonids (e.g., Keith et al. 1977). Summer and winter, adult mortality rates varied from year to year on Tran0y, but there was no significant relationship between spring-to-spring, adult mortality rates and changes in the breeding densities the next spring. Again, no support for the hypothesis is found in the data.
11.7.3 Spacing behavior An important hypothesis to explain regulation of breeding numbers of tetraonids is that there may be variations in spacing behavior from year to year, giving annual variations in size of territories, and hence in breeding densities (Watson & Moss 1979). Both phenotypic and genotypic factors may affect aggressiveness of birds and size of territories (Moss et al. 1974, Moss & Watson 1980). Studies of red grouse have documented that these birds establish territories in autumn and that nonterritorial birds may die (Watson & Jenkins 1968). This hypothesis cannot apply to the Tran0y willow grouse population, because there is no autumn territorial behavior in the area; most birds leave around 1 September (Myrberget 1975b). However, it is possible that spring territorial behavior can limit breeding numbers of tetraonids (Bendell & Zwickel 1978), but the following findings do not support this : (1) most adults seem to die in winter, when birds are on Senja, i.e., before spacing occurs; (2) when the lowest breeding density was reached in 1980, parts of the study area were not used at all for breeding, and the explanation was not a particularly low quality of habitat in these parts that year (Myrberget 1983); and (3) shooting experiments in May did not reveal any surplus population (Blom & Myrberget 1978).
11.7.4 Juvenile production On Tran0y, strong correlations were apparent between the annual rate of juveniles and changes in numbers breeding the following year, and in the percentage of yearlings among breeding hens. Similar relationships have been found in other willow ptarmigan populations (Bergerud 1970a), and in red grouse (Jenkins et al. 1967), rock ptarmigan (Watson 1965, Weeden & Theberge 1972), and white-tailed ptarmigan (May 1975). Following reduction of breeding numbers after the nesting failure in 1971 in Tran0y, breeding densities remained low during the 1970s. There was no significant difference between rates of juveniles in the autumn population, or in winter mortality rates, in the 1960s and 1970s. These findings provide strong evidence supporting the hypothesis that annual variations in productivity give rise to changes in breeding numbers. The data from Tran0y give no support to the various hypotheses that breeding numbers are limited by adult mortality. Rather, it seems that variations from year
416
S. MYRBERGET
to year in the number of juveniles determine changes in breeding numbers. Therefore, the major question is: Which factors determine variations in annual rates of productivity? 11.8 Summary In the studied population of willow ptarmigan, a 3- to 4-year population cycle was observed in the period 1960-80. The demographic parameter varying most consistently with the vole cycle was the percent juvenile ptarmigan in autumn. A schematic presentation of the ptarmigan 4-year cycle on Traney, starting with 101 breeding birds in a vole peak year and finishing with 97 birds in the next peak, is shown (Fig. 11.17). Breeding numbers tended to be highest in vole crash years. Autumn numbers were on average lowest in postcrash years. In this schematic cycle, maximum variation in spring numbers of old birds is 1:1.2, in total breeding numbers 1:1.4, in autumn numbers 1:1.4, and in autumn juvenile rates 1:1.8.
Fig. 11.17. Schematic presentation of the ptarmigan cycle in relation to the vole cycle based on actual numbers and rates from this study. Rates of loss are used to compensate for the unequal number of different types of rodent years.
WILLOW PTARMIGAN IN NORTHERN NORWAY
417
A summary of the most important correlations found in this study is presented (Fig. 11.18). It does not, however, include instances where nonsupportive evidence for various hypotheses was found. It must also be stressed that field data have not been acquired to test all relevant hypotheses concerning the causes of cyclic variations in ptarmigan. Extraordinarily high or low densities seemed to have affected clutch sizes, but no significant density-dependent effect was found for juvenile rates in autumn, or for winter losses of juveniles and adults. Annual variations in spring phenology affected egg-laying dates, clutch sizes, and characteristics of parental behavior. Vole cycles affected numbers of stoats. Thus, in some vole crash years, large numbers of ptarmigan eggs were taken by stoats. But predation was also high in some crash years with no stoats present. The weather immediately before the chicks hatched affected their subsequent rate of survival, probably through the quality of available food. The relationship between the quality of the chicks' food and their survival rate seems to be more complex, however, than suggested by Slagsvold (1975). No support was found for the hypothesis that varying intrinsic quality determines annual variations in production. Juvenile rates in autumn varied in a cyclic manner, i.e., with low rates in vole crash years. Variations in juvenile rates were affected by clutch sizes, rate of egg loss, and chick mortality. But not all these parameters gave low production in all vole crash years. Variations in juvenile rates seemed to determine changes in numbers of breeding birds and in the age composition of the breeding population. I suggest that a modified version of Kalela's (1962) production biology hypothesis combined with the alternative prey hypothesis (including the restrictive remarks given by Marcstrom and Hoglund [1980]) best explains ptarmigan cycles in Scandinavian. However, Kalela's inclusion of an intrinsic plant production cycle is not necessary. The important fact is that marked microtine peaks most often occur in years with favorable weather for plant growth, so that the likelihood of favorable weather is slight in vole crash years. Therefore, the weather, the predator situation, and the quality of the ptarmigan food supply are likely to be unfavorable in vole crash years, producing low ptarmigan juvenile rates in autumn. The hypothesis is consistent with the following observations: (1) The relative importance of proximate causes for poor production of ptarmigan in vole crash years varies from one cycle to another (see Table 11.12). (2) In some vole crash years the extrinsic conditions may be so favorable that there is no poor ptarmigan production (e.g., 1960 on Tran0y). (3) Poor production of ptarmigan may sometimes also occur in other phases of the vole cycle (Marcstrom & Hoglund 1980). (4) Ptarmigan (and microtine) cycles occur in synchrony over large areas of Norway, but many local exceptions to this general pattern have been observed (Myrberget 1974a, 1982a,b). (5) The intervals between peaks in ptarmigan and microtine populations have a variation, from 3 to 5 years in most Norwegian areas (Myrberget 1982a,b, 1984a).
418
S. MYRBERGET
h
Fig. 11.18. Correlation coefficients for the main relationships found in this study. For nest cover/incubation behavior, data for "old" hens are used. *0.05 > P > 0.01; **0.01 > P > 0.001.
WILLOW PTARMIGAN IN NORTHERN NORWAY
419
It must be emphasized, however, that all the conclusions drawn from the Tran0y data are not necessarily representative for Norwegian willow ptarmigan populations as a whole. For instance, the lack of autumn territories recorded in the study area may be a typical factor of northern Lagopus populations (e.g., Unander & Steen 1985), in contrast to more southern ones (Steen et al. 1985). The study area seems to represent an optimal breeding habitat for the species, as indicated by the high densities in certain years. Jenkins et al. (1967), too, have observed that the dynamics of red grouse populations were different for optimal and more-marginal habitats.
This page intentionally left blank
Part II Theory and Synthesis
This page intentionally left blank
12
A Genetic Explanation for Ten-Year Cycles of Grouse R. E. Page and A. T. Bergerud
12.1 Introduction With the exception of a few specific cases, there is still no general answer to the question "Why do populations of animals cease growing and not continue to increase without limit?" (Pielou 1977, Krebs 1978b). Population ecology as science has no universal paradigm under which it operates (Kuhn 1962, Lakatos & Musgrave 1965). Dennis Chitty (1958) proposed an encompassing theory, that all animals have the ability to self-regulate their population levels below resource limits. The theory was intuitively appealing, but initially was not sufficiently precise to provide much direction for research programs, or to allow falsification by empirical testing. Pitelka (1958, p. 247) criticized the concept on the basis that "it may be a strain on Occam's razor to suggest genetical hypotheses regarding such fluctuations as long as more directly ecological explanations can be invoked and tested." Since then, no "directly ecological" theory has emerged, yet the entire field of sociobiology has arisen with its explanations of behavioral traits based on just such genetical hypotheses (Wilson 1975). Chitty's theories developed further and became more rigidly specified as Chilly's Polymorphic Behavioural Hypolhesis (Chitty 1967), henceforth abbreviated to CPBH. Briefly, CPBH slales lhal al low densilies passive individuals lhal can tolerate crowding are selectively favored, allowing population buildup and increased density. Al high densilies, more aggressive, less viable individuals are favored. Inlerference or spacing behavior resulls in lower breeding densily
423
424
R. E. PAGE AND A. T. BERGERUD
and causes a population decline. The theory was developed to explain changes in abundance of animals that undergo regular periodic fluctuations, commonly called "cycles" (Elton 1942, Keith 1963). It would benefit the reader to be familiar with the review presented by Krebs (1978a). In the present paper, a study is detailed in which the logical consistency and falsifiability of the Chitty Hypothesis were investigated, leading to the definition of the central core of the hypothesis. The results were mathematically represented and simulated with empirical data from willow ptarmigan (Lagopus lagopus}. Chitty first suggested that the theory could be "falsified . . . by proving that there are no significant differences between expanding, stationary, and declining populations in the distributions of the properties of the individuals" (Chitty 1960, p. 99). Such a situation would in fact be a disproof, but it could never be expected to be realized. Only in the impossible situation where variations among individuals were totally irrelevant to fitness and were nonadaptive would changes in density not cause genotypic changes in density-dependent and frequencydependent alleles (Wright 1968, Roughgarden 1971, Smouse 1976, Poulsen 1979). "Differences in the properties of individuals" would always be present. The existence of these genotypic changes with density changes does not distinguish genetic viability as either cause or effect. To the present, testability has been enhanced but no absolute disproofs of CPBH have yet been conceived. Continual experimentation, particularly on microtine rodents, did not seem to significantly support or refute the theory (Krebs & Myers 1974), despite the fact that Chitty himself believed strongly in the Popperian view of a sophisticated falsificationist approach to science (Popper 1959, 1965, Lakatos & Musgrave 1965, Koertge 1979) and favored attempts to falsify the theory. The situation was that envisioned by Lakatos (1965) wherein the central theory forms a core surrounded by a belt of testable hypotheses. Empirical data do not necessarily impinge on the core theory itself, but if the belt is sufficiently riddled by falsified hypotheses, the belt collapses and the central theory is rejected. All tests of CPBH to date have been in the very outer fringe of the belt, owing to the great conceptual leap of inferring changes in the genotype from changes in the behavioral phenotype at a population level. There have been no recent advances in our understanding of CPBH. Many of Chitty's supporters have become disillusioned and have moved on to other studies, or are lending support to Keith's Hypothesis for cycling in snowshoe hare (Lepus americanus) (Keith & Windberg 1978). For these reasons, it appeared relevant to attempt a mathematical simulation of CPBH to determine which of the potential factors involved were truly necessary to explain cyclic populations (Krebs et al. 1973, Krebs & Myers 1974). To mathematically model the theory, it was necessary to closely investigate its structure and explicitly define its basic assumptions. By so doing, hypotheses close to the core were developed. The
GENETIC EXPLANATION FOR TEN-YEAR CYCLES
425
majority of this discussion deals with cycling animals where CPBH has been most extensively debated. Our emphasis is on grouse, particularly ptarmigan.
12.2 Defining the assumptions The Chitty Hypothesis has been couched in semantic terms that are vague and empirically imprecise. Chitty never did concisely define his theory himself. Some problems are obvious. We know that behavior is under genetic control, but how susceptible is it to environmental and other influences? What did Chitty mean by quality? What is viability? Does intrinsic mean strictly genetic? There are even a few logical inconsistencies. The most serious logical fallacy is that at high population levels a genotype would be favored that produces "individuals in a declining population that are intrinsically less viable than their predecessors." (Chitty 1960, p. 99). If viability and fitness are synonymous, by its very definition selection could not favor less-fit individuals. The existence of this paradox in the field data of voles is an artifact of the data-collection methodology and will be seen as caused by the confusion of population parameters of the total population with that of the breeding population in the selective arguments. We would paraphrase Chitty's definition of CPBH for cycling small mammals as follows: At low densities, individuals of low aggressive levels tolerate crowding such that population densities increase to the point where highly aggressive individuals are favored and the population declines. Within this definition are three inherent assumptions that are crucial to CPBH: (1) Level of aggression of an individual is primarily genetically determined. (2) Net population fecundity is inversely related to population level of aggression. (3) Aggressive individuals are more successful in breeding competition at high densities. Assumption 1 is explicitly stated by Chitty and must necessarily be met to consider a genetic explanation for these fluctuations. Assumption 2 has not been specifically identified previously. Only three methods are available to force population growth to cease and become negative at high densities. Either mortality increases, dispersal increases, or fecundity decreases (Pielou 1977). Measured mortality rates often do increase in the decline phase, but this condition is not necessary. Disappearance from the study site has often been confused with mortality. Overwinter mortality rates are relatively constant
426
R. E. PAGE AND A. T. BERGERUD
for one 10-year cycle in ptarmigan (Bergerud 1970a). Dispersal has been identified as a major population-regulation factor (Wynne-Edwards 1962) and was thought to be critical in voles, but the number of dispersers is greatest during the increase phase and drops significantly before the population begins to decline (Myers & Krebs 1971). There is usually no major dispersal in ptarmigan populations (Bergerud 1970, Myrberget 1972). Dispersal is not necessary to explain the decline. The only method remaining to cause the decline is a decrease in population fecundity. Yet litter sizes in voles and clutch sizes in ptarmigan are not correlated with changes in density (Zedja 1966, Bergerud 1970a, Keller & Krebs 1970, Weeden & Theberge 1972). The critical factor causing population declines from peak densities must be lack of recruitment to the breeding population. Given that aggression is correlated with population density (Moss & Watson 1980, Moss et al. 1984), assumption 2 is necessary. Assumption 3 is also implicit and necessary. Aggressives are at a selective advantage at high densities. For their fitness to be higher at high densities, they must pass more genes to the succeeding generations than must passives. They must either breed more often, or produce more offspring from each breeding. But given assumption 2, fecundity is lower; therefore, the option of producing more offspring is not available. Aggressives must breed more often than passives at high densities. For all intents, ptarmigan have only a single breeding period per year. Thus, aggressives must increase their fitness by dominating breeding competition such that some passives are excluded from successful recruitment. Are these three assumptions all that are necessary to explain cyclic fluctuations in abundance of animals? The criteria by which science accepts or rejects scientific explanations are the subject of continual philosophical debate (Lakatos 1965, Salmon 1979, Romesburg 1981). Predictive value, simplicity and goodness-of-fit to empirical data are all valid criteria by which to judge the worth of a theory; but unless we invoke the "Psychology of Research" (Kuhn 1965), resistance to falsification is the strongest, single criterion scientists possess (Popper 1959, Lakatos 1965, Koertge 1979). If a simple computer simulation incorporating only these three conditions could produce population fluctuations indistinguishable from real-world behavior, it would be strong evidence that CPBH is valid. Most important, the simulation could also help generate empirical falsification tests. To be represented mathematically, the assumptions must be more precise. It is necessary to restate assumption 2 in terms of individual selection. Given assumption 1, learning and accumulative population stress can be ignored. Thus, net fecundity is inversely related to parental level of aggression on an individual basis. Assumption 3 was generalized so that aggressives were always more successful in breeding competition, regardless of density. It was then possible to rigidly specify the assumptions so that they could be
GENETIC EXPLANATION FOR TEN-YEAR CYCLES
427
modeled. The parameters of aggressive level (i.e., success in breeding competition) and fecundity were specific to the genotype and constant over density changes.
12.3 The model The final form of the assumptions is: (1) Level of aggression is solely genetically determined by simple Mendelian selection of two alleles at one locus. (2) Recruitment is inversely related to female parental level of aggression (female genotype). (3) Aggressives are completely successful in breeding competition. The difficult simulation step was to formulate a competitive breeding system that was realistic. Willow ptarmigan proved ideal to simulate from this standpoint. Their breeding system is fairly well known because they are highly visible and breed in open habitats. Territories are required to attract a mate and both sexes are territorial. Males and females arrive in the breeding grounds separately and compete independently for territories (Chap. 10, Weeden 1959b, Hannon 1982). The simulation began in the spring, representing a finite breeding area in which territorial competition occurred. The most aggressive individuals arrived first and established territories. Other birds arrived in order of aggression and established territories until all had territories, or there was no area left. In the latter case, birds without territory joined a "waiting" flock. Packing of territories was perfect. Aggressive individuals also occupied larger territories with the result that a relatively small number of aggressive birds could dominate breeding. The process was repeated independently for females. All individuals with territories mated randomly. Fecundities were relative to female genotype, such that clutch sizes were equal, but passive birds had the greatest brood survival to 2 months. From fall to spring, survival rates were applied to arrive at the new breeding population. Juveniles were considered as capable of breeding as adults (Hannon & Smith 1984). The flow chart of the model is shown for one breeding period (Fig. 12.1). Initially, population information was stored in a matrix comprising two sexes by three genotypes by five age classes to investigate age-structure and sex-ratio effects. Gene flow was modeled by matrix manipulation. A compressed model was developed later, using equations of gene flow and considering only females.
428
R. E. PAGE AND A. T. BERGERUD
The model was run on an Apple II, 48K microcomputer in BASIC. Most of the more than 75,000 simulated years of runs were conducted on the Apple, but a FORTRAN IV version of an IBM 4341 provided graphical output. 12.4 Parameters of the simulation Whenever possible, population parameters are from Bergerud (1970a) for willow ptarmigan in Newfoundland. Most of the parameters necessary for the simulation differ between genotypes and could not be extracted from a single population statistic. Bergerud recognized nonnormality and bimodality in his data, and presented distributions for hatching dates and brood sizes. Clutch sizes are invariant with genotype, density, and period of the cycle, averaging 10.2 + 0.3 in Newfoundland. A clutch size of ten was used in the simulation.
Fig. 12.1. Flowchart of the model for grouse.
GENETIC EXPLANATION FOR TEN-YEAR CYCLES
429
Survival of chicks until August does not follow a normal distribution between broods. We considered that three distributions occur with means of 9, 6 and 4 (Fig. 12.2; from Bergerud 1970a). These were the parameters used for August brood sizes for passive, heterozygous, and aggressive genotypes, respectively. A survival rate of 0.3 was applied to the adult population and to juveniles after August, consistent with results of Newfoundland banding studies. The only parameter remaining is territory size. No published information was available on this aspect that would allow the separation of genotypes. Maps of
Fig. 12.2. Brood size distributions for Newfoundland ptarmigan over one cycle period (from Bergerud 1970a).
430
R. E. PAGE AND A. T. BERGERUD
territories of willow ptarmigan for three years for the Chilkat Pass in northern British Columbia were provided by Dave Mossop (see Fig. 10.19). Two basic territory sizes were apparent, with a ratio of smallest to largest of 1:6. A range this wide may be an exaggeration of the actual genotypic means, but in the simulation, territory size was the integrator of all possible aggressive interactions, so this range was used. Similar results have been found elsewhere (Pedersen 1984). In addition, stochastic variation in brood survival was added to represent the effects of weather. It has been noticed that brood survival is lower in decline years given the same weather conditions (Jenkins et al. 1963, Bergerud 1970a, Jenkins & Watson 1970). This was interpreted to mean that the aggressive genotype had lower brood survival than the passive genotype under the same, adverse environmental conditions. We simulated weather effects by using a random-number generator to vary August brood sizes around the mean for the particular genotype, with variances increasing with aggressiveness. The passive genotype had a variance in brood sizes of 10%, and the most aggressive genotype had a brood variance of 60%. These values seemed to fit what few data were available, even though variance in brood survival of 60% may seem somewhat high. It was the only way to get chicks/hen ratios low enough in the decline phase to fit the data.
12.5 Realism of the assumptions The assumptions have been specified in an extremely rigid manner in order to facilitate modeling. Obviously, assumption 1 is not completely realistic, as it denies any effects of the environment. Tests of the degree of heritability of aggressiveness in ptarmigan have recently been conducted. Aggressiveness in red grouse (L. /. scoticus) chicks was slightly better correlated with parental level of aggression of their father than with learning, maternal (nonparental), or sib effects (Moss et al. 1982b, cf. also Boag & Alway 1981). Though trying to avoid the obvious, red grouse workers in Scotland are inevitably concluding that level of aggression is primarily genetically determined (Moss et al. 1984, 1985). Moss and Watson (1980, p. 116) in speaking of changes in dominance, state that "these changes were inherited and could well have been genetic." Jacobs (1981) has discovered what may be the reason for the inability of workers to quantify the heritability of these behaviors. He defined an innateness index as the apparent degree of heritability of a measured behavior and found that spontaneous variation, i.e., environmental fluctuations, will significantly reduce the innateness index. Most important, the reduction is greatest when heritability is high. A behavior of high heritability will not appear so in a variable environment. Working with blue grouse (Dendragapus obscurus), Bergerud (Chap. 2) found that aggressive birds had a low amount of genetic variation and produced only aggressive offspring, whereas passive birds produced offspring exhibiting
GENETIC EXPLANATION FOR TEN-YEAR CYCLES
431
the full range of aggressive behaviors. This observation is interesting because it indicates that the environment may have very little effect in reducing the aggressive level of an individual that is genetically determined to be aggressive, but may increase aggressiveness in passive birds. Most emphasis by experimenters has been placed upon the means of increasing aggression in passive individuals rather than the reverse. Correlations between genotype and aggression have been identified in other birds. In their breeding displays, male ruffs (Philomachuspugnax) on the breeding grounds showed two distinctly different levels of aggression that are believed to be under genetic control (Hogen-Warburg 1966). Similarly, the distribution of territory sizes in male Arctic skuas (Catharacta skua) was not normal but clustered around a large and a small peak. Males with large territories ranked higher in tests of aggression and arrived on the breeding ground earlier, consistent with the model structure. In addition, aggressive males were melanistic, showing a pattern consistent with two alleles at one locus (O'Donald 1977). Testosterone is important in the production of melanin (Wydoski 1964), so the correlation between dark pigmentation and aggressiveness may not be spurious. It has been shown in many animals that aggression is affected by androgen levels (Alice et al. 1939, Watson 1970, Gandelman & Svare 1974, Dixson 1980, Wagner et al. 1979, Watson & Parr 1981). Androgen production provides the mechanism for aggressive levels to be determined by a simple Mendelian system of inheritance. If a gene significantly increased androgen production, the behaviors hypothesized here could be realized. With the expression of this gene for androgen production in both sexes, assumption 2 takes on new realism. A female exhibiting high levels of aggression probably would be a poorer mother. The female would incubate for shorter periods of time. Constant movement would increase nest detection by predators (Erikstad et al. 1981). Once eggs were hatched, an aggressive female would not brood as often or as steadily. This would explain greater susceptibility to bad weather of broods from aggressive mothers. Other mechanisms could be theorized by which increased female aggressiveness would lower chick survival and net fecundity. Female birds require high levels of testosterone for laying eggs and building nests, but levels must decrease to allow incubation (Silver et al. 1979). Female California quail (Lophortyx californicus) injected with testosterone stopped incubating (Collias 1950). An interesting case for the effect of testosterone on mothering occurs naturally in the Wilson's phalarope (Steganopus tricolor), in which incubation is undertaken by the male. Ovaries of female Wilson's phalaropes produce more testosterone than the testes of the males. Females do not incubate (Johns & Pfeiffer 1963). In a test of the genetic control of behavior, chickens could fairly easily be arti-
432
R. E. PAGE AND A. T. BERGERUD
ficially selected for breeding behaviors (McCollom et al. 1974). Even in these highly manipulated birds, breeding behavior was strongly controlled genetically (Craig et al. 1965). Finally, androgens stimulate RNA coding and could thus cause expression of genetically coded sexual behavior that would not otherwise be exhibited (O'Malley 1977, Beyer et al. 1979). Male ptarmigan must hold territories to attract females (Hjorth 1970), and territory size is correlated with level of aggression (Watson 1964, 1970, Watson & Parr 1981). Females chose males with larger territories and thus either directly or indirectly chose more aggressive males (Chap. 13, O'Donald 1977, Weatherhead & Robertson 1981). The model uses this principle at the limit, where any aggressive male will be chosen over any passive male, fulfilling Assumption 3.
12.6 Results 12.6.1 Basic factors The simulation was judged successful when a region was found of highamplitude, regular limit cycles of a 10-year period. The realistic parameters already outlined produced such cycles. The amplitude of the strictly deterministic model was increased by the weather effects introduced by the addition of variance to brood survival (Fig. 12.3). Cycle period then varied from 8 to 14 years, with a cycle missed every few hundred years. If the passive allele (designated as A) is considered partly dominant over the aggressive allele (B), an intermediate heterozygote would result as indicated by the Newfoundland brood-size distributions. Phase maps of the flow between the two alleles illustrate the genetic properties of the model (Fig. 12.4). Note that in the increase phase, at the top of the graph, gene A increases rapidly with a slow buildup of gene B at the peak. Gene A decreases rapidly, then gene B decreases in frequency to begin the cycle again. Buildup of gene B in the population occurs through the production of sufficient homozygous aggressives by reproduction of heterozygotes in the peak year to force a decline for many years thereafter. Changes in the breeding population are more dramatic than in the population as a whole. Cycles are driven by aggressive breeders, but variation in number of these birds is low. The effect of aggressives on passives is pronounced. Even though gene A is more abundant in the population, passives are excluded from breeding in the decline phase. The stability properties of this model have not yet been fully explored. There appears to be an unstable equilibrium point in the cyclic cases that exists outside the normal range of the model (Fig. 12.5). A very simple but less realistic series
GENETIC EXPLANATION FOR TEN-YEAR CYCLES
433
Fig. 12.3. Influence of weather effects on regularity of cycling. Weather imposed at year 100.
Fig. 12.4. Phase map of changes in frequency of the two model alleles.
434
R. E. PAGE AND A. T. BERGERUD
of differential equations has been solved for the same problem and may lead to a better mathematical understanding of this situation (Hung 1982). Territory size of the aggressive birds was six times that of the passives. A difference of three times produced a stable, polymorphic equilibrium (Fig. 12.6). Reducing variance in brood sizes because of weather effects predictably produced lower amplitude cycles. Fecundity and survival values were varied to represent other grouse species. With survival of 0.4 and smaller brood sizes, typical of rock ptarmigan (Lagopus mutus), 14-year cycles resulted, as in that species (Watson 1965). With survival rates of 0.7, typical of blue grouse, cycling ceased, as in the real-world species (Fig. 12.7). Weather effects produced small eruptions, as sometimes seen in blue grouse (Redfield et al. 1970). Effects of a forest fire on blue grouse was simulated by increasing the limit on territory available, followed by a decay in this parameter representing successional changes. Model behavior was typical of the real world.
Fig. 12.5. Phase map of frequency of model alleles for 200 simulated years.
GENETIC EXPLANATION FOR TEN-YEAR CYCLES
435
Fig. 12.6. Effect on model cycling of reducing territory size of homozygous aggressives by half. A stable polymorphic equilibrium is achieved.
Fig. 12.7. Simulated blue grouse population. First increase caused by initial conditions.
436
R. E. PAGE AND A. T. BERGERUD
12.6.2 Other factors Model output appeared realistic in the simple case, but other factors were incorporated to see whether realism could be increased, and effectively to test the importance of these other factors to the theory and to cyclic animals. Dispersal was represented by killing off the surplus or waiting birds. There was a slight increase in amplitude owing to the greater time required to recover from a decline. Passives initiate increases and they form the majority of the surplus. Similarly, predator switching was represented by increasing mortality rates from one year after the peak for various numbers of years. Although amplitude slightly increased, again no significant change in behavior occurred.
12.6.3 Realism of the results Is the surplus necessary? In this simple model, yes. If passives are not excluded from breeding, a polymorphic equilibrium will persist. In the more heterogeneous, real world, a parallel situation can be envisioned. If there is an area of prime habitat, aggressives will preferentially occupy it. Passives will be relegated to secondary habitats. At some level, as the population expands outward from the primary centers, habitat quality will be low enough that net recruitment for the individuals in that habitat will become negative (Wolff 1980). The effect of birds that breed beyond this limit on the breeders in the prime habitat is effectively the same as if they did not breed at all. The continuum of habitat types from positive to negative recruitment has been identified in Chilkat Pass, British Columbia, by Dave Mossop (Chap. 10), but it does not exist in all years (see also Pedersen 1984). Intraspecific strife is not necessary to prevent breeding. In a hierarchical system of resource partitioning, it is advantageous for an individual to determine its hierarchical position and if success is unlikely, to disperse with the hope that chances are better elsewhere (Lomnicki 1978). The absence of overt fighting cannot be used as indication of lack of aggression. Wiger has reached similar conclusions for the mechanisms of cycling in Clethrionomys (Wiger 1982). He believes that dominant females must occupy the limited number of breeding territories and identifies the relationship between population size and recruitment in his assumptions 1,2, and 3. Changes in total population parameters of the model are similar to those in the real world. Genetic makeup of the model surplus is similar to the dispersers in natural populations of voles (Myers & Krebs 1971). Changes in the variance of aggressiveness of the population replicate that found for red grouse (Moss & Watson 1980). There is a considerable volume of confirming evidence for the theory, but because the theory cannot be proved, this evidence serves mainly to attract the interest of other scientists. The test record of a theory, or its resistance
GENETIC EXPLANATION FOR TEN-YEAR CYCLES
437
to falsification, is critical. What falsification tests of CPBH are suggested by the model?
12.6.4 Falsification tests (1) The aggressiveness of an individual should not change with the stage of the cycle. (2) The fecundity of an individual should not change with the stage of the cycle. (3) Some passive individuals should fail to breed successfully after peak years, even though they may have bred successfully before. (4) Removal of predators should not stop cycling (except see 8). (5) Addition of food should not stop cycling. (6) Extreme removal of food should stop cycling, if it severely reduces fecundity. (7) Removal of aggressive individuals in the peak years should maintain high population levels. (8) Increasing survival rates (i.e., to 60% for willow ptarmigan) should stop cycling.
12.7 Conclusions More work is required to determine the underlying mechanisms that produce auto-correlations between aggressiveness and fecundity. Androgens may not provide the complete answer. The essence of CPBH is in the behavioral and reproductive success of individuals. More attention should be focused on individual differences and less on population parameters. We are convinced that Dennis Chitty found the key to cycling animals. It is hoped that this paper will open the lock.
12.8 Summary Chitty's Polymorphic Behavioural Hypothesis (Chitty 1967) was logically reduced to three main assumptions that were mathematically modeled:
438
R. E. PAGE AND A. T. BERGERUD
(1) Level of aggression is genetically determined by simple Mendelian selection. (2) Recruitment is inversely related to female parental level of aggression. (3) Aggressives are completely successful in breeding competition. The model utilized data from willow ptarmigan populations, but was generalized to other grouse species. Simulation results were indistinguishable from the behavior of real-world grouse populations, lending support to Chitty's hypothesis as the explanation of cycles. Eight tests that would falsify the model were identified.
13
Mating Systems in Grouse A. T. Bergerud
13.1 Introduction The evolution of mating systems can be best understood by determining the fitness costs and benefits of various reproductive options available to individual males and females (Wittenberger 1979), within the framework of sexual selection and parental investment theory (Darwin 1871, Trivers 1972). Such an approach permits the environmental constraints impinging on the options to be identified, leading to an integrated theory (Wittenberger 1979). Mating-system theory is well advanced in the study of birds, and the grouse (Tetraonidae) has received special emphasis (Wiley 1974, Wittenberger 1978, Bradbury 1981, Oring 1982). It is especially intriguing that all three ptarmigan species — white-tailed ptarmigan (Lagopus leucurus), rock ptarmigan (L. mutus), and willow ptarmigan (L. lagopus) — are monogamous; the six remaining North American grouse are polygynous. The most general explanation for monogamy in birds is that the male is needed for parental care (Lack 1968). But monogamous ptarmigan, like the polygynous grouse, are cursorial and nidifugous; males are not needed to feed the young. Removal of this constraint provides the potential for polygyny, provided that environmental conditions permit males to capitalize on this polygyny potential (Emlen & Oring 1977). Also of interest is that forest grouse, blue grouse (Dendragapus obscurus), ruffed grouse (Bonasa umbellus), and spruce grouse (Dendragapus canadensis) have adopted a system of dispersed polygyny where males display solitary and do not guard resources needed by the female, and where females come to the males for breeding. In contrast the steppe grouse, sage grouse (Centrocercus urophasianus), prairie chickens (Tympanuchus cupido), and sharp-tailed grouse (Tympanuchus phasianellus) have a system of males displaying at a communal 439
440
A. T. BERGERUD
arena (a lek). These males display a clumped polygyny system. Thus, each mating system—monogamy, dispersed polygyny, and clumped polygyny—has evolved in a different biome—tundra, forest, or steppe. The theme of this chapter is that there are different predation pressures in these three biomes and that this predation is the prime mover in the evolution of the diverse mating systems.
13.2 The monogamy model First I present a paradigm for the prolonged pair bond (monogamy) in ptarmigan integrating the following environmental variables: (1) few nest predators, (2) effective avian predators of adults, (3) open habitats with good visibility, and (4) long daylight hours. These factors, I suggest, have selected for the antipredator strategies of mutually exclusive, small prelaying ranges of females, and selection by females of males that are conspicuous in behavior and plumage and that have large territories with good nesting cover (Fig. 13.1). Conspicuous males should deflect predation pressure away from females, and females should select for a prolonged pair bond and male vigilance to reduce their risk from the effective avian predators.
13.2.1 Size of the prelaying range I define the prelaying range of females as that space a female travels on the breeding range after leaving winter flocks or the forest cover used in winter and before laying her first egg. Female ptarmigan, unlike other grouse, settle in the territories of males (Weeden 1959b, Jenkins et al. 1963). In the Bergerud and Mossop (1985) model these prelaying movements represent a nest-searching range, wherein females examine nesting cover and may evaluate the frequency of nest predators. Predation of nests accounts for most of the loss of young and potential fitness in ptarmigan and other grouse (Chap. 15). The choice of where a female places her nest to minimize detection by predators and thus hatch her eggs is the single most important decision within her control. One option in decision making is to acquire additional information before deciding among options (Wittenberger 1979). Prelaying ranges of monogamous ptarmigan are smaller than those of the polygynous grouse. Bradbury's (1981) review presents the home range of ptarmigan females before egg-laying as varying from 0.2 to 13 ha. The prelaying ranges of females of the forest polygynous species—ruffed grouse, blue grouse, and spruce grouse—were listed as 10-50 ha, a tenfold increase; the prelaying ranges of grouse of the steppe, which mate at leks—prairie chickens, sharp-tailed grouse, and sage grouse—were over 200 ha, or another tenfold increase. The differences in home-range size among the three grouse groups (tundraptarmigan, forest, and steppe) can be explained by the fact that ranges are posi-
MATING SYSTEMS IN GROUSE
441
Fig. 13.1. Proposed model (Bergerud & Mossop 1985) for the evolution of monogamy in ptarmigan.
lively correlated with the abundance of nest predators and negatively correlated with loss of nests by predators and the percentage of nests preyed on (Fig. 13.2). The correlation between the size of the prelaying ranges and nesting success for the 9 grouse species is r = -0.792 (see Fig. 16.2). Cursorial, ground-nesting birds should space their nests to reduce the risk of predation (Lack 1968). Nesting success of females declines in several groups of birds as densities increase —in ducks (Weller 1979, Livezey 1981), grouse (Bump et al. 1947, Boag et al. 1979), Phasianidae (Potts 1980), and in shorebirds and passerines (Horn 1968, Krebs 1971, Andersson & Wiklund 1978, Page et al. 1983). Reduced predation risk to nests should mean that ptarmigan are prepared to accept or to search for small prelaying ranges—nests can be relatively closely spaced and still hatch successfully. The more southern, forest and steppe grouse should search larger areas and invest more effort in the nest location because of increased contact with predators and greater importance of predation in their fitness. Ptarmigan live in the alpine-tundra biome where nest predators are scarce. The only common predators of ptarmigan are foxes (Vulpes vulpes, Alopex lagopus),
Fig. 13.2. Comparison of the percentage of nests destroyed by predators for steppe, forest, and tundra grouse. Each dot equals one study value, calculated by dividing nests destroyed by total nests found. Sources: Steppe—Gross 1930, Hamerstrom 1941, Keller et al. 1941, Lehmann 1941, Allred 1942, Schwartz 1945, Batterson & Morse 1948, Grange 1948, Hart et al. 1950, Patterson 1952, Baker 1953, Nelson 1955, Yeatter 1963, Gill 1965, Brown 1966b, 1967, 1968b, Silvy 1968, Bernhoft 1969, Klebenow 1969, Artmann 1970, Christen-
MATING SYSTEMS IN GROUSE
443
weasels (Mustela ermined), and crows and ravens (Corvus spp.) (Dixon 1927, Olstad 1932, Kristoffersen 1937, Weeden 1965, Bergerud 1970a, Giesen et al. 1980). At Chilkat Pass, British Columbia, we estimated that fewer than five mammalian predators/km2 were present. Jones and Theberge (1982) radio-tracked foxes at Chilkat Pass and reported that home ranges were larger there than for red foxes in more temperate regions. Arctic foxes and weasels also exist at low densities in the Arctic (Riewe 1975). A generalization is that the number of mammal species in North America decreases with increasing latitude (Fleming 1973, Wilson 1974, McCoy & Connor 1980). The polygynous grouse species that nest south of Arctic ptarmigan have more kinds of nest predators, and at higher densities. The forest grouse species may encounter any of six to seven genera of common nest predators. One estimate for ruffed grouse, based on an intensive predator removal, was ten predators/km2 (Bump et al. 1947). The steppe grouse that inhabit grasslands and sagebrush (Artemisia spp.) may have to contend with nine genera of predators, which may reach densities of 15/km2 (Balser et al. 1968, Chesness et al. 1968, Beasom 1974) or more if snakes and ground squirrels (Spermophilus spp.) are included. The percentage of nests destroyed by predators is correlated with latitude. South of latitude 40°, 45 ± 2.7% (± SE) (n = 11 studies) of the grouse nests were destroyed by predators. Between 40° and 46° north latitude, 36 ± 4.0% (n = 18 studies) were destroyed and in the north above 46°, 26 + 3.4% (n — 21 studies) of the grouse nests were lost to predation (see Fig. 15.7). Nesting success of passerines follows the same south-to-north sequence, and Ricklefs (1969) attributed the continuum to reduced predation. The percentage of nests destroyed by predators is significantly less for ptarmigan than for more southern, forest or steppe grouse (Chap. 15); it is likely that some of this improvement results from males assisting nesting hens (a circular argument) but much of it is due to the reduced abundance of predators. The small size of the prelaying ranges of ptarmigan, listed by Bradbury (1981), was based on the size of the territory that monogamous females shared with males in spring. Females of the polygynous grouse have prelaying ranges noncoincidental with the display territories of males; hence, their prelayng son 1970, May 1970, Bowen 1971, Pepper 1972, Schiller 1973, Horak 1974, Wallestad & Pyrah 1974, Rice & Carter 1975, 1976, 1977, Kohn 1976, Ramharter 1976, Sisson 1976, Horkel et al. 1978, Riley 1978, Petersen 1979, 1980, Sell 1979, Svedarsky 1979, Vance & Westemeier 1979. Forest-Fisher 1939, Bump et al. 1947, Grange 1948, Tanner 1948, Fallis & Hope 1950, Hardy 1950, Bendell 1955a, Kupa 1966, Neave & Wright 1969, Barrett 1970, Mossop 1971, Ellison 1974, Haas 1974, Maxson 1974, Zwickel 1975, Zwickel & Carveth 1978, Hoffmann 1979, Sopuck 1979, Robinson 1980, Keppie 1982, Rusch et al. 1984, pers. comm. Tundra — Kristoffersen 1937, Choate 1963a, Watson 1965, Bergerud 1970a, Gardarsson 1971, Weeden & Theberge 1972, pers. comm., Giesen et al. 1980, Hannon 1982, Myrberget 1983, pers. comm.
444
A. T. BERGERUD
ranges are probably associated with female preferences, as distinct from territorial boundaries defended by males. However, evidence indicates that female ptarmigan have a distinct affinity for prelaying ranges apart from the ones they acquire from joining males defending territories. Females commonly return to previous ranges, even if the prior mate is absent, and defend space (Choate 1963a,b, Watson & Jenkins 1964, Schmidt 1969, MacDonald 1970, Hannon 1982, 1983). Females may also desert apparent mates and move elsewhere (Schmidt 1969, Chap. 8). Females appear to assess the quality of the male's territory and sometimes visit several male territories before settling (Watson & Jenkins 1964). When males are removed from territories in experiments or by predators, the females remain (Gardarsson 1971, Chap. 9, Hannon 1983). If removals are done early in the season, new males may join the females (Hannon 1983, Table 10.10), or if no replacements are available, existing males can enlarge their territories to encompass more female ranges (Fig. 9.6). Again, the ranges of females appear centered on the nest sites and do not commonly adjoin, as do male territories (Jenkins et al. 1963, cf. Hannon 1983, p. 815). Hannon (1983) has documented that females defend prelaying ranges even in the absence of males. I feel that suitable habitat for nesting is an important component of female philopatry, and is distinct from an affiliation with the male.
13.2.2 Dispersion of males Bradbury (1981) has recently presented a tightly argued hypothesis that the spacing pattern of female grouse dictates the spacing pattern of males. If females have small home ranges, or are prepared to accept a small space resource, males can show a dispersed, uniform spacing pattern. Uniform spacing (territoriality) can be selected for only if resources are economical to defend (Brown 1964). The fitness of ptarmigan males will be enhanced if they disperse and defend the space and nest-cover resources sought by females; as a consequence more males will have the opportunity to breed. When males are clustered, as in the lek species, commonly less than 25% do most of the breeding (Kruijt & Hogan 1967, Wiley 1973b, Table 7.10). Ptarmigan males show a uniform spacing pattern in the spring and commonly over 70% of the males are paired (Jenkins et al. 1963, Watson 1965, Braun & Rogers 1971, Fig. 10.19). The resource that Bergerud and Mossop (1985) believe males defend is the territorial space (which must include nest cover) within which females will select nest sites. Given that males are uniformly spaced, the prelaying ranges of females still do not, theoretically, need to coincide with the territorial boundaries of males. A female could select a nest site within one territory but still range through the territories of several males. We believe that the concomitant overlap between male and female ranges results from the vulnerability of ptarmigan in the spring
MATING SYSTEMS IN GROUSE
445
to avian predators. Our model suggests that a female needs a male (mate) to act as a sentinel and deflect predators during the nest-searching period and while she lays and incubates her eggs.
13.2.3 Effective avian predators Adult rock and willow ptarmigan are hunted by red foxes, golden eagles (Aquila chrysaetos), gyrfalcons, (Falco rusticolus), and in some regions, goshawks (Accipiter gentilis) and peregrine falcons (Falco peregrinus). The most effective predator is the gyrfalcon. The range of the gyrfalcon coincides with the range of rock ptarmigan throughout the Holarctic region except in Newfoundland, Japan, and Scotland. The distribution of gyrfalcons overlaps 80% of the range of willow ptarmigan, and 40% of the distribution of white-tailed ptarmigan. The percentage of ptarmigan in the diet of some gyrfalcon populations has been reported as: over 89% in Norway, 66% in Finland, over 73% in Iceland, 90% in Alaska, and 100% in the Yukon (Hagen 1952a, Cade 1960, Bengston 1971, Pulliainen 1975a, Langvatn 1977, and pers. files). Thus many gyrfalcon populations depend on ptarmigan, the only common terrestrial, nonmigratory bird in the Arctic. Gyrfalcons nest in April (Fig. 13.3), when the north is still gripped in winter and before migratory birds return or hibernating ground squirrels (Spermophilus parryii) emerge. Gyrfalcons require ptarmigan for food during breeding —if ptarmigan are low in numbers the falcons cease breeding (Cade 1960, Barichello and Mossop 1983). The unusually early timing of the spring nesting behavior of gyrfalcons (Platt & Tull 1976) is due to the vulnerability of ptarmigan to gyrfalcon predation at that time. Ptarmigan, especially yearlings, are vulnerable to predation in the spring when they disperse from winter cover to contest for territorial space. Large numbers of birds have been found in the spring that were recently killed by raptors (Weeden 1965, Mercer 1967, Chap. 10). Darwin (1871) noted that when the snow has disappeared that this ptarmigan is known to suffer greatly from birds of prey before it has acquired its summer dress. There are several reasons for this mortality. First, the birds are partly white but substrates are becoming brown. Second, males, and to a lesser extent females, are conspicuous because they prospect in unfamiliar habitats. Third, these birds are dispersing and are no longer able to use flocking as an antipredator strategy (Gardarsson 1971, Chaps. 9, 10). Fourth, they are no longer able to roost in snow burrows and escape detection by avian predators (Chap. 10). And last, summer leaves are not yet available to use as cover. The female ptarmigan faces a major survival problem by molting her white plumage for pigmented summer plumage. Frequently there are only 2 weeks between the time that most of the ground is covered with snow and when she must
446
A. T. BERGERUD
be in brown plumage to be cryptic while nesting. Spring phenology is unpredictable, snow cover is patchy, and spring snowstorms are frequent. Further, she must prospect for nest sites on brown substrates while molting her white plumage. The control of nesting cover by males necessitates that females arriving later must localize for nesting in the vicinity of male territories. But by selecting the territory of a particular male she can have his undivided association to deflect predation pressure away from her. Also his territory provides the space so that the nest is far from others. Using the male is the female's primary tactic against effective raptors, and the male's mutually exclusive space enables the female to space her nest away from others to reduce the risk of destruction by nest predators.
Fig. 13.3. Comparison of annual breeding cycle of gyrfalcons with annual breeding cycle and vulnerability of ptarmigan, as expressed by plumage and behavior contrasts and presence of alternative prey (migratory birds).
MATING SYSTEMS IN GROUSE
447
The males of both rock and willow ptarmigan become conspicuous in the spring at the very time that females molt and become cryptic. Huxley (1938) was possibly the first to state that the conspicuous males deflected predation pressure away from females. The male rock ptarmigan delays the molting of white feathers for a month after the female molts in several populations exposed to gyrfalcon predation (Salomonsen 1939). The male's white body is conspicuous on brown "lookouts." The male willow ptarmigan molts his white winter feathers on his head, neck, and breast sequentially to take on a conspicuous breeding plumage of a reddish brown head, neck, and breast, and white feathers on the body, which have been retained (Stokkan 1979a,b). This two-toned, solid-color combination makes him equally conspicuous on snow or against bare ground. The females of both species molt their white feathers nonsequentially—newly acquired, pigmented feathers are scattered among the white, a salt-and-pepper stage that is cryptic.
13.2.4 Conspicuousness of males I compared the behaviors and plumages of paired, male willow ptarmigan with those of their females at Chilkat Pass, British Columbia, 15-23 May 1981, to investigate the hypothesis that males deflect predation from females. Two habitats were available in the study area, one with emergent willow (Salix spp.) stems, many with retained dried leaves, of which 90% was covered with snow, and one of dwarf birch (Betula glandulosd) on hilltops that was totally devoid of snow cover (i.e., 100% brown substrate). All the ptarmigan pairs (n = 93) were in or adjacent to the willow habitat. The females averaged 69% brown plumage on their heads and 35% brown plumage on their backs (Fig. 13.4). Brown feathers were scattered among the white, and the patches of brown approximated the size of the retained, dead willow leaves. These females were cryptic, matching the willow stems and leaves and the snow beneath. The males, by contrast, averaged 96% pigmented feathers on their heads and necks and 10% on their backs. When the male was on snow his reddish brown head was visible, and when he was on the brown patches under willows his white body was in contrast. Of the pairs in which the males were motionless and silent, we saw the male before the female in 62 of 68 pairs (Table 13.1). Nonadvertising males were first seen at an average distance of 56.6 + 3.8m. Males were conspicuous also because they were more often in the open than females and they frequented elevated lookouts. Females were usually under willows and in the denser parts of the vegetation stand (Table 13.1). The mean distance between males and females was 6.6 + 0.7 m (n — 61). One can argue that the birds sought willow habitats over birch because the former is a preferred food (Bryant & Kuropat 1980). However, birch is also nutritious (Gardarsson & Moss 1970). Willow could be the preferred habitat at Chilkat Pass because it provided a cryptic combination, white and brown, and provided
448
A. T. BERGERUD
Fig. 13.4. Comparison of the sequence of molt in the most nonpigmented female and th most pigmented female observed each day. Size of the female model is reduced when there is white plumage on a snow background and pigmented plumage against a bare background. For example, on 23 May the female with the least pigment had brown feathers on only 10% of the body (90% was white) and brown on 20% of the head and white on 80%; the most pigmented female that day was all brown (100/100).
MATING SYSTEMS IN GROUSE
449
Table 13.1. Conspicuousness of males and females in pairs of willow ptarmigan observed at Chilkat Pass, B.C., May 1981 Bird First seen (no calling) In open when first seen In willow when first seen Body in full view — no vegetation across body during encounter Mean number of willow stems counted crossing the body ( ± SE) Bird higher in elevation
n
Male (%)
Female (%)
68 pairs 79 pairs 93 pairs
62 (91%) 65 (82%) 21 (23%)
6 (9%) 14(18%) 72 (77%)
93 pairs
85 (91%)
8 (9%)
1 1 males & 26 females 51 pairs
3.4 ± 0.99 46 (92%)
7.8 ± 1.14 5 (8%)
more escape cover. Predation risk should be an important component in modifying optimal foraging with respect to nutrition (Chaps. 4, 10). When I approached a pair, the male made himself more conspicuous by giving warning calls, moving more into the open, and running on the snow ahead of the observer. The female, by contrast, crouched down and remained silent and motionless. Of the cocks in 87 pairs: 51 % of the males walked in front of the observer calling, away from the female (decoy behavior); 21 % flushed after the female flew; 14% walked after the female on the ground; 10% flushed but circled back, calling to the female; and only 5% flushed and left the female. My closest approach to the males in 84 pairs was 13.6 + 0.9 m and to the females was 8.5 + 0.6 m (P < 0.05). The mean distance of approach to single males was 23 + 3.2 m (P < 0.05). Clearly the tactic of paired cocks was to remain near the female with a ground approach and to be available to attract attention to themselves. I observed rock ptarmigan at Bathurst Inlet, Northwest Territories, in June 1981 and 1982. Fourteen paired males flushed only after the female of each flushed, at an average of 16.4 + 2.7 m, whereas 14 males without females flew at 54.2 + 8.8 m (P < 0.05). Males with females often made diversionary walks away from females and fanned their black tails. The white bodies of the males were in bold contrast to the brown substrates when they were on rock lookouts, and their black tails were in contrast when they were on snow. Again at Bathurst Inlet in June 1982,1 saw six male willow ptarmigan on the ground an average of 84 + 1.2 m away, but in all six observations the females were not seen until they flew later. Males flushed at 64 + 0.7 m (n — 13), whereas females flew at 15 + 0.3 m (n = 7). In this location willow cover was less than 0.5 m tall. The males made distraction flights and flew away from the females when I flushed them at a mean distance of 64 m. Only when I approached where the male had been first seen was the female flushed. Mossop secured further evidence of male deflection behavior when he flew a
450
A. T. BERGERUD
gyrfalcon over a pair of willow ptarmigan in June 1982 on the Dempster Highway, in the Yukon. The female flushed before the male, and the male flew behind and above the female. Mossop then repeated the experiments over undisturbed pairs of willow and rock ptarmigan by flying a kite model of a gyrfalcon. In all trials (n = 15) the willow ptarmigan males called and the females crouched. In four trials the female later flushed but flew low. The male then flushed and flew behind and higher than the female, where he would have been the first of the pair to be overtaken by a raptor coming from above. These birds did not attempt to outfly the kite. Rather, the female plunged first into dense willow cover and the male continued past her, before seeking cover. The males of rock ptarmigan pairs (n — 6) also flew higher than the females, but the pairs flew out of sight. Mossop and I have flushed many rock ptarmigan when we have been in helicopters, and their tactic seems to be to try to outfly the raptor (helicopter). Male rock ptarmigan seek higher vantage points in the level tundra in areas largely devoid of escape cover. They may possibly depend on a long lead time to escape raptors in habitats largely devoid of cover; they would be conspicuous if they simply crouched when on brown lookouts. Willow ptarmigan males remain with their hens during incubation and assist in raising the broods. I observed males with 38 broods on Brunette Island, Newfoundland, in 1979. Again, males were more conspicuous than females in defense of the brood (Table 13.2). When the hunting dog pointed at a hidden brood it was normally the male rather than the female that first emerged from vegetation cover, called first, flew first, or "led" away on the ground. Females most often remained crouched, hidden, and silent until the chicks were closely approached or until a chick was captured and gave distress calls. In these instances I suggest the male was attempting to attract attention away from the hidden chicks and female. These observations confirm that male ptarmigan act as both sentinels and decoys for their females. These observations are supported by the views of others (Weeden 1959b, Watson & Jenkins 1964, Schmidt 1969, MacDonald 1970, Watson 1972, Giesen & Braun 1979a, Hannon 1984). The vigilance of the male Table 13.2. Comparison of behaviors of male and female willow ptarmigan in pairs with chicks on Brunette Island, Newfoundland, June and July 1979
Behavior Called first Seen first in distractive behavior Flew first Mean distance (m) from observer when flushed ( + SE)
Broods in sample
Total male (%)
Total female (%)
21 36 37
17 (81%) 30 (83%) 32 (86%)
4 (19%) 6(17%) 5 (14%)
38
9.8 ± 1.09
6.5 ± 0.92
MATING SYSTEMS IN GROUSE
451
would be especially effective in the Arctic when all predation in May and June must occur in daylight.
13.2.5 Mortality of males The conspicuous behavior of males in attracting the attention of raptors away from females has a mortality trade-off. The annual mortality rate of adult rock ptarmigan males in Iceland was 65 %, whereas on average only 47 % of the females died (n = 1,724, t = 7.645, P < 0.01) (data from Gardarsson 1971). The annual mortality rates of rock ptarmigan adults in Alaska also hunted by gyrfalcons were 63% for males and 54% for females, (n = 1,008, t = 2.239, P< 0.01) (data from Weeden & Theberge 1972, and pers. comm.). Three ptarmigan populations exposed to gyrfalcon predation averaged 49 + 1.2% males, and six populations without gyrfalcons averaged 57 + 1.0% males (P < 0.05) (Choate 1963a, Jenkins et al. 1963, Watson 1965, Weeden 1965, Mercer 1967, Bergerud 1970a, Gardarsson 1971). During an 8-year study of rock ptarmigan, Gardarsson (1971, Chap. 9) recorded a 33% mortality rate of cocks in May and June on Hrisey Island, Iceland; he estimated that few females died during the same period (< 10%). He observed gyrfalcons taking nine territorial males. Others have remarked on the tendency of gyrfalcons to selectively take males in the spring (Wayne & Jolly 1958, Cade 1960).
13.3 Female choice in the model 13.3.1 Differences among species According to sexual-selection theory, females should generally select males on the basis of phenotypic traits that reflect genetic quality (Fisher 1958); but females also should select males on the basis of quality of the territories that males defend (Orians 1969). At present the relative merits of female choice for male quality versus territory quality are debated (Searcy & Yasukawa 1981, Weatherhead & Robertson 1981, Wittenberger 1981b, Kirkpatrick 1985). One problem with the view that females select male quality (aggressiveness) to have "sexy sons" and more grandchildren is that a father's fitness should have nearly zero predictive value for offspring fitness (Falconer 1960, review Cade 1984). I propose that the benefits for female ptarmigan selecting aggressive males relate to the immediate survival of the female and her current progeny, and thus avoid the "lek-paradox" (Borgia 1979, Taylor & Williams 1982) and "sexy-son" controversy. I can agree with Weatherhead and Robertson (1981) that female choice should involve both territory quality and male quality. Female ptarmigan must have both a safe nest site and an aggressive, vigilant male to maximize fitness. The relative
452
A. T. BERGERUD
importance of these two components of fitness should depend on the relative abundance of predators of adults with that of predators of nests, both between species and between populations. The white-tailed ptarmigan has the lowest nesting success of the three ptarmigan species and the highest adult survival (Table 13.3); also, males live significantly longer than females (Braun 1969). The willow ptarmigan has the highest nesting success of the ptarmigan and the shortest life span (Table 13.3). The rock ptarmigan is intermediate between the other two species. If the risk of raptor predation compared with that of nest predation influences a female's decision of territory versus male quality, white-tailed females should be the most involved of the three species with territory quality (nest sites), because of high nest losses, and be the least selective with respect to male quality (raptor predation). In contrast, willow ptarmigan females should be the most selective of the three species for male quality and least selective for territory quality. A greater investment by female white-tailed ptarmigan in nest sites than in male deflection behavior is suggested by the following: (1) the observations of paired females deserting their mates (Schmidt 1969, Chap. 8); (2) failure of white-tailed ptarmigan females to travel with the male in defense of boundaries (Schmidt 1969, Chap. 8)—a component of aggressiveness in female ptarmigan if male vigilance is important should be to prevent sharing her male with other females; and (3) the characteristic of female white-tailed ptarmigan to show fidelity to successful nest sites but sometimes switch males between years (Schmidt 1969, Chap. 8). Male white-tailed ptarmigan in Colorado and Montana invest less in partners than do willow or rock ptarmigan males in the Arctic. Males in these two whitetailed ptarmigan populations travel with their females only during courtship and egg-laying (Table 13.3); they desert their females during incubation, when the chance of a female renesting ceases. Because of the long life span of males in these two populations, males can afford to invest less in reproduction of the current season than in that of the future. The two other ptarmigan that inhabit highrisk, Arctic environments and have high mortality rates and reduced longevity cannot. Thus, in white-tailed ptarmigan, males and females invest less in each other than do the other two species. Monogamy can be explained as a by-product of the fact that the best nesting habitat is within the territory of the male. Further, the synchronous and brief nesting season and a breeding sex ratio in favor of males reduce the male's opportunity for polygamy (Emlen & Oring 1977). Female rock ptarmigan appear more aggressive than female white-tailed ptarmigan but less aggressive than female willow ptarmigan (Schmidt 1969, Watson 1972, Hannon 1983, 1984). Female rock ptarmigan do not actually defend territorial boundaries (Watson 1972), as do willow ptarmigan females (Hannon 1982), but they do aggressively interact with other females (Watson 1972). Also,
MATING SYSTEMS IN GROUSE
453
Table 13.3. Comparison of the abundance of predators, life-history parameters, and reproduction investments among the three ptarmigan species. White-tailed
Rock
Willow
Most Least
Intermediate Intermediate
Least Most
Edge of territory
Intermediate
More central
Life history parameters Clutch size Nesting success (%) Annual adult mortality (%)
5.5 64 39
8.2 73 57
9.1 78 68
Investments Male vigilance and decoy behavior
F(N)
F N (Y)a
F NY
Female agrressiveness to other females
Least
Intermediate
Most
Hypothesis of choice male vs. territory
Territory most
Intermediate
Male Most
Ecological parameters Abundance of predators Predators of nest Predators of adults Nest location
F =for female prior to nesting, N = for female during nesting. Y = for female with young, parentheses indicate that it was seen occasionally in some populations—after Wittenberger (1978). Sources: Nest location: Weeden 1959b, Choate 1963a, Schmidt 1969, Gardarsson 1971; Clutch size: Choate 1963a, Giesen et al. 1980; Nesting success: Weeden 1959b, Choate 1963a, Watson 1965, Bergerud 1970a, Braun & Rogers 1971, Weeden & Theberge 1972, Watson & O'Hare 1979; Annual adult mortality: Choate 1963a, Watson 1965, Braun 1969, Bergerud 1970a, Weeden & Theberge 1972, Myrberget 1976, Hannon & Smith 1984; Male vigilance: Wittenberger 1978; Female aggressiveness: Watson & Jenkins 1964, Schmidt 1969, Watson 1972, Hannon 1982, 1984.
rock ptarmigan hens sing less frequently than do those of willow ptarmigan (Watson 1972). In rock ptarmigan the investment of females in males is more than in whitetailed ptarmigan and less than in willow ptarmigan. The male provides vigilance and decoy behavior for the female during courtship, egg-laying, and incubation (Table 13.3), but his defense of territorial boundaries is less intense than that of willow ptarmigan. A few males defend the female and chicks when the latter first hatch (Watson 1972). Female willow ptarmigan select aggressive males with large territories in both Scotland and Chilkat Pass populations (Jenkins et al. 1963, Hannon 1983, Chap. 10). In willow ptarmigan males defend outer boundaries of territories to maximize space and thereby attract females that are ready to nest (Fig. 13.5); these territo-
454
A. T. BERGERUD
ries are generally contiguous (Chap. 10). Males with small territories commonly go without females. However, it is difficult to decide whether a female selects against a male's lack of aggressiveness and vigilance or the small area of nesting space he controls. The male willow ptarmigan remains with the hen through nesting and broodrearing periods, unlike male rock and white-tailed ptarmigan. When the willow ptarmigan hen is on the nest, the cock remains nearby in diurnal cover commonly overlooking the nest (Figs. 13.5, 13.6). Males at Chilkat Pass were commonly only 50 m from the nest (Hannon 1982). Hannon removed males from territories in each of 3 years. In 1 of 3 years the females that were without male assistance during incubation showed reduced nesting success compared with hens that could rely on male vigilance and decoy behavior. In addition, the distraction behavior of males accompanying broods is intense (Table 13.2, Mercer 1967). Willow ptarmigan broods use dense, tall willow and birch cover, whereas rock ptarmigan broods use low shrub cover; white-tailed ptarmigan females with broods go to rocky habitats with good visibility (Weeden 1959b, Schmidt 1969, Fig. 8.9). Ptarmigan chicks exposed to the low, Arctic-alpine temperatures commonly require brooding in inclement weather over 80% of the time during their first 2 weeks of life (Theberge & West 1973, Pedersen & Steen 1979). Brooding rock and white-tailed ptarmigan females can observe approaching predators, whereas female willow ptarmigan and their broods, in dense cover, would face increased risks in the absence of male vigilance when the females brooded the chicks.
13.3.2 Female choice of conspicuous plumage The conspicuous plumage of male willow and rock ptarmigan during the breeding season cannot be explained by sexual selection for these plumages in the absence of differential natural-selection pressure by predators. The white plumage of male rock ptarmigan represents the retention of winter plumage and is in place before spring pairing. Male willow ptarmigan establish territories in March and April, competition begins while they are still in winter garb, and many pair before acquiring their bright, pigmented plumage. In Scotland and Newfoundland, where gyrfalcons are not present, male rock ptarmigan molt their winter plumage synchronously with females (Watson 1973, and pers. observations). Similarly, in white-tailed ptarmigan populations not exposed to serious raptor predation, the spring molt is reasonably synchronous and both sexes are equally cryptic (Braun & Rogers 1971). The aggressive, conspicuous behavior of males in the spring has a physiological basis in testosterone levels. Males that were implanted with testosterone became more aggressive and increased the size of their territories (MacDonald 1970, Watson 1970, Watson & Parr 1981). A rise in testosterone level is known to initiate the molt of white plumage to conspicuous breeding plumage in willow
MATING SYSTEMS IN GROUSE
455
Fig. 13.5. Top: territories of male willow ptarmigan at Chilkat Pass, British Columbia, in 1957 (adapted from Weeden 1959b). Males defend outer boundaries to maximize space and attract females ready to nest. Bottom: territories of females at Chilkat Pass in 1981 (adapted from Hannon 1982). Females defend space surrounding their nests within male territories (not shown), and female territories commonly do not abut.
456
A. T. BERGERUD
Fig. 13.6. Diurnal lookouts of male ptarmigan at Chilkat Pass in 1957 (from Weeden 1959b).
MATING SYSTEMS IN GROUSE
457
ptarmigan (Stokkan 1979b), but in rock ptarmigan male testosterone levels inhibit the molt of white plumage (MacDonald 1970). Thus males of both species become more in contrast to their backgrounds as they become more aggressive. If females choose aggressive males on the basis of either behavior or the large territories males may possess, they will in turn select males with the most conspicuous plumage. By choosing a large territory a female also secures an aggressive male (Watson & Miller 1971) with conspicuous plumage. A male that is vigilant in watching for other males would have the necessary behavioral requisites for sentinel duty. But if he repeatedly flew to boundaries for defense, the female would need to accompany him before incubation to benefit from his vigilance for predators.
13.4 Food shortage and monogamy Wittenberger (1978) hypothesized that ptarmigan were monogamous because of a food shortage. He felt that males provided two contributions to female nesting success: (1) they defended territories that provided females with exclusive foraging areas and (2) they provided vigilance for predators, which allowed females to devote more time to feeding. He postulated that ptarmigan were faced with a food shortage because of the diverse nature of their diet. He argued that the promiscuous grouse had greater food supplies because of their monophagous food habits, depending on the leaves of dominant shrub or tree species. But ptarmigan are just as monophagous as grouse in food preferences. The dominant winter and spring foods of ptarmigan throughout the Holarctic region are the buds, leaves, and twigs of either birch (Betula spp.) or willow (Salix spp.). In some populations birds eat mostly birch, in others mostly willow. Usually one of these plant genera accounts for over 85 % of the diet preceding the growing season (Watson 1964, West & Meng 1966, Weeden 1969, Bergerud & Mercer 1972, May & Braun 1972, Moss 1973). Red grouse (L. /. scoticus), a subspecies of willow ptarmigan, eat mostly heather (Calluna vulgaris). Even southern ptarmigan populations, in which birds do not eat willow and birch, forage mostly on two to three species of common shrubs (Watson 1964, Bergerud & Mercer 1972). Also, in some populations ptarmigan females gain weight during the winter and spring (May 1975, Fig. 10.26). The low in the weight cycle for ptarmigan females, like other grouse, occurs during incubation (Fig. 16.11). Abundance of food was quantified in two studies when cyclic ptarmigan populations were at "highs"; neither population showed a food shortage (Gardarsson 1971, Figs. 9.9, 10.4). In both study areas birds spent only a small portion of their time each day feeding. If ptarmigan face local food shortage, this should be in midwinter when snow is deep and daylight is limited. There should be no food shortage in spring, when days are long and more plants are exposed by snowmelt. Also, all ptarmigan nest before the appearance of new growth in the spring. If
458
A. T. BERGERUD
food was a problem, they could delay nesting, as survival is high for late-hatched young (Bergerud 1970a, Parker 1981). Williams et al. (1980) quantified the activities of large numbers of male and female willow ptarmigan just before incubation. Males stood and watched 72% of the time, whereas females stood or crouched 47% of the time. Males fed 12% of the time and females 30%. These females could not have faced a food problem if they fed only 30% of the time and stood inactive 47% of the time. Female ptarmigan have the smallest prelaying ranges of any of the grouse. A monogamous red grouse (willow ptarmigan) female apparently can find sufficient heather for her well-being in an area only 1/1,000 the size of that traveled by a monophagous sage grouse hen. If food was in short supply for monogamous ptarmigan, their preincubation ranges should be larger than that of promiscuous species—not the reverse (Bradbury 1981). Wittenberger (1978, p. 131) also stressed male vigilance, and stated: "It would be interesting to know whether monogamy is associated with more open habitats because food availability is generally low, because predation vulnerability is generally high, or because of other factors. I can find no data to test these hypotheses." The data presented here support the predation-vulnerability hypothesis.
13.5 The female view of the pair bond Female ptarmigan should prefer to be monogamous and not share their male with another female. By being the sole occupant of a male's territory, she also has his space in which to select good nesting cover; the uniform spacing of male territories means that nests will be far apart (Fig. 13.5), unless a hen nests on a perimeter and the adjacent female does also. Generally, females select males with the largest territories (Jenkins et al. 1963, Hannon 1982, Chap. 10). The hypothesis that the sizes of territories are related to food quality has not been demonstrated (Lance 1978b, Miller & Watson 1978). Vigilance and deflection behavior of males should be unshareable with other females. A female first travels with the male while she looks for a nest site. I watched a female that was oblivious to danger scraping a nest; her watchful mate spotted a gyrfalcon, called, and the female crouched and remained motionless. Two females could not travel with a single male as inconspicuously as one. Females are often not in the same stage of molt (Fig. 13.4). The inconspicuousness of a pigmented female prospecting on bare substrates would be compromised by white or partly white females also looking in brown substrates. The conspicuous male remains apart from the prospecting female, where he watches for her; predators attracted to him would overlook the cryptic female nearby. Females also benefit from male vigilance by going to males when off the nest. These brief feeding bouts should be vigorous (Anglestam 1983) and vigilance
MATING SYSTEMS IN GROUSE
459
should be helpful. Two females would have difficulty visiting the male simultaneously, and three birds should be more conspicuous than two. Weeden (1959b) documented that when the female was absent, willow ptarmigan males selected diurnal cover that provided a vantage point and that had overhead cover. It would be difficult for a male to select a site that would provide an open view and cover, and yet also be in view of two widely spaced nests. This is especially true because females would select their sites at different times. Males that left their vantage point with one female could not watch for the second female that remained on her eggs. In polygynous willow ptarmigan trios, Hannon (1982) documented that the male generally left their territories with the first brood that hatched. Again, it would be impossible for a male to provide distraction and deflection displays for two separate broods. Gang broods do not form in ptarmigan until chicks are able to fly. Actually, females that share the same male with another female are infrequent in ptarmigan. Three examples of one male and two females in one territory are: 22 of 296 (7 %) female rock ptarmigan in Scotland (Watson 1965), 16 of 243 (7 %) female white-tailed ptarmigan in Colorado (Braun & Rogers 1971), and 5 of 57 (9%) at Chilkat Pass (Hannon 1982). Some females may attempt to return to previous nesting areas regardless of sharing if nesting losses in the population are high and they enjoyed previous nesting success. A possible tactic may be that the second hen in some of these trios joined the male after the first hen was incubating and no longer defending her nesting space (Fig. 10.21). Conceivably, the polygyny threshold could be exceeded in some large territories—two females might both be in good nesting cover and farther apart than if the second hen selected another male with a small territory. But even in a dense, red grouse population in Scotland, one female of each polygynous pair had low reproductive success (Miller & Watson 1978): most successful bigamous hen —4.3 young/female, n = 9 least successful bigamous hen—0.4 young/female, n = 9 monogamous (control) hens—4.1 young/female, n = 40 One of the bigamous hens died, three disappeared, three raised no young, and two fledged two chicks each. These results are consistent with the view that vigilance and decoy behavior are important components of female fitness and that this male quality is not shareable. The hypothesis that explains monogamy from the female's point of view is Wittenberger and Tilson's (1980, p. 200) hypothesis No. 2, "Monogamy should evolve in territorial species if pairing with an [available] unmated male is always better than pairing with an already mated male."
460
A. T. BERGERUD
13.6 The male view of the pair bond Males should attempt to be opportunistically polygynous. Hannon (1982) showed that males attempted to prevent their hens from attacking stuffed, dummy females. Miller and Watson (1978) documented that polygynous males enjoyed a higher breeding success (more young in autumn) than monogamous males, even though bigamous females produced fewer chicks/female than did monogamous females. If a male is to deflect predators and incur risk, he should fertilize females as soon as possible. Females faced with effective raptors should be amenable to breeding after the nest-site decision — early breeding should help keep the male monogamous to protect his investment. In contrast, in those ptarmigan populations where raptor risk is low but nest predation is relatively high, females should invest more time in the nest-site decision and withhold mating if male vigilance and deflection behavior are less crucial than the nest-location decision. After a male has mated, he should be more concerned with vigilance and deflection. If his female dies he has little chance for a second female, because of the shortage of females and because the best territories for attracting females are taken. Further, his breeding chances decline as the season advances because females become progressively tied to their nest site and are reluctant to shift (cf. Hannon 1983). Males should evaluate the trade-offs between survival and mating opportunities in deciding whether to display vigilant and deflective behavior. A male may opt for a safer life-style and try for future breedings, or invest heavily in his current female with the inherent risks. The effective, breeding sex ratio is the key variable in the decision. His wisdom is measured in descendants. Consider two options: a low-risk option where the female is on her own after mating, and a high-risk option of continued male vigilance and deflection behavior. I calculated a pair of recurrence relations to generate descendants in year n to year n + 1 where: Mn(Fn) = qm(qf) = y= s=
the number of male (female) descendants alive in nth year male (female) annual mortality rates male mating success the number of juveniles per nest (each sex) that survive till 12 months;
Then: Mn + 1 = (ys + (1 - 0m)) Mn + sFn and Fn + 1 = ysMa + (S + I - qf) Fn Now taking MQ = 1, F0 = 0, the number of surviving descendants in any year may be determined recursively. Assumptions in the equation generally met are:
MATING SYSTEMS IN GROUSE
461
no inbreeding (approximated for small n), juvenile males and females have similar mortality rates, and all living females breed. For the low-risk option I used the values from white-tailed ptarmigan studied in Colorado by Braun (1969), Schmidt (1969, Chap. 8) and Braun and Rogers (1971). Their populations live south of gyrfalcons and the birds have low mortality rates. In effect males take little risk because there are few raptors and the high, foggy mountains are safe habitats. The risk to females from raptors is also small, but females have a higher mortality rate than males because they have contact with more ground predators. I equated this higher breeding mortality of females with the higher losses of females in northern populations because of raptors, if the males in these northern populations would choose the low-risk option. Hence, we feel the 57:43 (n = 564) sex ratio of this Colorado population is an approximation of the differential sex ratio and mortality that might result if the low-risk option was chosen by males coexisting with gyrfalcons (cf. Hannon 1984). Values for the Colorado white-tailed ptarmigan population are: y = 0.73, qm = 0.31, qi = 0.46 and s = 0.44. The total numbers of descendants of a male in this population with the low-risk option are n = 1 to n = 5: 1.3, 1.8, 2.5, 3.4, and 4.6. At the other end of the risk continuum is the dangerous environment inhabitated by rock ptarmigan in Iceland. The sex ratio of adults there was 47:53 (n = 5,009) (Gardarsson 1971, Chap. 9). Using Gardarsson (1971) the values are: y = 1.00, qm = 0.65, qt = 0.47, o = 1.93, and s = 0.72. The numbers of descendants n = 1 to n = 5 are: 1.8, 3.3, 6.3, 11.8, and 22.3. The high-risk option in Iceland resulted in far more descendants for the male than the low-risk option, 22.3 versus 4.6, in n = 5. Males that risk their lives in dangerous environments can expect to have more descendants than males that face low-risk situations, because of more-balanced sex ratios in the high-risk situation and hence increased breeding opportunities for their male offspring. The territorial system of males and the selection by females of males that have large territories mean that a significant proportion of males do not breed in a monogamous system when the mortality rate of females is higher than males. Monogamy from the male viewpoint results from Wittenberger and Tilson's (1980, p. 200) hypothesis no. 4, "Monogamy should evolve even though the polygyny threshold is exceeded if aggression by mated females prevents males from acquiring additional mates."
13.7 Dispersed polygyny It is not a long step from white-tailed ptarmigan, which are almost polygynous, to forest grouse, in which males display a dispersed system of advertising sites and each male travels to a few display posts. Female forest grouse search a preincubation range about the same size as that of white-tailed females (Table 14.2)
462
A. T. BERGERUD
(nest predation is similar, Table 15.2), and male forest grouse can be spaced at densities similar to those of whitetails (Fig. 8.7, 14.12). Nor is there a long stride from the dispersed distribution of forest grouse males to the clumped distribution of polygynous males at communal display grounds (leks). Many authors have commented on the clumping of forest grouse males (Blackford 1963, Gullion 1967, Anderson 1973, Herzog 1977, McNicoll 1978, Little 1978, Lewis 1985, Figs. 2.16, 13.7, 14.12). One can see how a cluster of males might benefit from the "you first" principle relative to predation. I have noted that when a blue grouse hooter became silent when a raptor passed over, so did other males out of sight. McNicoll (1978) noted that blue grouse males sang in social groups. This could be "stimulus-pooling" or, simply, keeping up with the competition. The major difference between the mating systems of forest grouse (Fig. 13.8) and ptarmigan (Fig. 13.1) is that male forest grouse remain stationary, with females coming to them (Fig. 14.6), and do not accompany females or defend territorial boundaries (Fig. 14.7). Since the preincubation ranges of females are
Fig. 13.7. Distribution of male territories and the prelaying ranges of females on Stuart Island, Washington, in 1976. The males are clumped on ridges. The females are spaced and generally closer to one male than to any other male. One male moved to be closer to females when another male was shot.
463
MATING SYSTEMS IN GROUSE
relatively small (Table 14.2), it is conceivable that, energetically, males could defend the nesting habitat leading to monogamy (Fig. 13.1). Males could travel boundaries and use song to repel other males despite the concealing canopy. At least for spruce and blue grouse, mobile males could remain safe by using advertising posts in conifer cover. The ruffed grouse male might be at a disadvantage if he left his secure drumming log for less safe sites. Unlike that of ptarmigan on the tundra, vigilance of male forest grouse would not be helpful to females because of the short sighting distances in forest canopies (Fig. 13.8). Still, if the female and male traveled as a pair, which is the case for hazel grouse (Tetrastes bonasd) (Hjorth 1970), the male could deflect predation from the female, perhaps at a greater mortality price than that of male ptarmigan. However, two birds would be more conspicuous to ground predators than would one. Further, the female can still benefit from a nearby male that is advertising. He may attract raptors, and when he silenced, the female could be forewarned. REDUCED NESTING COVER
VARIABLE NEST PREDATION
AVIAN PREDATORS
LITTLE WARNING;
SEARCH INTERMEDIATE SIZE PREINCUBATION RANGES
VIGILANCE NOT HELPFUL TO
SECRETIVE, NEST AWAY FROM AND
REQUIRE SAFE SITES TO ADVERTISE
SELECT ONLY SAFEST BREEDING SITES NEAR NESTING RANGES
ONLY SOME AND
CLOSED CANOPY
ADVERTISE BREED WITH
SOME OF THESE
DISPERSAL POLYGYNY Fig. 13.8. Proposed model for the evolution of dispersed polygyny in forest grouse.
464
A. T. BERGERUD
Females dictate the dispersal pattern of males (Bradbury 1981). Female forest grouse wish to remain secretive, spaced away from other nesting females and advertising males who would display to them. Hence females should mate only with males who behave predictably, do not follow them, and stay away from nesting habitats. The nests of forest grouse are away from males (Bump et al. 1947, Lance 1970, Herzog 1977) and are more frequent between male territories (Ellison 1973). There should be a compromise of being away from males yet not so far that a female must trespass across the range of another female to reach the male (Fig. 13.7). If another female is between her and a male, a second male might be able to move between the ranges of the two females and succeed in breeding (Fig. 14.12). The second requirement of a forest female may be to breed with an appropriately placed male only if his display site is relatively safe. The display sites of males are known to predators, and ambush is possible. The safety of a new site may be an "unknown" and other things being equal, the female may choose the familiar to the novel. Since females show philopatry to nesting ranges they should revisit last year's advertising locations, and such sites should be at a premium in the ideal-dominance competition of males. Thus with immediate survival and nesting success at stake, there is a continuum of attractiveness of display sites and males to females. Yearling males cannot compete effectively. They were not present last season, they may arrive on the range later than adults; and they are less familiar to adult hens. Polygyny in forest grouse results because nearly all females, both yearlings and adults, try to nest but yearling males may not be in the effective competition. Hence there can be about 1.4 breeding females to each adult male in spruce and blue grouse in populations with a 30% turnover and a 50:50 ratio of males and females. Yearling males are more cryptic than adult males and resemble females in having duller plumage. One hypothesis to explain this delayed acquisition of adult plumage in passerine birds is that yearling males benefit by mimicry of females, thus deceiving adult males and being allowed to prospect for sites and mates on male territories (Rohwer et at. 1980). This explanation would not hold for forest grouse. Forest grouse do not have distinct defended boundaries, and nondisplaying adults commonly trespass. I have observed four adult blue grouse males come to a taped female call. Prospecting as a female mimic carries the penalty for the yearling of unwelcomed attention by adult males. The femalelike plumage is presumably the most cryptic, and this explanation is sufficient to explain the lookalikes. In ruffed grouse, which face the heaviest predation in the breeding season, even adult males display cryptic female color patterns and cannot be distinguished by plumage. However, the convergence of male and female color patterns in ruffed grouse does not explain why both males and females have ruffs that can be conspicuous. The cryptic color of yearling males results from the polygynous
MATING SYSTEMS IN GROUSE
465
mating system. Since yearlings have a reduced chance of breeding, they should not compromise their safety for more conspicuous plumage until it is needed for successful breeding as adults (Selander 1972).
13.8 Clumped polygyny There is not a great deal of difference between a trio of blue grouse males singing together in a montane habitat and a like number of prairie chickens or sharptails advertising at a communal arena at a forest interface. Blue grouse males are partly repulsed because they stay in trees or at other safe sites. The smallest groups of prairie chickens and sharptails also occur where forest meets steppe and the horizon and predator approaches are obscured by canopies (Amman 1957, Svedarsky et al. 1982). Mossop (pers. comm.) reports that displaying northern sharptails in the Yukon dance in small groups in forest clearings frequented by goshawks. The cocks appear nervous as daylight builds and leave early. We might expect that displays and group size would be tempered and more ephemeral with reduced sighting distances at forest edges. The dispersal continuum of polygynous males grades with a continuum in habitat visibility and safety. A number of hypotheses have been proposed to explain a lek mating system. First we can think of hypotheses that relate to the advantage for males: (1) the stimulation-conspicuousness hypothesis—males display collectively to attract more females (Lack 1939, Hjorth 1970); (2) the mutual benefit hypothesis—males benefit by traveling in flocks for foraging and predator vigilance (DeVos 1979). The advantages to females of leks might include: (1) the mate comparison hypothesis—females can compare between males at a lek and select superior males (Bradbury 1981); (2) the male vigilance hypothesis —females benefit from communal groups because males are vigilant and detect approaching predators (Wittenberger 198 la); and (3) the least costly male hypothesis —females select the least costly male that must be at communal arenas (Wrangham 1980). The male benefit hypotheses should not be valid. The average male does not benefit from a system where males contest for females at a lek. In contrast to a monogamous system where most males have breeding rights if the sex ratio is balanced, in clumped polygyny, extremely few males do the breeding (Table 7.9) and sexual selection is intense. Steppe males cannot defend economically the large preincubation ranges searched by females (Fig. 16.2) and thus force females into a prolonged pair bond as in ptarmigan. Communal polygyny must be explained in terms of female fitness. The least costly male model is the most parsimonious of the female hypotheses. With the exception of Partridge's (1980) study of mate choice in fruit flies (Drosophila spp.), there is precious little evidence that females select "good genes." The least costly male hypothesis avoids the problems of the lek paradox (Taylor & Williams 1982) and the sexy-son and handicap arguments (Kirkpatrick 1985, 1986),
466
A. T. BERGERUD
and is consistent with the synthesis of this book, of the primary importance of successful nesting in fitness and demography. In my model (Fig. 13.9), females must search large preincubation ranges of 80-800 ha (Fig. 16.2) to locate safe sites, given the high density of nest predators (Fig. 15.8). In addition, the heterogeneity of nesting cover and variation between years should increase searching ranges. Also, nesting females should avoid the habitats visited by males when the cocks are away from the lek feeding (cf. Rothenmaier 1979, Chap. 5). The ranges of steppe grouse are not large because females have food problems (Bradbury 1981, Oring 1982). If food was a problem, the ranges hens search during laying should be as large or greater than those of prelaying, but they are considerably smaller (Tables 6.9, 14.2). Again, when females are away from the nest during incubation they commonly repeat-visit certain locales where other hens may also be feeding. Feeding bouts are vigorous, birds stuff themselves with limited movements and quickly return to their nests. The smallest preincubation range of any grouse is that of willow ptarmigan (Table 14.2); willow ptarmigan in Scotland (red grouse) can live in only 5 ha of a monoculture of heather (Calluna vulgaris) yet eat only a fraction of the green heather growth—2-3% (Savory 1978). In addition, hens are unlikely to run short of food because they can lengthen their guts in response to poor food (Moss 1983). Food variability is not the explanation for difference in the size of prelaying ranges of the grouse species facing different pressures from predation in selection of nest sites (Fig. 16.2). Females of the steppe species must be especially secretive in the vicinity of the nest both because of mammal predators and because the open horizon permits egg-robbing by corvids (Bowen 1971, Autenrieth 1981, Gratson pers. comm.). Females cannot afford to have males display near nests or follow them when they return to nesting habitats. Nor is male vigilance helpful since corvids could orient their search near waiting males. Several nest predators of the steppe are nocturnal; vigilance could not assist hens, and the presence of the male would facilitate nest-searching when it was dark and hens could not take evasive action. Steppe females nest away from leks, approximately halfway between adjacent arenas (Fig. 13.10, 13.11). The lek of copulation or banding is commonly not the lek nearest the nest site (Fig. 13.10). Females thus can bypass the displaying males if such activity compromises the nest location. Such withholding of privileges would force males away from nesting ranges. The farther females travel to mate with males the more the males must be clumped (Fig. 13.12). Female choice is then a major force in the location of arenas and the total males present. An alternative explanation to predation for the locations of nests away from leks is that there is more space in ever-increasing concentric circles around an arena (Fig. 13.11). This space argument does not explain why the mean and mode distances of nests from leks are midway between display grounds (Fig. 13.10).
MATING SYSTEMS IN GROUSE
PAT CHY HERBA CEOUS NE ST CO\,'ER
LARGE PRELAYING RANGES SEARCHED FOR NEST SITE
EFFEL~TIVE AND C( OMMON NE ST PREDt ITORS
SECRETIVE BEHAVIOR AWAY FROM
467
OPi F/V HABITA TS AND Ft ~W RAP! -QRS
SEEKS SAFE BREEDING SITES AWAY FROM NEST SITES
CAN'T AFFORD TO DEFEND RESOURCES NOR FIND VIGILANCE NOT HELPFUL.
DISPLAY AWAY FROM NEST RANGES STIMULUS
POOLING
DISPLAY AT COMMUNAL ARENAS Fig. 13.9. Proposed model for the evolution of clumped polygyny (a lek mating system) in steppe grouse.
468
A. T. BERGERUD
Fig. 13.10. Distribution of the distances between nests of steppe grouse and leks, using nests found by radiotelemetry. Mean distance was the average that nests were from the nearest lek, or the lek of capture. The distance between leks indicates the mean distance between leks in that population. Nests are spaced away from leks and the lek of capture is often not the nearest lek to the nest. (Data from Bernhoft 1969, Christenson 1970, Wallestad & Pyrah 1974, Kohn 1976, Svedarsky 1979).
MATING SYSTEMS IN GROUSE
469
Fig. 13.11. Location of nests of sharp-tailed grouse around leks in North Dakota. Note the space adjacent to leks where few females nested. Sources: Bernhoft 1969, Christenson 1970, Kohn 1976.
470
A. T. BERGERUD
Fig. 13.12. Top: Males must be more clumped if females come long distances to breed. Bottom: Females should search for nest sites at maximum distances between leks. If the density of females in the center becomes too high, yearling females should move out beyond the leks to nest. Female progeny from yearling hens might pioneer even farther afield for nest sites.
MATING SYSTEMS IN GROUSE
471
The space argument would predict that the frequency of nests from the lek of copulation would continue to increase as one approached an adjacent lek that was somewhat near the lek of copulation compared to the overall mean distance between leks. This relationship did not hold for sharptails in North Dakota (Fig. 13.11). Actually, females return to the breeding range and search for nest sites between leks (Fig. 5.6), before going to males to breed (Fig. 5.6). By being choosy, a female can pick a lek sufficiently far so as not to attract males yet not so far that she must expose herself to extensive travel. Whether the stimulus pooling by males plays a part in her decision remains to be tested. Generally, leks with more males also attract more females, but cause and effect are not clear. If a population is stable in numbers females should concentrate their search between leks (Fig. 13.12) and males should establish some leks outside the preferred nesting habitat (Fig. 6.4, cf. Westemeier 197la), in low cover where they can see ground predators. But if the population is increasing, a point should be reached when females, avoiding both males and other females, become too concentrated in the center; then some females, probably yearlings, should move past the nearest leks and localize away from the female concentration. Yearling nests should be closer to leks than should those of adults (Table 6.4), and they should have reduced nesting success (Table 15.4). Returning progeny from the yearling hens could then pioneer even farther out into new nesting habitats (Fig. 13.12). Ultimately some yearling males should move beyond the outer females and form a new lek. The sequence may be one of leapfrog, females then males then females, moving past each other in their distributions, but males should avoid the best nesting cover for display purposes.
13.9 Summary The three North American ptarmigan species are monogamous, whereas the other six North American grouse are polygynous. In the Arctic there are few nest predators, which means that ptarmigan females should be more prepared to nest nearer each other than polygynous grouse, which lose more nests to predators. Hence ptarmigan females search relatively small prelaying ranges for nest sites. The small space requirement of females allows males to economically defend with territorial behavior the nesting resource (cover and space) that females will later seek, and require, and thus the fitness of a male is determined by the quality of the nesting resource he controls that results in selection by females. In the Arctic the open habitat, continuous daylight, and presence of effective avian predators have resulted in females also selecting conspicuous, vigilant males in a prolonged pair bond. These males reduce predation risk to females by deflecting predators away from females during nest searching and egg-laying in all three species, and from hens with chicks in willow ptarmigan.
472
A. T. BERGERUD
It is suggested that the grouse represent a continuum in mating systems between the most monogamous willow ptarmigan to the most polygynous sage grouse. This continuum of monogamy to polygyny is graded with an increase in the size of the preincubation range and increasing nest predation. In general, females must select the least costly male relative to predation pressure. By withholding breeding opportunities, females can dictate the distribution of males, resulting in dispersed, solitary males in forest cover and clumped males (lek system) in open steppe habitats.
14
Survival and Breeding Strategies of Grouse A. T. Bergerud and M. W. Gratson
14.1 Introduction The first rule of natural selection is not to maximize reproduction, but to stay alive. Grouse must choose activities that maximize survival before they can proceed to those that maximize lifetime reproduction. The solution to these two fitness problems, to live and to breed, depends upon the evolution of satisfactory life-history strategies and tactics through the survival of fit individuals. A strategy may be defined as "a set of rules that determines which of several alternative behavioral tactics is employed to solve a particular problem or to achieve a particular goal," and a tactic as "a behavior pattern within a a species repertoire that can potentially be employed in a specified context to achieve a particular goal" (Wittenberger 198la, p. 622). An emerging realization from sociobiology is that many aspects of animal behavior can be predicted on the basis of a few environmental variables (Wilson 1975, Wittenberger 1981a). For grouse, the problem for both sexes is that of living and reproducing within the constraints imposed by vegetative cover and natural predators. Grouse have successfully adapted to the weather and food regimes of their habitats. Their demographic challenge is that of keeping abreast with coevolving predator populations. If we can understand the strategies of individuals, we will be in a better position to understand the demography of populations. For the individual, the goal is fitness, the strategy is survival to reproduction, and the tactics are the behavioral options chosen for success. Behavioral options differ between the sexes. The male's primary concern is to advertise successfully for females. Breeding is only one of several problems for 473
474
A. T. BERGERUD AND M. W. GRATSON
the female; probably the most critical is determining where to nest. She has the sole responsibility of rearing the chicks; only the willow ptarmigan (Lagopus lagopus) male assists in care of the young. Thus, the female's investment is much larger than that of the male's (Trivers 1972), and her behavior is moderated by a large number of environmental constraints. It may often facilitate our understanding to consider the two sexes as separate species, because their investments are so disparate (Trivers 1972). This chapter formulates some general predictive hypotheses explaining the tactics that shape overall survival and breeding strategies of male and female North American grouse. It emphasizes the environmental constraints imposed by habitats and their predators, and the relationships of these to evolved tactics. 14.2 Male advertising strategies A male's fitness depends upon his ability to breed with females. However, the cost of advertising his presence to females is increased vulnerability and conspicuousness to predators. The male's personal hazard must be weighed against his chances for success. He should advertise when he will influence the female's mate-choice decision to the greatest degree, and at safe times and places that are conducive to attracting females.
14.2.1 Advertising displays The major advertising tactic of the male is the use of acoustical displays (wing, tail, vocal, etc.; see descriptions in Hjorth 1970) that carry long distances. Many biologists attribute both female-attraction and male-repulsion functions to these long-range sounds, but we believe that the latter—acoustical display as an "enforcer" of territorial boundaries—has been overemphasized. Rather, we suggest that the principal role of these signals is to advertise location, a continuous reminder to females (and as a consequence, other males) that "I am here and available." Depending on their interest, receivers of the signals can approach or avoid the male's location. Individuals can recognize neighbors by their songs (Falls & McNicholl 1979, Sparling 1981). From the perspective of the male sender, the display is primarily directed at the other sex. When blue grouse (Dendragapus obscurus) males approach experimental arenas (Chaps. 1,2), they commonly stop in view of the female dummy, spread their feathers in display, and hoot. If another male appears, the hooting male ceases calling, suddenly becomes "sleek" and horizontal, and often growls. Similarly, if a male observes himself in a mirror he becomes sleek and attacks his image, but once he steps out of the sight of his reflection his feathers spread again, the esophageal air sac is reinflated, and he hoots. Bergerud has frequently attracted blue grouse males outside their normal advertising range with tape-recorded calls
SURVIVAL AND BREEDING STRATEGIES
475
of females, and these birds hooted beyond their boundaries. Also, yearling males commonly approach female test arenas where there are already hooting adults. Because blue grouse males cannot distinguish the sex of conspecifics by plumage, beyond a few meters (McNicholl 1978, Jamieson 1982), these yearlings have little chance of being recognized and the sound of the hooting male does not deter them from investigating female calls within the range of the older bird. Once adult and yearling males are close (1-3 m), however, the advertising adult may assume a threat posture and may displace the yearling (Jamieson 1982). Similar to blue grouse, male spruce grouse (Dendragapus canadensis) switch from vigorous sequences of display and loud wing-claps when alone to a sleek and horizontal appearance when faced with another male (MacDonald 1968). Again, once the intruding male departs, the resident returns to advertising wingclaps. Observing ruffed grouse (Bonasa umbellus), Archibald (1976b) has shown that in a displaying duel between two males each attempted to outdo the other in advertising his presence to females, not in intimidating his rival. Even Bergerud's rooster at home is not intimidated in his infinite predawn duel with the despot next door, and his repetitious efforts do not flag when no hens appear. Like the long-range acoustical signals of the forest grouse, those of the prairie chicken (Tympanuchus cupido) and sharp-tailed grouse (Tympanuchus phasianellus) —"booming", and "cooing"—are female-oriented (Kermott & Oring 1975, Sparling 1981), and are in marked contrast to the more muted sounds and sleek postures that these grouse use in contesting site dominance (cf. Hjorth 1970). A male must assess the risk that results from advertising in relation to the expected gain in breeding success. He can minimize this risk by advertising from a safe site (see also section 14.2.6). His site selection is not "ideal free" in the sense of Fretwell (Fretwell & Lucas 1969, Fretwell 1972), because other males compete for those preferred sites that are safe and that attract females. If the perceived cost of display at an inferior site exceeds the expected benefits, a male may choose not to advertise and instead may wait for the vacancy of a preferred site arising from a competitor's death. Nonadvertising adult males have been documented for blue grouse (Lewis & Zwickel 1980), ruffed grouse (Little 1978), and white-tailed ptarmigan (Lagopus leucurus) (Choate 1963a), and reported in the spruce (Olpinski 1980) and sharp-tailed grouse (Moyles & Boag 1981, Giesen pers. comm.). These nonadvertising males are not necessarily doomed to die, as argued for red grouse (Lagopus I. scoticus) (Jenkins et al. 1967, Watson 1985), but may actually enjoy increased longevity by remaining inconspicuous (Lewis & Zwickel 1982). Nor should the disappearance of nonadvertising males be attributed to death (cf. Rusch & Keith 197ib); such males may have assessed their display options and moved elsewhere. Silence should be considered a survival tactic and should occur when and where the perceived cost of advertisement exceeds the expected gain.
476
A. T. BERGERUD AND M. W. GRATSON
14.2.2. Fidelity as an advertising tactic Fidelity to an advertising site is both a survival and a reproductive tactic of male grouse. The advertising location of males is their focal point of activity throughout much of the year. Ruffed grouse cocks remain near the logs where they advertise for females not only in the spring, but all year long. The mean annual distance of 24 males from their logs in Minnesota was only 98 m (Eng 1959), and the mean annual capture-recovery distance for drumming males in Alberta was 79 m (Rusch & Keith 1971b). The ruffed grouse is rather unique in having a "finegrained" mixture (i.e., small patches relatively close to each other) of food and escape cover near female nesting coverts. Thus, the drumming male has the opportunity to be relatively sedentary. Similarly, blue grouse males may remain near their advertising locations except during winter (Bendell & Elliott 1967, Chap. 2). Even in winter, the ranges of male spruce grouse in some areas include their spring advertising sites (Ellison 1973, Herzog 1977a, Herzog & Boag 1978). Male sharp-tailed grouse and prairie chickens visit their leks throughout the year except in July and August, when males are molting and females have broods (Hamerstrom & Hamerstrom 1949, 1951, Amman 1957, Chap. 5). Sage grouse (Centrocercus urophasianus) cocks commonly begin displaying in midwinter and do not abandon their leks until late June (Chap. 7). Ptarmigan frequent their territories from March through June, and if weather permits and there is competition with juvenile males, they are on station again in the fall (Mercer 1967). Red grouse set up their territories in September, abandoning them for winter flocking only when weather interferes (Jenkins et al. 1963). Male fidelity enhances site ownership and emphasizes the established male's reliable presence and location to females. After the original investment, the adult male appears "locked-in" to the site he has chosen. He usually remains faithful to it even if the habitat deteriorates and the population declines (Bendell & Elliott 1967, Gullion 1967, Chap. 2). Only 4 of the 231 ruffed grouse males shifted activity centers between years (Little 1978). That these birds lived to return speaks well of the safety of the sites. Probably such males also were successful with females and can expect continued success. With few exceptions, adult males of the forest and tundra species return to their advertising location of the previous year (Hale & Dorney 1963, Bendell & Elliott 1967, Gullion & Marshall 1968, McNicholl 1978, Lewis & Zwickel 1981, Pedersen et al. 1983, Steen et al. 1985, Unander & Steen 1985, Chaps. 8, 9). The only blue grouse male that shifted advertising sites on Stuart Island, British Columbia (Chap. 2), moved from one site where Bergerud never located a female to a site where he observed three hens (Fig. 13.7). The lek species also show fidelity to previous advertising sites. The last surviving heath hen was a male that returned to its lek on Martha's Vineyard, Massachusetts, for 4 years after the last male companion had vanished (Gross 1928).
SURVIVAL AND BREEDING STRATEGIES
477
There were only six recorded switches in advertising leks among 21 banded cocks observed at least two springs in Wisconsin (Hamerstrom & Hamerstrom 1949). Occasionally, however, all birds on a lek will shift to a new location (e.g., Chap. 5), and we have enticed sharp-tailed grouse leks and individuals from other leks to new locations where we had placed groups of female decoys (Gratson unpubl. data). In fact, short movements from year to year are common. The degree of fidelity to these advertising sites should depend upon the stability of adjacent, female nesting habitats, and if females are not encountered nearby, males should shift. If nesting habitats change, so will lek locations. The fidelity and continued presence of males at display sites are tactics for holding a location that has proved successful in attracting females. Also, to the extent that males are familiar with such sites, fidelity should confer a degree of safety that would not otherwise be present. The resident must be on hand also when the new generation prospects for sites, as his presence asserts his holding power.
14.2.3 Advertising in the fall The male of several species advertises his location again in the fall. This appears timed to the dispersal of juvenile males from broods. In Minnesota and Wisconsin, juvenile ruffed grouse males leave broods and start to disperse in September (Eng 1959, Godfrey & Marshall 1969, Rodgers 1980), and some are drumming in October (Eng 1959, Gullion 1967). A correlation between the weight of droppings on logs (peak 16 October) and dispersal of 4-month-old males (Eng 1959) indicates that adults are more frequently present at advertising sites when juveniles start prospecting for sites. The display postures and calls of male sharp-tailed grouse in the fall appear oriented more toward other males than females (Kermott 1982). At this time of year juvenile males are attempting to establish sites (Moyles & Boag 1981, Kermott 1982, Chap. 5) but females are uncommon visitors to leks. Kermott (1982) hypothesized that the resumption of display in the fall was primarily for the adult male to reassert ownership in the face of new competition. If this hypothesis is valid, adult males should not advertise in the fall except in those species or populations in which juvenile males do compete for sites at that time. There is good support for this (Fig. 14.1). Blue grouse and sage grouse do not advertise in the fall; white-tailed ptarmigan advertise only briefly. Juvenile male blue grouse depart from the summer range in July and migrate with females to winter ranges (Lance 1967, Sopuck 1979). They first prospect for advertising sites as yearlings the following spring, but few or none hold territories. Adult males do not leave their advertising ranges in the late summer until the yearlings have left (McNicholl 1978, Sopuck 1979). Sage grouse yearlings do not generally visit leks until late spring, after most of the females have copulated with adult
478
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.1. Comparison among the nine North American grouse of when males first advertise for females and whether fall display occurs.
males and left to nest (Dalke et al. 1963, Eng 1963, Hartzler 1972, Petersen 1980). At this time, yearlings have achieved only 86% of the weight of adults (Beck & Braun 1978). The yearlings of this new generation are probably not as sexually mature as juveniles their first autumn, having attained only 75% of their adult weight by October (Patterson 1952). White-tailed ptarmigan hatch later than other grouse, in July (Braun & Roger, 1971). Winter, however, arrives earlier in the mountains than at lower altitudes, in September and October. Juveniles would thus have only 8 weeks on the breeding range before migrating downslope; under these conditions there is no time for competition in the fall. This hypothesis of fall display in the face of competition from juveniles gains additional support from data within species. Juvenile male willow ptarmigan compete for display sites in the fall on Newfoundland, at latitude 47°N (Mercer 1967), but in northern Norway, at 69°N, they do not and adults do not display in the fall (Chap. 11). In northern Norway, migration occurs before the juveniles are mature. At lower latitudes in Norway, however, adults resume display in the fall (Phillips & Aalerud 1980, Pedersen et al. 1983), and we predict that juveniles
SURVIVAL AND BREEDING STRATEGIES
479
at these sites should prospect in the fall. Pedersen et al. (1983) reported more calling by adults in September 1981 following high chick survival than in 1980 and 1982 when production was lower and there was less competition with juvenile males. It is interesting that in the Capercaillie (Tetrao urogallus), a European species and the largest and most dimorphic grouse, yearlings do not seek territories until their second fall or later (Hjorth 1970, Wegge & Larsen 1987). And, in this species there is fall advertisement by adults. In summary, in those species and populations in which it occurs, fall display appears to be a tactic whereby adult males familiarize the new generation with their ownership of advertising locations in anticipation of joint competition for females in the coming spring. All the grouse of North America that do not advertise in the fall have in common a life-history schedule in which the new generation does not compete for sites at that time. Early winter conditions or a comparatively slow development to physical and sexual maturity, or both, can preclude the opportunity for juvenile males to seek advertising posts in the fall.
14.2.4 Tactics of daily advertising Male grouse usually advertise in the early morning and again in the evening; except during the height of breeding there is little midday display (Fig. 14.2). They generally begin at or just before first light, although ruffed grouse may drum on moonlit nights (Archibald 1976b). The morning display is so precise in willow ptarmigan that one can predict the commencement of becking on Newfoundland within a few minutes (Bergerud & Mercer 1966). Most cocks began within a minute or so of the first bird to display, similar to the short intervals reported for ruffed (Aubin 1970, 1972, Archibald 1976b) and blue grouse (McNicholl 1978). In the lekking species, males generally go to the lek in a flock from their night roosts and begin calling almost as soon as they disperse to their advertising locations. With these species, too, one can predict the arrival of the first cocks and enter a blind at the edge of the lek just a few minutes before they are expected. It is common for males to begin their displays earlier in the morning, in relation to sunrise, during the peak of the breeding season. Steppe grouse males may even roost overnight at the lek for an early start (Hamerstrom & Hamerstrom 1949, Hartzler 1972, Emmons 1980). Both sharp-tailed grouse and prairie chickens arrive on the lek earlier on those days when females appear for breeding (Evans 1961, Sparling 1979). Sharp-tailed grouse shifted arrival times one-half hour earlier on 20 April 1976, but prairie chickens in the same area in Minnesota shifted to earlier times on 30 April the same year. In 1975 and 1977 no abrupt shifts were noted; sharp-tailed grouse cocks arrived at the lek only gradually earlier, consistent with the pattern of female visits in those years (Sparling 1979). Hjorth (1968) felt that the later arrival times of male blackcock (Tetrao tetrix) early in the breeding season could relate to the contrast of the bird on snow; he showed an abrupt switch to earlier arrival times about mid-late April when the
480
A.T. BERGERUD AND M.W. GRATSON
Fig. 14.2. The crepuscular nature of advertising by males and time of copulation by females. Top left: willow ptarmigan males on Brunette Island (Bergerud & Mercer 1966). Top right: drumming by ruffed grouse males in Minnesota (Archibald 1976). Middle left: diurnal hooting frequencies of blue grouse males on Vancouver Island (McNicholl 1978). Middle right: morning arrival of sharp-tailed grouse males and females on a lek in Alberta (Evans 1961). Bottom left: copulatons by sage grouse in Montana (Hartzler 1972). Bottom right: departures by sage grouse males from a lek in Wyoming (Rothenmaier 1979).
SURVIVAL AND BREEDING STRATEGIES
481
snow melted (see also Hjorth 1970). From the data for sharp-tailed grouse and prairie chickens (Sparling 1979), however, a more general explanation than the cryptic hypothesis is that the earlier arrival of cocks is the result of increased competition, heightened by the expectation of visiting hens, and the necessary condition of being in position to advertise and demonstrate site fidelity before the arrival of females. Although daily variations occur, depending partly upon the likelihood of breeding females, a general explanation for the restriction of advertising by males to crepuscular periods is that this is an antipredator tactic of both male and female grouse. Recently, Sparling (1979, 1983) questioned this hypothesis and proposed that prairie grouse display in the early and late daylight hours, when wind velocity and thermal turbulence are often reduced, to maximize signal transmission. However, we believe that this is of secondary importance to the steppe grouse, and that this hypothesis is insufficient to explain the crepuscular displays in the forest and tundra species. With Sparling's hypothesis one might expect that birds would prolong their advertising on calm days and cease advertising early if wind velocities were high. However, birds on leks stay longer on cloudy days and vigorously display even in windy weather during the peak of breeding (Hamerstrom & Hamerstrom 1973). Also, ptarmigan frequently display all day in foggy weather, when sound transmission would be reduced, and at times on windy days. Cool, calm days do occur in the Arctic, yet ptarmigan do not advertise late into the day under these conditions. Visibility is a more consistent environmental variable acting to constrain when grouse display than are velocity and turbulence of air currents. Curtailment of morning display and visits by females also seem to vary more with predator risk than with signal transmission quality. Morning display ceases about the time important diurnal predators become active. Ptarmigan at Chilkat Pass, British Columbia, stop displaying 2 hours after sunrise (Chap. 10), before the increased hunting activity of gyrfalcons (Falco rusticolus). Sharp-tailed grouse in the Yukon cease vigorous display before the arrival of goshawks (Accipiter gentilis) (Mossop pers. comm.). Steppe grouse appear more watchful as the morning progresses. Sage grouse commonly remain on the lek until approximately 1.5 hours after sunrise (Emmons 1980), at which time eagles (Aquila chrysaetos) can use the thermal currents (Hartzler 1974). Eagles frequently flush sage grouse males that stay later in the day and these grouse do not return even if the weather is favorable for signal transmission. On Vancouver Island, British Columbia, blue grouse males commonly hoot until midmorning, whereas ruffed grouse in the same valley may stop drumming within 2 hours after sunrise. Blue grouse males have a much lower, annual mortality rate than ruffed grouse (Fig. 15.4) and appear generally more secure against predators than other grouse species. The longer daily display period of this species (Fig. 14.2, Stewart 1967) is consistent with these observations and supports
482
A. T. BERGERUD AND M. W. GRATSON
the view that the daily timing of advertisement is tuned to predator risk rather than signal transmission. The safety hypothesis predicts that birds relatively safe from diurnal predators will display longer into the day than those more seriously exposed to predation. Sharp-tailed grouse in Nebraska can be expected to display longer than the same species living with goshawks in the Yukon; and ruffed grouse at Cedar Creek, Minnesota (Meslow 1966), should display later into the day than ruffed grouse at Cloquet, Minnesota, where goshawks are again an effective predator (Gullion 1967). Commencement of the evening display period is less precise. Blue grouse commonly begin hooting by 1600 hours (McNicholl 1978), whereas ruffed grouse may not start in until 1700-1800 hours (Fig. 14.2). The open-country sage grouse start even later, approximately sunset (Hartzler 1974). Gratson (Chap. 5) noted that sharp-tailed grouse in three relic populations displayed in the afternoon and evening only at the largest leks, an observation one might not expect if the timing of advertisement depended on maximizing signal transmission. On Moresby Island, British Columbia, some blue grouse males start to hoot a halfhour or more before the majority, perhaps stimulated by the presence of a nearby female. There is a sharp peak of activity just before dark when male blue grouse leave their relatively safe and usual display sites to make noisy landings on open spots and roads. Aerial chases, hoots, and whoots usually follow, and when they do the latter confirm that a female is being actively courted on the ground (Stirling & Bendell 1970, cf. Jamieson 1982). When darkness falls, the activity abruptly ceases (cf. Stewart 1967). The male, in deciding to advertise or not, must weigh the potential cost of his conspicuousness — a greater vulnerability to predators — with the anticipated advantages of greater access to females now or in the future. In the polygamous grouse most females visit the display sites of males for the short period in which there is just enough light to see (Fig. 14.2). The female risks conspicuousness only when predator risk is minimal, whereas the male, in attempting to influence her decision, must risk his life much more often. We can hypothesize that males with safer sites will display longer than those without; that yearlings, which generally display at sites of higher risk and with probably less success with females, will advertise less than adults; and that the display effort should be maximized when females are near and receptive. The crepuscular nature of male advertisement is most consistent with the view that it serves as an antipredator tactic. The advertising period of the male is sandwiched between diurnal and nocturnal predation periods, a time in which the female is prepared to risk increased conspicuousness in order to breed.
14.2.5 Advertising near females An exciting hypothesis is that advertising males space themselves to maximize encounters with females (Bradbury 1981, Oring 1982). It follows from this
SURVIVAL AND BREEDING STRATEGIES
483
hypothesis that males should attempt to display near areas where females will later nest. This is the converse of conventional theory, which postulates that the advertising male is the basic spacing unit around which the females locate themselves. The idea of a "sphere of influence," referring to nests located around leks (Hamerstrom 1939, Schwartz 1945), and the concept of the "nesting radius of sage grouse leks" (Braun et al. 1977) are two such hypotheses. This new hypothesis also is in contradiction with the older view that forest grouse select habitats on the basis of the availability of desirable cover for males (Gullion 1967, 1970a,b,c). It is perhaps natural for most grouse biologists to assume that males are the basic spacing unit. Grouse are commonly counted in the spring by first locating the advertising males and biologists usually search for nests by first locating display areas. Furthermore, males arrive at the breeding ground before the females. However, if males evolutionarily track females and can recognize desired habitats for nesting, male competition for sites near these nesting areas would favor the early and synchronous return of males in order to compete for females attracted to these areas. Several lines of evidence indicate that males of the steppe grouse species form their leks near nesting females. First, we know that these males show a wide tolerance range in the "cover" that is chosen; they frequently display on roads, at airports, at cattle salt licks, and on plowed fields, oil pads, or pipeline rights-of-way (Schwartz 1945, Hamerstrom et al. 1957, Lutz 1979, Horkel & Silvy 1980). The sites are open, usually slightly elevated, yet there are many areas that meet these requirements and are not selected. Second, invariably leks are located near nesting habitats (Fig. 14.3), and males frequently travel long distances from them to feed (Wallestad & Schladweiler 1974, Chap. 5). Also, the number of males per lek is commonly correlated with the abundance of nesting cover (Fig. 14.3, Hamerstrom et al. 1957, Brown 1967, Westemeier 197Ib, Pepper 1972). It is also not unusual for lek locations to change from one year to the next (e.g., Kirsh et al. 1973, Kohn 1976, Chap. 5). Henderson and Jackson (1967) reported the following for sharp-tailed grouse: 104 grounds (leks) were active 4 years or less; 73 grounds were active 5 years or longer; 38 grounds were active 10 years or longer; 14 grounds were active 15 years or longer; and only three of the 178 grounds were active all 21 years of the study. Field workers have been impressed with the permanence of leks without mentioning the permanence of surrounding nest cover. If the hypothesis that males distribute themselves in relation to females is tenable, males should change display locations if females shift. Svedarsky (1979, Chap. 6) showed that females nested mostly southeast of a lek (Fig. 6.4); there was no "radius" of dispersal. Males at this lek declined in number from 1977 to 1978, from 16 to 6, and a new lek appeared closer to the nesting females. Gratson (1983, Chap. 5), also documented that a new, sharp-tailed grouse lek appeared
484
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.3. Number of sharp-tailed grouse males at leks in Saskatchewan was correlated with abundance of nesting cover found within 1.6 km (Adapted from Pepper 1972).
in the midst of female ranges. Shifts in locations of other leks on his study area also appeared to occur to areas that hens frequented. Wallestad (1975a) documented that the removal of sagebrush (Artemisia ssp.) used for nesting resulted in a 63 % decline in the number of males on an adjacent strutting ground, and Autenrieth (1969, pers. comm.) found that a lek where sagebrush was killed would continue to be used if the sage in the surrounding habitats used by females
SURVIVAL AND BREEDING STRATEGIES
485
remained intact. Sharp-tailed grouse in Montana established new leks and the mean number of males per lek increased when the surrounding nest cover improved, but when grassland cover declined many leks were deserted (Fig. 14.4). We know that prairie grouse males themselves do not require tall grass on the lek; in fact, both sharp-tailed grouse and prairie chickens commonly abandon leks in tall cover to establish leks in recently burned or plowed areas nearby (Anderson 1969, Horak 1974, Cannon & Knopf 1979, Sexton & Gillespie 1979). Both Kirsch et al. (1973, pers. comm.) and Westemeier (1973) have experimented with creating new nesting cover for prairie grouse. Females preceded males in using the new sites by a year or so and may have become familiar with the newly seeded grassland as a result of taking broods there when the vegetation was still sparse (cf. Fig. 16.19). In an interesting experiment, Rippin removed all the sharp-tailed grouse cocks from a lek. The next year a new arena appeared 400 m from the old location (Rippin & Boag 1974a). Nesting cover had not been altered. New birds chose the same general area for display, but because the previous owners died, the precise location of the lek changed somewhat. There are a number of examples in the literature of slight, finely tuned shifts in the precise location of display areas that likely are the result of cover factors affecting the vulnerability of displaying males themselves, such as increased plant growth, floods, or disturbance by domestic animals (Horak 1974). However, if the powerful magnet of nearby females remains, the males will not be far away. An experiment with prairie chickens provides further evidence that females are the basic spacing unit in prairie grouse. In 1976, hand-reared prairie chickens of both sexes were held in pens while male and female, wild prairie chickens were released (Toepfer 1976, pers. comm.). Some males displayed immediately adjacent to the captive birds, and 15 others established a lek not more than 1.6 km away. We suggest that the males remained nearby because of the captive females, and not because of the captive males. Most of the wild hens dispersed long distances, only two of seven that were radio-marked remained nearby. The following year there were 16 displaying males, but in a second release of wild females, only three often that were radio-marked remained nearby. Many of the females traveled extensive distances, up to 56 km, before radio contact was lost. The experiment suggested that males would remain if there were some females present, and that adult females were not sufficiently attracted by the presence of a nearby, active lek for nesting. There are a number of examples of experiments in which biologists have attempted to relocate leks by placing stuffed or silhouetted male decoys, accompanied by male calls, at alternate locations (Eng et al. 1979, Tate et al. 1979, Gratson & Anderson unpubl. data). These efforts have largely failed because, we believe, males will not vacate proven display grounds near nesting females. Females, however, are more flexible, and their relatively large, preincubation range reflects the opportunity to explore newly created, nesting habitats.
486
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.4. Top: Abundance of male and female sharp-tailed grouse at leks and desertion of leks between years with cover constant in relation to abundance of concealing nest vegetation measured within a 4.8-km radius of each lek. Light cover (20%) indicates the mean height of grass is sufficient to conceal 20% of the body of a grouse. Bottom: Changes in the numbers of males and females at leks were correlated with changes in the abundance of surrounding nest cover between years. (Adapted from Brown 1966a, 1968a,b.)
SURVIVAL AND BREEDING STRATEGIES
487
If females shift or colonize new nesting areas, males will not be far behind in establishing a new lek nearby. This is what we see when we establish groups of dummy females 100-200 m away from established sharp-tailed grouse leks and watch some or all of the males shift and reestablish new territories at the dummy locations (Gratson unpubl. data). Males of the forest species also appear to select advertising sites at places where they will encounter females. Eighty-four percent of the ruffed grouse males at Cloquet, Minnesota, advertised on upland sites, where 96% of 113 females nested (Eng 1959, Kupa 1966, Gullion 1967). Farther south, at Cedar Creek, Minnesota, 84% of 31 males chose lowland sites at which to drum (Meslow 1966); here, in contrast to Cloquet, many females nested in the lowlands (Maxson 1978). Gullion and Marshall (1968) noted that some advertising sites are used continuously over several years, despite the fact that individual occupants are often killed by goshawks (cf. Boag 1976b). They termed such sites "traps." These traps have in common the characteristic that they are near the nesting areas of females (Fig. 14.5). Males are apparently prepared to frequent dangerous advertising sites if they can expect to encounter more females there than they might at "safer" locations. Blue grouse males also prefer some advertising sites to others. These were termed permanent and transient sites by Lewis (1979, 1982, Lewis & Zwickel 1981). Although on Vancouver Island, nesting females were evenly distributed in areas adjacent to both types of sites (half as many permanent sites and about half as many females nested nearby), more females were seen near the permanent advertising sites than near the transient ones, though there were fewer permanent sites (Table 14.1). Here, males apparently preferred the permanent sites, regardless of the distribution of nesting females. However, because the permanent sites were farther from sites of equivalent elevation, females may have been more readily attracted to males there than at the transient sites. Males also lived longer at the permanent than transient sites, and females could have benefited if these sites were safer. The higher frequency of females near the permanent sites most likely explains the males' persistent use of them. Earlier, (Chap. 2), Bergerud recounted the way in which blue grouse males, introduced to Moresby Island, British Columbia, dispersed before they established advertising locations (Fig. 2.4). This dispersal probably occurred before females nested. Many of the females that were introduced failed to nest in 1971 and 1972 and even attempted to leave the island, despite the fact that most of the males were already localized and advertising. The remaining females nested primarily in the center of the island, away from the most aggressive and conspicuous males, but in later years five of these aggressive males shifted their ranges inland, toward the nesting hens (Fig. 2.4). Males of all the forest grouse commonly display from different sites within their advertising ranges. Blue grouse males averaged seven sites at Comox Burn
488
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.5. Locations of advertising sites of ruffed grouse males at Cloquet, Minnesota. Many of the advertising sites classified as persistent and short survival (i.e., persistently used by short-lived males) were near habitats used by nesting hens and within 2 km of nesting goshawks. (Adapated from Eng & Gullion 1962, Kupa 1966.)
(McNicholl 1978) and five sites at Moresby Island (Bergerud & Hemus 1975). Gullion (1967) reported that 44% of 168 ruffed grouse males that he studied shifted drumming locations within their activity centers. Multiple display sites in activity centers have also been recorded in Alberta (Boag 1976a). Male blue and
SURVIVAL AND BREEDING STRATEGIES
489
Table 14.1. Survival and replacement of male blue grouse at permanent and transient advertising sites, elevation of sites, and abundance of females nearby (Lewis 1979, 1981, Lewis & Zwickel 1981) Permanent sites
Parameter Number of sites 1969-78 % of sites where new males replaced removed males Total male replacements13 Mean survival of males (years) Distance (m) to nearest area of equivalent elevation Nesting females nearby Total females seen near sites (average per site) a b c
Transient sites
41
82
82 (11 )a 20 4.3
36 (11) 4 2.6
143 39 444(10.8)
60 74 301 (3.7)c
Number of sites in parentheses. Up to 3 males were removed from some permanent sites. 301 females seen in years when sites occupied.
spruce grouse often advance and change advertising locations when taperecorded female calls are played back to them (MacDonald 1968, Bergerud & Hemus 1975). We suggest that most of the natural shifts between display sites, possibly even between logs that lay at different angles, are made in order to move nearer to females or to enhance signal transmission aimed in the direction of a female, not to repel other males (cf. Archibald 1974). Another tactic of some ruffed grouse males is that of displaying at the border between open and dense habitats (Eng 1959, Meslow 1966, Little 1978). Such males may encounter more females than those in dense canopies, because females often prefer open canopy cover with greater ground-nesting vegetation (Maxson 1978). Displaying males that advertise in such fringe areas often suffer a higher mortality rate than those in more dense cover (Little 1978, Gullion 1984b), and probably compromise safety for increased contact with females. The female's nesting-site decision is a critical fitness choice; the breeding site is of secondary impact. The female appears to be the basic spacing unit in grouse, and the male selects display sites to capitalize on her range selections.
14.2.6 Selection of safe advertising sites A secondary consideration in the choice of a display site by a male, after that of localizing near females, is that the site should allow amplification of his advertising signal but not expose him unduly to predation. The advertising locations of males are inevitably known to females and predators alike. The spruce grouse male appears secure during advertisement in many popula-
490
A. T. BERGERUD AND M. W. GRATSON
tions. The dense stands of conifers in which males commonly display (Stoneberg 1967, McLachlin 1970, Anderson 1973) should make navigation by large owls and goshawks difficult. Tree densities adjacent to advertising sites range from 3,800 to 12,400 stems/ha (Stoneberg 1967, Haas 1974, Boag et al. 1979, Hedberg 1980). In contrast, stem densities in the nesting ranges of Cooper's hawks (Accipiter cooperii) range from 276 to 1,728 trees/ha and in those of goshawks 273 to 750 trees/ha (Titus & Mosher 1981, Reynolds et al. 1982). The dense forests used by spruce grouse should hamper the grouse hunting success of accipiters, an explanation that is compatible with the mortality rate for males of less than 35% in several populations (Stoneberg 1967, Anderson, 1973, Boag et al. 1979). Spruce grouse males have perfected their advertising displays to minimize vulnerability and maximize signal amplification. The male commonly perches on a low, protective limb of a conifer, approximately 1-2 m from the ground, and drops to the forest floor, clapping his wings (Anderson 1973). He remains on the exposed, open forest floor only momentarily before flying back up to protective cover where he once again produces his wing-beat signal. The sound, upon landing on the ground, is broadcast beneath the attenuating conifer foliage, but he is exposed for only a short time (Fig. 14.6). Blue grouse males are also secure in some populations, when they, too, select thick conifer cover in which to display (Martinka 1972). Males in many populations, however, display in open conifer forests (Bendell & Elliott 1966, 1967). They frequently choose lone trees on hills and knobs where their signal is unobstructed. On rocky hilltops, where there is little understory vegetation, they can see and court females. The males on Moresby Island, British Columbia, displayed in tall conifers, 20 to 30 m high, and perched close to the base of the trunk and several meters below the exposed top of the tree. A male's position next to the trunk while vocalizing created the effect of ventriloquism, and we often nad to search for 15 minutes or more before spotting the bird. When raptors flew over, the bird became silent and crouched—the trunk and conifer branch provided effective concealment. These birds were safe but could also look for females. Ruffed grouse range from Georgia to Alaska, in habitats varying from oak (Quercus spp.) to birch (Betula spp.), aspen (Populus spp.), and fir (Abies spp.), which are coinhabited by different arrays of predators. The assessment of costs and benefits associated with drumming locations by a male in any region must consider both predation risk and the need to be conspicuous and to amplify his signal to attract females. Males suffer high mortality in northern Minnesota and Michigan, which are within the breeding range of goshawks, especially if their activity centers are near this raptor's nests (Eng & Gullion 1962, Fig. 14.5). Here, where goshawks are prevalent, males commonly display at less exposed sites, often at the base of hills (99 of 139 males) or in lowlands (18 of 19 males) (Berner & Gysel 1969).
SURVIVAL AND BREEDING STRATEGIES
491
Fig. 14.6. (A) Spruce grouse males advertise in coniferous forests by flying to the forest floor using wing-claps and flying back to a lower limb with another clap. They are vulnerable to predators only briefly when on the ground. Males can leave their advertising sites and meet females and remain relatively safe because of adequate cover. (B) Male blue grouse advertise at safe locations high in conifers, but near a protective trunk, which baffles the sound; by moving around the trunk they can direct the sound. Like spruce grouse, they can meet females away from their protective advertisement locations because of conifer cover. (Q Ruffed grouse males advertise on logs in a protective shield of shrub cover. Females must come closer to the male than in blue and spruce grouse because of the males' vulnerability in dedicuous forests. (D) Ruffed grouse living south of goshawks often display on hilltops with less protective cover than where northern grouse advertise. Theoretically, these southern grouse should be able to leave their advertisement sites and meet females, unlike northern birds.
492
A. T. BERGERUD AND M. W. GRATSON
Ruffed grouse that live south of the breeding range of goshawks may opt for more conspicuous sites. Grouse in Tennessee, Iowa, Indiana, and Kentucky drum on exposed hilltops (Hardy 1950, Muehrcke and Kirkpatrick 1970, Porath & Vohs 1972, Taylor 1976a). Another tactic in northern areas is to drum in dense forest patches of young pole-size trees (Kubisiak et al. 1980), where owls and goshawks can probably not hunt as effectively because of the high density of vegetation. In more open forests, shrubs are needed near the log to deflect predators but also allow good visibility to detect their approach (Dorney 1959, Palmer 1963, Gullion 1967, Boag 1976a). Boag and Sumanik (1969) removed shrub cover from advertisement stages and found reduced recruitment of newcomers to those sites and that resident birds switched to alternate logs in their activity centers. The log itself may serve as a deflector if the bird has sufficient warning; drumming stages are commonly about the same height above the ground as the height of the bird, and we have seen males hop off the log at our approach and remain hidden behind it. In addition to choosing a site with a deflecting screen through which there is some visibility, birds appear to select locations where it is difficult for a predator to approach undetected (Meslow 1966, Gullion 1967). The drumming stage on a log is usually away from the ends, or if the log has obstructing roots, the display stage may be near that end (Meslow 1966). Gullion (1967) reported that ruffed grouse choose sites where they can see at least 18m, and that males avoid drumming sites near pines and other low-visibility and ambush cover. Display sites also have an unobstructed avenue of escape (Edminster 1947). Departure most often follows a set route, suggesting familiarity with surrounding cover. Although few males die at advertising logs (e.g., three of 358, Gullion 1963; one of 37, Meslow 1966), the remains of many males have been found close by (Meslow 1966, Gullion 1967, Gullion & Marshall 1968, Rusch & Keith 1971). This suggests that predation risk is higher during the approach and departure from display sites than on the logs themselves. Because the drumming stage appears safer than the surrounding area, ruffed grouse males may often be more immobile than other grouse at the approach of females (Fig. 14.6) (Gladfelter & McBurney 1971). Blue grouse males, by contrast, move toward the female as soon as she has been detected, allowing her less choice in selection of both the male and the precise courtship area. The conspicuousness of the ruffed grouse is also affected by snow cover; a brown bird running or flying from his log against white snow should be vulnerable. The start of the ruffed grouse's drumming season correlates closely with the disappearance of snow. For example, at Cedar Creek, Minnesota, in 1970, snow disappeared rapidly on 7 and 8 April, and males started drumming on 9 April (Archibald 1976b). At Rochester, Alberta, the ground was 50% free of snow on 4 April 1966, 1 May 1967, and 23 March 1968; drumming was first heard 6 April 1966, 30 April
SURVIVAL AND BREEDING STRATEGIES
493
1967, and 24 March 1968 (Rusch & Keith 1971b). Petraborg et al. (1953) considered the disappearance of snow the main environmental stimulus for the start of spring drumming. In northern Minnesota, spruce grouse males began to display 1 May, 1970 and snow disappeared on 3 May; the next year the snow disappeared 21 April and males advertised on 22 April (Anderson 1973). Blue grouse show a gradual increase in spring advertising. We have heard blue grouse males hooting when the weather was warm in March, 6-8 weeks before actual breeding, and lek grouse begin to visit the lek in mid-March or earlier, a full month before breeding (Hartzler 1972, Sparling 1979, Svedarsky 1979, Emmons 1980). In contrast, the less secure ruffed grouse commonly begins displaying only 2 weeks before females are ready to breed (Aubin 1970, Archibald 1973). Males of the lek species select locations where their signal will be amplified and where they can observe the approach of predators. They appear secure on their open leks (Berger et al. 1963, Hamerstrom et al. 1965, Hartzler 1974, Sparling & Svedarsky 1978), although predators are doubtlessly aware of these locations. Horak (1974) observed an instance in which prairie chickens displayed in a sown grainfield, but when the grain reached a height of 38 cm in each of 2 years, the birds deserted the lek. Anderson (1969) moved trees 200 m closer to a prairie chicken lek, and the birds abandoned the ground. Such trees would block the view of the horizon, and provide cover for an ambushing raptor during the early-morning arrival of the birds. In another experiment, when fire reduced nearby cover, prairie chickens shifted their lek from 38-cm grass to the newly burned site, apparently because the new one provided better visibility (Anderson 1969). Occasionally, one can find the leks of both prairie chickens and sharp-tailed grouse in relatively tall grass, shrub, or even tree cover (Svedarsky 1979, Mossop pers. comm., Hamerstrom pers. comm., pers. obs.). Apparently sharp-tailed grouse will tolerate taller grass and shrub vegetation than prairie chickens, e.g., 0.5-0.75 m compared with 0.4 m (Ammann 1957, Sparling 1979), and more tree cover. Young males commonly display on the edges of leks. Although the frequency of ambush is apparently low at these display grounds, some males are killed (Hartzler 1972, Gratson unpubl. data). The yearlings of all the nonmonogamous species, both forest and steppe grouse, usually must settle for second-best sites in terms of security, if they wish to advertise. The decision by these yearlings to display or not must consider the risk of predation relative to the possibility of breeding females. Yearling sharp-tailed grouse and prairie chickens often breed. Ruffed grouse yearlings may advertise, but the risks are high (Gullion 1966, Gullion & Marshall 1968). For the forest grouse the risks are commonly higher than the expected reproductive benefits, and many yearlings forgo advertising (Herzog 1977a, Little 1978, Sopuck 1979, Gullion 1981).
494
A. T. BERGERUD AND M. W. GRATSON
14.3 Advertising tactics of yearling males Breeding opportunities for yearling males of the ptarmigan, which are generally monogamous, differ from those of the polygynous grouse. Yearling rock and willow ptarmigan males commonly obtain females by occupying nesting habitat needed by females and then advertising (Chap. 13). In two white-tailed ptarmigan populations, in which males predominated, many yearlings neither competed for space with adults nor advertised. On the other hand, in five willow or rock ptarmigan populations in which males also predominated, the yearlings did display (Jenkins et al. 1963, Watson 1965, Mercer 1967, Bergerud 1970a, Unander & Steen 1985). In these latter populations, females selected those males defending the largest territories, and territory size in two of these studies was positively correlated with male aggressiveness (Watson 1964, 1965, Watson & Miller 1971). Yearlings were sometimes as aggressive as adults and secured both territories and females (see also Pedersen 1984, Steen et al. 1985). Yearlings of the polygynous grouse show all levels of participation in advertising. Possibly 75% of the ruffed grouse yearlings in some northern populations advertise (Little 1978, Gullion 1981), whereas generally less than 30 % of the blue and spruce grouse yearlings, at 11 months of age, display (Herzog 1977a, Sopuck 1979). Yearlings of the steppe grouse illustrate yet another pattern. The majority of these apparently attend leks; changes in lek counts show a correlation with fall age ratios of the previous year (Chap. 15). However, these yearling males often visit a number of leks or advertise alone or in twos and threes at female locations (Brown 1966b, Emmons & Braun 1984, Gratson unpubl. data). The repertoire of yearling behavior can be explained as the various tactics yearling males have adopted to enhance their lifetime fitness, given that male-male competition results in adults generally occupying the most-advantageous advertising locations; researchers are only beginning to investigate and understand this variation. The most generally favored hypothesis regarding nonbreeding, yearling males is that of sexual selection, i.e., that females select the more vigorous or experienced adults rather than naive yearlings (Wittenberger 1978). An alternative to this idea proposes that yearling males do not breed because they are sexually immature, and that sexual development is delayed so that body size can increase. Males of large size presumably benefit because of reduced mortality. Adult males with improved longevity in an absence of yearling competition should be more successful in breeding, and lifetime fitness is presumably enhanced through delayed sexual maturity (bimaturism) (Wiley 1974). These competing hypotheses may be tested by removing the adult males from a population. The hypothesis of sexual selection is supported if the yearling males breed; the hypothesis of sexual bimaturism is supported if the yearling males do not breed and if the females remain unproductive and/or move elsewhere. Removal experiments have clearly shown that yearling males do take part in
SURVIVAL AND BREEDING STRATEGIES
495
breeding (Bendell 1972, Zwickel 1972, Keppie pers. comm.) when females no longer have the option of choosing adult males. Wittenberger (1978) also showed that the proportion of males in grouse populations decreased as sexual dimorphism increased, whereas the bimaturism hypothesis predicts the reverse—i.e., that the proportion of males in the population will increase, due to increased longevity, when there is a greater disparity between the sizes of males and females. Most devastating to the bimaturism hypothesis has been the recent rinding that large, adult sage grouse males have a shorter life expectancy than either the immature males or females, which are smaller (Braun 1979). The sage grouse is the best example of deferred maturity coupled with dimorphism in North American grouse and the best evidence against the bimaturism hypothesis. The theory of sexual selection, on the other hand, predicts that large body size, although a positive influence on intraspecific competition and mate choice, may have the disadvantage of increased conspicuousness and vulnerability to predation (Darwin 1871), consistent with the sage grouse data. Yearling males are physiologically capable of breeding; even the slowly maturing, sage grouse males produce viable sperm their first spring (Eng 1963). Athough a yearling male does not have a free choice, he has the option of weighing the increased predation risk always associated with advertising with his chances of attracting females that probably prefer adults, for a variety of reasons. Inexperienced yearlings lack both knowledge and access to the principally adultheld sites, which are relatively secure and well-frequented by females. One tactic may be to refrain from advertising and to remain inconspicuous in the first breeding year, while becoming more familiar with the distribution of females and the security of the various sites. It is not an ideal-free tactic; if a yearling could be assured of breeding, he should change tactics and advertise. This is precisely what happens when adult males are experimentally removed, and yearlings have access to safe sites and females. However, yearling males of polygynous species, because of their smaller size and inexperience generally, cannot compete with adults; to make the best of a bad situation, they adopt a conditional strategy (Dawkins 1980) that generally yields reduced results (e.g., Rubenstein 1980, Howard 1984, review Dominey 1984).
14.3.1 Early prospecting by juvenile males Natural selection should favor early sexual maturity in males if this enhances their reproductive fitness. Young males in some populations begin advertising in their first September and October. Ruffed grouse are probably the most precocious, some drum at 4 months of age (Eng 1959, Gullion 1967). Juvenile spruce grouse in Michigan have been reported prospecting for advertising sites (Robinson 1980), and young sharp-tailed grouse have been observed displaying in their first fall (Hamerstrom & Hamerstrom 1951, Kermott 1982, Chap. 5). On Brunette Island, Mercer and Bergerud counted a total of 75 territorial males in three springs
496
A. T. BERGERUD AND M. W. GRATSON
and 115 males in October, 6 months later; at least 40 were juveniles (Bergerud & Mercer 1966, Mercer 1967). In Scotland, juvenile red grouse males are very active in contesting advertising territories in their first fall (Jenkins et al. 1963, Lance 1978a). Watson (1965) noted that rock ptarmigan also displayed during their first fall in Scotland. In most instances noted above, with the exception of red grouse, the percentages of juveniles that displayed in fall were small. These early birds may be early-hatched young, slightly ahead of other members of their cohort in physiological development (Gullion 1967, Moyles & Boag 1981). Blue and sage grouse juveniles do not display in their first year. These two grouse species are the largest and the most sexually size-dimorphic in North America. The first-year growth rate in both species is such that yearling males in the fall are considerably smaller than adult males (Patterson 1952, Redfield 1978), and even at 12 months, sperm production by a yearling is less than that of adults (Eng 1963, Hannon et al. 1979). Generally, not as many juveniles in more northern populations display in the fall as in more southern populations of the same species. Juvenile red grouse advertise in the fall in Scotland, as do willow ptarmigan in Maritime Newfoundland, but ptarmigan in the Arctic and northern Norway do not. Juvenile prairie chickens visit leks and display in Missouri and Oklahoma (Schwartz 1945, Copelin 1963), but fewer do so or do so less frequently in Wisconsin (Hamerstrom 1939). Ellison (1967) noted only one juvenile spruce grouse male displaying in a sample of 101 sites in Alaska. All the ruffed grouse males in Kentucky occupied logs in the fall, but in northern Minnesota, 46% of 140 waited until spring to occupy and display in sites unoccupied the previous fall (Hardy 1950, Little 1978). A juvenile bird apparently must reach a certain point in sexual development in order to advertise. In northern habitats birds hatch slightly later and the departures for winter ranges begin earlier in autumn. The north-south continuum in the proportion of juveniles that display appears to reflect the reduced development and times between birth and the onset of snow cover in northern areas; the increased conspicuousness that results from snow apparently precludes the option for fall display from being profitable.
14.3.2 Advertising in spring The percentage of yearling males that advertise at 10 months of age in the tundra in spring varies from nearly 100% in willow and rock ptarmigan populations, to possibly only a few percent in some white-tailed ptarmigan populations. Choate (1960, 1963a) located only one of 15 tagged, yearling white-tailed ptarmigan displaying. In the same population one 2-year-old male waited until his third year to advertise. The so-called silent males among the forest grouse are primarily yearlings. The percentage of silent males varies among populations and from one year to
SURVIVAL AND BREEDING STRATEGIES
497
the next (Gullion 1981). Over a ten-year period the percentage of nonadvertising ruffed grouse varied from 0% to 38% in Alberta (Rusch et al. 1984, pers. comm.). Silent males varied from 3% to 37% of total males for ruffed grouse over a 7-year period in Minnesota (Gullion 1966b, 1981). On Moresby Island, British Columbia, there were four advertising and approximately eight silent yearling blue grouse in 1971. Near Campbell River, Vancouver Island, Bendell and Elliott (1967) reported that 26% of the 57 yearling blue grouse advertised, whereas at Comox Burn practically all yearlings and many 2-year-old males failed to advertise (Lewis & Zwickel 1981). Herzog (1977a) radio-tracked spruce grouse yearlings in Alberta and found that two of six localized and advertised, two localized but they did not display, and two showed large ranges and did not advertise. Males of the forest grouse defend advertising sites and not space. There are no clearly defined boundaries (Aubin 1970, Anderson 1973, Archibald 1976a), and adjacent advertising ranges seldom have a common boundary (Fig. 14.7; cf. McNicholl 1978, Bendell & Zwickel 1978, Herzog & Boag (1978). Yearling males do have the option to display and may squeeze in between other males. On Moresby Island, two males repeatedly and simultaneously advertised from the same tree—perhaps an extreme, but clearly a useful example. We should question the idea that yearling males may be social outcasts and are prevented from breeding by other males. All evidence suggests that an equally likely explanation is that the yearlings themselves decide whether or not to display, and that this "decision" is made on the basis of their expectation of success and the risks incurred. The percentage of male steppe grouse that do not advertise has not been thoroughly documented. Most radio-tracked males in many studies have been captured on leks and were already displaying. A sample of 1,039 banded, male prairie chickens trapped in the winter showed a juvenile:adult ratio of 1.3:1, whereas the ratio of yearlings to adults on leks was 1.7:1 (n - 626) (Hamerstrom & Hamerstrom 1973), suggesting most yearlings showed up on leks. Hamerstrom reported (1981) that of 80 immature males banded in winter 1955-56, 23 were never seen again, and 13 were seen but not on a lek. Brown found fewer yearlings on leks (1.7 yearling/adult, n — 586) than on winter ranges and fewer than in samples provided by hunters in the fall (Brown 1966b, 1967, 1968b). Removal experiments with sharp-tailed grouse also suggest that some yearlings had not localized until they were 18 months old (Moyles & Boag 1981). When adult prairie grouse were removed from leks they were replaced by yearlings (Rebel 1970, Rippin & Boag 1974a). These may have been nondisplaying birds or birds that shifted from other arenas or that displayed alone. Yearling sage grouse males that were radio-tracked visited an average of 3.9 leks before localizing at one display ground (Emmons 1980). Thus, yearlings are mobile and often move between leks as they prospect for potential advertising sites.
498
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.7. Males of the steppe species advertise and defend small territories with precise boundaries on leks. Physical fighting is common. Tundra grouse defend much larger territories, within which females nest. Males commonly patrol boundaries, "walk the line," and engage in aerial chases and fighting. Forest grouse advertise at one to several locations within a home range, and the intensity of the male's aggressiveness to other males decreases with increasing distance from the display site. Trespassing by yearlings is common, and fighting is infrequent.
The steppe grouse suffer little predation when advertising; display grounds are safe because of good visibility and experienced adults are present. The cost of advertising should thus be low, but the immediate, expected breeding benefits are also low for yearlings because of female preferences for more vigorous and experienced males. However, a yearling's lifetime fitness may depend on his seniority at a lek (Moyles & Boag 1981, Kermott 1982). These yearlings probably spend some time deciding where to display, but theoretically a higher percentage of them should make an early decision and advertise than should the forest grouse, in which sexual selection is less intense and probably more of the total population of adult males have an opportunity to breed females.
SURVIVAL AND BREEDING STRATEGIES
499
14.3.3 Tactics for securing an advertising site Yearling males, unlike most adults that have established sites, are not committed to specific advertising spots. Yearlings can be expected to spend considerable time prospecting. They can be opportunistic and move to newly created habitats, such as logged areas, as do blue grouse (Redfield 1972), or regenerated grasslands, as do prairie grouse (Westemeier 1973). Although opportunities for sites exist, there is uncertainty of predation risks and attractiveness to females of such areas. It appears that, except for red grouse in Scotland (Lance 1978a), yearling males only rarely attempt the tactic of vigorously contesting advertising sites with established adults. Occasionally, however, exceptional yearlings do oust older birds. Bergerud observed a yearling, male blue grouse approach a dummy female, dominate the 5-year-old resident, and mount the stuffed hen. Also, Kermott (1982) observed a banded, yearling sharp-tailed grouse secure a central location on a lek. If females prefer adults, one tactic for yearling males is to acquire knowledge of female locations and potential advertising sites. While prospecting he should remain inconspicuous and thus relatively safe from predation. Yearling blue grouse have elaborate tactics for securing advertising sites. This has recently been documented by radio-tracking young males at Comox Burn, on Vancouver Island, and on Hardwicke Island, just off Vancouver Island (Sopuck 1979, Jamieson 1982). Yearling males return from the winter range at the same time as females. Adult males precede both yearlings and females (Bendell & Elliott 1967, Bendell et al. 1972). There should be little advantage for yearlings to return before female occupancy, because they could not assess the distribution of females. Early in April when yearling females search large ranges for nest sites, yearling males are also wide-ranging (Hannon 1978, Sopuck 1979, Hannon et al. 1982). Ranges of these males have averaged 28.5 ± 5.2 ha (n = 30) (Sopuck 1979). After females localized near their future nest sites (approximately 6 May), yearling males also reduced their ranges (mean = 11.9 ± 6.9 ha, Sopuck 1979). These yearlings traveled through an average of 4.4 ranges of advertising adult males (Sopuck 1979). Sopuck recognized a wide-ranging, yearling phenotype (n = 13) and a more localized, yearling phenotype (n = 17). Jamieson (1983) and Jamieson & Zwickel (1983) delved further into these patterns and found that a wide-ranging phenotype in their study area visited the vicinity of one to four advertising males (Fig. 14.8). These males visited areas an average of 57 + 1.9 m from the nearest adult males. Small-ranging males showed restricted ranges and localized at vacant, but formerly occupied, advertising sites, but did not advertise. These vacant sites averaged 75 + 2.3 m (n = 7) from the nearest male. Both Sopuck (1979) and Jamieson (1982) also identified a third yearling phenotype: birds of this kind ranged very widely and appeared attracted neither to a vacant site nor to the sites of currently advertising males. Movement patterns of yearling males in these two studies suggest a basic
500
A. T. BERGERUD AND M. W. GRATSON
strategy of assessing where females frequent, and then locating potential advertising sites in the vicinity. Wide-ranging males should be near the advertising sites of two to three adult males and a similar number of females (cf. Hannon 1978). Between breeding seasons, one of the adult males and one of the females should die (i.e., 33%) (Zwickel et al. 1983). The wide-ranging yearling, now an adult, can shift his activity center and advertise within the range that he learned as a yearling, in relation to the new realignment of surviving males and nesting females (Fig. 14.8). In contrast, the small-ranging yearling has his spot for next year assured, but he is committed and cannot make major adjustments later to maximize the new arrangements of males and females. The male that ranges widely could be characterized as remaining inconspicuous and tolerant to densities of surrounding males. The small-ranging male is not as secretive and could be termed density-intolerant—he preferred a vacant site, and one that was farther from the nearest, other displaying males. The bird that ranged very widely (the third phenotype) had seemingly no site attachment and followed a pattern that would allow him to colonize newly created habitats in his second year. The three phenotypes recognized in blue grouse may also occur in spruce grouse. Again, radio-telemetry studies suggest three movement options. Herzog (1977a) found two yearlings that selected unoccupied sites and advertised; their ranges were small and similar to those of adults—approximately 1 ha. Two other yearlings ranged over areas of 6.6 ha and 2.3 ha, spent more time nearer adults, and did not advertise. Herzog's evidence of a very wide-ranging type consisted of two yearlings that moved over 10.5 and 14.8 ha and showed little site attachment. The three phenotypes described in each of the three studies (Herzog 1977a, Sopuck 1979, Jamieson 1982) appear to encompass three diiferent tactics for the dispersal of males to advertise for females. Ruffed grouse males show two of the site-prospecting tactics of blue and spruce grouse. Wide-ranging yearlings that prospect near many adult males (Bump et al. 1947, Marshall 1965, Godfrey & Marshall 1969), and yearlings that localized at unoccupied sites (Little 1978), have been documented. Unlike blue grouse, however, ruffed grouse may begin their activities as juvenile birds in September and October (Eng 1959). The proportion of males that opt to search widely near other males and the proportion that localize at unoccupied sites can be estimated for one population in central Minnesota (Little 1978). Little found that during 4 years 55 yearlings took advertising sites in the spring after the sites became vacant following the disappearance of resident adults over the winter. Conversely, 28 yearlings drummed in the spring at activity centers not occupied the previous fall when these males had probably visited the areas. In addition to these yearlings were birds that did not advertise (silent males), even though the population was declining and sites formerly occupied were continuously becoming vacant. When these birds became adults 21 took over the locations of drummers that died and 26 established
SURVIVAL AND BREEDING STRATEGIES
501
Fig. 14.8. Top: Yearling blue grouse males commonly prospect for future advertising sites in the vicinity of one to four advertising adults. One of six yearlings established a display site and advertised. Bottom: One yearling that prospected near the sites of three adults occupied one of the sites the following year when the adult was dead (bird /). Another yearling advertised the following year after both males that he had visited the year before had died (bird 2). A third yearling remained silent as an adult when both the males he prospected near as a yearling reoccupied their sites the following year (bird3). (Adapted from Sopuck 1979, Jamieson 1982, Jamieson & Zwickel 1983.)
502
A. T. BERGERUD AND M. W. GRATSON
new advertising centers (Little 1978). If these additional advertising adults represent new drummers, and not birds switching locations, the ratio of males using the tactic of ranging widely near adults (density-tolerant) to those ranging over smaller areas at unoccupied sites (density-intolerant) was 1:0.71 (76:54). Steppe grouse yearlings can establish and advertise at new leks, or elect to display at established leks with adults. Some yearling, sharp-tailed grouse males visit the periphery of display grounds in September; at first they mostly watch and crouch (Moyles & Boag 1981). These younger birds often arrive later and leave earlier in the morning than adults (Kermott 1982). Slightly later in the season some yearlings display, but back down and walk away when adults approach. Finally, they are no longer challenged and display in peripheral positions, and in time they in turn chase newer birds that arrive. Birds that succeed are those that are both faithful to the lek and persistent (Moyles & Boag 1981, Kermott 1982). The birds that Moyles and Boag observed in Alberta continued to visit display grounds on clear and warm days throughout the winter. Most young joined the establishment between September and December, but some did not take up positions until April. Both studies showed that birds generally worked their way toward the center position of the lek as older, interior birds disappeared (see also Rippin & Boag 1974b). On average, the sooner a yearling begins to display, the earlier he should become an inside bird; generally the more central males on a lek do the majority of breeding (Robel 1966, Hartzler 1972, Wiley 1973b, Sparling 1979, Kermott 1982, Chap. 7). Also, joining early would allow yearlings the opportunity to learn female preferences. Early persistence at established leks could advance seniority and improve fitness. This tactic is analogous to the approach used by some forest grouse yearlings of visiting advertising adults and waiting for vacancies. Other yearlings in steppe species follow the option of forming original leks. New leks commonly appear when populations are expanding (Brown 1967, Kohn 1976, Svedarsky 1979, Cannon & Knopf 1981). Because adults seldom change locations (Hartzler 1972, Hamerstrom & Hamerstrom 1973, Hamerstrom 1981), unless there are extreme shifts in the distribution of females, a new lek should usually be formed by yearlings. The establishment of new prairie chicken and sharp-tailed grouse leks by yearlings has been observed (Robel 1964, Rippin & Boag 1974a). Robel (1964) observed seven yearling males arrive at a lek and leave together. The only banded birds seen at three new, sharp-tailed grouse leks in Montana were yearlings, and on another lek that increased from four to 16 males, ten males were yearlings (Brown 1967). Some daily counts at leks reveal more males early in the season, before most females arrive (Artmann 1970, Bowen 1971). In other populations the peak male count generally coincides with female attendance (Hamerstrom & Hamerstrom 1973, Svedarsky 1979). In the former populations, numbers were increasing— the early, high counts could reflect yearling males prospecting at display sites
SURVIVAL AND BREEDING STRATEGIES
503
early in the season, but later leaving to settle on other leks (Brown 1966b), or establish new ones (Gratson unpubl. data). The emphasis in the literature on yearlings initially displaying at established leks is that of a new bird trying to compete with his seniors and be "admitted to the club" (Moyles & Boag 1981, Kermott 1982). However, how can an adult prevent a yearling on the edge of a lek from becoming established? The yearling need only move farther away if challenged. Another interpretation is that the yearling is evaluating his options, and based upon the likelihood of breeding females at that lek, he will make a decision whether to display there or elsewhere. The intensity of intrasexual competition would be only one aspect of this decision. Possibly some males display alone in their first year, and if they are unsuccessful because females ignore them, in their second year they establish themselves at a lek. If they follow this scenario, however, they will have lost the advantages that might accrue to yearlings that establish themselves at a lek by their first spring. Yearlings that establish new leks would probably be favored if a population was expanding and females were moving to new ranges. Yearlings that opt to go to established leks should have an advantage if a population is declining and only the established leks are visited by females. Yearling steppe grouse males should assess the distribution of females as part of a decision to select an advertising site. This hypothesis is consistent with the observation that there were fewer males at sharp-tailed grouse leks in Montana where the surrounding, residual nest cover declined between years (Brown 1967, Fig. 14.4), and leks had more males when there was tall nesting cover in surrounding areas. A change in the number of cocks on leks between years should result primarily from variable recruitment of yearlings. In Montana, there were larger percentages of yearlings at leks near dense, tall nesting cover than at leks with low- or medium-quality nest cover nearby (Fig. 14.9). Also, a greater number of females visited leks that were near dense nesting cover. The ratio of adult males to females was 1:2, compared with a ratio of approximately 1:1 at leks near low- and medium-quality nesting cover (Fig. 14.9). Thus, leks chosen by yearlings offered improved opportunities for yearlings to breed. Brown (1966b) indicated that 22% of 106 banded yearlings had central positions on leks (cf. Hjorth 1970). Also, at two leks there was high mortality of males, yet they continued to attract yearlings (Brown 1966b). A higher proportion of females (35%) visited these leks than others (Brown 1967), similar to the large number of visits of blue grouse females to persistent advertising locations (Table 14.1). All the evidence presented above is consistent with the idea that yearlings assess the distribution of females and that this assessment is part of their site-selection process.
14.4 Female nesting strategies The most important reproductive decision a female must make is where to locate her nest. Grouse commonly lose 50% of their clutches (Hickey 1955, Table
504
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.9. Top: Yearling sharp-tailed grouse males were more common on leks that had heavy grass cover within a radius of 4.8 km. On such leks a higher proportion of females was observed than on leks surrounded by less grass cover. Yearling males, by selecting leks with more visiting females, should have a better opportunity to breed in the same or future years. Bottom: The ratio of females to adult males was 2.2 females/adult male compared with 0.7 female/yearling male on leks when the surrounding grass cover was at least 60%. (Adapted from Brown 1966a, 1968a,b.)
SURVIVAL AND BREEDING STRATEGIES
505
15.2). A female's potential or realized loss in fitness is greater from this single factor than from any other. A short life expectancy allows but a few opportunities to perfect nesting options. When her clutch hatches, the female may lose chicks because of inherent weakness, inclement weather, food shortages, and predation. These losses are less predictable and more beyond her control than nest losses. Selection should result in major behavioral tactics developing in populations and species to positively affect the nesting outcome. The female's decision should relate primarily to where the nest should be located to avoid detection by predators. It may be difficult to select the nest site to maximize future chick survival by nesting near favorable brood-rearing habitat. Nests hatch 35-45 days after the female's commitment to a site by depositing eggs. During this interval there are rapid increases in plant growth and changes in abundance of invertebrates; what may have been assessed as optimal habitat for chicks in May might no longer be the best place in June and July. That broods commonly make long movements after hatch (see section 14.6.1) suggests that the female's first concern is to have a successful nest, using the available options in April and May. The importance of the nest location decision argues that a female should begin to search for a site as soon as such action would be rewarding (Brewer & Harrison 1975). She could prospect broadly at first and fine-tune her decision later when she has a maximum choice of locations following snowmelt in spring, and when the most current information on the concealment value of cover becomes available. Some juvenile males seek advertising sites in the fall (section 14.3.1), and it is possible that females might also start their assessment of potential nesting sites then if there is adequate time before winter weather persists (cf. Weise & Meyer 1979). Herzog (1977a) reported that female spruce grouse were beginning to space themselves in November. Ruffed grouse females have a tendency to nest where they were found the previous fall (Eng 1959). In the most southern grouse, Attwater's prairie chicken, females have the largest range in November and December (Silvy & Lutz 1978). Prospecting may already have begun. As a generalization, yearling females are less successful nesting than adults in all the grouse except ptarmigan (Table 15.4). The improved nesting success of the six other grouse species between the first and later years supports the view that experience provides knowledge on how better to proceed. The nesting strategy in grouse may include the following tactics: (1) Females should research areas to ascertain the abundance of predators and the general availability of suitable nest sites. (2) Females should attempt to nest at the time and place of maximum cover, yet still allow for the option of renesting if necessary. (3) Hens should remain inconspicuous and avoid other females near nest sites. (4) If a female was successful last season, she should show philopatry; if she was unsuccessful and the responsible factor is likely to recur, she should shift nesting areas. (5) Females should abandon current clutches and try again else-
506
A. T. BERGERUD AND M. W. GRATSON
where if the certainty of a current loss outweighs the unpredictability of the loss of a future effort.
14.4.1 Assessing predator abundance and nesting sites The prelaying ranges of females—the area traveled between the time they leave winter ranges and flocks and the time of egg-laying — are generally largest for grouse in steppe environments and smallest for the tundra species (Table 14.2). The size of these ranges is inversely related to nesting success and is larger where there are many predators (Fig. 16.2). The steppe grouse must contend with about 13 groups of predators. Some of these can occur at high densities, such as ground squirrels (Citellus spp.) and skunks (Mephitis spp.). Other predators are extremely mobile, searching large areas, such as magpies (Pica pica), corvids (Corvus spp.), and coyotes (Canis latrans). Forest grouse have approximately nine groups of predators in the conifer and deciduous forests of North America, whereas ptarmigan have the fewest kinds and numbers of nest predators. We propose the hypothesis that the prelaying range of the female reflects her movements while prospecting for nest sites; she may assess both the number and distribution of predators, the availability of nesting cover, and the spacing of other nesting females that may add to her conspicuousness. If she frequently encounters a predator, she should range farther and in other directions. Few data are available to directly evaluate the prediction that females are able to evaluate the distribution of predators. Seldom have researchers radio-tracked predator and prey simultaneously. We do know that females may move considerable distances after losing nests (Christenson 1970, Robel et al. 1970a, Schiller 1973, Svedarsky 1979, Chap. 6). Svedarsky found that prairie chickens preferred seeded habitats of redtop (Agrostis spp.) and brome grass (Bromus spp.) and that nesting success was higher in planted habitats than in more diversified, native vegetation. These planted grasslands had less litter than native communities and possibly attracted fewer invertebrates and voles (Microtus spp.), upon which skunks and foxes (Vulpes vulpes) prey. Sharp-tailed grouse also had reduced nesting success when there was more litter (Schiller 1973). Prairie chickens in Illinois selected grasslands that were burned 2-4 years previously over unburned sites; densities were 0.41 nests/ha compared with 0.27 nests/ha in unburned sods (Westemeier 1972). Nest success of prairie chickens is commonly higher in burned than in unburned habitats (Lehmann 1941, Kirsch & Kruse 1973), despite the often reduced density of the cover. Fewer predators may search in burned habitats with reduced faunal diversity. Maxson (1978) found that six of seven ruffed grouse nests located near water were successful, whereas overall nesting success was only 59%. Attwater's prairie chickens commonly lose nests from flooding because of the selection of lowland sites, but if such wet areas hinder some predators, the expected success could be greater than at alternative sites.
SURVIVAL AND BREEDING STRATEGIES
507
Many predators hunt in predictable patterns (Curio 1976). Also, areas near dens may be searched almost daily (cf. Sargeant 1972). Grouse may be able to respond to these patterns and correlated patterns of particular habitats. The hypothesis that the size of the prelaying range relates to a nest-site decision is consistent with the finding that yearlings generally travel over a greater range than adults and narrow their search pattern as the time of egg-laying approaches (Herzog 1977a, Hannon 1978, Sopuck 1979). Adults, with more experience, can quickly avoid large patches of unsuitable habitat. Yearlings are less familiar with habitats and, at least for blue and spruce grouse, avoid areas occupied by adults. The sizes of prelaying ranges vary among species within steppe and tundra grouse (Table 14.2), and potentially should vary within the forest group. Sizes of prelaying ranges of all nine species are positively correlated with the loss of nests to predators (Fig. 16.2). The different abundances and kinds of predators in the habitats of grouse species within groups may be a factor contributing to this, but the distribution and morphology of specific nesting cover should also affect nesting success and sizes of prelaying ranges. Generally, heterogeneity of vegetative cover increases in time and space in climatic regions that receive reduced rainfall (Wiens 1974). Plant cover in such arid regions is more unpredictable (Wiens 1976, Knapp 1984). Grouse in drier environments may thus need to search larger areas than those in areas with more precipitation. Of the species in steppe environments, the sage grouse is the largest and most conspicuous of the three lek species. Females search for nest sites in dry habitats under sagebrush canopies (Patterson 1952, Klebenow 1969, Wallestad & Pyrah 1974, Peter sen 1980). These steppe habitats, unlike the grasslands, are largely devoid of dense, herbaceous growth to hide nests from a horizontal view. It is relatively easy for field workers, and presumably predators, to find nests, even though large areas must be searched. The sage grouse should take the longest time and cover the largest area in the site-evaluation process. We can also expect prairie chicken and sharp-tailed grouse females to search larger areas for nest sites in places like Kansas and Montana, where the climate is generally dry and unpredictable (Knapp 1984), than in Minnesota and Wisconsin, where nest cover should be more homogenous in distribution (Table 14.2). Of the forest grouse, the blue grouse is the largest, and usually nests in more open habitat than spruce or ruffed grouse (Bump et al. 1947, Kupa 1966, Zwickel 1975, Sopuck 1979, Redmond et al. 1982). Blue grouse generally should range over areas larger than spruce or ruffed grouse. Blue grouse on Vancouver Island had ranges similar in size to those reported for ruffed and spruce grouse in mainland, interior populations (Table 14.2). However, these estimates may not be representative of other blue grouse populations, because the data from Vancouver Island are from the dense populations at Comox Burn, where there are generally few nest predators. The ruffed grouse is smaller than the blue grouse and frequents deciduous woodlands with more herbaceous ground cover. This cover ad-
Table 14.2. Average ranges (ha + SE) of female grouse during Prelaying
Grouse groups and species Forest Spruce grouse
Yearling
Adult
14.4 ± 2.9 (10) 18.5 ± 4.6 (7) 13.5 ± 4.2 (3)
4.1 ± 0.6 (10)
6.4 (3)f 12.1 (9)
Ruffed grouse
Blue grouse
12.0 ± 8.8 (27) 16.6 ± 1.2 (4) 16.1 ± 2.5 (18)
Steppe Sharp-tailed grouse
Prairie chicken
34
± 15
Tundra White tailed ptarmigan
2.9 ± 0.4 (33) 2.7 + 0.1 (30) 5.6 (27)
c d e f
(3)
82 85 194 182 882
(5) (7) (2) (2)
14 (1) ± ± ± ± (7)
14.4 5.2 5.4 20.9 2.3 5.1
Willow ptarmigan
b
464 ± 145
(2)
Sage grouse
a
4.6 ± 2.7 (10) 6.3 ± 3.2 (5) 6.6 ± 1.3 (7)
16 37 97 17
± 2.9 (7) ± 0.3 (92) ± 0.4 (33) (90) (86) ± 0.3 (11) 3.2 ± 0.3 (43) 2.9 ± 0.2 (28) 7.2 (47)
Estimates as reported or as calculated from sources listed below; in some instances SE could not be computed. For ptarmigan this estimate is the size of male territories. Chicks less than 2 weeks old. Chicks at least 2 weeks old. Early and late combined. Yearlings and adults combined. Sources: Spruce grouse—Alberta, Herzog 1977a; Minnesota, Haas 1974; Alaska, Ellison 1973. Ruffed grouse- Minnesota, Archibald 1975; Minnesota, Maxson 1974; Minnesota, Godfrey 1975. Blue grouse-British Columbia, Sopuck 1979; British Columbia, Hannon 1978; British Columbia,
different phases of the breeding cycle. 3 Number of females in parentheses. Brood-rearing
Laying and incubation
Early c
Late"
29.0 ± 4.8 (6)e
4.2 ± 1.0(10)
8.0 +
1.5 (3)
19.8 ± 4.2 (11) 42.1 ± 17.0 (8)
1.9 ± 0.5 (3) 8.4 ± 0.6 (8)
4.9 ±
15.9 ± 2.7 (4)
0.6 (8)
12.8 ± 2.7 (4) 14.0 (8) 4.1 (9)
45.3 ± 26.6 (7)
14
± 5 (6)
48 ± 36 (4) 1 0 + 4 (6) 11 (9) 54 19
31 (9) 31 ± 6 (11) 42 ± 57 (13)
±
3
± 12
55 ± 20 4 4 + 7 94(5)
(3) (5)
108
± 35
(7)
11 59
± 2 ± 46
(4) (7)
(9)
(8) 67 (10)
1 8 + 4 (9) 115 + 106 (8) 134
± 72
(10)
127 56
± 30 ± 15
(3) (3)
86(13)
27
7.3 ±
± 5.4 (18) 14.3 ± 7.9 (15)
3.4 (15)
65.8 (12)
Armleder 1980; British Columbia, Ash 1979; British Columbia, Hines 1986a; Montana, Mussehl & Schladweiler 1969. Sharp-tailed grouse—Wisconsin, Gratson 1983, Chap. 5; Wisconsin, Ramharter 1976; Minnesota, Schiller 1973; Minnesota, Artmann 1970; North Dakota, Christenson 1970. Prairie Chicken-Texas, Sell 1979; Minnesota, Svedarsky 1979; Texas, Silvy & Lutz 1978; Missouri, Arthaud 1968; Kansas, Robel et al. 1970a; Texas, Taylor & Guthery 1980. Sage grouse—Idaho, Connelly 1982; Idaho, Autenrieth 1981; Wyoming, Berry & Eng 1984; Montana, Wallestad 1971. White-tailed ptarmigan-Colorado, Schmidt 1969. Rock ptarmigan-Iceland, Gardarsson 1971; Scotland, Watson 1965; Svalbard, Unander & Steen 1985. Willow ptarmigan — Scotland, Miller & Watson 1978; British Columbia, Mossop pers. comm; British Columbia, Hannon & Smith 1984; British Columbia, Weeden 1959b; Norway, Erikstad 1985a; Norway, Pedersen 1984.
510
A. T. BERGERUD AND M. W. GRATSON
vantage for nesting suggests to us that ruffed grouse females should range over smaller areas than blue grouse. An alternative hypothesis for the differences in the size of prelaying ranges of grouse is that birds with larger body size will have a wider variety of foods but will take only the highest quality items; this selectivity will lead to larger home ranges. Further, a shift to a monoculture diet such as sagebrush leads to greater selectivity for plant parts and phenological stages, and thus to wide-ranging foraging patterns (Bradbury 1981). However, all the grouse are largely monophagous regardless of the size of their ranges or their body sizes (Robinson 1969, Gardarsson & Moss 1970, Pendergast & Boag 1970, Wallestad et al. 1975, Ellison 1976, Herzog 1978, Williams et al. 1980, Chaps. 4, 9, 10). No one has shown differences in quality of food in relation to range sizes. The red grouse, with the smallest prelaying range, has one of the most monophagous and poorest diets (Moss & Hanssen 1980). Birds occasionally die from starvation (Jenkins et al. 1963). Blue and spruce grouse eat needles of generally lower nutritional quality than do ruffed grouse that utilize high-quality buds (Huff 1970), but the ranges of all three species are roughly similar in size (Table 14.2). Blue grouse are twice the size of spruce grouse yet they eat similar foods and their ranges are of comparable size. The suggested correlations among body size, food quality, and home range size are not convincing. Another alternative to the hypothesis that the first priority of females is a safe nesting site is that in the spring females move from winter habitats to male advertising locations for breeding, and then they search for their nest site. This model had the important implication that the male location may be the focus of the female's search for a nest location. But this sequence does not provide enough time for a yearling female to adequately assess potential nest sites, because females generally deposit their first egg within a few days of breeding with males (Brander 1967, Christenson 1970, Svedarsky 1979, Chap. 6). The problem is not so acute for returning adult females, especially if they were successful the previous year; they may "know" where they are going. However, most returning adults and generally all yearling females should prospect for nest sites before selecting males. Radio-telemetry studies of forest grouse have verified that females do select nesting sites before breeding. Four yearling, blue grouse females localized on their nest ranges 13 + 0.8 days before the egg-laying period, and five adults localized 12 + 3.8 days before laying eggs (Hannon 1978). Female spruce grouse visited males just before laying eggs but had established their nesting areas in mutually exclusive ranges several weeks before they began laying eggs (Herzog 1977a). Particularly good data are available for ruffed grouse at Cedar Creek, Minnesota. One female visited her future nest site 15, 19, and 23 April and laid her first egg on the 24th; a second female visited her site 13, 16, 17, 24, 26, and
SURVIVAL AND BREEDING STRATEGIES
511
27 April and laid her first egg on the 28th; and the last female visited her site on 25, 26, 27 and 29 April, and on 1, 2, 4, 5, 7, and 8 May, and laid her first egg on 9 May (Maxson 1974, 1977). Edminster (1954) had said much earlier that ruffed grouse lay their first egg a week after they make the nest, and Brander (1967) had shown that females laid eggs 1-7 days after visiting several males. These females had to have selected their nest sites before they visited males. Steppe grouse often travel a long and relatively straight course toward the nest site after leaving the lek following breeding; this suggests a goal-oriented movement. Some hens return to areas where they had successfully nested previously (Berry & Eng 1985, Chap. 6, Gratson unpubl. data). The time is too short between copulation and deposition of the first egg for these females to adequately select sites after breeding (1-3 days for sharp-tailed grouse, Schiller 1973; and 3-5 days in prairie chickens, Chap. 6). An interesting twist to the pattern can be seen in sage grouse: yearling females average 9.5 days between breeding and egg-laying and adults dally a mean of 7.6 days (Petersen 1980). But Petersen found that most radio-tracked hens took only a day to localize after leaving the lek. One hen took 2 days to travel the 11 km from the lek to her nest area. Petersen (1980, p. 156) suggested that "this small amount of variability could possibly be due to most hens having chosen the nest location prior to breeding." Thus, these female sage grouse arrived an average of 8 days in advance of laying their first egg at the nest site. What does the female do during this period? Predation on nests is extensive in sage grouse (Table 15.2), hens are large, and herbaceous cover is lacking at the nest site. It may be that a sage grouse hen delays her commitment to a precise nest location, until just before sperm viability begins to decrease (see Eng 1963), so that a last-minute and very local assessment of predator activity and vegetation suitability can be made. In summary, the assessment strategy of nest-site selection should require more time where predators are mobile, abundant, and of many kinds. Increased heterogeneity of suitable nest cover in time and space, such as in steppe environments, should also result in more time invested. A reasonable tactic of a female would be to put some distance between herself and an area of high risk. If the potential of nest loss is high, birds should increase the probability of success by increased effort and selectivity in selecting nest sites.
14.4.2 Timing and inconspicuousness Another tactic should be to nest at a time that provides a hen the option to renest if she loses her first clutch (cf. Cartwright 1944). Grouse could nest in June and July when there is maximum, new plant growth. Rather, they nest earlier and do not initially use the new, annual vegetation. Sage grouse nest under sagebrush, ptarmigan under Krumholz or ericaceous shrubs; forest grouse have little herbaceous cover in shady forests, but nest at the base of trees. Prairie grouse seek
512
A. T. BERGERUD AND M. W. GRATSON
residual cover from the previous year, but later there is new plant growth for renesting (Christenson 1970, Schiller 1973). All the grouse normally have overhead vegetation that provides some cover from corvids, which rely on vision, but the dense, low cover for concealment from mammalian predators seems inadequate since a number of ground predators depend more on scent than vision. A universal tactic should be to nest when there is a maximum amount of nesting habitat that predators must search to find nests. Hens should start their search for nest sites as soon as snow cover has disappeared. Some of the most secure sites could be in depressions and shrub thickets, which may lose their snow cover last. Also, brown females would be conspicuous if snow cover was complete. Byrkjedal (1980) placed dummy clutches in the field in May when bare patches were small and in the same scrapes in June when snow was gone. Predators (mostly foxes) found more nests in May, when their search areas were smaller, than in June when more of the area was brown and they had to search larger areas. The timing of breeding and nesting commonly varies a maximum of 2 weeks in grouse populations (Table 14.3). Birds nest earlier in years when there is little snow in March and April and later when snow lingers in April (Fig. 14.10). In Minnesota, spruce grouse bred 2 weeks later in 1970, when 61 cm of snow fell in mid-April than in 1969 (Haas 1974). Spruce grouse in Alberta nested later in 1967 than in 3 other years; the delay followed a winter of much snowfall and deep snow on the ground (McCourt 1969). Ruffed grouse in the same area also delayed nesting. Peaks in drumming frequency of ruffed grouse males are correlated with the number of female visits, and females begin to lay eggs shortly after breeding (Brander 1965, Kupa 1966, Maxson 1974). Aubin (1970, p. 27) said of one year that "the entire study area was blanketed by approximately two inches of snow until the 7th of May. This appears to be responsible for the delay in peak drumming until May 17" (the normal date for 4 years was 3-4 May). In British Columbia, blue grouse breed approximately a week later each year at Tsolum Main than at Comox Burn, corresponding with the later disappearance of snow at the former site (Zwickel 1977). The relationship between breeding and nesting and the disappearance of snow cover also holds for the steppe and tundra grouse. The mean date of nesting by prairie grouse in North Dakota is correlated with snowfall in April and May (r = 0.531, n — 16 years, data from Kobriger pers. comm.). Sharp-tailed grouse severely curtailed advertising and nesting 10 April-4 May in North Dakota in 1970, when 66 cm of snow fell (Christenson 1970). The date of peak numbers of breeding prairie chickens in Wisconsin and the date of nesting by females in Minnesota are both correlated with spring snow conditions (Fig. 14.10, Table 14.4). Sage grouse in Colorado nested later in 1979, after a winter of deep and persistent snow, than in 1978 after a mild winter (mean nesting date of 16 versus 1 April, respectively, Emmons 1980). Patterson (1952) suggested that sage grouse wait on the range for the snow to recede before breeding.
SURVIVAL AND BREEDING STRATEGIES
513
Table 14.3. Variation among years in breeding or hatching nests Grouse groups and species Forest Spruce grouse Alberta3 Blue grouse Vancouver Island b Ruffed grouse Central B.C. Central Alberta Steppe Sharp-tailed grouse North Dakota Prairie chicken c Wisconsin Sage grouse Montana Tundra White-tailed ptarmigan Colorado Rock ptarmigan Alaska Willow ptarmigan Norway 3 b
c
Hatching date
No. of years
Mean
13
3
Difference (days)
Earliest
Latest
30 June
20 June
8 July
18
12
18 June
7-8 June
~ 1 July
10
12 6
1 1 June 10 June
6 June 5 June
17 June 21 June
11 16
22
20 June
14 June
26 June
12
14
23 April
14 April
28 April
14
10
13 June
7 June
21 June
14
12
19 July
5 July
26 July
21
9
22 June
18 June
1 July
13
21
23 June
18 June
2 July
14
Mean or median date reported. Zwickel & Bendell (1967) and Zwickel (1977) present hatch by weeks; values in table are only approximate. Represents peak numbers of copulations. Sources: Spruce Grouse — Smyth & Boag 1984; Blue grouse — Zwickel & Bendell 1967, Mossop 1971, Zwickel 1977; Ruffed grouse - Central B.C., Davies & Bergerud, Chap. 3; Central Alberta Rusch et al. 1984, pers. comm; Sharp-tailed grouse — Kobriger pers. comm; Prairie chicken — Hamerstrom & Hamerstrom 1973; Sage grouse — Wallestad & Watts 1972; White-tailed ptarmigan — Gieson et al. 1980; Rock ptarmigan — Weeden & Theberge pers. comm; Willow ptarmigan — Myrberget Chap. 11.
Ptarmigan also breed later in years of late springs (Watson 1965, Mercer 1967, Bergerud 1970a, Fig. 14.10). Indeed, the generalization appears to apply to all northern grouse. There is a synchronization of the disappearance of snow and the start of nest searching in grouse. This synchronization allows females maximum residual cover in which to select nest sites, reduced conspicuousness, and results in maximum space that predators must search. An alternative hypothesis to explain the close synchronization between spring weather and nesting dates is that in years with an early disappearance of snow,
514
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.10. Annual dates of nesting are correlated with disappearance of snow cover. All snow statistics shown above were gathered at government weather stations, often at some distance from the study areas. (Data from McCourt 1969, Hamerstrom & Hamerstrom 1973, Giesen et al. 1980, Davies & Bergerud Chap. 3, Myrberget Chap. 11, Keppie pers. comm., Kobriger pers. comm.)
females can begin feeding on new plant growth and develop eggs sooner than in a later year (Lack 1968). Contrary to this hypothesis is a lack of correlation between the appearance of spring plants and the date of nesting in grouse. In the
SURVIVAL AND BREEDING STRATEGIES
515
Table 14.4 Breeding and nesting by prairie chicken hens in Minnesota in relation to spring weather and weight of hens (adapted from Svedarsky 1979)
Parameters
Late spring (1975)
Intermediate spring (1976)
Early spring (1977)
Mean date went to lek for breeding
15 May (5)a
30 April (12) (12)
19 April (12)
Mean date of first egg
15 May (8)
29 April (10)
1 May (7) (7)
Clutch size (first nests)
± 0.56 (6) 13.3 +
13.5 ± 0.43 (10)
14.0 ± 0.62 (7)
Weight when captured on lek (g)
934 + 12 (5)
944 + 21 (6)
856 ± 14 (7)
April minimum temperature (°C) (/ o/^\b C) b
a b
-16.1 -16.1
-3.9 -3.9
April mean temperature (°C)
2.3
8.3
9.5
End-of-March snow depth (cm)
18 18
0
0
March snowfall (cm)
35
31
10
8.9 8.9
Number of hens in parentheses. Weather data from Crookston, Minnesota.
far south, by the time the birds nest, growing seasons are well under way for the Attwater's prairie chicken in Texas and for ruffed grouse in Georgia and Kentucky. As one moves north, there is progressively less time between the start of new plant growth and the egg-laying period. Ruffed grouse in central Alberta hatch about 7 June (Rusch et al. 1984, pers. comm.), and the growth of green plants begins about 25 April. Here egg-laying is concomitant with new plant growth and too late to affect the prelaying weight and condition of the hen. Similarly, willow ptarmigan in Newfoundland hatch their chicks the last week of June, but egg-laying begins in mid-May at the time mean temperatures first reach values needed for plant growth. Rock ptarmigan in Scotland start laying just as new plant growth is initiated (Watson 1965), and ptarmigan in the high Arctic must lay eggs before the start of the growing season (cf. MacDonald 1970). Most populations show variations in the timing of nesting, but plant growth in many places is relatively constant. In many other places birds must nest before the growing season starts, thereby precluding any variability in herbaceous plant nutrition affecting laying hens. Also in contradiction to the nutrition hypothesis is that weights of females in
516
A. T. BERGERUD AND M. W. GRATSON
a series of springs are either constant or vary with winter nutrition, and do not reflect gross availability of spring foods. The average dates first eggs were laid by prairie chickens in Minnesota, from 1975 to 1977, were 15 May, 29 April, and 1 May (Svedarsky 1979, Table 14.4). Neither the date of copulation nor the start of egg-laying covary with the weight of females captured at breeding leks (Table 14.4); thus, hens that were heavier did not begin sooner. The mean hatching date of ruffed grouse in central Minnesota was 23 April in 1971 and 28 April in 1972 (P < 0.01), whereas on average females were lighter in the earlier year (517.4 ± 19.0 g versus 544.3 ± 14.9 g, P < 0.01) (Maxson 1974). March and April weights of Wisconsin prairie chickens also showed no correlation with copulation peaks, (Hamerstrom & Hamerstrom 1973, Hamerstrompers. comm., Figs 14.10, 16.10). Yet another hypothesis states that the start of breeding in ptarmigan is correlated with numbers. As breeding density increases, nesting is progressively later (Moss et al. 1974). However, we find no significant correlation between the density of birds and date of hatch in Scotland's red grouse (r = 0.139, n = 9; Jenkins et al. 1963), rock ptarmigan in Alaska (r = - 0.199; « = 9; Weeden & Theberge 1972, pers. comm.), nor for willow ptarmigan on Brunette Island, Newfoundland (r = -0.752, n = 4; Mercer, 1967), or for Newfoundland as a whole (r = 0.230; n = 5; Bergerud 1970a). Earlier, Watson (1965) had said that rock ptarmigan hens nest earlier in warm Mays and also that breeding was delayed by an unusually late disappearance of snow (e.g., 1951) and by heavy snowfalls in April and May. A final hypothesis proposed to explain variability in nesting dates is that females nest at a time that ensures a maximum abundance of insect food when the young hatch (Lack 1954). Several studies have shown that the number of young per adult in summer and fall is higher after early-nesting springs than after late ones (Jenkins et al. 1963, Marcstrom & Hoglund 1980, Chap. 11). One possible factor is that clutches are larger in early years. We found no studies that documented a peak in insect abundance coinciding with hatching, but two studies of blue and spruce grouse have shown that insect numbers do not necessarily align with peak dates of hatching (Elliot 1979, Armleder 1980). Evidence also indicates that late-hatched young may survive as well as young that hatch early (Bergerud 1970b, Parker 1981 b). Eggs may be larger in late years in willow ptarmigan (Myrberget 1977, Parker 1981b, but see Erikstad et al. 1985). Yearlings generally nest slightly later than adults and have fewer chicks, but differences in brood size can be explained on the basis of clutch size (Zwickel et al. 1977, Keppie 1982); the chicks of yearlings survive as well as those of adults. A major problem for chicks at hatch is their inability to thermoregulate. Early-hatched young face colder temperatures and require more brooding time, leaving less time for feeding (Erikstad & Spidso 1982, Erikstad & Andersen 1983). Later-hatched young have been shown to grow faster than early-hatched
SURVIVAL AND BREEDING STRATEGIES
517
young in blue grouse (Redfield 1978), spruce grouse (Quinn & Keppie 1981), and willow ptarmigan (Myrberget 1977). Females initiate egg-laying approximately 40 days before their eggs should hatch. The environment during laying is brown; temperatures are variable and there are no environmental clues that would clearly relate to the predictability and abundance of insects more than a month later. The considerable variation in chick survival in the first 14 days of life may be balanced by the advantages of nesting early; females can be expected to improve their fitness by experience, and nesting early permits sufficient time to use this experience in a second attempt, if necessary. The timing of the entire nesting sequence relates, in our view, to successfully hatching the nest, and only secondarily to chick survival. The female has more control of the nesting decision and less control of the more unpredictable mortality factors that operate after her chicks hatch.
14.4.3 Spacing and other behaviors to remain inconspicuous Females may reduce the likelihood of their nests being discovered by remaining motionless and in place as predators pass nearby, by performing distraction displays at appropriate times, and by spacing away from conspecifics. Also, females should remain away from the nest site when feeding (cf. Lance 1984, p. 79). We all recognize that females are cryptic; they have a general "female coloration pattern" (Hjorth 1970). Females do have contrasting colors that they usually keep concealed. The black ruff, the black rectrices under brown coverts, and the white shoulder spots of ruffed grouse are examples (Hjorth 1970, Lumsden 1970). Females commonly walk lower to the ground than males, and in general behave in ways that emphasize inconspicuousness (cf. Bendell & Elliott 1967). A female's behavior should be most inconspicuous near the nest unless she is discovered and displays to distract attention away from it. In general, female grouse come and go secretly to their nests, and travel some distance before they fly. They often "sneak" back after landing. Females nesting without overhead canopies move to and from the nest most frequently in the crepuscular hours when light intensity is low and few diurnal predators are about. Females use a host of tactics that we believe help to prevent the nest from being discovered (Table 14.5). Giesen & Braun (1979a) reported that white-tailed ptarmigan most frequently left the nest during foggy weather. Sage grouse females may wander farther from the nest on darker days than on bright ones (Petersen 1980). Generally, females "sit tight" during the late stage of incubation, and risk their own safety to keep the eggs concealed. In black and white photographs of the nests of ruffed grouse, sage grouse, and prairie chickens, the eggs are much more conspicuous when the hens are absent (Bump et al. 1947, Patterson 1952) than when they are incubating (e.g., Gross 1930, Schwartz 1944, Hamerstrometal. 1957). But early in the nesting period, when females are possibly still prepared to renest, females may favor
518
A. T. BERGERUD AND M. W. GRATSON
their own safety and sneak or fly from nests when predators are still at long distances (see also section 14.4.6). Between these extremes females may perform distraction behaviors near nest sites to decoy predators away. However, field workers seldom disturb nesting females and there is little information on the quality or variability of tactics, or of the environmental constraints that affect a female's options. For example, Gross (1930) reported that female prairie chickens seldom used broken-wing displays, but Schwartz (1945) reported that they performed decoy behavior; Svedarsky (pers. comm.) thought prairie chicken females mostly flew away or 'sat tight." In Manitoba, Gratson has observed some sharp-tailed grouse hens decoy humans away from nests by wing-dragging and tail-flicking behaviors and by clucking, but most simply flew away. One possible consideration in assessing the potential cost of distraction in relation to the alternative—sitting tight—is the stage of chick development. Predators may prefer protein-rich egg yolk and albumen in lieu of bone and feathers of partly developed chicks. Foxes prefer eggs, and Sargeant (pers. comm.) has obTable 14.5. Behavior of nesting grouse hens a Spruce grouse: May or may not cover eggs with vegetation during laying phase; few feathers in nest bowl; feeds 20-200 m from nest; deposits eggs at 1315, 1700, 1100, 1200, 1305, 1630 hours; during incubation phase feeds about 32-83 m away, walks a few paces from nest and flies; feeds rapidly and deposits "clocker" droppings while feeding in trees, or < 10 m from nest, flies back to nest site and walks last meter to nest bowl; reported feeding times were 15.2 min/bout 3-6 x/day, 16 min/bout 2-3 x/day and 53 min/bout 1-3 x day, between 0600-1015 hours and 1715-2025 hours (McCourt et al. 1973, Haas 1974, Herzog 1978, Keppie & Herzog 1978, Robinson 1980). Blue grouse: During incubation phase in AM may walk to feed, walk back, fly to feed, fly back, or a combination; in PM consistently flies both ways after walking away from or landing a meter from the nest (Lance 1967). Ruffed grouse: Female lays eggs at all times of daylight —0530-1600 hours; spends several hours laying eggs; eggs left uncovered; during incubation phase feeds 15-30 min/bout 1-3 X/day, 0400-2100 hours; 88% of absences in AM before 1000 hours and 75% PM absences after 1700 hours; yearlings spend longer time feeding than adults (Bump et al. 1947, Kupa 1966, Barrett 1970, Maxson 1977). Willow ptarmigan: Lays eggs < 10 min. after arriving at nest; covers eggs with vegetation before leaving; during incubation conceals white primaries; rarely accumulates feathers in nest bowl; clocker droppings away from nest, where it feeds in male's territory; female scans before leaving and does not cover eggs; walks 20-40 cm, flies to feeding area; 53% of feeding absences between 0600 and 1800 hours and 48% between 1800 and 0600 hours; averages 52.5 min/bout 3x/day (Weeden 1965, Pulliainen 1978, Bergerud pers. files). Rock ptarmigan: Early laying phase, covers eggs completely; during incubation phase no covering; feeds between 2200 and 0200 hours in "dusky hours of Arctic night"; averages 24.1 min/feeding bout; returns by walking and flying partway, feeds quickly with head and tail low,
SURVIVAL AND BREEDING STRATEGIES
519
served a fox burrow under an incubating duck to steal her eggs! The food value of the eggs should further decrease as incubation progresses. Predators may search more for, or more easily locate, nests during egg-laying than when most females are incubating. Their preference for eggs over flesh may be one of the factors that has allowed sitting tight to be an option. If a grouse can feign death, like a mallard (Sargeant & Eberhardt 1975), a female may still escape even when discovered (Sargeant pers. comm.). The risk of distraction may also vary with different kinds of predators, their experience, and the vegetative cover. Ptarmigan often give distraction displays at the feet of feeding caribou. Here there is no risk to the female herself, but egg destruction is a real possibility. To fly or sneak off is not an option. Gross (1930) recorded a prairie chicken using the broken-wing display on a domestic cow, yet the same female sat tight as a local farmer passed by. It is conceivable that females Table 14.5. continued concealing white underparts, then repeats "run-feed" closer to nest; at the last minute runs and settles on clutch; eats white feathers in nest bowl; will perform distraction displays —hissing, "trembling," and dragging lower wings and tail (Weeden 1965, MacDonald 1970, Watson 1972). White-tailed ptarmigan', Deposits eggs 0830-1630 hours; lays eggs immediately upon reaching bowl, then becomes inactive until leaving nest; feeds in male's territory 100-300 m away from nest; eggs completely covered with vegetation while absent; during incubation phase feeds mostly before sunrise and after sunset 300 m from nest; deposits clocker droppings at feeding; may swallow white feathers at nest; 4 of 8 times left nest during daylight when there was fog, rain, or snow; distraction-hissing while exposing white carpal patches and "advance-retreat" behavior (Schmidt 1969, Giesen & Braun 1970a, Giesen et al. 1980, Chap. 8). Sharp-tailed grouse: During incubation leaves nest to feed 20-45 min/bout 1-2 x/day 0800-1200 hours and 1600-1900 hours; feeds < 200 m from nest—mostly 1 or 2 directions away from nest site; eggs uncovered during incubation phase (Hart et al. 1950, Schiller 1973, Gratson 1983). Prairie chicken: Some hens cover eggs during laying phase; lays eggs 0800-1400 hours; incubating hens average 38 min/bout off the nest feeding in AM and 50 min in PM, departing 0500-0800 hours and 1630-2230 hours; eggs uncovered during incubation phase; distraction displays include "broken-wing" and flying slowly away and circling back to nest; walks in crouched position away from nest a few meters after surveying area, hesitates, then flies to feeding site; flies back and stops 15-30 m away before stealthily approaching nest and settling on eggs (Gross 1930, Lehmann 1941, Schwartz 1945, Silvy 1968, Bowen 1971, Svedarsky 1979). Sage grouse: Feeding sites of incubation hens 0.3 km from nest; feeds before sunrise and after sunset, averages 15-30 min/bout 1-2 x/day; often goes through the sagebrush to exit the nest (Patterson 1952, Nelson 1955, Peterson 1980, Autenrieth 1981). a
Data are unavailable on all behaviors for each species.
520
A. T. BERGERUD AND M. W. GRATSON
may be able to distinguish foxes from skunks, and hunting, "sharp-set" hawks from resting, recently fed ones and adjust their strategy of inconspicuousness. One reason why some ruffed grouse females nest in vegetation edges (Bump et al. 1947) may not only be to have a safe exit, but to be able to assess the identity and character of any approaching danger. A bird nesting where there is little cover at ground level, or a sage grouse under a sagebrush, may have to sit tight and take a greater risk than a prairie chicken in the center of a grass tussock, where the bird can sneak away or appear at a distance with decoy antics. We have not even scratched the surface of the variability of and the factors that influence female decision-making in preventing her nest from being discovered. Female willow ptarmigan are in a unique position because the male helps with distraction behavior. However, males stay away from the nest in diurnal cover when not helping (Fig. 13.6, Weeden 1959b). Pulliainen (1978) reported that males never came closer than 10 m when the female was incubating. Females do appear to benefit from the presence of males. The mean nesting success of females that had males compared with a group in which some had no male (males removed) was 79% (n = 66) and 68% (n = 56) respectively (Hannon 1982). Mossop observed at Chilkat Pass, British Columbia, that when there was a dense population, male ptarmigan were usually obvious visually and also called when females were nesting and he disturbed them. By contrast, in a sparse population in Newfoundland in the 1960s, males were seldom seen; however, they were more obvious in one year (1962) when the population was dense and aggressive (Mercer 1967). When populations are dense, territories are small and wandering intruders pass closer to nesting females than when populations are sparse and females are farther apart. Foxes may be habituated to the constant decoy behavior of male ptarmigan. Possibly a factor that we do not appreciate is that males running on the ground or "walking in circles" (Watson & Jenkins 1964) may leave a maze of scent trails that masks the route of the female when she leaves the nest. What is clear is that female willow ptarmigan have a high nesting success to which the male may contribute significantly. A major tactic of females of the promiscuous species should be to avoid males during nesting (Brown 1964, Crook 1965, Wittenberger 1978, Chap. 5) (Figs. 14.11, 14.12). Males are conspicuous and can be expected to display if they encounter females. The separation between females and males is clearly apparent in forest grouse. The nests of spruce grouse are generally between or away from the location of advertising males (Ellison 1971, 1973, Herzog & Boag 1978, Nugent & Boag 1982, Fig. 14.12). Forty-seven percent of 484 ruffed grouse nests were more than 125 m from drumming males (Bump et al. 1947). Nesting ranges of female blue grouse on Stuart Island, British Columbia, in 1976 showed little overlap with male ranges (Fig. 13.7, Bergerud & Butler 1985). When Lewis (1984) removed territorial, blue grouse males he found that females increased
SURVIVAL AND BREEDING STRATEGIES
521
Fig. 14.11. Sizes of advertising ranges of males increase from steppe to forest to tundra; sizes of prelaying ranges of females decrease steppe to forest to tundra. Females of polygynous species visit males for breeding but nest away from these conspicuous males and on the edges of their prelaying ranges. Nesting forest grouse females also avoid other females. Females of the monogamous ptarmigan exclude other females, and possibly their males, from the immediate vicinities of their nests.
522
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.12. Spruce grouse wintered in a small area of "winter" cover. In spring, yearling females moved farthest from the cover and spaced themselves away from adult females and advertising males; adult females moved away from advertising males; males moved the least distance from their winter cover. Spring prelaying ranges of females were mutually spaced, and nests were commonly located on the edges of their ranges. Females with broods traveled over wider, overlapping ranges in the summer. (Adapted from Herzog 1977a).
SURVIVAL AND BREEDING STRATEGIES
523
their ranges—here is experimental evidence that females will avoid the ranges of males. The mean distance of nests from males of the lek species is most commonly equal to half the distance between leks (Fig. 13.10) (Christensen 1970, Schiller 1973, Wallestad & Pyrah 1974, Chaps. 5,6). Females also appear to avoid the locations used by males when the latter are feeding away from leks (Chap. 5). In several grouse species, e.g., in sharp-tailed grouse, prairie chickens, and ruffed grouse (Hamerstrom 1939, Hjorth 1970), males probably cannot easily recognize females by morphology. In other species, such as blue grouse, they probably cannot recognize females until they are only a few meters away (e.g., Jamieson 1982). This sexual indistinguishability permits females to escape the attention of males unless they reveal their identity by postures or calls. It should increase female and nest inconspicuousness to also avoid other females, and to have nests dispersed (Tinbergen et al. 1967). Dispersed nests should suffer lower predation rates by crows than clumped nests (e.g., Sugden & Beyersbergen 1986). This dispersion has been documented in forest grouse but not in steppe grouse. In both spruce and blue grouse, nesting ranges of females may be mutually exclusive and female aggressiveness and territoriality have been documented (Lance 1967, 1970, Stirling 1968, Herzog&Boag 1977, 1978, Hannon 1978, Nugent & Boag 1982, Bergerud & Butler 1985, but see Hines 1986a). The spacing of ruffed grouse females has not been determined; no one has tagged all or most of the individuals in a prescribed space. Females have, however, appeared evenly spaced as they flushed on walking transects (Rusch & Keith 197la) and "ranges of adjacent hens usually had little overlap suggesting some spacing mechanism" (Maxson 1978, p. 65). The spacing of steppe grouse females is not yet clear. Females are commonly seen together, especially in sage grouse. Nests of sage grouse and of prairie chickens are often found near each other in optimum habitats (Schwartz 1945, Baker 1953, Gill 1965, Klebenow 1968, 1969, Stewart 1975). For example, in 1903 a large prairie fire exposed many prairie chicken nests. These nests were clumped around sloughs and some were 15-20 m apart (cf. Johnson 1964). Does such clumping result from the loss of nesting habitats or is it a naturally-evolved behavior to utilize patchy, heterogeneous environments? The habitats in 1903 should have been relatively pristine. Females of the lek species search prelaying ranges that are too large for females to defend. They are also in open habitats where the predation risks associated with overt, aggressive behavior may be higher than in forest habitats. Still, they could, through mutual avoidance, disperse themselves. Indirect evidence of spacing can be seen in a comparison of prairie chicken populations in Minnesota, where recently they expanded their range, compared to those in Wisconsin, where habitat was mostly unsuitable surrounding relic populations. In Minnesota, as total numbers increased, the mean number of males
524
A. T. BERGERUD AND M. W. GRATSON
per lek and lek density remained relatively constant (Fig. 16.1); as the population increased new leks were formed and spaced away from other leks. If male prairie chickens form leks in areas to be near females, as we argue earlier, then this should mean there is a limit to the number of females within a given space and that females space themselves, thus expanding into new habitats. In Wisconsin, however, the mean number of males per lek was correlated with the total male count (r = 0.757, n = 22; Hamerstrom & Hamerstrom 1973). There, females were "packed into" more restricted and less variable acreage of habitat and spacing was reduced. Alternatively, under some as yet unknown conditions clumping may be the most inconspicuousness pattern for females and nests. A clumped distribution, where nests within a cluster are close but there are long distances between clumps, might reduce the probability of discovery by predators where space is extensive. Dummy nests in a clumped dispersion in Attwater's chicken habitat were preyed on less than nests in a dispersed pattern (Horkel et al. 1978). Horkel et al. (1978) argued that the rare patches of nests may preclude the formation of a "searching image" (Tinbergen et al. 1967, Croze 1970, Krebs 1973). All female grouse are less mobile while laying eggs and incubating (Table 14.2). Reduced ranges should help them to avoid conspecifics and predators. The ranges of nesting, female spruce grouse overlap less with displaying males than with yearling males (9% versus 42% overlap), and they overlap on average only 6% with other nesting females (Herzog & Boag 1978, Fig. 14.12). The locations of advertising males would be predictable, but because yearling males move unobtrusively they would be more difficult to avoid.
14.4.4 Avoiding nest ambush Females may often be faced with the decision of whether to flush from a predator and have the eggs located, or to sit tight and risk being killed. If females can select nest sites with reduced risk to themselves, they can afford to remain on the nest longer and give greater preference to the concealment of the eggs. The female herself provides the best camouflage of the clutch. One behavior that should reduce risk to the female on the nest is for her to range on only one side of the nest during egg-laying and incubation periods. Clocker droppings and ground scent would then be restricted to this route. Because adjacent, female nest ranges seldom overlap, at least in forest grouse, it is unlikely that another female would travel near the side that the resident hen does not visit. Predators could pass near the nest on at least one side with a reduced probability of detecting the female's or her neighbor's scent. With the exception of willow ptarmigan all grouse appear to nest most often on the outside perimeter of their nesting ranges (cf. Herzog & Boag 1978, Hannon 1978, Chaps. 5, 6, 8, 9). Pheasants (Phasianus colchicus) also show this tac-
SURVIVAL AND BREEDING STRATEGIES
525
tic (Dumke & Pils 1979). Female willow ptarmigan commonly locate their nests within the center of the male's territory (Weeden 1959b, Erikstad 1978, Hannon 1982, Erikstad 1985a,b, Chap. 13, but see Steen et al. 1985). The ptarmigan cock stays with the female during nesting and prevents other males (and thus other females) from coming near the nest. If communal foraging areas occur, they are at the edges of territories (Pedersen et al. 1983). Thus, all the grouse have tactics of reducing activity near the nest. Forest grouse commonly reinforce their edge position by having a backstop—a tree or stump adjacent to and behind the nest (i.e., females face away from it). Bump et al. (1947) indicated that 52% of 1,158 ruffed grouse nests were at the base of trees, 15% at the base of stumps, 14% under bushes, and 14% beside logs or rocks. Haas (1974) reported 8 of 11 spruce grouse nests at the base of trees, and Redmond et al. (1982) found 88% of 104 nests within 10 cm of a conifer. Similarly, Bendell and Elliot (1967) reported that most blue grouse hens nested under logs, near stumps, or beside small conifers. These backstops permit the female to focus her attention outward, to that part of her range with which she is familiar, and where she is most likely to be detected because of scent. If she is detected from behind, the predator can not spring directly on the hen because of the obstructing backstop. By being next to a tree or shrub, grouse also benefit by having maximum overhead cover to prevent detection by raptors and corvids (Sugden & Beyersbergen 1986). What predator other than a squirrel would travel a search route that went from tree to tree? If tree squirrels are a potential problem, as are red squirrels (Tamiasciurus hudsonicus), the tactic may cost the hen an occasional egg. Nesting next to vertical stems has the additional advantage of deflecting wind currents, allowing the hen to nest in a wind shadow. If the nest is on the lee side of the backstop, predators that hunt by scent would less frequently detect the hen. Sage grouse nest under sagebrush and sharp-tailed grouse commonly nest under bushes if available (Patterson 1952, Artmann 1970, Schiller 1973). Ptarmigan nest beneath prostrate conifers and shrubs, and white-tailed ptarmigan often nest beside rocks. Only the prairie chicken usually nests without benefit of either shrub, tree, or some backstop (e.g., Svedarsky 1979). Twenty-three of 35 prairie chicken nests in Wisconsin had no backstop or overhead shrub cover, based on photographs taked by F. J. Schmidt (University of Wisconsin files). Shrubs and rocks should often prevent a successful ambush, but in grassland habitats where shrubs are rare, prairie chicken hens and sharp-tailed grouse hens should avoid nesting under lone shrubs and instead attempt to become lost in a "sea of grass" (cf. photographs in Hamerstrom et al. 1957, pp. 19, 39). In one study of prairie chickens a worker marked nests with poles and found that corvids sat on the conspicuous marks and found the nests (Bowen 1971). Authenrieth (1981) marked sage grouse nests with conspicuous flags, and ravens (Corvus corax) investigated the objects and found the nests.
A. T. BERGERUD AND M. W. GRATSON
526
The lack of backstops in grassland habitats has cost some prairie chicken females their lives (Table 14.6). Hens apparently hesitate to flush until the very last moment, and without cover that helps deflect predators they are often killed. Except for sharp-tailed grouse and prairie chickens the frequency of ambushed grouse hens is generally less than 1% (Table 14.6).
14.4.5 Philopatry of nesting females Females that nested successfully generally should return to the same area the next year to attempt to nest. The safety of the site has been proven, and the uncertainty Table 14.6. Percentages of grouse hens killed on the nest and that abandon their nests Type of grouse
% die on nesta («) All studies
0
Telemetry
% abandon nestb (n) All studies0
Telemetry
2. 0 (2/102) 4. 6 (3/65) 2. 1 (2/97)
Forest Spruce grouse Blue grouse Ruffed grouse
0.0 (0/53) 0.0 (0/58) 0.0 (0/33) 0.0 (0/92) 0.3 (4/1,598) 0.0 (0/23)
4.1 (6 /148) 10.6 (27/256) 1.0 (17/1,629)
Steppe Sage grouse Prairie chicken Sharp-tailed grouse
0.4 (1/281) 9.1 (11/121) 1.3 (2/159)
13.9 (71/510) 7.3 (64/876) 7.8 (24/308)
Tundra White-tailed ptarmigan Rock ptarmigan Willow ptarmigan
0.0 (0/72) 4.7 (2/43) 0.0 (0/115)
a
b
c
0.0 (0/14) 7.8 (5/64) 4.3 (2/47)
16. 1 (9/56) 9. 9 (14/142) 6.7 (10/150)
2.9 (2 /70) 3.1 (3 /96) 10.0 (12/115)
Number that reportedly are killed on the nest -=- total hens observed for which there is sufficient information to determine their fates, x 100. Number that reportedly abandon nests of undamaged eggs -H total hens (nests) observed for which there is sufficient evidence that an author reported abandonment if it occurred, x 100. Sources: Spruce grouse—Ellison 1974, Hass 1974, Keppie & Herzog 1978, Robinson 1980, Keppie 1982; Blue grouse-Zwickd & Lance 1965, Mossop 1971, Weber 1975, Zwickel & Carveth 1978, Sopuck 1979; Ruffed grouse-Bump et al. 1947, Grange 1948, Kupa 1966, Neave 1967, Barrett 1970, Maxson 1974; Sage grouse-Batterson& Morse 1948, Patterson 1952, Gill 1965, Klebenow 1969, May 1970, Wallestad & Pyrah 1974, Petersen 1980; Prairie chicken-Gross 1930, Hamerstrom 1939, Lehmann 1941, Schwartz 1945, Grange 1948, Baker 1953, Ammann 1957, Yeatter 1963, Silvy 1968,Bowen 1971, Horak 1974, Rice & Carter 1975, 1976, 1977, Sisson 1976, Horkel et al. 1978, Riley 1978, Sell 1979,Svedarsky 1979, Vance & Westemeier 1979; Sharp-tailed grouse-Gross 1930, Hamerstrom 1939, Grange 1948, Ammann 1957, Brown 1966b, 1967, 1968b, Bernhoft 1969, Artmann 1970, Christenson 1970, Pepper 1972, Schiller 1973, Rice & Carter 1975, 1976, 1977, Kohn 1976, Ramharter 1976, Sisson 1976; White-tailed ptarmigan Choate 1963b, Schmidt 1969, Giesen et al. 1980; Rock ptarmigan-Watson 1965, Weeden & Theberge 1972, pers. comm. 1982; Willow ptarmigan — Bergerud 1970a, Hannon 1982.
SURVIVAL AND BREEDING STRATEGIES
527
associated with new sites should impart a greater risk. If, however, females encounter current, unfavorable information, such as low vegetation suitability and high predator activity, about the old site when they return in spring, they may elect to move. Philopatry of hens has been documented for all the grouse groups (Choate 1963a, Berry & Eng 1985, Chaps. 6, 8). What remains to be determined are its general frequency, variability, and factors that influence it. Adult hens usually move shorter distances between wintering areas and breeding ranges than do juvenile females and return earlier (Table 14.7). This difference suggests less searching in spring by returning adults than by naive yearlings. Adults have invested time in learning an area—this experience should be incorporated into their nesting strategy. Although successful hens should show philopatry, it is less clear if unsuccessful hens should again invest in the same site if unsuccessful. If failure was a result of weather, flooding, burning, or filching of eggs, a suitable stimulus to shift may be lacking (see also section 14.4.6). These factors are also unpredictable in their recurrence, and a hen may be "betting" for the site when she returns after an unsuccessful year. But if nest loss is associated with a strenuous encounter with a predator, the female may reduce her current risk by changing areas. It is conceivable that the distance a female moves between nesting efforts from one year to the next may relate to the mobility of the predators that caused the nest failure. For example, red-winged blackbird (Agelaius phoeniceus) females move farther when they lose their clutches to magpies than when mice take their eggs (L. Rotterman pers. comm.). Many of the subtleties of philopatry as a tactic need to be clarified. 14.4.6 Nest abandonment and renesting Females should abandon their current clutch if disturbed and try again if the certainty of a current loss outweighs the unpredictability of the loss of a future effort (Fig. 14.13). Abandonment may occur during laying or incubation phases. Dawkins and Carlisle (1976) argue that the decision to abandon the nest and renest should depend more on future prospects than on the loss of current investment. Nest abandonment in grouse decreases as investment increases throughout the nesting schedule. Eleven of 12 (92%) blue grouse females deserted their nests following human disturbance when they had three or fewer eggs, whereas only 9 of 22 (41 %) deserted when the number of eggs was from four to seven (Zwickel & Carveth 1978). A female whose clutch is near hatching has almost succeeded and the probability of success, even if disturbed, may still outweigh starting over. The probability of further disturbance and the severity of the encounter should also be considered in the female's decision on whether to abandon her current investment. We believe that grouse may nest earlier than is optimal for early chick survival in order not to prohibit the opportunity to try again—abandonment
528
A. T. BERGERUD AND M. W. GRATSON Table 14.7. Movements of banded grouse,
Grouse groups and species Forest Spruce grouse
Ruffed grouse
Blue grouse
Steppe Sharp-tailed grouse
Movement type
summer to winter summer to summer summer to fall winter to spring winter to spring summer to winter summer to fall summer to fall summer to fall summer to fall summer to fall summer to spring summer to winter
winter to fall winter to fall winter to spring
Prairie chicken
Sage grouse
Tundrad White-tailed ptarmigan
winter to spring fall to spring spring to fall summer to fall spring to winter
summer to fall summer to winter winter to summer summer to winter
Willow ptarmigan a
b
c d
summer to fall
Refers to distance moved; A = adult, Y = yearling, J = juvenile, B = brooding, NB = nonbrooding. Mean + 1 SE (n) of each sex and age class for which a trend is shown; in some instances SE could not be computed. Approximate; computed from reference given. No estimates were located for rock ptarmigan.
SURVIVAL AND BREEDING STRATEGIES
529
emphasizing trends of sex and age classes
Distances (km) b
Trend3
a < A J < J A < J A < J < A <J <
A < A < J < J a
A
< J
< A
A
< J
< J
< A
J
< J
A < J < A J < J A < Y A < A
< J
J
<J
< J <
A
,J
A
< J
<J
J
< A
< a <j
A
1.7 1.6 0.9 0.8 2.2 0.3 0.1 0.6 2.7 0.4 4.4 6.7 0.9 0.3 1.9
± 0.5 (4); 5.0 ± 1.0 (9) ± 0.5 (14); 3.2 ±0.5 (12) (9); 3.2 (13) c ± 0.2 (19); 1.8 ± 0.9 (9); ± 0.4 (46); 2.3 ± 0.5 (19) (15); 1.9 (6)c (11); 0.5 (15); 0.7 (82); 1.2 (40)c ± 0.1 (36); 1.0 ± 0.3 ( 1 1 ) ; 1.8 ± 0.1 (125) ± 1.1 (3); 2.9 ± 0.5 (3) (50); 2.9 (43)c ± 0.9 (10); 5.4 ± 1 . 2 ( 1 1 ) ; 13.9 (1) ± 3.8 (6); 13.7 ± 4.2 (12) (24); 1.4 (42) (17); 0.3 (16); 0.5 (21); 0.7 (24) (34); 2.3 (76)
6.9 ± 0.8 (34); 9.1 ± 0.8 (87); 10.6 ± 1.8 (32); 21.6 ± 3.7 (80) 4.2 ± 0.3 (9); 6.1 ± 0.7 (33); 13.6 ± 1.8 (42); 15.8 ± 3.9 (18) 1.2 ± 0.3 (5); 2.2 ± 0.5 (5); 2.8 ± 0.6 (3); 48.5 (1) 2.7 (297); 3.7 (369); 4.7 (98); 7.9 (125) c 1.2 + 0.3 (27); 4.1 ± 1.3 (5) 3.4 ± 1.1 (4); 8.8 ± 2.2 (9) 6.1 ± 1.1 (8); 8.1 ± 2.1 (12) 9.4 (68); 17.5 (10)c
2.3 3.5 2.4 6.2 1.5 6.4 1.3
± 0.2 (53); 3.0 ± 0.2 (68)c ± 0.4 (6); 5.1 ± 0.4 (9)c ± 0.5 (14); 3.8 ± 0.7 (18); ± 0.8 (28); 10.4 ± 0.9 (39)c ± 0.4(4); 3.8 ± 1.2 (2); 6.0(1); ± 0.5 (5) (17); 1.4 (9); 2.2 (14) c 3.4 (15)
Sources: Forest grouse-Herzog & Keppie 1980, Robinson 1980, Ellison 1973, Schroeder 1985, Hale & Dorney 1963, Dorney & Kabat 1960, Godfrey & Marshall 1969, Rusch & Keith 1971b, Sopuck 1979, Zwickel etal. 1968, Mines 1986a,b; Steppe grouse - Robel etal. 1972, Sisson 1976, Hamerstrom & Hamerstrom 1973, Copelin 1963, Campbell 1972, May 1970, Beck 1975; Tundra grouse-Giesen 1977, Hoffman & Braun 1975, Herzog 1980, Bergerud 1970a.
530
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.13. The decision to abandon a nesting effort is hypothesized to be influenced by the certainity of a current loss as perceived by experience and moderated by stage of investment and female risk. Renesting is a tactic in which the decision is hypothesized to be based on predictability of future success, which is moderated by plant cover, relocation space, and lateness of the season.
should be considered a tactic to remove some of the uncertainty of success. Abandonment and renesting can correct for an initial error in nest location or unpredictable circumstances. With the exception of willow ptarmigan and possibly blue grouse, nest abandonment is higher for steppe grouse than for the other groups (Table 14.6). The sage grouse hen has little concealing, herbaceous growth in sagebrush nesting habitats, but has a backstop and is rarely ambushed. If predators are active near
SURVIVAL AND BREEDING STRATEGIES
531
her nest the probability of eventual discovery is high—the eggs are not concealed during egg-laying. She also has a large preincubation range and can move to familiar locations far-removed from the initial nest site. Abandonment and moving may be a better tactic than awaiting almost certain loss. Because prairie chickens and many sharp-tailed grouse do not have backstops or overhead cover to prevent ambush, they should frequently abandon the nest, rather than wait and later be killed on the nest (Table 14.6). They too can move to renesting locations within their large prelaying ranges (Table 14.2). On the prairies and savannas, new June plant growth may provide more cover for second nests of prairie grouse than residual cover did for the initial nests (Schiller 1973). Nest abandonment is generally less frequent in forest grouse than in steppe grouse (Table 14.6). For these species there is less nest predation and females are seldom killed on the nest. Thus, the probable costs of staying are lower than for steppe grouse. Forest grouse also have small preincubation ranges, and it is more difficult to put distance between themselves and imposing risks while remaining in a familiar habitat. In tundra grouse the frequency of nest abandonment varies. Nesting efforts varied during a 10-year cycle of abundance of willow ptarmigan in Newfoundland (Bergerud 1970a). Nine of 41 hens abandoned clutches in 1962, preceding a population decline, but few deserted their nests in other years (2 of 24). In the monogamous ptarmigan, nesting females benefit from male vigilance and decoy behavior. Thus, an assessment of the potential trade-off when abandoning a current clutch should include consideration of what males may do. In white-tailed and rock ptarmigan, cocks may abandon females during nesting. After abandonment, renesting attempts could thus lose the benefits of a vigilant male. Also, there are fewer predators of nests in the arctic than in the more southern latitudes (Fig. 15.7), and nesting success is higher. Abandonment in ptarmigan probably results more often from weather factors than from predator disturbances. Few data exist to determine if nest abandonment varies with the age, and thus the experience of the female. An adult should normally have more experience with predators than a yearling, but she may also have strong site tenacity because of prior nesting success. Yearlings have less experience with predators but probably abandon more readily. One prediction is that an adult nesting in a new site will have a higher probability of abandoning her nest than an adult showing philopatry at a site where she was previously successful. Renesting is a major tactic that quickly incorporates knowledge from an initial failure in order to succeed in the same season before hen mortality intercedes. A major theme of Chapter 15 is that predation of nests is density-dependent. Grouse may increase nesting success with second attempts because nesting densities are lower, as noted for passerine birds (Nolan 1963). Gratson (unpubl. data) has found that dummy nests in July were significantly more successful than were dummy clutches placed out in May and June at the peak of nesting by grouse and
532
A. T. BERGERUD AND M. W. GRATSON
ducks (see also Sugden & Beyersbergen 1986). Another consideration is that late renesting birds could be out of synchrony with denning mammals like foxes, which may switch later to more abundant but less nutritious foods. The decision to renest in the same year may be influenced by the life expectancy of the female. To postpone until the next year depends on the gamble of her being alive. Grouse with long life expectancies should renest less often than those with short life expectancies. The usual timing of death should also affect the decision. Females in some populations die primarily during the nesting and brood-rearing periods (e.g., Maxson 1974, Svedarsky 1979, Angelstam 1984). Females in other populations may face more dangers during the winter (Chap. 4, Huempfner pers. comm.) or in early spring (Mercer 1967). A prairie chicken may have a better probability of being successful later in the season than to postpone her attempt until next year on the chance that she will still be alive; but a female spruce grouse in Alberta might profit more by waiting until the following year, as her probability of overwinter survival is fairly high (Keppie 1979). Blue grouse on Vancouver Island, spruce grouse in Alberta and Minnesota, and whitetailed ptarmigan in Colorado generally show long life expectancies and low renesting frequencies (Sopuck 1979, Haas 1974, Giesen & Braun 1976, Giesen et al. 1980, Keppie 1982, Sopuck & Zwickel 1983, Chap. 16). Sharp-tailed grouse in North Dakota, Minnesota and Manitoba; prairie chickens in Wisconsin and Minnesota; and willow ptarmigan in Norway and Newfoundland show short life expectancies, and frequently renest (Bergerud 1970a, Christenson 1970, calculated from Hamerstrom & Hamerstrom 1973, Schiller 1973, Erikstad 1978, Svedarsky 1979, Parker 1981b). But there are many exceptions; short-lived ruffed grouse in Minnesota and New York seldom renest, nor do rock ptarmigan in Alaska or short-lived spruce grouse in Alaska (Bump et al. 1947, Ellison 1974, Maxson 1974, Weeden & Theberge 1972, Weeden & Theberge unpubl. data). Steppe grouse should generally renest more frequently than forest grouse, for reasons similar to those explaining abandonment (Fig. 14.13). Forest grouse have small preincubation ranges and rather constant vegetative cover; they can use neither distance nor a major increase in cover to improve their second attempt. Steppe grouse have large prelaying areas from which to choose a second nest site, and new herbaceous growth to conceal eggs has appeared during the laying period. Early grouse workers, such as Hamerstrom (1939) and Ammann (1957), felt that renesting frequencies were low in prairie grouse, but they probably missed renesting attempts without the aid of radio telemetry. Schiller (1973) reported that one sharp-tailed grouse hen moved 19 km between first and second attempts. This renesting hen would surely have been missed had she not been radio-tagged. Gratson has observed that third nesting attempts are common for sharp-tailed grouse radio-tagged in Manitoba, and fourth attempts do occur. In ptarmigan, the sorting out of renesting tendencies, like abandonment
SURVIVAL AND BREEDING STRATEGIES
533
proclivities, are confounded by cyclic changes in breeding behavior. Ptarmigan in Newfoundland readily renested when the population was increasing from a cyclic low, but had reduced renesting in 1962 before the crash (Bergerud 1970a). There was little renesting in rock ptarmigan in 3 of 10 years of low initial failure in 1962, 1963, 1964 (Table 15.3, Weeden pers. comm.). Two studies suggested that late-hatched young of second nests had good survival rates until fall (Bergerud 1970a, Parker 198 Ib). Still, it seems intuitive that the smaller juveniles of late-hatched clutches would be vulnerable to the effective avian predators of the Arctic. In Newfoundland, late-hatched young were extremely vulnerable to hunting (Bergerud & Huxter 1969b). 14.5 The strategy of selecting a male for breeding The higher parental investment of females than males in the reproductive process should result in female rather than male choice of mates (Bateman 1948, Maynard-Smith 1958, Trivers 1972, Bateson 1983). We argue that the mateselection decision is not as critical as the nest-location decision to a female grouse, but if she can show discretion without compromising her nest decision, she should select a male that best enhances her fitness prospects. Females normally mate with only one male, but males of the steppe species may breed more than one female (Hamerstrom & Hamerstrom 1955, Robel 1966, Wiley 1974); commonly only very few of the males mate with most of the females and some males obtain no copulations at all (Table 7.9). Even in the forest grouse, because many yearling males do not breed, some adult males will breed more than one female. Males localize and advertise before mating; clearly it is the female that moves in and closes the distance between the male and herself and makes the choice. Her mate-choice decision in the nonmonogamous species should be based on: (1) assuring fertilization, (2) maximizing expected survival during courtship and coitus, and (3) maintaining an inconspicuous nest site. In addition, in monogamous ptarmigan, the female should select vigilant males who satisfactorily control nesting sites (Chap. 13). Whether or not it is theoretically possible that the female also or solely selects mates on the basis of a correlate of high genotypic quality and thereby possibly improves the fitness of progeny is currently under contention (e.g., Borgia 1979, Weatherhead & Robertson 1979, Lande 1981, Wittenberger 1981b, Kirkpatrick 1982, 1985, 1986, Taylor & Williams 1982, Bateson 1983, Heisler 1985). Females of the three forest grouse should prefer to remain inconspicuous during breeding. Visibility is restricted under canopies, reducing the distance at which birds can detect predators. The female will be at a disadvantage during copulation because ambush is possible and frequently there is little undergrowth in which to hide. Males display at relatively safe sites, but their locations are probably known to predators; and they are frequently killed adjacent to display
534
A. T. BERGERUD AND M. W. GRATSON
stages. A female's inconspicuousness is lost once the male recognizes her as a female and begins courtship. A female may use many tactics to maintain her inconspicuousness. If a hen travels to a distant male, she will often trespass in other females' ranges. If discovered she may lose her conspicuousness by the aggressive actions of a resident female (Hannon 1978, Herzog & Boag 1978, Bergerud & Butler 1985). If the females' ranges she must pass through are unfamiliar this should also increase her risk. Unless there are additional costs a female may often simply choose nearby males. Male and female blue grouse on Stuart Island, British Columbia, and spruce grouse at Gorge Creek, Alberta, were spaced such that there was one male considerably closer to each female than were other males (Bergerud & Butler 1985, Fig. 14.12). At Gorge Creek, Alberta, nine yearling, spruce grouse females were on average 303 ± 60 m from the nearest male, compared with 177 + 53 m for 11 adult females (from Herzog 1977a). Yearlings there may have had to pass near adult females to reach males, but adult females generally had at least one male nearby that could be reached without conspicuous encounters with other females. The three forest grouse vary in the distance the two sexes will travel for mating (Fig. 14.6). If the male ruffed grouse will not leave his safe advertising log, the female must go to him (Allen 1934, Barrett 1970). Thus, at least for ruffed grouse, it is possible that the female may not pick the spot of coitus. In contrast, female blue grouse may lead males considerable distances on the ground. Such females may have more control than the ruffed grouse female in selecting safe sites for breeding. Female forest grouse select adult rather than yearling males (Olpinski 1980, Jamieson 1982). There may be several reasons for this. First, because of inexperience and adult competition, yearling males may occupy sites that are not as safe as those of adults. Second, yearling courtship displays seem less perfected (Olpinski 1980), and yearlings may simply be less adept at attracting females. Third, male yearlings range widely and their locations are less predictable than those of adults. The female's selection of adult males in preference to yearlings in polygynous grouse species is the basic theoretical reason for the existence of nonbreeding yearling males (Wittenberger 1978). Female choice of males in the steppe grouse, which advertise at communal display grounds, is especially contentious (Bradbury & Gibson 1983). Does she choose the genetically fittest male and thereby produce superior offspring (Trivers 1972), or does she mostly select a mating site contested between males (Wiley 1973)? Hartzler (Chap. 7) and Wiley watched sage grouse at the same lek and reached different conclusions. Wiley (1973b) concluded that females selected mating sites rather than individual males, and that males competed for these sites. Hartzler and Jenni (Chap. 7), however, provide convincing documentation that females do not go to mating spots per se, but rather select males that strut the most
SURVIVAL AND BREEDING STRATEGIES
535
vigorously. In contrast, Kermott (1982) has concluded that female sharp-tailed grouse select mating spots. Hjorth (1970) diagrammed a sharp-tailed grouse lek watched by Brown (pers. comm.) where a dominant male was removed and three yearlings expanded their advertising stages to include his site. Females came to the former site and a yearling was successful in breeding. Again, however, others believe that the males themselves are chosen (Hamerstrom pers. comm., Emlen & Oring 1977, Sparling 1979). The 3-4 days or more that each female steppe grouse on average visits the lek suggests that females may be evaluating males (Wiley 1973b, Hamerstrom & Hamerstrom 1973, Robel & Ballard 1974, Hamerstrom 1980). Emlen and Oring (1977, Oring 1982) suggest that females patronize male clusters because it provides a forum for evaluating male quality, and this implies that male quality is an important determinant of female choice. It is important to decide if the selection of males by females is really a means of altering the genetic constitution of the offspring (Williams 1975, Harpending 1979). If this is true, then additional problems are those of resolving the lek paradox question (Borgia 1979), the "sexy-son" controversy (Weatherhead & Robertson 1979, Wittenberger 1981, Taylor & Williams 1982, Kirkpatrick 1985), and run-away selection (Fisher 1958, Lande 1981, O'Donald 1983, Arnold 1983). Most workers have documented that the more central males do most of the breeding (Scott 1942, Robel 1966, Wiley 1973); the Hamerstroms (1973) reported that 31 % of 555 copulations were by exterior males. A problem is how to define peripheral and central males; on a small lek all males are peripheral and no one is completely bounded by other males. The mean lek size in Wisconsin over a 22-year period was 8.3 males (n = 6,611) (cf. Hamerstrom & Hamerstrom 1973). With leks of this size one could then expect a ratio of interior to exterior cocks of 1:7, assuming similar configurations as in Rippin and Boag (1974b), Sparling (1979), and Robel (1966). Thus, based on their frequency in the population, exterior cocks should have performed approximately 88% of the breeding if there was no female choice. In fact, however, the Hamerstroms reported that only 31 % of the copulations were by exterior males. Thus the Hamerstroms' data also agree with data of other workers —i.e., that interior males breed more often than peripheral males. Observations of females are consistent with the view that females want undisturbed breeding. Evans (1961) reported fewer attacks between sharp-tailed grouse males when females were present. This might explain the daily hen peaks reported in prairie chickens (Hamerstrom & Hamerstrom 1973). Hens arriving ahead of time could wait for more hens and less disturbance. Females spend more days on large leks, and large leks are generally less stable in social climate (Hamerstrom & Hamerstrom 1955). Sage grouse hens are much smaller than males (1:1.8 ratio) and harassment of hens often occurs. These hens move through the lek as a pack and this should reduce harassment from males (Chap. 7). Oring (1982) felt that birds on leks were safe from predators, but that leks
536
A. T. BERGERUD AND M. W. GRATSON
were dangerous for females because of aggressive male behavior. He suggests that females avoid the lek centers at first, as well as other areas of overt aggressiveness, and that only after male relations are stable and aggressiveness subsides do the females move in to mate. This is consistent with Rebel's (1972) finding; after the removal of three dominant, male prairie chickens he reported aggression increased among resident males and successful copulations declined from an average (6 years) of 34 per year to three. Copulations are commonly interrupted by neighboring males (Evans 1961, Lumsden 1965, 1968, Hamerstrom & Hamerstrom 1973, Chap. 7). The Hamerstroms (1973) judged that based on behavior, 23% of 590 copulations by prairie chickens were unsuccessful. Mating was interrupted in 14% of 511 attempts by sage grouse (Chap. 7). Dominant cocks commonly knock subordinate males off females in coitus. By visiting the lek before the time when she is ready to copulate, a female can watch the activities of males and may learn where she will be least disturbed. By this hypothesis a female may benefit by using male dominance within his territory as a shield to reduce harassment. Although females can be expected to choose undisturbed breeding situations—a phenotypic advantage of certain males or particular spots—we can also expect that, in an ultimate and proximate sense, females choose males on the basis of other phenotypic traits. Our hypothesis is that females should select those males at leks that are least likely to interfere with the maintenance of an inconspicuous nesting site. This is an extension of the least-costly-male hypothesis of Wrangham (1980), which attempts to explain the evolution of the lek mating system. Favored males should be those which move around less than other males, move out to meet incoming females less often than neighbors, and follow females to a lesser degree than other cocks. Peripheral males commonly follow females from leks; central males remain more in place (Oring pers. comm., Hartzler 1972, de Vos 1979, Gratson unpubl. data). Female black cocks prefer to breed when there are male clusters (de Vos 1979, 1983). In Texas, prairie chicken hens successfully bred at leks that had central males—10 copulations in 86 observations (12%)—but rarely bred at linear leks that had no central males — 1 copulation in 76 sightings (1%) (Horkel & Silvy 1980). Males from linear leks commonly followed females away from leks (Horkel & Silvy 1980), and nesting success was low in that population (Chap. 15). Small, central territories may thus be used as a preliminary index of the tendency of males to "wait" for females rather than to seek or follow females into nesting habitats. Hartzler and Jenni (Chap. 7) elegantly documented that female sage grouse select males that strut the most frequently. Their findings are consistent with our stationary-male or waiting-male hypothesis, because the mobility of strutting males is reduced (Hjorth 1970). Such dominant males assured of future matings (hens even line up in sage grouse) should not be interested in following females. The stationary-male hypothesis we have briefly outlined here does not pre-
SURVIVAL AND BREEDING STRATEGIES
537
cisely predict how females choose, but which and why males are selected. The sampling strategies and cues females use may be many and varied, and include other females themselves (Gratson unpubl. data). In a system where males can gain no increased, relative fitness benefits from being dispersed (cf. Bradbury 1981), females should enforce the male dispersion pattern by exerting extreme choice. Perhaps the occasional copulation that occurs with peripheral lekking males reduces the tendency for these generally unsuccessful males to desert leks and seek females away from arenas in nesting cover. By initially spending more time with central males, females may force peripheral males to contest for sites toward the center of the lek (Rippin & Boag 1974b, Kermont 1982); the male dispersion is driven inward and the locations of leks are kept in place and prevented from expanding outward into nesting habitats, where males appear to impose costs on nesting females.
14.6 The strategy for improving chick survival After the female finally succeeds in hatching her clutch, she must continue her investment; 50% of the chicks commonly die between the time of hatching and their independence from the hen in August and September (Fig. 15.1). Tactics to improve the survival of progeny should include: (1) frequenting areas where young chicks can optimally forage, and thus maximize growth during this critical stage in their life history; (2) avoiding predators by remaining inconspicuous; and (3) defending the brood from predators, if necessary. 14.6.1 The trade-off: Optimal foraging versus antipredator tactics It is common for broods to move long distances in the first few hours and days of life (Fig. 14.14, Cebula 1966, Viers 1967, Silvy 1968, Schladweiler 1968, Barrett 1970, Schiller 1973, Haas 1974, Ramharter 1976, Maxson 1977, Sopuck 1979, Bakke 1980, Armleder 1980, Chap. 6). Long brood movements frequently take ptarmigan outside the territory defended by the male (Bergerud & Huxter 1969a, Erikstad 1978, 1985a, b, Steen & Unander 1985, Chap. 8). These long initial movements suggest that many nests are not placed in optimal locations with respect to food and cover for the chicks. Nests are generally in dense growth and shaded canopies, because they are more successful where concealed. Young chicks, however, require plant communities with abundant insects, warm temperatures, and a plant structure that provides concealment but does not hinder movement; but females should not compromise nest security by locating nests near such communities. Thus, where there is not a good interspersion of dense (nest) cover and open (brood) cover, young broods, of necessity, must move. Yearling forest grouse females commonly move farther from nest sites to
538
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.14. (-4) Prelaying movements of a yearling blue grouse hen suggest that she was searching for a nest site and narrowing her search area with time. (B) Prelaying range of an adult blue grouse hen indicates that she nested on the edge of her range. (C) Movements of a male and female willow ptarmigan with their brood. The birds left the territory defended by the male and made one long movement between two more localized ranges. (D) The long, initial movement of a yearling blue grouse hen with chicks after hatching suggests that the nest was not located near early brood habitat. (Adapted from Erikstad 1978, Sopuck 1979, Armleder 1980.)
brood ranges than do adults. This may also be true for sage grouse (Berry & Eng 1985). Yearling blue grouse females with broods traveled an average of 484 + 333 m (n = 16) after 1 week, whereas four adults went 205 + 81 m (Sopuck
SURVIVAL AND BREEDING STRATEGIES
539
1979, Armleder 1980). Some yearlings of ruifed and spruce grouse also make long movements (Herzog 1977a, Maxson 1977). These longer travels of yearlings may result in part from nest-site selection without consideration of where brood habitat is located, but probably another factor is the lack of knowledge about where brood cover is located once the nest has hatched. Further, yearling hens may have had to select inferior habitats to nest because of competition with adults in the spacing of nests. Sopuck (1979) showed seven yearling females nesting in second-growth forests; all would have had to take chicks long distances to locate more open habitat. Young broods commonly move faster and farther than older broods (Bump et al. 1947, Hungerford 1951, Kupa 1966, Godfrey 1975). Godfrey (1975) and Schiller (1973) recognized an early-brood interval, the first 2 weeks of life, and a late-brood period, the remainder of the summer. The earlier time includes the critical days when chicks cannot thermoregulate and when many die. This is a period when chicks require a high-protein diet of insects, and warm, open habitats to forage in without constant brooding (Erikstad & Spidso 1982, Erikstad & Andersen 1983, Jorgensen & Blix 1985). Brood ranges in the first 2 weeks are often larger than later ranges (Table 14.2). One hypothesis is that mobility is inverse to insect and forb abundance (Fig. 14.15). This has been documented for the Hungarian partridge (Perdixperdix) in England; broods move farther in areas with insecticides sprayed and less in those left unsprayed (Southwood & Gross 1969). Erikstad (1978, 1985a), working with willow ptarmigan in Norway, and Godfrey (1975) and Barrett (1970), using ruffed grouse, have also shown that movement of broods increases as insect abundance decreases. Movements may also be influenced by the dry ness of the season (Fig. 14.16, Giesen 1977). In dry years, forb and insect distribution may be more patchy, resulting in longer movements and increased concentrations of broods (cf. Wallestad 1971). An implication of this general hypothesis is that chick growth should vary between years, depending on the variability of insect abundance and warm weather, and that within years, broods that move the least should show faster growth than those which travel more in response to reduced insect abundance. Redfield (1978) found that blue grouse chicks had different growth rates in different years. Chicks that hatched late in the season when the weather was warmer and insects presumably common, had faster growth rates than early-hatched chicks (cf. Myrberget et al. 1977). Spruce grouse chicks also grew faster in a year of early phenology than in a late year (Quinn & Keppie 1981). Willow ptarmigan showed a negative correlation between growth and distance traveled in Norway (Erikstad, 1978, 1985a), and Mercer (1967) documented a similar sequence in Newfoundland. Growth of ptarmigan chicks was significantly slower in 1962 than in 3 other years, and this reduced growth was accompanied by increased movement (Fig. 15.35).
540
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.15. Model of the hypothesized role of the environment in the movement and conspicuousness of broods, which influence their risk of predation.
An alternative explanation for continuous and long-distance movements by broods is that these movements represent behavioral adaptations to avoid encounters with predators that bring prey to a temporarily fixed place, such as a nest, and return to the last capture location to hunt again (Sonerud 1985). Goshawks, red-tailed hawks (Buteojamaciensis), and gyrfalcons probably hunt in this manner. After encounters with humans, broods do move (e.g., Chap. 6), and chicks can be expected to be led away by the hen if she observes nearby raptors. As noted by Sonerud (1985), what is needed as a test of this hypothesis is a comparison of movements of broods provided with adequate insect supplies with those of broods with inadequate insect abundance, under the same predator conditions. If brood movements reduce the probability of detection by predators, broods with lots of food should move just as often and far as broods with less food.
SURVIVAL AND BREEDING STRATEGIES
541
Fig. 14.16. Summer ranges of blue grouse hens with broods on Moresby Island, British Columbia (Chap. 2). In a wet year, 1974, broods were more centrally located in dry, upland habitats than in 1975, when the weather was drier. Brood ranges of yearling females with chicks did not correspond with ranges the yearlings had traveled the year before, when they were chicks.
542
A. T. BERGERUD AND M. W. GRATSON
14.6.2 Inconspicuousness of females and broods Young chicks are especially vulnerable to predation before they can fly, and females should seek habitats with protective cover. But habitats used by broods are commonly more open than many nearby habitats. Small chicks can obviously find cover under much shorter vegetation than can mature females. The use of the open habitats, optimal for chicks but not for the larger female, thus results in a parent-offspring conflict (Trivers 1974). The conspicuous female must compromise her own safety so that her chicks can forage in warm, insect-rich habitats to maximize growth and hurry through the critical, first 2 weeks of life. Hens that lose their chicks generally seek denser cover than hens that still have broods (Haas 1974, Maxson 1978, Sopuck 1979). Ten of 11 radio-tagged, blue grouse females that lost their broods went to denser cover than did brood females (Sopuck 1979). Sopuck also found that 80% of 150 observations of broodless hens were in thickets, whereas only 24% of 133 observations of brood hens were in the same cover. Some unsuccessful, white-tailed ptarmigan and blue grouse females may actually leave the summer range and move to heavier cover if they lose their broods (Mussehl 1963, Giesen 1977, Sopuck 1979). The same pattern has been documented for ruffed (Maxson 1977) and spruce grouse (Ellison 1973, Haas 1974, Herzog 1977a). The increased conspicuousness of brood hens in chick habitat has its cost, these hens may often have a higher summer mortality rate than broodless hens (Christenson 1970, Maxson 1977, Table 15.7). One tactic that females could use to remain more inconspicuous is to avoid other hens and broods. The ranges of broods commonly overlap (Fig. 14.16), but for forest grouse this overlap is spatial rather than temporal (Keppie 1977, Maxson 1978). Although broods may use the same ranges to forage, they may still avoid each other, using the same areas but at different times. Keppie (1977a) found only 8 of 428 sightings of broods within 50 m of each other. Females with broods should elect to show mutual avoidance of other broods since aggressive interactions between females would increase the conspicuousness of chicks. An additional disadvantage of broods being in close proximity is that females risk the chance of mortality with distraction display for chicks that are not their own. On Brunette Island, Newfoundland, in 1962, when there were 20 broods per square kilometer (Mercer 1967), adults of up to three broods often responded simultaneously with distraction displays to distress calls from chicks that were not immediately nearby. Keppie (1977a) showed that mixed broods occurred mostly when the maternal hen died; 10 of 11 chicks that joined a new hen lived till autumn, but such mixed broods may decrease the survival of the hens's own progeny, especially if they are younger than the adoptees. Since female distraction behavior is related to chick age and number (cf. Andersson et al. 1980), such displays may be inappropriate for the size and number of a mixed brood. Also, an increase in brood
SURVIVAL AND BREEDING STRATEGIES
543
size and the unsynchronized stage of developing young should increase the conspicuousness of mixed broods. The escape tactics of maternal chicks of the adopting hen must also differ from the newcomers. This implies that mixed broods could decrease survival of the "adopting" hen's progeny, and should be avoided. Many authors have noted that hens with chicks move to denser and taller cover as the chicks mature (e.g., Ammann 1957, Bergerud & Huxter 1969, Barrett 1970, Christenson 1970, Haas 1974). Ruffed grouse commonly seek dense, lowland, alder-conifer habitats (Fig. 14.17, Grange 1948, Eng 1959, Schladweiler 1965, Kupa 1966, Barrett 1970, Maxson 1974, Pietz & Tester 1982), where large owls infrequently visit (Nicholls & Warner 1972). After the critical, early chick stage is passed, the female no longer must compromise her safety to allow optimal foraging by her brood, and she should then seek denser cover. If predators are generally not a problem in the summer—as, for example, for spruce grouse at Gorge Creek, Alberta (Keppie 1979, Boag et al. 1979) —females could remain in relatively open habitats. These habitats may have a greater variety of foods than those with more protective cover. The priorities should be that optimal foraging for chicks takes preference over safety for the hen in the early brood period, and that safety takes preference over feeding in the late brood period. The trade-off can be reduced by using areas with sufficient vertical stems for cover and with reduced horizontal cover so that light reaches food species at ground level (Frandsen 1980). Patchiness in microhabitats provides the best of both worlds, safety and nutrition. In all the grouse, females are smaller than males, with the possible exception of white-tailed ptarmigan (Johnsgard 1973, 1985). This dimorphism is accepted as being partly a result of intrasexual selection for increased size in males, but apart from this there could also be selection for reduced size in females. Small females would be less conspicuous if they visited effective cover when they had young broods. Effective cover is defined as cover that is sufficiently tall and dense to hide females bent over while feeding, but also short enough to permit the female clear visibility while standing (Mussehl 1963). Smaller females would have an advantage in survival compared with larger hens in short brood cover. The white-tailed ptarmigan female weighs approximately 320 g in July, and on high mountains these females use small rocks as cover (Weeden 1959b, Braun & Rogers 1971, Chap. 8). The rock ptarmigan is larger, at an average of 420 g, and frequents the short, dwarf-shrub communities (Weeden 1959b). Next in the continuum of both cover height and female height is the willow ptarmigan, at 450 g; she frequents the taller shrub communities (Weeden 1959b). This continuum in larger body size also occurs in willow ptarmigan as one goes from north to south (Salomonsen 1972). Salomonsen argues that smaller birds in the northern environment would be favored because of the harsh environment, but Bergman's rule would suggest the converse. Also, ptarmigan in the far north commonly migrate and leave their harsh winter ranges. May (1975)
544
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.17. Nests of ruffed grouse hens at Cloquet, Minnesota, were dispersed on upland sites, but brood locations were clumped in lowland situations, where there was probably more cover from predators (data from Kupa 1966). A similar pattern of dispersed early-brood locations (near nests) and clumped locations of older broods was described for blue grouse by Zwickel (1973).
SURVIVAL AND BREEDING STRATEGIES
545
documented that the small, 300 + g white-tailed ptarmigan is physiologically well adapted to its harsh mountainous environment. This size continuum both among and within species of ptarmigan may be the result of selection to match effective cover. The spruce grouse is the smallest of the forest grouse and inhabits the most dense, coniferous cover, with the shortest shrub and herbaceous cover. Ruffed grouse are slightly larger, and herbaceous vegetation in the more open deciduous forests is taller. The blue grouse is the largest and frequents open canopies with tall ferns and herbs in the brood season (Bendell & Elliot 1967). The sage grouse is the largest of the steppe grouse, and sagebrush is commonly 0.5 to 1 m tall. The prairie grouse do not seem to conform to the effective-cover model. Prairie chickens are slightly larger than sharptails, but often frequent shorter cover. The lesser prairie chicken is smaller than the greater, but its cover is taller (Jones 1963). However, the original range of the prairie chicken may have included the more eastern mesic, tallgrass prairies and brushlands, and that of the sharp-tailed grouse the more western and shortgrass savannas and plains (Johnsgard 1985). Thus, they too could have evolved a morphology to help minimize the parentoffspring conflict; i.e., the chick's need for low, open cover and the female's need for effective cover. A number of authors have reported the roosting behavior of broods. Before chicks can fly or thermoregulate, they must be brooded during the night. Broods presumably roost in the most secure habitats available. Older prairie chicken and sharp-tailed grouse broods may even roost in lowlands and wet areas, where cover is usually dense (Ramharter 1976, Chap. 6). At this stage the risks of predation if they were elsewhere may be greater than costs associated with cold and wet microenvironments. The broods of some of the open-dwelling grouse abandon the tactic of inconspicuousness in August and join gang broods. Gang broods have been documented in sharp-tailed grouse (Ammann 1957, Sisson 1976, Brown pers. comm.), prairie chickens, sage grouse (Wallestad 1975b), and white-tailed ptarmigan (Choate 1963a, Giesen 1977, Chap. 8). The brood creche occurs at the same time that distraction behavior by hens is curtailed. These gang broods represent the first use of the flocking tactic as an antipredator strategy (see section 14.8.2. The increased vigilance permitted by this tactic leads one to expect that it should occur mostly in open-habitat grouse. Gang broods should also occur only after the young are capable of strong and prolonged flight.
14.6.3 Reducing interactions with predators If a female with chicks encounters a predator, she might be expected to make a long movement to minimize the risk of another contact (see also section 14.6.2). Brood ranges may thus partly reflect the frequency and severity of such contacts with predators (Sonerud 1985, Table 14.2). A plot of brood sites often shows
546
A. T. BERGERUD AND M. W. GRATSON
clusters of points separated by long distances (Barrett 1970, Schiller 1973, Erikstad 1978, Fig. 14.14). A sharp-tailed grouse brood moved 28 km in the fifth week after hatch, after it had previously used a radius of only 1.0 km (Schiller 1973). One spruce grouse female with older chicks ranged over 52 ha, whereas eleven other females ranged over a mean area of 16.6 + 2.5 ha (Herzog 1977a). Svedarsky (Chap. 6) showed that prairie chicken broods remained in small areas for several days and then inexplicably shifted long distances. It seems improbable that foraging ranges should suddenly be abandoned if food was gradually diminishing. These large ranges and long-distance movements shown by some hens are probably tactics to avoid further encounters with a previously observed predator. Predators can be expected to return to sites where prey were located; to move seems a prudent strategy on the part of the hen (see also Sonerud 1985). The large brood ranges of female steppe grouse may be partly in response to the higher density, diversity, and mobility of steppe predators (Fig. 15.8). Prairie grouse raise large broods and often lose fewer chicks than forest grouse, despite the high, inherent predator risk associated with steppe environments. Mobility, though costly in energy, may be a tactic to reduce predation risk.
14.6.4 Defense of young against predation A female should be able to improve the survival prospects of her offspring by defending them against predators, but this entails some risk. Females should be prepared to risk more when the young are the most vulnerable, and when the brood is concentrated and all chicks are thus in danger (Andersson et al. 1980). Most young grouse die in the first 2 weeks of life. This is the period of extreme vulnerability, as they can neither fly nor thermoregulate, and they are often together, being brooded. Female defense should be greater in this early stage. Apparently the only data available are for blue grouse. Distraction displays are most intense during the first 2 weeks of life, and are less intense when there are fewer chicks together (Kristensen 1973, Fig. 14.18). Kristensen also showed that brood defense is less intense when the chicks are very young, 1 to 4 days old; during these first few days they are still receiving nourishment from the yolk sac, and should not be as vulnerable to windchill when they are separated. Intensity of defense by the hen should also vary in relation to danger to the young compared with danger to herself. Birds can probably distinguish among predator types (Simmons 1955, Curio 1975, Veen 1977). Killdeers (Charadrius vociferus) commonly fly at cattle and horses, who might trample the young but pose no danger to the hen. With more dangerous mammals, such as foxes, more impeded flights are shown (Simmons 1955). When willow ptarmigan, ruffed grouse, and blue grouse chicks are captured, and give distress calls, females commonly risk approaching humans to within 2 m; in these instances the danger to the young is extreme. If grouse could discriminate between naive and experienced
SURVIVAL AND BREEDING STRATEGIES
547
Fig. 14.18. Comparison of the intensity of distraction behavior among hens with chicks of different ages and among hens with different numbers of chicks. The two study areas are Comox Burn and Tsolum Main Road, Vancouver Island (Kristensen 1973).
548
A. T. BERGERUD AND M. W. GRATSON
predators, they could evaluate the risk and grade their responses. Grouse commonly "lead" if they are followed (Watson & Jenkins 1964, MacDonald 1970) and they may change their tactics depending on the predator's response. Ruffed grouse, blue grouse, and willow ptarmigan hens often run in front of humans with their tails fanned and flicking back and forth. White-tailed ptarmigan hens do not fan their tails (Chap. 8), and the two prairie grouse species most commonly remain hidden and run without the tail fanned. Grouse species that fan their tails all have black markings on the rectrices; this makes the birds much more conspicuous. The tail of ruffed grouse is especially long (a ratio of wing to tail of 1:0.8) and includes a black, subterminal band. Ruffed grouse, especially, fly with their tails maximally fanned. In contrast, the steppe grouse have relatively narrow tails, which are mostly brown and white-mottled. These markings probably add to the crypticness of the bird, rather than emphasize the tail. The black markings on the tails of some grouse likely act as deflectors (Fig. 14.19). There are common deflecting adaptations on the tails of several animal groups (Powell 1982, Caldwell 1982). An attacking predator is directed away from the body and to the less essential tail. Fanning and flicking of the black rectrices permit the female to increase the effectiveness of distraction, without increasing her own risk.
14.7 Brood disbandment and migration There appear to be two distinct movements by grouse in the fall: (1) the dispersal of brood members and (2) the movement to cover on winter ranges. The first movement may involve an assessment of breeding opportunities by juveniles, whereas the second, involving both juveniles and adults, is clearly survivaloriented.
14.7.1 Brood disbandment Members of ruffed grouse broods in Minnesota, sharp-tailed grouse broods in Wisconsin and Manitoba, and prairie chicken broods in Kansas initially disperse in September-October (Eng 1959, Godfrey & Marshall 1969, Caldwell 1976, Bowman & Rebel 1977, Chap. 5). Juvenile males leave the brood before juvenile females in ruffed grouse, and this is also suggested for sharp-tailed grouse and prairie chickens (Chap. 5). Willow ptarmigan broods in Newfoundland and rock ptarmigan broods in Scotland also disband in September (Watson 1965, Mercer 1967); juvenile males search for territories and some advertise. No one has provided evidence that the aggressiveness of the adult hen is a factor in brood disbandment; nor is this suggested by laboratory work (Alway & Boag 1979). In fact, the hen commonly leaves the brood before the juveniles disband and disperse (Chap. 5). In natural populations where fall dispersal has been observed, juvenile males had time to prospect for advertising sites between the attainment of sexual
SURVIVAL AND BREEDING STRATEGIES
549
Fig. 14.19. Willow and rock ptarmigan have partly black tails, which may act to deflect the attack of predators to less-essential tail extremities. White-tailed ptarmigan live on high mountains, above most avian predators, and in areas often with low cloud cover, where the white tail aids them in remaining cryptic.
550
A. T. BERGERUD AND M. W. GRATSON
maturity and the arrival of winter. Further, juvenile males could remain cryptic (brown body on brown substrates) while prospecting. Brood breakup and dispersal apparently does not occur in the fall in many other grouse populations. Failure to disband in the autumn in some species may result from late attainment of sexual maturity. Blue grouse juveniles do not reach full sexual maturity their first fall, and there is no brood disintegration; instead, in some areas they migrate to winter ranges with the adult hens (Lance 1967, Sopuck 1979), but in other locations females may leave the brood before migration (Hines 1986a). Apparently, once the period of rapid growth is over in July, it is advantageous for them to leave the open, logged, or burned habitats to return to the relatively safe, coniferous forests. Nor do sage grouse mature by autumn. Juveniles in broods join gang broods that later form the nucleus of winter flocks (Patterson 1952, Dalke et al. 1963, Wallestad 1975b). A second reason that some populations do not show brood disbandment is the simultaneous arrival of both snow and sexual maturity. The birds are brown against a white background, and cannot afford the risk of prospecting in open, unfamiliar habitats. This occurs in ptarmigan in northern or alpine environments (Giesen 1977, Chaps. 10, 11).
14.7.2. Fall migration During the breeding season both males and females expose themselves to risk in order to frequent relatively open areas to advertise or to nest and rear chicks. As winter approaches, the tactics of survival take precedence over those of reproduction—birds must "beat" the winter in order to be alive the following spring to breed again and repeat the investment cycle. A common explanation for fall migration of grouse is that birds move to places where their survival will be enhanced by a more plentiful food supply (e.g., Schorger 1944). An alternative explanation is that grouse migrate in order to find cover that will reduce predation risk (see also Chap. 10). Two environmental changes appear to correspond with fall migration—leaf fall and snow cover. Gratson (Chap. 5) showed that sharp-tailed grouse begin to move to winter cover at snowfall. Mossop (Chap. 10) watched willow ptarmigan throughout the winter as their ranges shifted. As the snow cover increased, the birds became more restless, shifting downslope to taller, emergent willows only after they had lost most of their cover. Each time they departed, ample willow buds for food still remained; in fact, as snow depth increased, it allowed the birds to reach more abundant bud supplies on the upper branches. Nonetheless, they moved. Eng (pers. comm.) reported that sage grouse migrate at the same time that snow appears, usually long before snow depth reduces the supply of available food. Some sage grouse populations are migratory and others more sedentary (Patterson 1952, Dalke et al. 1963), but all move to dense, tall sagebrush (Patterson 1952, Dalke et al. 1963, Eng & Schladweiler 1972, Beck 1977). These migra-
SURVIVAL AND BREEDING STRATEGIES
551
tory and nonmigratory populations occur because of differences in the nearness of sagebrush of sufficient height to conceal birds as snow depth increases (Fig. 14.20). Populations that are not migratory have tall sagebrush nearby; those populations that migrate have sagebrush lower in height, or sagebrush inundated with snow and therefore ineffective as cover. Even the terminal branches of sagebrush emerging from deep snow cover should provide ample food. Food availability appears to be an insufficient stimulus to explain migratory movements in most grouse species. White-tailed ptarmigan in Colorado migrate downslope and seek willows approximately 0.5 m in height (Braun & Schmidt 1971), and usually arrive with the
Fig. 14.20. Fall movements of grouse are generally to areas with denser "winter" cover. Birds in some populatons of blue grouse shift uphill when cover is reduced by snow on breeding ranges. Some sage grouse populations move downhill to find more cover. Ptarmigan generally move downhill to reach taller cover and brown backgrounds if they are still in pigmented plumage. In Iceland, ptarmigan may migrate uphill in the fall when they are in white plumage to find white backgrounds (Gardarsson 1971). In these fall and winter movements, males tend to remain closer to breeding ranges and use sparser cover than do females.
552
A. T. BERGERUD AND M. W. GRATSON
first major storm in September or early October (Braun et al. 1976). At this time birds are mostly in pigmented plumage. The first winter storms would not reduce their food supplies, especially on windswept, alpine slopes. These ptarmigan feed in the crepuscular period, appearing watchful and moving quickly across openings in search of denser, more centrally located vegetation. Mobility increases with reduced snow cover. At times, white-tailed ptarmigan even return to the breeding ranges in the winter if fair weather prevails, but only after they acquire their white, winter plumage. This movement would be away from more food in the valley, to less food in the alpine regions. Even more elucidating are the movements of rock ptarmigan in Iceland. They move to alpine habitats with less plant biomass than the areas downslope that they leave, and find white backgrounds that match their white plumage, despite the lesser food supplies in the higher elevations (Chap. 9). Blue grouse males on Vancouver Island migrate uphill as well, but in July, leaving open areas with a rich diversity of nutritious foods for a monotypic diet of needles in coniferous cover (King 1971, Sopuck 1979). Grouse that move long distances become partly segregated into male and female flocks. Invariably, males move to vegetation lower in cover value and in harsher climates, yet nearer their breeding ranges than females (Figs. 14.20, 14.21, 14.22, Chaps. 5, 10, Irving et al. 1967b). But even in these environments, with reduced food and cover in comparison to the locations of females, males maintain their fall weights as well as females (Fig. 16.10, 16.11). The prairie chicken apparently migrated in the 1800s (e.g. Schorger 1944). Because it is the most granivorous grouse, it may have trouble finding seeds in deep snow and is thus the most likely candidate to demonstrate food-induced migrations. But the migrations of prairie chickens also took place at the same time cover characteristics changed, owing to leaf fall and snow, and before the time when food was in short supply. During these migrations, females moved farther than males, exactly what we see in the other grouse species. In the past, sharp-tailed grouse have also made spectacular fall movements. In the autumns of 1892 and 1932 thousands moved out of the Hudson Bay lowlands in October and November (Snyder 1935). In both years, these populations were probably at cyclic highs (Keith 1963). No adequate explanation has been proposed for these emigrations (Snyder 1935, Hanson 1953, Keith 1963), but birds may have moved south to seek winter cover when large flocks first formed. Such enormous flocks would have been especially conspicuous in October and November after leaf fall, and when snow cover canceled their crypticness. The sex ratio of birds that reached the farthest south in the course of these migrations was 48 females to 17 males (from Snyder 1935). This segregation suggests a normal winter stimulus; males probably remained closer to the northern breeding grounds than females. Large areas of coniferous forest lie between the Hudson Bay lowlands and the birch and aspen forests north of Lake Superior.
SURVIVAL AND BREEDING STRATEGIES
553
Fig. 14.21. Relative length of movements between breeding and nonbreeding seasons. Adult males move the least, followed by juvenile males; adult females and juvenile females move the farthest. These differences between sexes and ages can be explained by: (1) juveniles must travel farther than adults in finding a suitable breeding site, because adults show philopatry and juveniles are inexperienced, and (2) males are less prepared than females to travel long distances from breeding ranges; females seek cover that is optimal for survival, even if they must move far away.
554
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.22. Flock sizes, movements, and vegetation cover sought by female (top) and male (bottom) sage grouse in two winters of different snow and cover conditions. (Adapted from Beck 1975, 1977.)
Once the sharptails left the lowlands, they would have had to travel long distances before finding a suitable food and cover combination. Possibly this massive emigration was a winter shift to cover, and became a migration because birds found unsuitable habitat and were forced to continue their journey.
SURVIVAL AND BREEDING STRATEGIES
555
14.7.3. Movements between the breeding seasons Investigators band adults and juveniles to learn of their movements and philopatry between seasons. Findings from these studies generally show a consistent pattern: adult males are the most sedentary; juvenile males and adult females move farther; and juvenile females travel the greatest distances between breeding seasons (Table 14.7, Fig. 14.21). This consistency can be explained by the different tactics of adults and juveniles and of males versus females. A major component of this movement is a result of searching for advertising and nesting sites (Fig. 14.21). Because adults of both sexes generally show philopatry, their searching movements are shorter than those of juveniles. Adult males move shorter distances than adult females because males are returning to last year's advertising sites, whereas females, though likely returning as well, may shift farther. This will be especially true if the previous year's nest site was unsuccessful and another must be chosen. Some juvenile males may have located their sites the previous fall, but others are making a final decision in the spring and may therefore move farther than adult females. Juvenile females probably accumulate little nest-location information from the previous fall. They should delay their decision until they can assess spring cover and possibly predator activity, and they must also avoid adult females. This results in large movements. Included in the total distance traveled between seasons is the distance a bird moved to find winter cover. Females generally move farther than males. The combination of searching for breeding sites and for winter cover results in the four different distances of adults versus juveniles, and females compared with males. In effect, there is an attraction between males because some yearling males located themselves near proven advertising sites; and there is a repulsion between females, who most commonly avoid each other when nesting. These tendencies, coupled with the tactic in the winter of males staying near advertising sites because of intrasexual competition, and the female's strategy of searching for optimal winter cover and food to prepare for next year's nesting season, provide the four distinct movement modes.
14.8 Strategies for winter survival The major problem for grouse in winter is to avoid predation. Many of the alternate prey of their predators have migrated south or live under snow. Forest and steppe grouse are conspicuous against their white background of snow. Deciduous leaves have fallen, shrubs are snowed in, and the ultimate predator of grouse—the goshawk—may range far and wide, from the prairie provinces of Canada south to Kansas. Commonly, 40% to 50% of the grouse die between August and the next spring, when they are counted (Chap. 15). Some populations are relatively secure over winter; for example, spruce grouse in Alberta do well in coniferous cover,
556
A. T. BERGERUD AND M. W. GRATSON
and white-tailed ptarmigan survive well in tall willows (Braun 1969, Keppie 1979). However, many other populations have less secure cover and/or more effective predators, and suffer high winter losses as a result; examples are ruffed grouse in Minnesota (Chap. 15), rock ptarmigan in Iceland (Chap. 9), and willow ptarmigan on Brunette Island, Newfoundland (Mercer 1967). The major tactic of grouse to escape detection by predators is to move to cover. There appears no exception to the general rule that birds are found in denser cover in winter than in the breeding season. Even in winter, males are generally in more open cover than are females and they travel shorter distances from breeding ranges to find winter cover (Table 14.7). Male sage grouse in one population traveled only 5 km in a mild winter and used cover 30 cm tall, whereas females in this population traveled 16 km to cover 36 cm tall (Beck 1977, Fig. 14.22). In a winter of deep snow, both males and females traveled farther to find cover, 11 km for males and 19 km for females. Yet males used cover that averaged only 19 cm, whereas females remained in cover averaging 24 cm (Fig. 14.22). In the three ptarmigan species, the male stays closer to the breeding range and in shorter cover than the female (Weeden 1964, Braun 1969, Chap. 10). Robel et al. (1972) found more sharptailed grouse males in one study area (Kadoka), which contained mostly low grass cover, and more female sharp-tailed grouse at another study area along the Missouri River, where grass and shrubs were taller. King (1971) traveled the open subalpine region in search of blue grouse. Breeding males were on ranges used in both summer and winter, but females had left to seek denser cover. The differences in cover requirements between males and females is probably the least for spruce and ruffed grouse, both of which travel only short distances between breeding and winter ranges. That males and females show different movements to winter cover, and that they often select cover of different heights and densities, should result in some segregation of the sexes in winter. Sexual segregation is most common in ptarmigan (Weeden 1964, Braun 1969, Chap. 10) and in some steppe grouse populations (Robel et al. 1972, Beck 1977). Spruce and ruffed grouse are the least segregated in winter, owing to restricted movement (cf. Doerr et al. 1974, Herzog 1977a).
14.8.1 Winter feeding tactics Generally, grouse do not have winter food problems. Spruce and blue grouse may spend days, even weeks, in the same conifer (King 1971, Herzog 1977a). There they can both forage on needles and roost in the safety of a concealing fir or pine. The overlap of cover and food is also complete, for example, for sage grouse, which feed on sagebrush, ptarmigan, which use willows and birches, and red grouse, which both forage and hide in heather (Calluna vulgaris). Speaking of the winter foods of grouse Doerr et al. (1974, p. 614) said: "These species [one
SURVIVAL AND BREEDING STRATEGIES
557
to two major foods] seem to have several common attributes; they are often abundant, easily available, can be fed on rapidly and efficiently [large leaves and buds], and are nutritious if consumed in large quantities." There may be constraints on which plants or plant parts may be taken, such as the level of antiherbivore resins produced by some shrub or tree species, certain ages, or parts (Bryant & Kuropat 1980). But in general, as Bryant and Kuropat (1980) have concluded about subarctic grouse (ruffed, spruce, and ptarmigan) and their foods, the browse species utilized by grouse are abundant in space and time, and grouse reside in and are adapted to, nutritionally and physiologically, these habitats. The conspicuousness of dark-bodied grouse increases abruptly with leaf fall and snow cover. In summer and early fall their brown bodies are cryptic against a brown background—but in a matter of days, or after one heavy snowfall, they are in bold contrast to their physical environment. To minimize conspicuousness they must change feeding tactics abruptly; and if they must remain and be conspicuous, they seek heights from which they may watch for predators. Gratson (Chap. 5) reported that sharp-tailed grouse formed larger flocks after snowfall, became more mobile, and visited taller and more varied food sources (see also Marshall & Jensen 1937). In Europe, capercaillie and black grouse also feed less on the ground and more in trees just after snowfall (Koskimies 1957). Keppie (1977b) showed that a significant increase in feeding in trees by spruce grouse occurred only after the first snowfall (Fig. 14.23). Another example is ruffed grouse, which began to feed in aspen trees on 5 December 1968 at Rochester, Alberta—the same day as the first heavy snowfall. The next year, the first snow was 6 weeks earlier, on 22 October; and on 24 October the birds were again in trees budding on aspen (Doerr et al. 1974). Grouse subject to heavy predation pressure limit their feeding to short bouts during crepuscular periods when fewer avian predators are about. Mossop (Chap. 10) documented how ptarmigan emerged from their snow burrows in the evening and fed vigorously, then returned to snow cover. The sequence was repeated in the morning just before there was enough light to see. Braun and Schmidt (1971) watched white-tailed ptarmigan feed from 1730 to 1900 hours. They normally engaged in intense feeding activity. No longer cautious, they would run or fly from one Salix bush to another. Their feeding appeared almost desperate and individual birds were observed "standing on their toes" or jumping off the ground to reach buds overhead. At such times they exhibited little fear or alarm. In contrast, in the middle of the day birds were mostly sedentary, and when they were active they moved slowly and cautiously, with their bodies positioned low to the ground. Braun and Schmidt (1971) called this "inconspicuous creeping." The feeding schedule of ruffed grouse appears more complicated than that of ptarmigan. Ruffed grouse live in forests without clear views, and are faced with an effective nocturnal predator—the great horned owl, and an especially effective diurnal predator as well —the goshawk, an early riser. Ruffed grouse also feed
558
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.23. Spruce grouse quickly shift from ground locations to tree cover when snow cover becomes complete and they are no longer cryptic against white backgrounds. (Adapted from Keppie 1977b.)
for short periods of from 15 to 30 minutes in early morning and late evening (Doerr et al. 1974, Chap. 4). A ruffed grouse probably needs less light intensity to relocate a large, vegetable food source than does a goshawk to pinpoint a mobile grouse in a new location, and it is within this brief margin of time that the grouse makes its move to feed. Ruffed grouse feed in aspens later in the evening in winters with little snow cover for roosting (Chap. 4). This would allow them to leave with full crops, and the full crops would permit them to feed less the following morning when predator risk is high. It may also take the birds longer to fill their crops if they must remain more alert because of the lack of snow cover. Huempfner and Tester (Chap. 4) also noted that birds usually began to feed in the lower branches of the aspen, and that the rate of moving higher in the tree varied between years with different snow cover (Fig. 14.24). Blue and spruce grouse generally do not have the winter predation problems of ruffed grouse. These two grouse live in dense conifer forests and eat needles. Their lower, annual mortality rates reflect this cover advantage; on average only 35 % die over winter compared with 55 % of the ruffed grouse (Chap. 15). Ellison (1971) reported that spruce grouse did not select spruce trees to feed in for either
SURVIVAL AND BREEDING STRATEGIES
559
Fig. 14.24. Winter feeding behavior of ruffed grouse in aspens, based on one hypothesis of the findings presented in Chapter 4. In AM (morning) feeding bouts, birds remain higher in the tree in a winter with good snow (year ]) than in a winter without "escape" snow (year 2). In winters of good snow, birds in PM feeding bouts feed high in the tree, but depart sooner because snow-roosting helps them to conserve stored food and because they can feed in relative safety the next AM, owing to the nearness of escape snow. In a winter of little snow, they feed longer in the PM because it is a relatively safe period and they can leave with a full crop to reduce the length of feeding the next morning, when they will be more vulnerable to predators because of inadequate snow cover.
560
A. T. BERGERUD AND M. W. GRATSON
crown density or stand density. Birds are sometimes conspicuous on branches, but they can quickly move into the center of the tree where the tight whorls of branches provide concealment and obstruction when predators are near. Spruce grouse also commonly use the trunk of the trees for loafing and for roosting cover (Ellison 1971). It is often argued that aspen, because of its nutritional qualities, is essential to ruffed grouse (Gullion 1970a,b,c, 1977a,b, 1982). Yet many ruffed grouse populations live outside the range of aspen. In southwest British Columbia there is no aspen, and ruffed grouse live in alders (Alnus spp.). In some winters in Minnesota, ruffed grouse feed prodigiously on aspen, in others they feed mostly on other species (Svoboda & Gullion 1972, Chap. 4). In the winter of 1968-69, ruffed grouse fed primarily on beaked hazel (Corylus cornuta), as aspen production was low; nonetheless, the grouse population increased in 1969. Also, production of aspen buds was the highest in eight years in 1964-65, yet the spring grouse population declined in 1965 (Gullion 1970a, 1981, Svoboda & Gullion 1972). Ruffed grouse should use aspen when they can combine optimal foraging tactics with reduced predation risk. Birds can fill their crops quickly on aspen, since buds are larger, and then retreat to cover. This is perhaps its greatest advantage. Aspen grows in clones, and birds feeding on the periphery may be provided with more visibility than those feeding in a more uniform birch forest. Doerr et al. (1974) failed to show that the aspen stands selected by ruffed grouse had buds of higher nutritional quality than stands that were not utilized. Even the findings of Bryant and Kuropat (1980), which revealed a close association between the preferred foods of ruffed grouse and the lowered resin concentrations of those foods, have shown that selective pressures other than simply nutritional and energy demands are operating in the grouse-forage relationship. Aspen also provides sufficient height for birds to dive into snow below for escape (Fig. 14.24). Birds do not use aspen if the trees are lower than the surrounding canopy. Especially interesting is the finding that grouse stopped using aspen for 3 weeks after a sleet storm in Minnesota (Svoboda & Gullion 1972). Svoboda and Gullion felt that the grouse could walk on the snow crust, although against the white snow they would be quite conspicuous. The crust probably prevented them from using the snow as escape and roosting cover and the aspen was no longer as safe as being elsewhere. Huempfner and Tester (Fig. 4.1) showed earlier that use of aspen by ruffed grouse was much higher in the winter of 1971-72, a winter of good escape snow, than it was in two other winters when the snow cover was inadequate for escape. The exclusive use of aspen appears to depend on the presence of two conditions: goshawks and sufficient snow for escape and roosting cover. Aspen's primary benefits may be that it allows short feeding bouts and its structural characteristics permit its use as an antipredator tactic.
SURVIVAL AND BREEDING STRATEGIES
561
14.8.2 Winter flocking Flocking by birds is generally explained as an adaptation to gain information on food distribution, or as a tactic to increase the speed at which predators may be detected. In the latter hypothesis, presumably the survival advantage of early warning outweighs the flocks' conspicuousness. Flock size or the proportion of flocked birds would thus vary with the need for vigilance; flocks should be larger or more birds should be in flocks in habitats where birds are more conspicuous and must react early to the approach of predators. If flocking is food-related, flock size should generally increase with the difficulty of finding food. Flocking in grouse appears to represent a need for mutual vigilance rather than for finding patchy food supplies. Grouse feed in the winter on ubiquitous sagebrush, spruce, pine and fir needles, willow, birch, and aspen buds, all of which are abundant and can be found in large, extensive patches. Variability in flock size changes not with the availability of food but with the conspicuousness of the birds. Large flocks are characteristic of the steppe and tundra grouse which forage in open habitats where they would benefit from mutual vigilance. Open habitats allow birds to flush at long distances from predators and escape by flight. The forest grouse commonly form only small, temporary groups of two to four birds (Ellison 1972, Doerr et al. 1974). Hines (1986a) found 43% of the blue grouse in winter alone and 25% in duos (n = 1,068). Such small flocks would allow the birds to remain relatively inconspicuous, and to use the tactics of crouching and avoiding detection, rather than flight, which may be less effective as an escape tactic at shorter detection distances (cf. Chap. 4). Flock sizes of sharp-tailed grouse, which commonly feed in trees and shrubs in winter, decrease when birds can frequent cryptic backgrounds, or when deep snow permits the birds concealment in snow burrows (Chap. 5). Prairie chickens formed tighter packs as winter progressed and cover diminished (Hamerstrom & Hamerstrom 1949). Flocks are largest when birds are most conspicuous, after snowfall but before deep snow allows snow-burrowing. The flocks of sage grouse, which feed on the ground and use sagebrush for cover, are larger in winters with less snow, and correspondingly more cover and food, than in winters when snow is deep and cover is more restricted (Beck 1977, Fig. 14.22). The differences in flock sizes between open-land and forest grouse—cryptic and noncryptic grouse—are most consistent with the hypothesis that the flocking habit and flock size represents a trade-off between the benefit of mutual detection and the cost of increased conspicuousness. Koskimies (1957) stated that the first line of defense in black grouse was to be cryptic, to crouch and to escape detection, and that flight to escape was the second line of defense. This may not apply to sharp-tailed grouse and prairie chicken on top of the snow, because they usually cannot avoid detection in large flocks and may flush while predators are far off (Brown 1966a). Because they are
562
A. T. BERGERUD AND M. W. GRATSON
not totally synchronous in feeding, some birds are usually alert while the others feed. Prairie grouse rely on a long lead-time to reach cover; the predator should know it has been detected and that pursuit is usually unprofitable (Baker & Parker 1979). The crouch-or-fly decision for all grouse should involve an assessment of: (1) the seriousness of the predator threat, (2) the escape options, and (3) the predictability of detection (Fig. 14.25). Grouse are able to distinguish among predators; for example, prairie grouse commonly crouch at leks when marsh hawks (Circus cyaneus) appear, and flush "hard" at the approach of goshawks and peregrine falcons (Falco peregrinus) (Berger et al. 1963, Sparling & Svedarsky 1978). Several possible costs must be weighed against the benefits of flushing; these may include increased conspicuousness and flight energetics, among others. The crouch-or-fly decision may also be influenced by intrinsic differences in approachability; blue grouse and ruffed grouse, for example, may be polymorphic in their approachability (Chaps. 1, 2, 3). Grouse at cyclic highs may be more reluctant to flush than when the populations are low and increasing (Keith 1963, Chap. 4). The means of escape should vary with the availability of escape habitat. Koskimies (1957) distinguished between a "cover-taking response," in which birds fly low and for short distances if they are cryptic against the background and cover is nearby, and an "escape by flight," in which birds fly long distances at high elevations. If ruffed grouse can plunge into the snow, they should take the first option. Mossop (pers. comm.) has watched rock ptarmigan attempt to fly above approaching raptors. He has also observed willow ptarmigan flush at the approach of a helicopter; the birds flew low and downhill and headed for taller vegetation, where the female took cover first. In Iceland, rock ptarmigan used a fence in the local village to avoid gyrfalcons (Chap. 9). Willow ptarmigan generally fly low and duck into the most available cover. The ptarmigan's black rectrices may provide a final strategy, which can result in a strike at the tail and some missing rectrices—but the bird often escapes (Fig. 14.19). The escape options of grouse may be restricted if they face a variety of predators. In Newfoundland, willow ptarmigan flushed at farther distances in foggy weather, when they were relatively safe from raptors. The crouch-or-fly decision under these conditions could thus be directed solely toward the approaching ground disturbance. To assist in the capture of blue grouse females on Moresby Island, British Columbia, it often helped to imitate the call of a red-tailed hawk. When the females heard the call they would crouch and freeze, allowing us a closer approach so that we could "noose" them. Mossop's observations of ptarmigan in snow burrows successfully bursting from their roosts to avoid a fox, only to be captured by a gyrfalcon (which was probably watching the fox), illustrate the problems that arise when a normally effective tactic is used against odd combinations of predators. Ptarmigan at Chilkat Pass would sometimes not flush from
SURVIVAL AND BREEDING STRATEGIES
563
Fig. 14.25. Grouse should consider the seriousness of the predator threat and the probability of being detected in comparing the benefits and costs of the option of crouching and the option of flying when predators are detected. There are also intrinsic differences among individual grouse in the tendency to flush Chaps. 1, 2, 3).
willow cover because of possible gyrfalcon predation, but this permitted a golden eagle to land and walk into the willow and snatch a grouse. Is there any longtime grouse biologist who has not seen a goshawk mysteriously appear when he or she had captured or flushed a grouse (e.g., Gratson 1981)? Grouse must compromise
564
A. T. BERGERUD AND M. W. GRATSON
escape tactics when faced with a number of predators with varying hunting repertoires. Flocking may not only increase vigilance, but it may also spread the risk among flock members and help to confuse predators as well. In Iceland, large flocks of ptarmigan may fly in close formation and wheel together, possibly hindering the stoop of gyrfalcons (cf. Chap. 9).
14.8.3 Winter roosting tactics Grouse spend the long, winter nights roosting. If they are detected, they may be in trouble with nocturnal predators. Many daylight hours are also spent in nonfeeding activities, at which time inconspicuous behavior should be a high priority and should not be compromised by moving about. Roosting places should be selected in areas not searched by predators and where there are good opportunities for escape if detected. Grouse commonly burrow under snow cover in order to roost. One explanation for snow-burrowing is that it provides a warmer microclimate, so that heat energy may be conserved (Gullion 1970a, Hoglund 1980), although Grange (1948) believed it served an antipredator function. Although the snow certainly conserves heat, perhaps a burrow's greater benefit is as cover to avoid predators. By insulating the bird, snow cover makes it possible for a grouse to stay in the burrow for many hours, slowly digesting the food gathered during brief feeding bouts and stored in the crop. Huempfner et al. (pers. comm.) found that ruffed grouse that used snow burrows were able to restrict feedings to only two daily periods, during crepuscular hours. Otherwise they remained under cover and avoided goshawks. When they were unable to snow roost, they fed at least three times daily, and the third bout occurred in the middle of the day when goshawks were hunting. Bump et al. (1947) provided anecdotal evidence that predators, specifically humans and foxes, could capture ruffed grouse in snow burrows. But Huempfner (pers. comm.) examined 1,100 snow burrows of ruffed grouse and could not find a single burrow that showed signs of a successful predator attack. Mossop followed a partly tame fox at Chilkat Pass as it hunted; the fox never succeeded in capturing a ptarmigan in a burrow. Invariably, as the fox became too close to a burrow, first one and then all the ptarmigan would burst forth and escape. Ruffed grouse apparently do not need to snow-burrow to survive cold weather. Our evidence for this is that the mortality rate of ruffed grouse was not correlated with the suitability of snow for roosting in an 8-year study in Alberta, or for a 13-year period in Minnesota (Gullion & Marshall 1968, Little 1978, Rusch et al. 1984). The average overwinter mortality rate of ruffed grouse males during two winters in Minnesota, during which there was little snow for roosting (1958-59 and 1960-61; Fig. 14.26) was 55% (n = 95) (Gullion & Marshall 1968). In con-
SURVIVAL AND BREEDING STRATEGIES
565
trast, the mortality rate estimated from six other winters in which there was adequate snow for roosting was 53% (n = 483) (Gullion & Marshall 1968). Ruffed grouse actually showed a weight gain in spring 1961, after a winter without snow (Fig. 14.26). Above-average mortality of ruffed grouse occurred in Minnesota in three winters-1963-64, 1972-73, and 1973-74 (Fig. 14.26). In these three winters, the numbers of days of snow at least 20 cm in depth (adequate for burrowing) were only 40, 61, and 67 days, respectively, whereas the mean for a 13year period was 78 days. These 3 years were also those in which goshawks came south to Minnesota (Mueller et al. 1977). Grouse, conspicuous against their white backgrounds, were forced to feed between the relatively safe, crepuscular periods. These birds did not die from physiological or nutritional stress — spring weights were not low (Fig. 14.26). Rather, they died because they foraged in the daytime, at a time when goshawks were abundant, and when snow was frequently not deep enough to permit escape dives. If snow-burrowing is an antipredator strategy, the blue and spruce grouse should snow-burrow less frequently than other grouse, because they can stay in the cover of conifers. Ruffed grouse should also seek dense conifer cover for roosting when snow is not available. Some blue grouse at Mt. Washington, British Columbia, occasionally spend several days in conifer forests without snowroosting (King 1971); however, some do burrow. The alpine forest at Mt. Washington is open, and all locations may not provide cover. Furthermore, some grouse may have to move to snow as an escape tactic. Apparently, spruce grouse in New Brunswick, Alberta, and Montana, and on the Kenai Peninsula, Alaska, usually do not snow-roost (Stoneberg 1967, Ellison 1971, Hedberg 1980). The winter range in Alberta and Montana, in particular, is comprised of dense forest stands (Stoneberg 1967, Boag et al. 1979). On the Kenai Peninsula, the snow is often crusted, preventing snow-burrowing (Ellison 1971). Spruce grouse snowburrow in Michigan, where they feed in more-open stands of jack pine (Pinus banksiand) (Robinson 1980), and where the sparsely branched and scattered trees may not protect them at night from horned owls. Ruffed grouse usually roost in dense conifers near available food supplies (Grange 1948, Barrett 1970, Doerr et al. 1974). At Cedar Creek, Minnesota, ruffed grouse roosted at the base of cedars (Thuja occidentalis) when snow was not available, and were occasionally captured by horned owls (Huempfner pers. comm.). Thus there are a number of examples of grouse roosting successfully in conifers in cold regions. Grouse often use snow burrows or snow depressions in relatively warm weather, and in the daytime. Rock ptarmigan have used snow for cover even in the summer. The birds cover their bodies with snow and leave their heads exposed. This has been reported for rock ptarmigan (Watson 1972), willow ptarmigan (Mossop pers. comm., Chap. 10), ruffed grouse (Bump et al. 1947), sage grouse (Patterson 1952), sharp-tailed grouse (Chap. 5), and prairie chickens (Hamerstrom et al. 1957). These burrows or depressions are not needed for heat
566
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.26. Annual mortality rates of ruffed grouse males in relation to densities, spring weights, temperatures, and days suitable for snow-roosting. (Grouse data from Gullion & Marshall 1968, Little 1978; weather statistics from Cloquet and Pine River, Minnesota; goshawk data from Hofslund 1973, Mueller et al. 1977.)
SURVIVAL AND BREEDING STRATEGIES
567
conservation; rather, the birds have camouflaged their brown bodies and can now watch for predators. Snow burrows should be located where predators are seldom present. Rock ptarmigan burrow on steep slopes next to rocks (Watson 1972); sharp-tailed grouse prefer to roost in marshes, where there may be fewer alternative prey for fox (Chap. 5). Other antipredator behaviors also indicate that snow is used for cover. Sharp-tailed grouse often tunnel far from the snow opening, making it difficult for a fox to know where to pounce (Chap. 5). Burrows are also spaced apart, and once one bird flushes the others can also flush away from the intruder (Chap. 5, 10). Snow-roosting may at times allow birds to remain close to food supplies, thereby reducing mobility and conspicuousness. Ruffed grouse stay predominantly in deciduous forests where they can both snow-roost (Grange 1948) and feed at nearby aspen (Svoboda & Gullion 1972, Chap. 4). The snow provides cover when not foraging and also escape cover while feeding in aspen. That some populations do not use suitable snow for roosting, even in cold weather, is not consistent with the thermoinsulation hypothesis. Cold—the hypothesized cause—is present, but the expected effect—snow-roosting—is not. Birds may roost in trees instead. Cold is not, therefore, a sufficient cause for snow-roosting.
14.9 Polymorphic spacing strategy In an ideal world, a bird would select a habitat type for nesting and advertising on the basis of its inherent fitness prospects (Levins 1968). However, the idealfree world would not consider competition. Fitness should decline within habitats as density increases (Fig. 14.27). One can visualize two options: (1) A bird can settle in the intrinsically best habitat; but this habitat will also be selected by others, thus detracting from its utility. (2) A bird can settle in lower-quality habitats, where there will be fewer good sites but also fewer competitors. The trade-off is that the quality of the best habitat is reduced by competition for nesting and advertising sites, and the increased conspicuousness of the birds to predators. In the second option birds will be farther apart, less conspicuous, and may face fewer predators. The evolutionary advantage of spacing is never without qualification —dispersion and aggregation confer conflicting advantages and disadvantages (Waser & Wiley 1979). Grouse do have density options. Females space themselves relative to nesting sites (Figs. 13.7, 14.12); the hen's territory is spatiotemporal (Wilson 1975), and her defense space travels with her. Male forest grouse defend advertising sites and, as with females, there are no distinct boundaries with adjacent birds; rather, there are areas of little use and no clear ownership (Fig. 14.7) (Zwickel & Bendell 1972, Archibald 1975, Herzog & Boag 1978). Even in the vigorously defended
568
A. T. BERGERUD AND M. W. GRATSON
Fig. 14.27. Two behavior morphs are density tolerant and density intolerant. (A) When a population is low, the density tolerant will be clumped in the best cover. (B) When the population increases, the habitat will be more uniformly occupied. (C) Bird A, reared in good cover, is intrinsically more density intolerant than its brood mate B. Bird A prospects for a breeding location in a lower density-less cover situation (A\ loA2). The fitness of B improved (Bi to 82) with the departure of A. If the density is reduced in the best cover by a removal experiment before A is committed to a breeding site, the bird may shift back (time 3) to the better cover to improve fitness. (D) The space between both morphs decreases as the population increases (Fig. 2.17).
territories of ptarmigan, the size of the male's territory is correlated with the number of male competitors (Fig. 15.24, Bergerud et al. 1985). Thus, new birds can squeeze in. In steppe grouse, males can join the perimeter of established leks or
SURVIVAL AND BREEDING STRATEGIES
569
move elsewhere. It is the bird's option to select the site for maximum fitness. Avoidance of high-density situations is an alternative to competition. Genetically influenced and determined, polymorphic strategies in spacing behavior could occur in heterogeneous environments when the resources are distributed, coarse grain, and the birds settle in patches (Levins 1968, Gadgil 1971, Gillespie 1974, Hedrick et al. 1976, Wiens 1976, Austad 1984). Polymorphic spacing behavior has been shown for fish, mice, toads, frogs, and others (Williams 1966a, French et al. 1968, Krebs et al. 1976, Wells 1977, Davies & Halliday 1979). Alternative reproductive behaviors and evolutionarily stable strategies (ESS) are currently very active areas of research in other animal groups (see American Zoologist 24, p. 307). Grouse settle in patches, and patch sizes and quality appear hierarchical. When prairie chicken populations declined in Wisconsin, the reduction was most apparent in marginal habitats (Hamerstrom et al. 1957). When the willow ptarmigan population in Newfoundland reached a cyclic high, densities were comparatively homogeneous; the second-best hunting barrens, Portugal Cove, had numbers comparable to the prime habitat at St. Shotts (Bergerud 1970a). But at the cyclic low, the Portugal Cove population declined more than that at St. Shotts. Ruffed grouse populations in optimal habitat in Ontario showed little population change through a provincewide cycle (Theberge & Gauthier 1982), and ruffed grouse densities at Watch Lake, British Columbia, were more stable in a small study area than in a wide area (Chap. 3, Fig. 3.3). Another example is blue grouse, which were patchily distributed on Moresby Island, British Columbia, in 1971 when there were only 26 males, and more uniformly distributed in 1976 when there were 42 males that selected advertising sites (Chap. 2). A number of workers have removed territorial birds in optimum habitats (Jenkins et al. 1964, Bendell et al. 1972, Zwickel 1972, Hannon 1978, 1983). Generally, birds from other areas arrived to repopulate the high-quality habitats (Fig. 14.27). Although we would not predict such replacements in cyclic populations starting to increase (Chap. 15), these replacement birds verify that birds are dispersed in patches and suggest that they can improve their fitness by exercising particular density options. We can recognize at least two morphs: (1) a highdensity, tolerant morph, which is adapted to squeeze into habitats with abundant resources (more nest and advertisement sites per area) and adapted to living in close proximity with others and (2) a low-density, intolerant morph, more inclined to seek low-density situations where resources are more dispersed. These morphs may be analogous to Chitty's (1967) two types: n selected and a selected animals (review in Krebs 1978). Our knowledge of, phenotypic polymorphism in spacing strategies is most advanced in forest grouse. All three forest grouse—ruffed, spruce, and blue—live in habitats liable to forest succession. Blue grouse colonize newly logged, coniferous forest sites (Redfield et al. 1970) or following fire successions (Mar-
570
A. T. BERGERUD AND M. W. GRATSON
tinka 1972, Bendell 1974). Spruce grouse are frequent in early successional stages of lodgepole and jack pine forests (Szuba & Bendell 1982). The boreal forest occupied by spruce grouse is a fire-driven ecosystem (Kelsall et al. 1977). Advertising, ruffed grouse males find young aspen stands safe as soon as the canopy closes (DeStefano & Rusch 1982). Aspen is another species of early successional stages and is adapted to regenerate following forest fires (cf. Doerr et al. 1971). The temporal favorableness of these seres for grouse lasts longer than the birds' generation time. A genetically based, polymorphic spacing system would allow birds to persist in older successional stages, and at the same time permit their progeny to pioneer the newly created, unoccupied habitats (Gadgil 1971). This system would be the result of individual fitness strategies. The pioneering work on polymorphism in forest grouse was Mossop's study of the behavior of birds at three locations — Copper Canyon (CC), Comox Burn (CB), Middle Quinsam (MQ)—on Vancouver Island (Chap. 1). He showed that birds in a population at high densities showed behavior that reduced their conspicuousness to predators. Birds had long flushing distances, were nonaggressive, and were seldom seen on the ground. A population at low density (MQ) was comprised of birds that were more mobile, aggressive, and more conspicuous in spacing behavior. Comox Burn birds were intermediate. Bergerud captured birds from these three stocks and released them on three islands: Moresby Island received birds from all three stocks (CC, CB, MQ); Portland Island received founders from CC: Stuart received birds from MQ (see also Chap. 2). To review briefly the conclusions of this experiment: the founders retained their distinct behaviors throughout their lifetimes in these new and very different habitats on the islands. Also, progeny raised on Stuart acted like their (originally) MQ parents. The progeny on Portland behaved like the parent stock from CC for half the birds, but the other half was much more conspicuous and aggressive, like MQ stock. The conspicuous MQ and the inconspicuous CC types of progeny both occurred in the population on Moresby, as well as an intermediate type. These behavioral types of progeny occurred in ratios that suggested a heritable basis. The more constant behavior of MQ progeny versus CC progeny suggests that CC stock had more underlying genetic variations than did MQ. These high- and low-density morphs have different dispersal patterns. Males and females of the low-density phenotype were mobile on Moresby and sought low-density habitats on the island's edge. Birds of the high-density phenotype settled near each other and had small advertising and preincubation ranges. Both types of birds were present in the closed system on Portland Island. But in the open populations on the mainland, where Bergerud had originally captured the birds, he could not find both types near each other. At Copper Canyon, where the founders of Portland Island were captured, he found only the high-density type. We believe that the low-density type in this open system had dispersed away from the high-density sites occupied by their parents, similar to what occurred
SURVIVAL AND BREEDING STRATEGIES
571
on Moresby Island, with mixed stock. On Vancouver Island Bergerud (pers. files) found a high-density population at the top of a mountain near Cowichan. These birds were closely spaced, flushed at long distances, and could not easily be captured. Surrounding these birds, but downhill, were birds spaced farther apart and in denser vegetation cover. In contrast to the hilltop birds these could be approached closely and would come to arena tests and fight their mirror images. Bergerud also found at Mt. Washington, Vancouver Island, widely spaced birds that would fight their mirror images, this time on the subalpine habitat (pers. files). Hannon (1978), working at Comox Burn immediately beneath Mt. Washington, found blue grouse more closely spaced and less aggressive than the widely dispersed birds on Mt. Washington. Hand-reared progeny from lowdensity, blue grouse populations fought their mirror images more, and were more mobile pacers than chicks raised from parents who lived in higher-density situations (Chap. 2, Cooper 1977). Hines (1986b) recently radio-tracked blue grouse and classified birds as dispersers and nondispersers. Both types had similar reproductive success and mortality rates. There is ample evidence of suites of tactics grouped within alternative reproductive behaviors (ARBs sensu Austad 1984), which permit blue grouse to select either high- or low-density situations. The genetics at one locus (Ng) of colonizing blue grouse have been documented in two populations (Redfield 1974, Zwickel et al. 1977). At Alberni, Vancouver Island, Redfield (1974) found an excess of homozygous, yearling blue grouse settling in vacant, newly logged areas. At Comox, Vancouver Island, Zwickel et al. (1977) removed birds from second-growth forests 9 years after logging. At Comox, the frequency of Redfield's (1974) colonizing genotype was similar for the removed birds and the colonizing replacements. Although densities were essentially zero at both Alberni and Comox after removal, homozygotes could adapt to the reduced cover and extensive space at Alberni, whereas heterozygotes should have sought in the more optimal and advanced plant cover at Comox. These results are consistent with the spacing hypothesis (Fig. 2.26) outlined earlier to explain the existence of blue grouse types on the Gulf Islands. Together these studies suggest that: (1) density-tolerant birds most often occur in resourcerich habitats; (2) more density-tolerant birds are heterozygous; (3) densityintolerant, dispersal types seek lower densities in more resource-limited habitats; and (4) more density-intolerant birds are homozygous. Spruce grouse also appear to show behavioral polymorphism. Yearling males and females exhibit a wide or a restricted nest (females) and advertising site (males) searching pattern (Herzog 1977a). Keppie (1981) documented sedentary, wide-ranging, and intermediate types of juvenile birds dispersing from the same broods. Herzog and Keppie (1980) and Schroeder (1985) recognized that a complex ingress-egress (sorting) occurred in both fall and spring; some birds left as others came into the study area. Juveniles in one area dispersed in the fall even though Keppie (1979) had removed all the adults. The complex movements of
572
A. T. BERGERUD AND M. W. GRATSON
spruce grouse (Fig. 14.28) can be explained by five processes: (1) their study area was of optimal habitat, where density-tolerant types should breed; (2) in the fall, progeny of both density-tolerant and density-intolerant birds sought winter cover nearest to their breeding ranges; (3) birds in spring dispersed either into the study area or left the study area, according to whether they were density-tolerant or intolerant types; (4) there was positive assortment of these two morphs, as argued for blue grouse (Fig. 2.26); and (5) the density-tolerant morph produced both density-tolerant and density-intolerant phenotypes. Keppie (pers. comm.) has confirmed that his study area was of more nearly optimal habitat than the surrounding land. His study also suggests that the extensive shuffle in fall and spring, which has been documented for forest grouse, may be exaggerated when we select small, high-density areas to study populations in open systems. Ruffed grouse have a similar pattern of fall dispersal. Some birds leave and others arrive (Rusch & Keith 197 Ib). These movements cannot be explained by social explusion; rather, birds must be seeking their own best-fitted environment. Behavioral studies of ruffed grouse have shown that most gray-colored males flush at longer distances than red-colored ones (Chap. 3). In that study, the same red-phase birds fought trespassers (mirror images) in their advertising ranges and at their drumming logs more vigorously than did gray-phase birds. Also, redphase females were more conspicuous than gray females in distraction displays (Fig. 3.13; see also Haas 1974 for spruce grouse). In Minnesota, red-phase birds have a shorter life expectancy than gray-phase birds (Gullion & Marshall 1968, Little 1978), possibly because of their conspicuous behavior (Chap. 3). In the Minnesota studies the red phase increased in frequency as numbers increased (Gullion & Marshall 1968, Little 1978). This was also true for two periods of increase in 15 years at Watch Lake, British Columbia (Chap. 3). Differences in spacing behavior thus seems to vary according to phenotype (color), and these differences are more striking between years than between places within years (cf. Gullion 1981). Whereas the spacing patterns of phenotypes in the forest grouse may be bestadapted to different stages of forest succession, climate may be the dominant extrinsic variable that affects spacing in steppe grouse. In the prairies and savannas, drought cycles affect the abundance of herbaceous nesting cover. An abnormally wet or dry year that varies 25% above or below the mean can be expected, on average, every 2.5 years in the shortgrass prairie, every 3 years in the mixedgrass prairie, and every 4 years in the tallgrass prairie (Wiens 1974). This variable template means that steppe grouse should be flexible in searching behavior. In some years they will have to search large areas to find suitable nesting patches; at other times nest cover will be abundant. Some females should be prepared to nest near each other when the environment is patchy, and to be widely spaced when conditions permit. Density-tolerant and density-intolerant types may exist in the steppe grouse. Gratson (Chap. 5) has recognized a bimodalism in move-
SURVIVAL AND BREEDING STRATEGIES
573
Fig. 14.28. Fall and spring movements of juvenile spruce grouse documented by Herzog and Keppie (1980). This complex movement pattern can be explained by assuming that most of the breeding birds in the study area were density tolerant (suggested because the study area had a high density —Keppie pers. comm.) and that these birds produced both density-tolerant and density-intolerant progeny. These movements are explained by: (1) both morphs seeking the nearest winter cover (fall movement) and (2) each morph showing positive assortment the next spring (spring movement), when they space themselves to establish nesting and future advertising sites.
574
A. T. BERGERUD AND M. W. GRATSON
ment patterns in male sharp-tailed grouse. One would expect alternative mating tactics in colonizing tendencies. When populations are increasing, the best tactic might be for yearlings to establish new leks near new, female nesting ranges. When the population is decreasing, it would be a better tactic to contest lek positions at established display grounds. The grouse cannot predict if the population is increasing or decreasing, and both types may be maintained where different phenotypes are favored at different times. The tundra grouse have the most constant physical environment of the three grouse groups. Plant succession is minimal on the arctic-alpine floor. Vegetation is not buffeted by wet and dry growing seasons as it is on the prairie. The major environmental variable that changes is the number of birds, because nesting success is usually high (Chap. 15). Many ptarmigan populations show regular changes in number-cycles. The aggressive behavior of rock ptarmigan varies between increasing and decreasing phases in cyclic abundance (Watson 1965). Aggressive spacing behavior has been shown to be at least partly innate (Theberge & Bendell 1980), and heritable (Moss et al. 1982, 1984, 1985) in ptarmigan. Mercer (1967) documented a suite of parameters that changed between years in an insular population of willow ptarmigan as it declined, and Bergerud has noted that birds were more conspicuous and vulnerable to hunting at high densities than when populations declined and were low (Bergerud 1972). Henderson's (1977) work has shown that more colonizing red grouse were homozygous than birds living in higher densities. These marked differences in behavioral tactics in ptarmigan at high and low densities, together with high probability of a heritable basis, suggest the occurrence of a genetic polymorphic system in which one morph is fittest al low densities and another at high densities, as predicted in Chitty's (1967) polymorphic behavior hypothesis (Chap. 12). There is a gambit of behavioral tactics for tolerant-intolerant density morphs. Each set of tactics (a strategy) will be ineffective unless there is positive assortment of phenotypes. Dispersal is the assortment mechanism. Phenotypes seek a density habitat that contains "like-minded" birds. The density-tolerant tactic of spacing near other advertising males in habitat with good vegetation cover would be counterproductive if a neighbor were more aggressive or more mobile, and monopolized females. Density-tolerant females are sedentary and docile, to remain inconspicuous while nesting in high numbers. Their strategy would be ineffective if an aggressive, more intolerant female were to settle in their midst. Mutual avoidance of different phenotypes by sorting into preferred density habitats would result in optimal spacing for individuals and maintain divergence of aggressiveness in the population. The temporal and spatial sequence of sorting of these behavioral types may vary among populations and species with different life expectancies. The behavior of the long-lived blue grouse at Copper Canyon varied little from 1970 to 1980, yet densities were much less in the latter years. Conversely, there were ma-
SURVIVAL AND BREEDING STRATEGIES
575
jor temporal changes in aggressive behavior in a short-lived, ruffed grouse population as densities changed (Chap. 3). For the long-lived species, these behavioral polymorphisms seem more striking between areas than between years. In the shorter-lived species, changes between years appear more striking (Jenkins et al. 1963, Watson 1965, 1967, Mercer 1967, Bergerud 1970a, Watson & Miller 1971). The aggressive phenotype of blue grouse is found in situations where densities were low, whereas the aggressive phenotypes of ptarmigan and ruffed grouse appear more frequently at high densities and when populations decline. But, blue grouse are noncyclic, whereas ruffed grouse and ptarmigan populations in northern areas show 10-year cycles of abundance (Chap. 15, Figs. 15.30, 15.32). A major distinction between cyclic and noncyclic species may be the spatial-temporal pace of sorting morphs moderated by longevity. In noncyclic populations, phenotypes orient themselves in space, whereas polymorphisms in cyclic populations occurs temporally. There has been a discussion of whether density and demography are causal in behavior, or if behavior is causal in demography (e.g., Chitty 1967, Charlesworth & Giesel 1972, Gaines 1978). Like many earnest debates, the truth may lie between the poles. Aggressive and spacing behavior have a heritable basis in grouse (Moss et al. 1982, 1984, 1985) independent of density, but these polymorphic suites of tactics preadapt birds to maximize fitness in different density environments (Fig. 2.27). "Two fundamental antithetical sets of behavior are always present between individuals, repulsed behaviour to maximize resources and thus separate from others, and attraction behaviour to maximize current resource availability and thus congregate where resources are most abundant" (Taylor & Taylor, 1977, p. 418). These conflicting, fundamental forces provide the template of the behavioral, polymorphic sorting system in grouse.
14.10 Summary This chapter reviews the strategies and tactics grouse have adopted to maximize their fitness, first by enhancing their survival and second by maximizing breeding opportunities. The male's primary concern, besides survival, is advertising for and breeding with females. The female is less concerned about breeding, which is easily accomplished. Her most important decision is where to nest to minimize predation of her clutch. A second important decision for her is where to take the chicks so that they can rapidly grow and yet escape detection by predators. Males have adopted several tactics to successfully advertise for females; they show fidelity to sites they have selected, and in many populations they display in the fall, months before breeding, to assure ownership of these locations. Males also select safe sites near nesting females and commonly display in the crepuscular periods to avoid diurnal predators. Yearling males are capable of breeding but generally are not chosen by females. Thus, yearling males devote their time to
576
A. T. BERGERUD AND M. W. GRATSON
prospecting for future advertising sites that are safe and will attract females in later years. Females prospect early for suitable nesting sites. Factors that may influence a female's decisions are: extent of nesting cover, local abundance of predators, and locations of other females and males that may "blow her cover" by overt display directed at her. Females avoid other females to reduce the chances for density-dependent nest predation to occur and to minimize conspicuous, aggressive interactions, which may also attract predators. Females of all the grouse, except some prairie chickens and sharp-tailed grouse in pure grasslands, nest adjacent to logs, trees, or rocks, or under shrubs to avoid being ambushed on the nest, which permits them to "sit-tight," concealing their eggs. Laying and incubating hens show a variety of behaviors on and near the nest that minimize nest conspicuousness. Females of some species often abandon nests and renest, and the factors influencing these decisions are discussed in this chapter. The phenotypic cues and ultimate factors that females use to select males are still contentious. We believe that choice is made on the basis of factors that will assure the female a safe, successful coitus and will minimize the conspicuousness of herself and her nest site. We suggest that females of the steppe species generally select those males on leks that are less likely than other males to approach and follow females off the lek and into nesting habitats. This reduces the conspicuousness of nests in habitats with many nest predators. Forest grouse females may select males displaying at intermediate distances from their nest sites. A male that is too close may bring unwanted attention to the nest site. If a female moves to a male far away, she increases her conspicuousness. Ptarmigan hens select males with large territories that contain ample space, to maximize distances to other nests, and high-quality cover to hide nests. The brood-rearing strategy of hens includes moving from the nest site to find optimal foraging habitats for young chicks. Later, when the chicks can thermoregulate, hens take their broods to denser cover, decreasing their conspicuousness to predators. The intensity of distraction displays by brooding hens is proportional to the potential loss in fitness from chick mortality and wanes after the risk to chicks is reduced at 2-3 weeks of age. Following the breeding season, grouse are concerned primarily with survival. Migrating in the fall to denser and taller cover, flocking, and snow-roosting are all viewed as tactics that reduce predation-risk rather than tactics to cope with weather or food factors. In winter, males will compromise their safety more than females to remain nearer to breeding areas and thus enhance breeding opportunities the following spring. This chapter closes with a discussion of two alternative suites of breeding tactics that grouse may have evolved in a cover-space trade-off. This polymorphic spacing system involves a density-tolerant morph and a density-intolerant morph. Density-tolerant birds select habitats of high intrinsic cover value for breeding
SURVIVAL AND BREEDING STRATEGIES
577
and are prepared to compromise space in a high-density context for this ideal cover. These birds have adopted behavioral tactics that promote inconspicuousness and close-living, at the expense of social interactions to enhance space. Density-intolerant birds, in the face of competition, elect to compromise cover value to select sites with more space to avoid predators. This morph is more aggressive and conspicuous than the density-tolerant birds, and more adept at prospecting widely to find satisfactory sites for nesting and advertising. It is hypothesized that these two traits of attraction and repulsion are present in all grouse groups. The variation in the frequency of these phenotypes will be primarily in space for noncyclic populations, and positive assortment will occur through dispersal; for cyclic populations the frequency of the morphs will vary with time.
15
Population Ecology of North American Grouse A. T. Bergerud
15.1 Introduction The first synthesis of the demography of North American grouse was thirty years ago when J. J. Hickey (1955) reviewed the then current literature. His review emphasized census methodology, the age and sex structure of populations, and the question of fluctuations and cycles in the numbers of grouse. Johnsgard (1973) also reviewed the current literature, but emphasized life-history characteristics of species. In the past ten years, wildlife biologists have been actively counting grouse, determining the sex and age composition of the living and the dead, calculating mortality rates, and searching for nests. Radiotelemetry has allowed biologists at last to find nests, to evaluate the use of space, and to investigate and document the factors that cause death. Unfortunately, many of these data on grouse are unpublished—in government reports or in doctoral and master's theses. I used these sources because not to have done so would have excluded many important findings and prohibited a current synthesis. This review chapter of the population ecology of North American grouse is my interpretation of the literature and does not, therefore, necessarily represent the views of the authors of this book nor those from whose works I have collected the statistics. The annual change in the number of grouse can be considered to begin when yearlings and adults arrive on the breeding range in year 1. The first potential influence that could reduce the intrinsic rate-of-increase (rm) is the percentage of hens that nest. Next, there is variability in the size of clutches and in hatching success. After hatching, some chicks die. Further, some adult birds die during the summer. Indexes to production in year 1 include the mean size of broods in autumn, and the ratio of juveniles per adult in the harvest. These indexes of 578
POPULATION ECOLOGY OF GROUSE
579
production are not valid measures of recruitment, because juveniles die over winter at greater rates than adults (Hickey 1955), but the fall juveniles:adult ratio (as a breeding success index) remains the common yardstick, in the literature, of the abundance of the new generation. The juveniles, yearlings, and adults then face the winter season, and more losses ensue. Dispersal should not be considered, a priori, a variable in demography; birds that have left probably remain alive, and the impact of the loss of such birds on the dynamics of a population says much about the adequacy of the size of the study area. After winter we have a new breeding population in the spring of year 2. Recruitment and mortality have failed to balance and the size of the breeding population has changed. What drives these changes and what prevents the population from continuing to increase is the subject of this chapter.
15.2 Percentage of hens nesting Darwin (1871, p. 417) was one of the first to suggest that all females do not nest. He asked "How is it there are birds enough to replace immediately a lost mate of either sex?" His interpretation was based on the removal of territorial birds and the observation that replacements appeared. A major theory of population regulation of grouse is, simply put, that females prevent other females from breeding. This notion has been advanced to explain changes in numbers of willow ptarmigan (Lagopus lagopus) (Watson & Moss 1972, Hannon 1982, 1983), blue grouse (Dendragapus obscurus) (Zwickel 1972, Hannon & Zwickel 1979), and spruce grouse (Dendragapus canadensis) (Boag et al. 1979). The main evidence for this theory is that females arrive to replace birds that are removed, as Darwin noted (Watson & Jenkins 1968, Bendell et al. 1972, Zwickel 1972, Zwickel 1980, Hannon 1983). These experimental results provide convincing evidence that territorial behavior spaces breeding birds, as argued by Lack (1966). They do not demonstrate that some females are prevented from breeding by social interaction. Birds may simply be relocating in habitats with reduced densities, resulting from removal, to improve their fitness prospects (Fig. 14.27); this sequence has been shown for great tits (Parus major) (Krebs 1971). Two removal experiments conducted in closed systems (no ingress or egress) with blue grouse males and willow ptarmigan males and females failed to find surplus nonbreeding birds (Chap. 2, Blom & Myrberget 1978). Myrberget and Blom provided evidence that ptarmigan pairs in fact shifted from where they would have nested to the vacancies created. The replacement females observed in the removal experiments of Zwickel (1980) and Hannon (1982, 1983) appeared relatively early in the season, when some hens were still searching for sites. After the hens had begun to nest, little replacement occurred (Zwickel 1980, Hannon 1982, 1983). The failure of hens to appear later is consistent with the idea that
580
A. T. BERGERUD
they are nesting elsewhere. No studies have documented large numbers of nonbreeding females during the peak of incubation. It seems unlikely that nonbreeding females could be available just before incubation and then disappear 2 to 3 weeks later. The most convincing method to detect nonbreeding females is to radio-track females throughout the preincubation period. For the test to be valid, these females must be captured before they visit males. Forest grouse have been monitored extensively by telemetry, and the results suggest that nearly all adult females breed and nest and that most yearlings attempt to do so. All 18 adult, blue grouse females that were radio-tracked on Vancouver Island nested, and of 46 yearling hens that were followed, 38 nested, 5 more were believed to have nested, 1 was killed too early to decide, and 2(4%) definitely did not nest (Hannon 1978, Sopuck 1979). The results of Herzog's (1977a) study of spruce grouse in Alberta are less clear; Herzog and Boag (1978, p. 867) said, "We have no direct evidence to indicate all females do not breed," but later Boag et al. (1979) said that only 60-70% of 47 females nested. Herzog (pers. comm.) said that of the 11 questionable yearlings, some indeed had nested, but that others, based on evidence from their movement patterns, probably had not. In Minnesota, Haas (1974) radiotracked 17 female spruce grouse and Maxson (1974) tracked 13 yearling ruffed grouse (Bonasa umbellus), and both concluded that all had nested. Evidence from studies of the steppe species indicates that nearly all female sharp-tailed grouse (Tympanuchus phasianellus) and prairie chickens (Tympanuchus cupido) nest (Table 15.1). In the longer-lived sage grouse (Centrocercus urophasianus), some yearling females probably do not breed (Stanton 1958, Eng 1963). I can find no evidence of nonbreeding females in rock (Lagopus mutus) or willow ptarmigan (Table 15.1). Conceivably there could be a shortage of males in these monogamous species, owing to predation by gyrfalcons (Falco rusticolus) on displaying males. When this occurred in Iceland, however, some females simply mated and nested with the same males (Gardarsson 1971). In Svalbard, one cock was seen to mate with four different hens (Unander & Steen 1985). Choate (1963a), working with white-tailed ptarmigan (Lagopus leucurus), observed females moving into his area late in the nesting season and called these females "nonbreeders"; again these may have been unsuccessful females that had tried to breed elsewhere. Other workers who have studied white-tailed ptarmigan have reported that all females breed (Table 15.1). Conceivably, it might benefit a yearling of a species with a long life expectancy to delay breeding until 2 years of age. The greatest mortality of hens occurs during nesting and the brood season; if a female waited for better nesting conditions as an adult, she might live longer and thus contribute more offspring than commencing with a yearling breeding option. Blue grouse are one of the longest-lived
POPULATION ECOLOGY OF GROUSE
581
grouse species, and F. Zwickel has provided enough statistics on yearling and adults to test the delayed-option strategy. If 100 yearlings delayed breeding until 2 years of age, they would still contribute fewer young throughout the lifetime of the cohort than 100 yearlings that first bred at 1 year of age—800 compared with about 875 young (calculations based on Zwickel 1975, Zwickel et al. 1977, 1983). This would occur even if the first-year survival rate of nonbreeding yearlings that delayed breeding was 80% and only 69% for yearlings that nested. Theoretically, all females should breed in their first year. Their life span is too short to delay breeding, and the experience they accumulate as yearlings should benefit their later attempts. The demographic data indicate that, indeed, nearly all adult females nest, but also that at times there are some nonbreeding yearlings in the long-lived species: blue grouse, spruce grouse, sage grouse, and whitetailed ptarmigan. Even if some females do not breed, this does not necessarily suggest that they are prevented from doing so by other females, nor that such nonbreeding would be a mechanism that regulates numbers.
15.3 Clutch size in grouse Lack (1947, 1954) proposed that the clutch size of each species of bird has been set by natural selection at that which provided the maximum number of surviving young. For the nidifugous species, Lack (1968) hypothesized that selection favored the clutch size that provided eggs of high quality and hence, survival of precocial young; also clutch size, proximately, should be affected by the hen's physical condition as influenced by her diet. An alternative ultimate hypothesis is that species have evolved their characteristic clutch sizes to minimize losses from predation (Skutch 1949). Large clutches require more trips to the nest by a female and thus extend and increase the vulnerability of the nest to predators. Also, large broods would be more conspicuous and female defense behavior could entail higher risk; thus, the increased vulnerability does not end after the eggs have hatched. Although this hypothesis has been discounted by some (Klomp 1970), it remains an alternative to the food hypothesis for nidifugous species (Cody 1966, Johnsgard 1973, Safriel 1975, Perrins 1977, see also Slagsvold 1985). A valid hypothesis of clutch size in grouse must explain: (1) the wide variation in the mean number of eggs laid among species, ranging from five to six eggs in some of the spruce grouse and white-tailed ptarmigan populations to 12 to 13 eggs in prairie grouse, and (2) the extreme variation between populations within species. For example, the mean clutch size of spruce grouse shows wide variation (Rand 1947); in Alaska the mean clutch is 7.5 eggs (Ellison 1974), but in Minnesota it is only 4.7 (Haas 1974). Rock ptarmigan in Scotland have a clutch of 6.6 eggs (Watson 1965), but in Iceland the average is 11 eggs (Gardarsson 1971).
582
A. T. BERGERUD
15.3.1 Maternal condition and clutch size If clutch size both ultimately and proximately is principally determined by maternal nutrition, it should vary among places, years, and individuals in relation to the physical condition and spring foods of females. However, there was no correlation between the weights of prairie chicken females in Minnesota in the spring and their clutch sizes (Table 14.4). Ruffed grouse females in Minnesota laid an average of 11.9 ± 0.39 ( ± SE) eggs (n = 11) in 1971 and 10.5 ± 0.58 eggs (n = 10) in 1972, but females weighed significantly more in 1972 than in 1971 (Maxson 1974). Myrberget (Fig. 11.7) found no correlation between the weights of willow ptarmigan hens on Traney Island, Norway, and their clutch sizes for 11 years (1965-76), but in 1977, 1978, and 1980, the clutch was two eggs larger than the mean of the previous 11 years, and this increase was accompanied by a major increase in female weight. Table 15.1. Comments in the literature on the proportion "Greater than 90% of the females nest." Tundra grouse Willow ptarmigan: "Probably all hens attempted to nest once" (Bergerud 1970a, p. 305). "So surplus birds apparently did not exist at that season" (Pedersen et al. 1983, p. 267). Rock ptarmigan: "No non-breeding females were observed" (Gardarsson 1971, p. 97). "Specimens collected . . . failed to uncover evidence that only female(s) did not breed (Weeden 1965, p. 340). "Polygamy . . . ensured that almost all hens bred every year (Unander & Steen 1985, p. 204). White-tailed ptarmigan; "There were no unmated females on the study area" (May 1975, p. 197). "The evidence obtained indicated that all females attempted to nest" (Braun & Rogers 1971, p. 39). "Field observations indicated that all females nested" (Giesen 1977). Forest grouse Ruffed grouse: no references found. Blue grouse: "Practically all hens on the summer range . . . were breeding birds." (140 of 143 had broods or brood patches-Zwickel & Bendell 1967:829), "96 percent, 66/69 of adults hens . . . were breeding." (Bendell & Elliot 1967:53), 4 of 43 yearlings may not have nested and 2 of 88 adults (1974-74) may not have nested. On Moresby Island 84 of 86 tagged adults nested, as did 27 of 34 yearlings (pers. files). Spruce grouse: More than 90 percent of the hens nested (Ellison 1972). "We found brood patches on all females examined in late May and June with one exception." (Robinson 1980:60); 174 of 186 spruce grouse hens nested (Keppie 1975b). Steppe grouse Sage grouse: 97 percent (379/390) of the females had ovulated (Stanton 1958, Dalke et al.
POPULATION ECOLOGY OF GROUSE
583
Generally, the mean size of clutches for a particular species increases with latitude. In prairie grouse, for example, the correlation between clutch size and latitude is r = 0.573, using 29 studies. Females in many of the northern populations of grouse begin laying eggs before or coincident with new plant growth. Presumably, these populations would have little variation in clutch size if variability resulted from changes in the spring condition of females following spring "green-up." Contrary to this hypothesis, clutch size varied one to two eggs among years in willow ptarmigan populations in Newfoundland (Bergerud 1970a) and rock ptarmigan in Alaska (Weeden & Theberge 1972, pers. comm.). Prairie chickens lay three to four fewer eggs in Texas and Missouri than in Minnesota (Lehmann 1941, Arthaud 1968, Svedarsky 1979). Egg-laying in Minnesota is coincident with the start of the growing season (Table 14.4), but the southern
of females that nest (excluding radiotelemetry studies) 1963) (this does not necessarily mean all nested); "All field observations indicated that the entire female population conducted at least one nesting attempt." (Patterson 1952:103); 242 of 256 adults ovulated (Braun 1979). Sharp-tailed grouse and prairie chicken: no references except Robel and Ballard (1974), who believed some female prairie chicken might not get bred. "Less than 90% of the females nest." Tundra grouse Rock ptarmigan: "Some non-breeding was suspected but this was not certain." (Watson 1965, p. 159). Forest grouse Ruffed grouse: "Non-nesting varied from 0 to over 25 percent of all females . . . indifferent years" (Bump et al. 1947, p. 359). Spruce grouse: "We estimate only 60-70 percent (yearling females) bred" (based on brood patches, n - 47 and radio-tracked females —Boag et al. 1979). But Herzog (1977a, p. 42) said of the 4 "nonbreeding" radio-tracked hens: "Movements were localized and nests may have been destroyed before discovery." Blue grouse: Substantial number of sub-adults (nonbreeders?) of mixed sex were found during the breeding season (Boag 1966). "76 percent (29/38) of yearling hens were breeding" (Bendell & Elliot 1967, p. 53). 44% of yearling females were classified as nonbreeders (Hannon and Zwickel 1979). Steppe grouse Sage grouse: "5 of 20 (sub-adults) failed to ovulate (Stanton 1958:64); 119 of 139 yearlings ovulated (Braun 1979).
584
A. T. BERGERUD
populations have several weeks of new, spring growth before egg-laying. They should, according to predictions of the nutrition hypothesis, have larger clutches than the more northern races; this is the converse of what occurs. We have provided evidence that nesting commences with the disappearance of snow (Fig. 14.10); under these conditions annual variations in nesting schedules should nevertheless start from a common-food baseline. The food hypothesis predicts that females should delay nesting in late springs until they can accumulate sufficient nutrient and energy reserves; contrary to the hypothesis, hens begin laying in the north before new and nutritious foods are available, even though chicks from late initial nests and renests appear to have good survival (Bergerud 1970a, Parker 1981). Females that nest early generally lay more eggs in their initial attempt than do females initiating nests later in the same year (Jenkins et al. 1963, Bergerud 1970a, Svedarsky 1979). This applies to willow ptarmigan, in which yearlings and adults have similar clutch sizes. Presumably, females that start nesting later in a given year assimilate more higher-quality foods than do early hens. If these late hens were in poor physical condition initially, one might still expect that their improved physical condition owing to their delay would compensate, and clutch sizes would be similar—but this does not occur. Further, in the aviary, latestarting red grouse have smaller clutches even though food quality and availability are held constant (Moss et al. 1981). The food hypothesis is neither a necessary nor a sufficient explanation of the variability in clutch size. It provides no explanation for differences in clutch size between populations of the same species in which birds are eating similar foods and initiating nests at approximately the same time. All the populations in which grouse have long life spans (i.e., annual survival rate > 60%) show clutches of fewer than seven eggs (section 15.3.3), yet plant phenology and food show no consistent pattern among populations. The largest North American grouse, the sage grouse, and the smallest, the white-tailed ptarmigan, have a similar clutch size; but the egg expressed as a percentage of the total weight of the sage grouse is 3.4% and of the white-tailed ptarmigan is 6.4% (Johnsgard 1973). The food hypothesis has been related only to maximizing fitness in the current breeding season, and ignores that grouse should evolve life-history traits to optimize lifetime production (Williams 1966b, Charnov & Krebs 1974).
15.3.2 Clutch size determined by chick survival Safriel (1975) proposed that for birds with precocial young, clutch size was dependent on the optimal brood size that could be defended by the parents as modified by the density of food. Parents would have more difficulty defending chicks while searching for a dispersed food supply. The survival of grouse chicks in broods, however, is largely independent of clutch size; commonly 40-50% of the chicks die before autumn (Fig. 15.1), regardless of whether the hens laid large
POPULATION ECOLOGY OF GROUSE
585
or small clutches. Also, the frequency distribution of brood sizes in the autumn is essentially normal for most grouse unless cyclic polymorphism is involved (Bergerud 1970a). This suggests that survival of chicks from different clutch sizes is approximately equal. Willow ptarmigan chicks have mortality rates similar to the other grouse even though both parents provide protection. Also, the mean sur-
Fig. 15.1. Top: Summer mortality of chicks is not correlated with clutch size. Bottom: Hence, brood size increases linearly with clutch size (data from 38 studies). References for this and other figures with many data points are found beneath tables (see especially Table 15.2).
586
A. T. BERGERUD
vival rates of chicks do not increase with latitude, even though ground predators are less abundant in the north. Safriel's (1975) hypothesis is not supported in the tetraonids. He makes the interesting point that if the young themselves take protective action, parents could successfully raise more young. He offers this as an explanation for the larger clutches of gallinaceous birds compared with those of shorebirds.
15.3.3 Clutch size determined by life span and predation risk of eggs The mean clutch size of the various grouse increases as the annual mortality rate of adults increases, or stated conversely, clutch size declines as survival improves (Fig. 15.2). This correlation between fecundity and longevity is most precise in the arctic, where man has not upset natural predator systems, altered nesting habitats, or harvested grouse populations to the same degree as in southern latitudes. In my view the more independently evolved variable in this correlation is the mortality rate; mortality is not explained as the density-dependent result of a maximized reproductive rate as argued by Lack (1966). Nor do these parameters simply covary fortuitously or covary in response to a common, environmental resource antecedent as argued by Ricklefs (1977). Species evolve a clutch size that optimizes the trade-off between the benefits of high reproduction and the cost of reproductive risk. By prorating its reproductive effort over its
Fig. 15.2. Mean clutch size of grouse in 25 populations regressed against annual mortality rate. Natural mortality rates of birds in heavily hunted populations (indicated by arrows along data points) are lower than those shown.
POPULATION ECOLOGY OF GROUSE
587
greater life span, a long-lived individual can afford to invest less in annual reproduction than can a short-lived individual (Williams 1966b, Murphy 1968, Gadgil & Bossert 1970). Experimental confirmation of this trade-off is the recent work of Rose et al. (1984), who showed a reduction in ovary weights of youngadult fruit flies (Drosophild) in populations exhibiting postponed senescence. The greater independence of mortality than reproduction is explained by the fact that each grouse species inhabits a fairly specific biotype; sage grouse live in the distinct sagebrush biome; ruffed grouse have taken a niche in deciduous forests; and spruce and blue grouse have evolved adaptations that permit their exploitation of coniferous forests. These distinct habitats provide different escapecover characteristics, and are coinhabited by different arrays of coevolved predators. This distinct predator-cover complex, coupled with the specific antipredator tactics of each grouse species, has shaped the mortality rates of populations and species. Risk associated with reproductive activity may contribute to mortality, but the predator-cover dyad will be the major determinant of annual mortality rates and is mostly independent of fecundity and density (section 15.6). My model for the clutch size of tetraonids distinguishes between the factors that influence clutch size differences among species and among populations within species. The differences in clutch size among species are hypothesized to be caused by different reproductive optimization schedules that result from the
Fig. 15.3. Environmental factors that determine clutch size. The realized clutch size for birds in each population may be less than the optimal for each species, based on longevity, in order to reduce conspicuousness and predation loss.
588
A. T. BERGERUD
unique longevity of individuals of each species. Life span is the result of the species-specific habitat (Fig. 15.3). This habitat includes a characteristic type of escape cover and is coinhabited by specific coevolved predators. I have termed this the predator-cover complex. The fine-tuning of populations within species results from different levels of egg predation, and possibly predation of females. Hens in populations faced with heavy predation risk individuals could reduce their clutch so that the length or the egg laying period is shortened, thereby reducing the number of trips to the nest and thus the conspicuousness of the clutch (Johnsgard 1973). The clutch sizes of grouse can be segregated into high and low modes (Fig. 15.4). The low mode, from 5 to 8 eggs, is characteristic of grouse with annual survival rates greater than 50%, and the high mode, from 9 to 13 eggs, is characteristic of grouse with survival rates equal to or less than 45% per year. Within these two reproductive strategies, yearling hens lay significantly fewer eggs than adults only in the low-clutch group, and the difference between yearlings and adults is greatest where nest predation is low within this group (Fig. 15.5). These differences in clutch size between yearlings and adults are consistent with the optimization hypothesis that females of species with long life spans invest less in the initial clutch because of increased risk resulting from lack of experience and a longer life span over which they can maximize their effort. Yearlings in the high-survival group are, for the most part, less successful in nesting than adults (Keppie 1975a,b, Herzog 1977a, Zwickel et al. 1977, Petersen 1980, Smyth & Boag 1984). If nest predation is high in the long-lived group, the nests of both adults and yearlings will be at high risk, and adults may have clutches closer in size to those of yearlings. In general, the clutch-size data fit the hypothesis that long-lived grouse can afford to invest less in each attempt and to commit more resources to maintenance and later reproductive efforts than can birds expected to live only a short while. The validity of this hypothesis requires a significant loss of eggs during egglaying (see Myrberget 1985, Boag et al. 1984). Admittedly, there are few relevant data; most nests are found only after incubation has begun. Two of 22 radiomonitored, ruffed grouse hens lost their incomplete clutches during the laying period (Maxson 1974). Four of 37 sage grouse hens lost their clutches while still laying (Petersen 1980), and Svedarsky (1979) reported 5 of 36 prairie chicken hens that lost incomplete clutches. For bobwhite quail (Colinus virginianus), Klimstra and Roseberry (1975) found a 6% predation rate per day for the first 3 eggs laid; 3% per day for the remainder of egg-laying; and a rate of 2% per day during incubation (n = 863). Most eggs were lost to foxes (Vulpes vulpes). Mortality of waterfowl nests in North Dakota was generally highest during the early laying period, moderately high for later eggs, and lowest during incubation, according to Klett and Johnson (1982). They speculated that the presence of hens at nests reduced predation.
POPULATION ECOLOGY OF GROUSE
589
Fig. 15.4. There are two distinct modes of clutch sizes — a low clutch mode of five to eight eggs in populations in which birds have an annual mortality rate of 21%-41%, and a high clutch mode often to 14 eggs, corresponding to annual mortality rates of from 51 % to 75%. Natural mortality rates of heavily hunted populatons would be less than the reported mortality rates. (Many references to clutch size in literature cited in Table 15.2.
Another consideration is that eggs are easy to steal when hens are laying and generally absent from the nest. An incubating female probably defends her clutch against several important predators including squirrels (Spermophilus spp.), snakes, and corvids (Corvus spp.). Predators may also search more for undeveloped eggs, which may be more nutritious than eggs with developing embryos (Sargeant pers. comm.). The evidence is substantial that females should improve their nesting success by reducing the period of egg-laying and the resulting vulnerability of unattended eggs.
590
A. T. BERGERUD
Fig. 15.5. The difference in mean clutch size between yearlings and adults is greatest in populations with the low clutch mode, and the difference is greatest in habitats where relatively few nests are destroyed by predators. (For references see Table 15.2.)
The principal evidence for the hypothesis that clutch size is primarily determined between populations by the risk of predation is that most grouse have smaller clutch sizes in those populations exposed to heavy nest predation (Fig. 15.6). A similar sequence has been reported for the passerine birds (Ricklefs 1969). In earlier papers, Ricklefs (1969, 1970) attributed the difference in clutch size to predation pressure, but later he supported a resource-oriented explanation (Ricklefs 1977). A typical finding is that the commonest clutches are smaller than the clutch size that yields the most chicks per hen (Klomp 1970, Hussell 1972), contrary to the Lack hypothesis. In red grouse, hens that lay clutches of 10 to 11 eggs rear more young than females with clutches of the most common size, 7 to 8 eggs (Jenkins et al. 1963, Moss et al. 1981). The red grouse moors are managed to reduce nest predators, but if predators were common, the field clutch mode of 7 to 8 eggs might be more productive than the larger clutches because of its improved chance of escape from predators. The most productive clutch size in grouse should be evaluated based on survival of eggs, rather than on the survival of chicks from successful nests. The clutch size of grouse decreases with increased longevity of adults and with reduced nesting success (Figs. 15.2, 15.6). I found no examples of a population in which birds have a long life expectancy (qx < 0.50) and also have a mean clutch size of greater than 7 eggs, regardless of nesting success. But populations in which birds have short life expectancies (qx > 0.50) have large clutches (> 9 eggs) if nesting success is low, but may have either large clutches (> 9 eggs)
POPULATION ECOLOGY OF GROUSE
591
Fig. 15.6. The mean number of eggs per clutch is low in populations where hens can expect a high probability of nest predation, and is greater where hens can expect a low rate of nest predation. (For references see Table 15.2.)
592
A. T. BERGERUD
or small clutches (< 9 eggs) if nesting success is high. High survival is a sufficient but not necessary condition for the small-clutch mode. This suggests that longlived birds cannot further reduce clutch size to make their nests more inconspicuous if predation risk is high, but individuals with low survival may be able to afford some reduction in their generally larger clutches, to mitigate risk if nesting success is relatively low.
15.4 Nesting success I define nesting success as the percentage of nests that hatch at least one young. There are many biases in nesting-success investigations, including those that result from combining clutches in different stages of development (Miller & Johnson 1978). Biases also result from unrepresentative samples because the most conspicuous nests were found; attraction of predators to nests; human-induced desertion; and small sample sizes. In this review I could not arbitrarily exclude many studies. Any statistical comparison involving nesting success that is statistically significant should be conservative because of the great variability in the data. The literature search was for broad differences between grouse groups— forest, steppe, and tundra—and between species and populations within these groups. All the grouse are extremely vulnerable to predation because of their relatively large clutches and their habit of nesting on the ground. Very early, M. Nice (1942) reported a 55% combined nesting success for the phasianiae and tetraoninae (n — 5,597). J. J. Hickey listed a 58% success figure in 1955 for seven grouse studies. A real problem in studying tetraonids is finding large samples of nests. Only the mammoth New York State study of ruffed grouse had a sample of nests in excess of 1,000 (Bump et al. 1947). In this review, nesting success was 58% (n = 5,422) the same as that reported in 1955. Predation on average accounted for 79% of the unsuccessful nests (Table 15.2). The second most frequent cause of loss was nest desertion, which is especially prevalent in the sage grouse and prairie chickens. Furthermore, prairie chickens in particular lost nests to flooding, fires, and agricultural practices (Lehman 1941, Bo wen 1971, Svedarsky 1979). In addition, desertion and predation losses are correlated. Species with low nesting success desert more frequently (Tables 14.6, 15.2). The abundance and diversity of predators decreases with increasing latitude (Fleming 1973, Wilson 1974, McCoy & Connor 1980), and as the abundance of predators increases, the nesting success of hens decreases (Fig. 15.7); this has also been documented in the passerines (Ricklefs 1969). Hens south of latitude 40 °N averaged a loss from predators of 45 %, whereas hens nesting north of latitude 46 °N had an average loss from predation of only 26%. Grouse in the southernmost regions face extreme predation problems. Hardy
POPULATION ECOLOGY OF GROUSE
593
Table 15.2. Nesting success and predation of grouse nests based on total nests for each species and unweighed means from different studies (excludes studies without normal predators) Nesting success (%) successful total
Predation rate (%) preyed on total
Studies
Nests
Studies
Nests
Studies
1,783 148 260
13 5 4
62 55 53
61 ± 5.4 62 ± 11.3 53 ± 3.8
33 (86)a 39 (87) 36 (76)
30 + 4.7 33 ± 12.5 34 ± 6.4
Totals and means
2,191
22
60
60 ± 4.0
34 (84)
32 ± 4.0
Steppe grouse Sharp-tailed grouse Prairie chicken Sage grouse
540 934 699
16 22 12
54 49 35
54 ± 4.4 48 ± 3.1 38 ± 4.0
37 (73) 39 (77) 50 (76)
36 ±4.2 38 ± 3.1 47 ± 6.3
Totals and means
2,173
50
46
± 2.5
42 (76)
40 ± 2.6
70 213 775
2 2 6
60 70 71
65 ± 5.0 73 ± 3.5 72 ± 4.5
39 (93) 20 (68) 21 (75)
35 ± 5.0 22 ± 1.5 22 ± 3.9
Totals and means
1,058
10
70
72 ± 3.3
22 (76)
24 ± 3.3
Totals and means
5,422
82
57
59
36 (79)
31
Sample size
Grouse groups and species
Nests
Forest grouse Ruffed grouse Spruce grouse Blue grouse
Tundra grouse White-tailed ptarmigan Rock ptarmigan Willow ptarmigan
a
Percentage of unsuccessful nests destroyed by predators. Sources for nesting success and for many clutch sizes: Ruffed grouse — Fisher 1939, Bump et al. 1947, Grange 1948, Hardy 1950, Fallis & Hope 1950, Tanner 1948, Banasiak 1951, Kupa 1966, Neave 1967, Barrett 1970, Cringan 1970, Maxson 1974, Ruschetal. 1984. Spruce grouse-Ellison 1974, Haas 1974, Herzog & Keppie 1978, Robinson 1980, Redmond et al. 1982, Keppie 1982. Blue grouse-Mossop 1971, Zwickel 1975, Zwickel & Carveth 1978, Sopuck 1979, Hoffman 1979. Sharp-tailed grouse-Gross 1930, Hamerstrom 1939, Grange 1948, Hart et al. 1950, Blus & Walker 1966, Bernhoft 1969, Christenson 1970, Artmann 1970, Pepper 1972, Schiller 1973, Rice & Carter 1975, 1976, 1977, Sisson 1976, Kohn 1976, Ramharter 1976, Caldwell 1976. Prairie chicken-Gross 1930, Hamerstrom 1939, Lehmann 1941, Schwartz 1945, Ammann 1957, Copelin 1963, Yeatter 1963, Silvy 1968, Arthaud 1968, Watt 1969, Bowen 1971, Horak 1974, Rice & Carter 1975, 1976, Sisson 1976, Drobney & Sparrowe 1977, Horkel 1979, Horkel et al. 1978, Riley 1978, Sell 1979, Svedarsky 1979, Vance & Westemeier 1979. Sage grouse-Dargan and Keller 1940, Keller et al. 1941, Batterson & Morse 1948, Patterson 1952, Nelson 1955, Gill 1965, Carr 1967, Klebenow 1969, May 1970, Wallestad & Pyrah 1974, Petersen 1980. Whitetailed ptarmigan — Choate 1963a, Giesen et al. 1980. Rock ptarmigan—Watson 1965, Weeden & Theberge 1972. Willow ptarmigan - Hagen 1935, Kristofferson 1937, Bergerud 1970, Watson & O'Hare 1979, Hannon 1982, Hannon & Smith 1984, Myrberget Chap. 11).
594
A. T. BERGERUD
Fig. 15.7. The percentage of nests destroyed by predators decreases with a decrease in number of predator species and in populations farther north.
(1950) found that only 3 of 12 (25%) ruffed grouse nests hatched in Kentucky. Snakes were a serious predator. The southernmost grouse, the Attwater's prairie chicken, hatched only 17 of 54 (31%) nests (Horkel et al. 1978). In that study, the loss of hens was as serious a problem as the destruction of nests. Nine of 59 radio-tracked hens were lost during egg-laying; 22 additional hens were killed during incubation. The sex ratio of Attwater's prairie chicken is one of the most unbalanced in the grouse literature-approximately two males per one female (Horkel & Silvy 1980). The southern distribution of grouse appears limited by predation of nests and of females. Grouse in the middle latitudes fare better (Fig. 15.7), but both sage grouse and prairie chickens have generally low nesting success, approximately 35% and 49%, respectively (Table 15.2); in addition, prairie chicken hens are commonly
POPULATION ECOLOGY OF GROUSE
595
killed on the nest (Table 14.6). Desertion and renesting of adults are tactics both species use to reduce predation loss (Fig. 14.13). The ptarmigans enjoy significantly higher nesting success than do either the forest or steppe grouse. One reason is that they face fewer kinds of predators of nests (Fig. 15.8), and these predators are at lower densities. In addition, ptarmigan are monogamous, and the males help distract predators from the nests (Chap. 13). Little information exists on the total density of predators in the three habitat types —steppe, forest, and tundra. The available evidence, discussed in Chapter 13, indicates that the steppe has the most predators. The comparative sequence is: Nest predation: Number of predator species: Density of predators:
steppe > forest > tundra steppe > forest > tundra steppe > forest > tundra
In addition to abundance of predators, there are other reasons for the low nesting success in the steppe. The extent of grass and sagebrush communities has been reduced by agriculture—this concentrates nesting females and thus reduces the size of areas that predators need to search for prey. Furthermore, nesting cover has been reduced by grazing and herbicides. We have also aided the predator's access to nesting habitats by providing perching posts, travel lanes, and culverts to be used as dens. Further we have increased the abundance of alternate prey with domestic animals and by roadkills. Man has upset the predator-prey adaptive race more in the steppe than in the other two habitat zones.
15.4.1. Frequency of renesting The renesting decision of female grouse should be based on the benefits and costs of the increased parental investment in laying a second clutch (Fig. 14.13). Early grouse biologists doubted the importance of renesting (Hamerstrom 1939, Bump et al. 1947, Ammann 1957). Even as late as the 1970s, Johnsgard (1973), in a review of the literature, reported little renesting in the tetraonids. Now we can measure accurately the frequency of renesting, based on radiotelemetry. Renesting can also be judged by the distribution of copulations for lek species and by hatching distributions. Using these distributions is valid because females revisit males at leks between first and second nestings to secure viable sperm (Lake 1975, Parker 198la) and because hatching frequencies in the absence of predation are essentially normally distributed (Fig. 15.9, Hannon 1982). Skewed hatching distributions can result from renesting. I have calculated the renesting frequency for seven species using the presence of late-hatched broods or second peaks in the frequency of matings (Table 15.3).
596
A. T. BERGERUD
Fig. 15.8. Top: For the three grouse groups, the diversity of predators that kill adults is similar, as are mean mortality rates. Bottom: The greatest diversity of nest predators is in steppe habitats and the least diversity is in the tundra. Nesting success in steppe < in forest < in tundra.
POPULATION ECOLOGY OF GROUSE
597
Fig. 15.9. The hatching curve of nests of willow ptarmigan on Brunette Island, Newfoundland, was normally distributed because there was no predation of nests by mammals and little renesting. On the southern shore, Newfoundland, hatching distribution was skewed to the right-to the later dates —because most females renested after their first nests were destroyed by foxes and weasels (Mercer 1967, Bergerud 1970a).
An assumption of this method is that nesting success in first and second attempts is the same. This assumption may not be valid for some species, especially those nesting in the steppe, where cover improves with the growing seasons. Forest grouse appear to renest less frequently than steppe species (Table 15.3). These grouse face effective predators in forests with little herbaceous growth— they also have small preincubation ranges, with less space than the lek species to move nest locations (Table 14.2). Radiotelemetry data are not yet available for tundra grouse. Parker (1981b) found that 14 of 56 (25%) willow ptarmigan hens renested when their eggs were removed during incubation. Tundra grouse probably have a high inclination to renest—but it has escaped notice since hens are usually successful with their first
598
A. T. BERGERUD
nests. Also, corvids and weasels (Mustella spp.) commonly steal the eggs of ptarmigan one at a time; such a gradual, nontraumatic loss may go undetected by the hen. Grouse of the steppe show a high propensity to renest (Table 15.3). Nesting Table 15.3. Estimations of renesting based on radio-tracking females that lost their nests and the sighting of late-hatched young or late matings Grouse groups and species
% of females that renested From telemetry
Tundra grouse Willow ptarmigan Rock ptarmigan
88 13
White tailed ptarmigan Forest grouse Ruffed grouse
Based on sightings
30 (18/138)
22 (9)c
26 (149/1473)
Spruce grouse
10(10)
20
Blue grouse
26(31)
Steppe grouse Sharp-tailed grouse Prairie chicken
83 (6)
72 (609/1655)
41 (17)
c
(32/89)
43 ( = 52/343)
Sage grouse
b
(3/18)
86 (14)
59
a
(55/249) a ' b (8/201)
References
Bergerud 1970a Weeden 1965, pers. comm. Choate 1963a
Maxson 1974, Barrett 1970, Bump et al. 1947 Ellison 1974, Haas 1974, Keppie 1975b Sopuck & Zwickel 1983
Schiller 1973, Christenson 1970 Svedarsky 1979, Hamerstrom & Hamerstrom 1973 Robel & Ballard 1974 Hartzler 1974 Petersen 1980
Sample size: 55 females with late-hatched chicks, and 249 females with chicks hatched at normal time. Calculations based on 0.75 nesting success: (a) 249 x 1.00 = 0.75 x = 332 females started 1st nests; (b) 332 - 249 = 83 females failed 1st nests; (c) 55 x 1.00 = 0.75x = 73 females failed 1st nest and started 2nd nests; (d) 73/83 = 88% of the females that failed first nests started 2nd nests. A check of method is: If 100 hens in Newfoundland hatched first nests at 0.75 (Bergerud 1970a), and 25 unsuccessful hens renested at 0.88 and also had a renesting success of 0.75, females seen with broods in August should be 91.5% (75 hens + 16.5 hens (25 x 0.88 x 0.75) = 16.5; 91% of the hens seen 1955-65 had chicks (n = 841, Bergerud 1970a, p. 308). 9 hens radio-tracked
POPULATION ECOLOGY OF GROUSE
599
success was higher in second attempts than first attempts for sharp-tailed grouse in Minnesota and North Dakota (Christenson 1970, Schiller 1973). Robel (1970) argued that early-nesting prairie chickens had a higher success than later-nesting prairie chickens in Kansas. If drought conditions existed, it is conceivable that residual growth could be denser than new growth. But in years with normal rainfall, nesting success should improve as new cover appears. Schiller (1973) provides convincing arguments of how increased growth can help disperse and conceal renesting attempts. In general, yearlings renest less often than adults. This has been documented for blue grouse (Sopuck 1979), spruce grouse (McCourt 1969), and sage grouse (Wallestad 1975b). The initial nesting success of radio-tagged, yearling sage grouse in Colorado was 34% (5 of 16) compared with 33% for adults (4 of 12) (cf. Petersen 1980), but the success of both age classes, based on the molt sequency in the harvest for 5 years, was 32% for yearlings and 58% for adults (Braun 1979). This sequence suggests greater renesting in adults. Yearling hens are generally less successful in concealing their nests (Sopuck 1979, Redmond et al. 1982) and generally have reduced nesting success in forest and steppe populations, which are faced with effective nest predators (Table 15.4). It might benefit the inclusive fitness of yearlings of long-lived populations that lay small clutches to forgo renesting. As adults the next season, they would face less intraspecific constraint in the selection of nest sites. Also they could lay at the optimum time. Yearlings, at least in the blue and spruce grouse, lay their eggs about one week later than do adults (Keppie 1975b, Hannon et al. 1982, Smyth & Boag 1984). As nesting becomes progressively more difficult it could be expected that yearlings would fare less well than experienced adults (Fig. 15.10).
15.4.2 Annual variations in nesting success All the grouse show variation between years in nesting success. These differences may reflect annual differences in weather as in white-tailed ptarmigan (Braun & Rogers 1971); changes in age structure of nesting females, as in sage grouse (Eng, pers. comm.); and variation between years in predator searching and nesting cover (Chap. 14). The percentage of yearling females in a population changes from year to year based on prior survival. In a year with few yearlings, nesting success may be higher than in a year when there are many inexperienced yearlings (Eng. pers. comm.). This can be an important density-dependent, dampening mechanism; i.e., the difference between adult and yearling nesting success in Table 15.4 is correlated with the nesting success figures in Table 15.2 (r = -0.837, n = 8). The variables in this correlation are not strictly independent, but the correlation is still instructive. If sage grouse have a good nesting season in one year, there is more than an even chance that they will be less successful the next. The more
600
A. T. BERGERUD
Table 15.4. Nesting success compared between adults and yearlings based on nests, broods seen, or molt sequence % females successful
Grouse groups and species
Sample size
Forest grouse Ruffed grouse
Adults
Yearlings
24 nests
71
61
-10
Blue grouse Blue grouse Blue grouse
682 females 647 females 911 females
81 73 76
71 67 53
-10 -6 -23
Blue grouse Spruce grouse
56 nests 257 females
38 41
58 30
+20 -11
Spruce grouse Spruce grouse
48 females 36 nests
82 48
44 7
-38 -41
64
49
-15
(no studies) 20 nests 22 nests
61 77
43 44
-18 -33
Sage grouse Sage grouse
35 nests large wing sample
41 77
39 61
-3 -16
Sage grouse
large wing sample
58
36
-22
63
44
-19
209 females
78
49
-29
49 nests
58
61
+3
57 nests
82
76
-6
Rock ptarmigan
315 females
77
60
-17
Willow ptarmigan Willow ptarmigan
50 nests 49 nests
83 85
94 74
+ 11 -11
77
69
-8
Mean Steppe grouse Sharp-tailed grouse Prairie chicken Sage grouse
Mean Tundra grouse White-tailed ptarmigan White-tailed ptarmigan Rock ptarmigan
Mean
Difference References
Maxson 1974, Barrett 1970 Redfield 1972 Hoffman 1979 Zwickel et al. 1977 Sopuck 1979 Keppie 1975b, McCourt 1969 Robinson 1980 Redmond et al. 1982
Svedarsky 1979 Wallestad & Pyrah 1974 Petersen 1980 Wallestad & Watts 1972 Braun 1979
Braun & Rogers 1971 Giesen pers. comm. Weeden & Theberge, pers. comm. Weeden pers. comm. Myrberget 1970c Hannon & Smith 1984
POPULATION ECOLOGY OF GROUSE
601
Fig. 15.10. The reduction in nesting success of yearlings from that of adults is greater in species that have lower nesting success. The X and Y axes are not strictly independent. (For references to differences between adult and yearlings see Table 15.4, and for species nesting success see Table 15.2.)
problems a species encounters in nesting, the more yearlings will lag behind experienced adults in being successful, and the more prior production will dampen population changes. Nesting success should vary between years with the number of predators searching for nests. The New York State inquiry on ruffed grouse documented that nest predation increased when mice populations were low, and predators — mostly foxes —were probably more inclined to hunt for clutches. Weasels are the major predator of nests in many rock ptarmigan populations, and weasel predation was more severe in some years than others, probably in response to a shortage of mice (Weeden & Theberge 1972). Willow ptarmigan in Norway had high losses of eggs in 1963, 1967, and 1979—these losses followed years with high populations of rodents and probably involved more predator searching and reduced nest cover from rodent eat-outs (Chap. 11). A similar sequence has been documented in black grouse (Tetrao tetrix] (Angelstam 1983). Prey switch-over can add significant variation in nesting success between years.
602
A. T. BERGERUD
Herbaceous nesting cover should vary between years, depending on spring weather, and thereby influence the concealment of nests and rates of predation. This variability should be most important for the steppe species that rely on herbaceous plant cover to hide nests. The prairie grouse commence nesting in residual cover from the previous growing season. While they are incubating, new plant growth becomes increasingly important as the dead material falls and decays. Soil moisture on the prairies varies between years and should be a primary variable in plant growth (Williams & Robertson 1965) because it affects cover and nest concealment. I evaluated the role of rainfall in breeding success for the sharp-tailed grouse in North and South Dakota and Minnesota by calculating a soil-mixture index. The soil-moisture index was based on summing the monthly precipitation from September in year 1 until July in year 3 (23 months). The 23-month total was needed to account for both residual cover and new cover. I compared this moisture scale to an index of the percentage of successfully nesting hens from statistics of sharp-tailed grouse in North and South Dakota. This index was calculated: [(juveniles/adult in harvest) X 2 4- mean brood size in late summer]. Assumptions of this index were (1) that all hens attempted to nest and (2) that the proportion of males and females in the population was equal. Next I compared the soil moisture to juveniles/adult ratios in Minnesota. Last, the moisture index was tested by correlation analysis against the late-summer brood sizes of sharp-tailed grouse in both the Dakotas. Sharptail production was significantly correlated with the soil-moisture index in both North and South Dakota (Fig. 15.11), but not in Minnesota (not shown). The importance of moisture in the variability of plant growth increases as one moves down the rainfall gradient going from east to west (Wiens 1974). Grasslands in Minnesota are less influenced by variations in moisture than those in the central and western Dakotas. The moisture index was correlated with both the successful nesting hens and chicks/brood statistics for sharptails in the Dakotas (Fig. 15.11). My interpretation is that annual vagaries in plant cover affect predator/grouse interaction both at the nest (nesting success) and with the brood, but that the annual changes in nesting success contribute considerably more to the variability in total chick production than does variation in the survival of chicks in broods. Additional evidence that annual moisture-cover changes affect nesting success in prairie grouse is that there is a higher proportion of adult males in years of low production (Fig. 15.12). This correlation was also satisfactory if I used juveniles per adult male rather than the juveniles/adult figures; the latter would be biased by female deaths. Prairie chicken females in disturbed habitats commonly get killed on the nest (Table 14.6) and this type of ambush could be more frequent when cover was short and sparse or patchy during dry cycles. An extensive literature substantiates the view that soil moisture affects nesting
POPULATION ECOLOGY OF GROUSE
603
Fig. 15.11. Breeding success of sharp-tailed grouse is correlated with soil moisture, which influences plant growth and hence also residual cover. Vegetative cover influences both nesting success and survival of chicks in broods in South and North Dakota. The correlation for juveniles/adult statistics and the moisture index were for South Dakota, r = 0.608, n = 30 (1949 to 1978), and for North Dakota, r = 0.553, n = 32 (1949 to 1980).
604
A. T. BERGERUD
Fig. 15.12. Four grouse populations that showed significant declines in breeding success with an increase in percentage of adult males in the harvest. Summer predation of hens probably occurs in these populatoins and results in unbalanced fall and winter sex ratios of adults. (Grouse data from Hamerstrom & Hamerstrom 1973, Hillman & Jackson 1973, Linde et al. 1978, Schulz 1981.)
cover and the dynamics of grouse. Hamerstrom (1936) and Errington (1935, 1937a) recognized 50 years ago that drought cycles would reduce vegetation cover and increase predation losses of pheasants. Janson (1951) and Wallestad (1975a) mention declining populations of prairie chickens and sage grouse during dry cycles. Breeding success for sage grouse increased in wet years (Carr 1967,
POPULATION ECOLOGY OF GROUSE
605
Blake 1970). Several workers have commented on the concentration of nests or broods in dry seasons or following overgrazing. Pheasants in South Dakota showed low productivity in warm, dry springs (Martinson & Grondahl 1966). Wagner (1957) emphasized the low production of pheasants and deaths of hens in dry, warm springs, but felt these losses were caused by physiological stress. However, more recent studies show that females do get killed when nesting cover is sparse. The loss of eggs from dummy pheasant nests to crows in South Dakota was highest early in the season before new growth appears (Grondahl 1956, Mitchell 1957). Mallard (Anas platyrhynchos) females are commonly killed on the nest by foxes (Johnson & Sargeant 1977), and the production of dabbling ducks, as in sharptailed grouse, is related to soil moisture (Boyd 1981). A major annual variable in sharp-tailed grouse production, and probably in prairie chickens and sage grouse as well, is the abundance of nesting cover as it is influenced by soil-moisture conditions in regions of low rainfall.
15.4.3 Nesting success and densities The proportion of young in grouse populations in the fall generally declines as spring numbers increase (Fig. 15.13). This principle of inversity (Errington 1945) is, I believe, commonly a result of density-dependent nest predation. Relevant data to directly evaluate the hypothesis are meager in the grouse literature, but evidence is extensive for other ground-nesting birds that do not defend their nests. Nesting success declined as the spacing of nests diminished in dabbling ducks (cf. Weller 1979, Livezey 1981, Cowardin et al. 1985) and in phasianids (Wagner etal. 1965, Gates 1971, Potts 1980), as well as in shorebirds and passerines (Horn 1968, Krebs 1971, Andersson& Wikund 1978, Page etal. 1983). Tinbergen et al. (1967) conducted the first experiments that demonstrated that dispersion of nests can affect survival from predation. Lack (1968) felt that predation was the driving force in the dispersion of breeding birds in habitats vulnerable to predators. Taylor (1976) has modeled the advantages of the prey's spacing out to avoid predators that search randomly. Changes in clutch size do not sufficiently explain inversity. Spruce and ruffed grouse populations show inverse rates of gain with density (inversity) (Fig. 15.13), but clutch size is relatively constant (Bump et al. 1947, Keppie 1975b). Lack (1954) and Hickey (1955) felt that changes in adult mortality did not explain inversity. Invariably, those grouse populations that had production indexes negatively correlated with density were populations that had nest losses greater than 55 %. Ptarmigan populations seldom show inversity and they commonly have nest losses less than 40% (Table 15.2). The percentage of ruffed grouse and spruce grouse hens with broods (% successful nests) decreased significantly with increased spring numbers (Fig. 15.13,
606
A. T. BERGERUD
Fig. 15.13. Six grouse populations that demonstrated a decline in breeding success as spring densities increased —i.e., inversity. (Data from Bump et al. 1947, Hillman & Jackson 1973, Linde et al. 1978, Boag et al. 1979, Watson & O'Hare 1979, Autenrieth 1981.) The correlation no longer held for spruce grouse when observations were extended to 1981 (Smyth & Boag 1984). In later years, red squirrels (Tamiasciurus hudsonicus), which probably do not hunt in a density-dependent manner, were important predators (Boag et al. 1983).
Bump et al. 1947, Boag et al. 1979). But the percentage of female blue grouse with young was lowest when birds first colonized newly logged habitats on Vancouver Island, and increased with density, contrary to inversity principles
POPULATION ECOLOGY OF GROUSE
607
(Zwickel & Bendell 1967, Redfield 1972, Zwickel et al. 1977). This contrary trend is consistent with the density-dependent predator hypothesis; because these habitats were newly logged when first colonized, nest cover increased coincident with population growth and increased density. The annual nesting success of prairie grouse also declined in three populations as spring numbers increased (Fig. 15.14). The decline was more rapid at high than low densities for the two sharptail populations. For sharptails, the mean proportion of successful nesting hens was 55 + 4% (n — 18) in North Dakota and 66 + 3 % (n = 28) in South Dakota; for prairie chickens in South Dakota, the mean proportion was 64 + 6% (n = 19). All three populations generally increased after a year of above-average success and declined when hens were less successful (Fig. 15.14). Another generalization was that populations that showed inversity had large nest losses from canids. Examples are spruce grouse at Gorge Creek, Alberta, with 30% nesting success and coyotes (Canis latrans) the suspected predator (Keppie & Herzog 1978), and ruffed grouse in New York, with 62% success and red fox the major predator. Bump et al. (1947, p. 314) said, "There is some evidence that when high breeding density exists certain animals, particularly the fox, often happen on enough nests to associate the flushing of the bird with the possibility of dining on omelet." If predators of nests—particularly canids—cause inversity, then areas without effective foxes and coyotes should not show inversity. The red grouse populations are intensively managed in Scotland and fox populations are reduced. The red grouse population shows little inversity relative to summer gains and density (Jenkins et al. 1963,1967). In Ireland, however, where nesting success is approximately 20% lower and foxes are more common, the summer gains of red grouse were inverse to density (r = - 0.989, n = 4; from Watson & O'Hare 1979). In summer, the ruffed grouse population at Connecticut Hill, New York, demonstrated a nesting success that was negatively correlated with breeding numbers (Fig. 15.13), but when foxes and other predators were removed during a 4-year study this summer rate-of-gain was not correlated with density (r = - 0.146, n = 4; from Bump et al. 1947, p. 349). The number of juvenile blue grouse in the harvest was inverse to adults in Washington—where foxes are a potential predator-but on Vancouver Island, where foxes are rare, inversity has not been documented in blue grouse (Zwickel & Bendell 1967, Redfield 1973, Zwickel et al. 1977). The forest grouse can benefit the most, in terms of nesting success, by dispersing under canopies with nest backstops. In steppe habitats, overhead cover is paramount for these species and may compromise spacing between nests in patchy cover, thereby facilitating searching images in predators and leading to reduced nesting success. Inversity should be more apparent in steppe species when populations are high in dry years, when hens are forced to nest closer together.
608
A. T. BERGERUD
Fig. 15.14. Left: In three prairie grouse populations, the proportion of successful nesting females was correlated with density (% successful hens calculated by: [(juveniles/adult x 2) -=- mean brood size]). Right: The proportion of successful females was also correlated with changes in population size. (Data from Hillman & Jackson 1973, Kobriger 1975, 1981, Linde et al. 1978.)
POPULATION ECOLOGY OF GROUSE
609
The percentage of yearlings in a populations will affect overall nesting success since yearlings are less successful than adults. As populations increase in density, there will be a greater proportion of yearlings and their productivity will contribute to an inverse relationship between breeding numbers and summer rate-ofgain. This influence is greatest in the steppe grouse where the success rates of yearlings and adults are most disparate (Table 15.4). In summary, two explanations for inversity in grouse are: (1) nesting success decreases with increasing spring density because the nests are closer together and attract more predator attention and (2) nesting success declines when the proportion of the population comprised of yearlings increases. The first explanation of reduced space between nests seems a dominant factor in forest grouse and the second explanation based on naive yearlings is more appropriate for steppe grouse. However, annual changes in nesting success correlated with density are the common explanations of inversity in both grouse groups. 15.5 Mortality of chicks Chicks in their first summer die at high rates and the survivors represent production. When the survivors are expressed at the autumn proportion of juveniles to adults (J:A) the parameter is termed breeding success (Jenkins et al. 1963) or the production ratio (Bump et al. 1947). To understand breeding success, the mortality of chicks must be quantified and the mortality factors determined. I estimated chick mortality by comparing the mean number of eggs per completed clutch with the mean number of chicks in broods at about 2 months of age. This index ignores the fact that some eggs in successful nests do not hatch and that some females successfully hatch their eggs but lose all the chicks. Infertility of eggs is quite constant among species, and represents approximately 8% of all eggs (Johnsgard 1973). Failure to quantify the loss of entire broods means chick mortality is underrepresented by this method.
15.5.1 Mortality rates The mean mortality rate of chicks between hatch and autumn for the nine species is 44% (28 studies) (see Fig. 15.1). There is no significant difference among the three groups of grouse: forest grouse incur 46.6% + 3.7% mortality; steppe grouse 40.5% + 1.9%; and tundra grouse 40.3% ± 3.3% mortality. The only major differences I have found in mortality of chicks among species within the groups were in spruce grouse (37% + 2.7%) compared with blue and ruffed grouse (56% ± 6.8%). Within the steppe grouse, prairie chickens appear to have higher mortality rates than either sharp-tailed or sage grouse. The only prairie chicken mortality estimate less than 40% was from Michigan (0.27), based on small samples of clutches and broods (Ammann 1957). In comparison, only 2 of 10 studies of sharp-tailed and sage grouse showed chick losses greater than 40%.
610
A. T. BERGERUD
15.5.2 Mortality factors The common causes of death of chicks mentioned in the literature are: (1) chilling from weather, (2) predation, and (3) starvation. There is no consensus on the importance of these extrinsic factors. An acceptable explanation must account for the high early loss of chicks that universally occurs before 3 weeks of age (see Fig. 1.3), during the period chicks cannot thermoregulate. Some examples are blue grouse hens, which lost 40% of their charges within 2 weeks of hatching (Zwickel & Bendell 1967); spruce grouse hens had about a 40% loss of chicks when counted at about 40 days posthatch (Keppie 1975b). White-tailed ptarmigan broods were reduced 20%-44% by 2 weeks of age (Choate 1963a, Braun & Rogers 1971). One problem in determining the causes of death is that few dead chicks are found. In the 13-year, intensive study of ruffed grouse in New York, biologists found only 83 chicks of approximately 4,300 (2%) that died (data from Bump et al. 1947). A second problem in determining the cause of mortality is that many of the decimating agents are interrelated. Cold weather may cause chilling (Myhre et al. 1975), but it may also prevent chicks from feeding while they are brooded (e.g., Boggs et al. 1977, Erikstad & Spids0 1982), or it can affect plant growth for food or escape requisites and the availability of insects. Thus for all three grouse groups, the evidence directly linking weather and substantial losses is tenuous, especially when we consider that females should be able to brood chicks and insulate them from the extrinsic environment. Cold spring weather is a common explanation for early mortality before chicks can thermoregulate, especially for ptarmigan (Slagsvold 1975, Marcstrom & Hoglund 1981, Choate 1963a). Dead ptarmigan chicks were found after inclement weather in Alaska and Iceland (Weeden 1965, Gardarsson 1971). But other workers have found no correlation between weather and mortality (Myrberget 1972, Bergerud 1970a, Watson 1965, Weeden & Theberge 1972, Jenkins et al. 1963, and Braun & Rogers 1971). Theberge & West (1973) discounted weather as a factor in mortality for rock ptarmigan young, noting that young chicks need to feed only 6 minutes per hour and can thus be brooded by their mothers at least 90% of the time. Complicating the argument for ptarmigan is that female fidelity in brooding and distraction behavior may vary between hens and cohorts, possibly with chick behavior, independently of weather vicissitudes (Jenkins et al. 1963, Mercer 1967). Also confounding this is evidence that the intrinsic viability of ptarmigan chicks may vary between years (Jenkins et al. 1963, Theberge & Bendell 1980, Bergerud 1970a, Watson & Moss 1980). A further problem is that cool weather could make insects more sluggish and easier to capture, therein compensating for the greater need for brooding (P. Hudson pers. comm.) The role of unfavorable weather in causing deaths in forest grouse is also argued. Zwickel (1967) felt that cold, wet weather was not a factor in the deaths
POPULATION ECOLOGY OF GROUSE
611
of young, blue grouse chicks. Storms were thought to result in the loss of some ruffed grouse chicks, but predators and other "undetermined causes" were considered more important (Bump et al. 1947). Survival of ruffed grouse and spruce grouse chicks was higher in warm springs than in cold ones (Larsen & Lahey 1958, Ritcey & Edwards 1963, Dorney & Kabat 1960, Robinson 1980, Fig. 3.8). The steppe grouse actually showed improved survival of chicks in wet springs, which is especially damaging to the hypothesis that weather causes chick mortality in all cases (Fig. 15.11). Again, heavy rain may cause direct losses (Svedarsky 1979). Predators certainly take large numbers of chicks; mortality that occurs beyond the thermoregulating interval can possibly be ascribed to predation, but ground predators would be most serious before chicks can fly and thermoregulate. Small chicks are commonly together and thus this period of vulnerability to predators coincides with the known and universal heavy loss of chicks prior to 3 weeks of age. Chicks are killed mostly by raptors and foxes. Bump et al. (1947, p. 335) listed 46 of 64 ruffed grouse chicks taken by raptors and 16 of 64 taken by foxes. A crow (Corvus brachyrhynchos) and weasel each took one chick. A fox disturbed a prairie chicken hen that was brooding her chicks, and the brood hen was lost. The chicks then died from chilling (Svedarsky 1979). Dorney and Kabat (1960) reported larger broods of ruffed grouse in areas of Wisconsin where foxes are less common. On Vancouver Island, where foxes are rare, few blue grouse chicks disappeared after 2 weeks of age (Zwickel & Bendell 1967). Juveniles/adult ratios were correlated with the adult sex ratio in several studies (Fig. 15.12, Dorney & Kabat 1960). The unbalanced sex ratios probably result from the loss of females to predation. The correlation between fewer females and fewer chicks could be explained by the predation of females, resulting in the death of both the hen and the entire clutch. Predators may kill nesting or brood hens, and the eggs or brood will die from exposure; when predation is heavy, females and chicks may be killed independently. These correlations between sex ratios and breeding success suggest that predation can be a sufficient cause of changing survival of chicks. If predation was a serious factor, one might predict that brood mortality would increase with clutch size. Large broods would be more conspicuous and more difficult to defend. But the mortality of chicks was not correlated with clutch size (Fig. 15.1). A large sample of brood sizes for sharp-tailed grouse in North Dakota was normally distributed (Kobriger 1981). Chicks from larger-thanaverage clutches survived as well as chicks from smaller clutches. A female can expect to raise one additional chick if she can afford the cost of two additional eggs (Fig. 15.1). Some studies have been conducted in areas where predators have been reduced; others have been conducted on islands where there were few predators.
612
A. T. BERGERUD
In all cases a considerable percentage of chicks disappeared between hatching and the fall (Fig. 15.15). Further, I regressed summer mortality of chicks against nest predation (%) for 34 studies and found no correlation (r = 0.268); there was also no correlation between chick mortality and increasing latitude and reduced predator diversity (r — 0.184, n = 34). Predation of chicks is a sufficient but not necessary explanation of the high, constant mortality of chicks in the first 3 weeks of life. Many authors must feel that grouse chicks starve—students commonly measure the abundance of arthropods. Bump et al. (1947) felt that invertebrates for chicks were superabundant at about 75 /m 2 . Barrett (1970) was more concerned about a shortage of insects for ruffed grouse even though he measured 210 invertebrates/m2 in habitats where young broods traveled. If females seek to nest as early as inconspicuousness permits (Chap. 14), it does not necessarily follow that chicks would hatch coincident with a flush of chick food (cf. Elliot 1979). But the vagaries in insect foods may be a major influence on chick survival. The growth rate of chicks has been shown to vary between years for willow ptarmigan (Mercer 1967, Myrberget et al. 1977, Erikstad 1985), blue grouse (Redfield 1978), and spruce grouse (Quinn & Keppie 1981). These differences in growth rates appear to relate more to posthatch weather than to prehatch factors, i.e., female nutrition (Myrberget et al. 1977, Redfield 1978, Quinn & Keppie 1981). Temperature can be expected to influence insect abundance as well as the competition of thermoregulatory processes for metabolizing energy (Ricklefs 1972). When insects were scarce for ruffed grouse and ptarmigan, the broods moved farther (Barrett 1970, Erikstad 1978), and brood survival decreased with increased mobility, in both ptarmigan and blue grouse (Armleder 1980, Erikstad 1978). Wet, cool springs may provide more cover and insects for the steppe grouse; warm, dry springs may do likewise for forest grouse living in more shaded, mesic communities. I found in my blue grouse study (Chap. 2) that chicks survived less well on Portland Island, where there was thick undergrowth, than on Moresby Island, where logging had occurred. Blue grouse populations generally declined as canopies closed in (Redfield et al. 1970), and brood sizes were generally lower at Middle Quinsam than at Comox Burn, when Middle Quinsam was at a late stage of succession (see section 15.7.5). Armleder (1980) documented fewer insects in blue grouse habitats as forest canopies increased. The detailed laboratory studies of J0rgensen and Blix (1985) of ptarmigan chicks explain much of the weather-food paradox in chick survival. They showed that chicks had the same browsing time whether the temperature was 2°C or 12°C, and this browsing time exceeded the lower critical minimum of 6 min/hr (Theberge & West 1973). Chicks always tried to fill their crops, and thus they were unable to compensate for increased energy expenditure and a reduced food quality in a spring with late phenology and reduced insect abundance. J0rgensen
POPULATION ECOLOGY OF GROUSE
613
Fig. 15.15. The survival of chicks in four populations where predators were reduced was similar to that in four closely related populations where predators were left undisturbed (Bump et al. 1947, Jenkins et al. 1963, Mercer 1967, Bergerud 1970a, Zwickel et al. 1977, Watson & O'Hare 1979, Chap. 2).
614
A. T. BERGERUD
and Blix concluded that the availability of high-quality food was critical for growth and survival, and low ambient temperatures, if they did not affect food quality and intake, were of little consequence. The reduction of insect populations has occurred following chemical spraying programs. The English workers (Southwood & Cross 1969, Potts 1980) saw a clear connection between production in Hungarian partridge and DDT. Early survival was reduced in ruffed grouse broods in areas in New Brunswick that were sprayed for spruce budworm, compared to unsprayed areas in control blocks (Neave & Wright 1969). Grouse should have evolved physiological and life-history schedules to hurry through critical life-history stages (Williams 1966a). The need for a high-protein diet for thermoregulation and rapid growth is not liable to compromise (Wise 1982). Variations in weather, maternal behavior, and intrinsic viability may all influence survival, but the abundance of insect supplies or other high-quality food in good cover may be the necessary pacemaker to successfully pass through the critical, first 2 weeks of life.
15.5.3 Density-dependence of mortality If chick mortality was density-dependent, it would help to buffer population changes. Another possibility for density-dependent mortality is that yearlings may be less successful mothers. In high populations there would be more yearlings, who might raise fewer chicks per female; however, the difference in brood size noted between yearlings and adults appears explicable on the basis of differences in clutch size, rather than different brood survival (Zwickel et al. 1977, Keppie 1975b). One explanation for density-dependent mortality of chicks relates to the spacing of nests. At least in spruce and blue grouse, females space their nests; at high numbers more females may be in more marginal habitats (Sopuck 1979) and may need to take their broods farther to find suitable insect and cover habitats. Yearling blue grouse females travel farther initially with their broods than do adults (Sopuck 1979, Armleder 1980). Increased movement might stress the chicks, or it might result in contact with more predators (Fig. 14.15). Blue grouse had increased mortality with increased movement and large home ranges (Armleder 1980). This explanation for a relationship between density and chick survival would seem most applicable to forest grouse, and least to steppe grouse, which may nest near each other and often take their chicks long distances (Schiller 1973, Svedarsky 1979). Chick mortality is not correlated with breeding densities in rock ptarmigan (r = -0.224, n = 11; Weeden & Theberge 1972); willow ptarmigan (r = 0.105, n = 9; Bergerud 1970a) or ruffed grouse (r = 0.452, n = 13; Bump et al. 1947, pp. 527, 539). Also, there is no correlation between size of the brood and density
POPULATION ECOLOGY OF GROUSE
615
in white-tailed ptarmigan (r = - 0.406, n = 18; Braun & Rogers 1971); spruce grouse (r = -0.380, n = 14; Smyth & Boag 1984); or prairie chickens (r = 0.009, n = 18; Sisson 1976, Robertson 1979). Nor is there a correlation for sharp-tailed grouse in Nebraska (r = 0.209, n = 18; Sisson 1976, Robertson 1979, Mitchell & Vodehnal 1982); North Dakota (r = 0.275, n = 18; Kobriger 1975, and government reports); or South Dakota (r = 0.411, n = 24; Hillman & Jackson 1973, Rice & Carter 1976, Linde et al. 1978). These data provide little evidence for density-dependent mortality of chicks. 15.6 Mortality of juveniles and adults This discussion to this point has centered on parameters that affect breeding success: (1) percentage of hens nesting, (2) clutch size, (3) nesting success, and (4) survival of chicks in their first summer. All contribute to the abundance of the new generation and recruitment (mA). These additions are balanced in stable populations by the deaths of juveniles—birds from 4 to 10 months of age—and of adults. This section focuses on mortality rates of juvenile and adult birds (qx). I emphasize mortality rather than survival rates (sx), because the factors that cause death are more tractable than those which contribute to the absence of death and continued survival.
15.6.1 Mortality factors Grouse die from many causes, but predation accounts for 85% of reported mortalities (Table 15.5). The importance of predation has been further documented in numerous telemetry investigations. There may be biases —a bird that is weak from disease or starvation may be more susceptible to predation, or a bird with a radio harness may not be the equal of a bird without. Still, many workers have argued that results from radio tracking are representative—at least after initial losses — and one can always compare mortality rates between tagged and untagged birds to evaluate these assertions. Accidents are the second most common cause of mortality. These fatalities are usually the direct intrusion of humans into the birds' environment with power lines, fences, highways, and other constructions. No cause of death was determined for a large class of dead birds. The birds may have died from disease—generally such birds are not emaciated. Documented cases of starvation are exceedingly rare. Apparently red grouse die from both disease and starvation—females have even been found dead on the nest (Jenkins et al. 1963, 1967), but these red grouse populations live in extremely dense populations where the natural predators have been reduced. These results cannot be extrapolated to more natural populatons with normal assortments and densities of predators.
616
A. T. BERGERUD
Table 15.5. Causes of mortality of North American grouse (excluding hunting and accidents from research)
Grouse groups and species
Sample size of birds
Tundra grouse Willow ptarmigan
50
Rock ptarmigan
278
Rock ptarmigan White-tailed ptarmigan Forest grouse Ruffed grouse Ruffed grouse Ruffed grouse Ruffed grouse Ruffed grouse Ruffed grouse Spruce grouse Spruce grouse Blue grouse
9 44
% of mortality Raptor Mammal Unknown Accidents predation predation predation
mostly a few by foxes gyrfalcon 96% gryfalcon and foxes _ 100 2 59
642 186 72 107 13 26 19 18 77
70 72 79 73 77 54 95 39 57
2 12 10 2 23 12 5 4 4
Steppe grouse Sharp-tailed grouse Prairie chicken Prairie chicken Sage grouse
28 10 20 10
4 30 16 10
50 70 43 30
Approximate means
1,591
58
27
—
0
4
Other
—
-
4
-
-
9
11
-
3 8 4 2 35 6 -
3 3 4
40 40
20
5 -
-
10
5
8 1
5
-
4
35
46
Sources: Willow ptarmigan — Mossop Chap. 10; Rock ptarmigan (278)—Weeden 1965; Rock ptarmigan (9)-Gardasson 1971; White-tailed ptarmigan-Braun and Rogers \91\\Ruffedgrouse (642)Bump et al. 1947; Ruffed grouse (186) — Eng & Gullion 1962; Ruffed grouse (72) — Huempfner pers. comm.; Ruffed grouse (707)-Rusch & Keith 1971b; Ruffed grouse (13)-Ma\son 1974; Ruffed grouse (26) — Hager 1954; Spruce grouse (19) — Ellison 1971, 1974; Spruce grouse (18) — Robinson 1980; Blue grouse-Hmes 1986a; Sharp-tailed grouse-Earl et al. 1950; Prairie chicken (10)Svedarski 1979; Prairie chicken (20)— Schwartz 1945; Sage grouse—Nelson 1955.
15.6.2 Mortality rates I calculated the mortality rate of juvenile birds 4 to 10 months of age in populations based on: (1) band returns, (2) counts of juveniles in the fall versus yearlings in the spring, and (3) by combining adult mortality and the fall breeding success (J/A) needed to stabilize numbers. The formula was [1.00 - (annual adult mor-
POPULATION ECOLOGY OF GROUSE
617
tality/juvenile per adult needed to stabilize breeding numbers)]. To illustrate method 2, the rock ptarmigan at Eagle Creek had an average of 395 juveniles/year from August 1960 to 1970 and 136 yearlings the subsequent springs. J<jx = (395 - 136)/395 = 0.66. Based on method 3, the juvenile mortality rate was 0.69; the number of August young/adult needed to maintain breeding numbers was 1.87, and the adult annual mortality rate was 0.58: [1.00 - (0.58/1.87)]. All three mortality methods are based on the assumption that birds still alive the next spring will return and be identified. This assumption is probably violated in some studies of open populations. The mortality rate of adults was calculated (1) from yearlings/adult ratios that result in stable populations (Hickey 1955), (2) based on band recoveries (lifetable analysis), or (3) comparing the number of adults and yearlings in year 1 with number of adults in year 2. None of the methods permit a distinction between natural and hunting deaths. The adult mortality calculations are more robust than the juvenile rates since yearlings and adults that have bred once can be expected to return in later years if still alive (Chap. 14). The mortality rates for juvenile birds over winter averaged 50%-70% (Table 15.6) and were generally about 15% greater than adult mortality rates. Juvenile birds are naive when broods disband and should be more conspicuous than adults in seeking winter cover, joining winter flocks, and finding their first breeding locations. Also juveniles may be more vulnerable than adults to fall hunting. This is especially relevant to ruffed grouse (Dorney 1963, Rusch & Keith 1971b). The most unusual mortality rate for juveniles was a low 13 percent winter loss of spruce grouse at Gorge Creek, Alberta (Table 15.6). This mortality rate was based both on banded birds (the survival of 123 of 141 juveniles in three winters (Keppie 1979) and on the stabilizing recruitment method. By the latter method, adult mortality equaled 32.5 (Boag et al. 1979) and stabilizing recruitment for the years 1966-73 was 0.37; hence [1.00 - (0.325/0.37)] the juvenile mortality was 0.12. Conversely, this population had one of the highest known losses of nests for forest grouse, 45 of 64 nests (70%) (Redmond et al. 1982). The population was able to persist in the face of this high nest loss because of the high survival of juveniles and adults. Mortality rates for adult grouse ranged from about 25% to over 75% (Fig. 15.4). Unlike with egg losses there was no significant difference in the mean mortality rates of grouse between steppe, forest, and tundra groups, and all the grouse face similar assortments of predators (Fig. 15.8). Hunted populations had higher losses than unhunted populations (Fig. 15.4) suggesting that hunting was additive (see section 16.3).
15.6.3 Bimodalism of mortality rates The mortality rates of grouse fall into two distinct modes: a low mortality mode of less than 45% and a high mode of greater than 45% (Fig. 15.4, Table 15.6).
618
A. T. BERGERUD
Table 15.6. Comparison of clutch size and mortality rates for grouse Mortality rates
Total qx of new to = loss of
Clutch size
Nest
Chick
Tundra grouse: White-tailed ptarmigan (Colo.) White-tailed ptarmigan (Mont.) Rock ptarmigan (Iceland) Rock ptarmigan (Alaska) Willow ptarmigan (Nfld.) Willow ptarmigan (B.C.)
5.9 5.2 11.0 7.5 10.2 7.3
0.43 0.30 0.08 0.31 0.25 0.12
0.42 0.44 0.35 0.47 0.35 0.33
0.60 0.58 0.81 0.69 0.79 -
0.41 0.44 0.46 0.56 0.50 0.54
0.86a 0.83 0.92 0.85 0.90 0.85
Forest Grouse: Ruffed grouse (Minn.) Ruffed grouse (Alta.) Spruce grouse (Alta.) Spruce grouse (Mich.) Spruce grouse (Alaska) Blue grouse (B.C.)
10.9 11.0 4.8 5.7 7.6 6.4
0.39 0.23 0.71 0.22 0.19 0.45
0.67 0.47 0.33 0.32 0.28 0.47
0.55 0.13 0.71 0.77 0.50
0.66 0.64 0.34 0.55 0.57 0.31
0.89 0.88 0.86 0.81 0.85 0.90
Steppe Grouse: Prairie chicken (Wise.) Sharp-tailed grouse (Mich.) Sharp-tailed grouse (N.B.) Sage grouse (Colo.)
12.0 12.1 12.8 7.0
0.50 0.30 0.46 0.59
0.36 0.39 0.29
0.63 -
0.56 0.70 0.63 0.37
0.91 0.88 0.90 0.89
8.6
0.35
0.38
0.62
0.51
0.88
32%
49%
27%
31%
22%
4%
Grouse group and species
Means Coefficient of Variation a
Juvenile Adult
100 females lay 580 eggs of which 290 are females; 41 adult females die. Total mortality of new females to equal annual loss of adult females (290-41)7290 = 0.86. Sources: Tundra grouse — Choate 1963a,b, Braun 1969, Bergerud 1970a, Braun & Rogers 1971, Gardarsson 1971, Weeden & Theberge 1972, Giesen et al. 1980, Hannon 1982, Mossop pers. comm. Forest grouse-Kupa 1966, Gullion & Marshall 1968, Gullion 1970, Ellison 1974, Zwickel 1975, Zwickel et al. 1977, Keppie 1979, Boag et al. 1979, Robinson 1980, Zwickel et al. 1983, Ruschetal. 1984. Steppe grouse—Hamerstrom 1939, Ammann 1957, Hamerstrom & Hamerstrom 1973, Kohn 1976, Braun 1979, Kobriger 1981, Petersen 1980.
These mortality rates are for populations rather than for species. There are ruffed grouse, spruce grouse, and blue grouse populations with mortality rates below 45 %, and other populations with mortality rates above 45 %. Sage grouse females in Colorado die at a rate < 40%, but in Montana, where the clutch size is larger mortality rates are 60% (Wallestad 1975b, Braun 1979). The two white-tailed ptarmigan populations that have been studied in Colorado and Montana both fall into the low mortality mode (Choate 1963a, Braun & Rogers 1971), but white-
POPULATION ECOLOGY OF GROUSE
619
Fig. 15.16. Comparison of the escape features of three low-risk habitats and three highrisk habitats.
tailed ptarmigan farther north, in British Columbia and Alaska, occupy ranges with gyrfalcons and can be expected to have higher mortality rates. Young whitetailed ptarmigan broods in northern British Columbia averaged seven chicks; hence, clutches would have to be eight or nine, and mortality may be greater than 45% (Fig. 15.2). The low and high mortality modes can be explained by differences in the predation of adults. The low-mortality populations live either where there are few raptors or in habitats where they are relatively safe from these predators (Fig. 15.16). High-mortality populations face more raptors and/or live in less secure habitats. Spruce grouse populations may be in either the high or low mortality mode. Spruce grouse in southwestern Alberta face high predation of nests—70%—but have low mortality rates of adults—less than 33% (Boag et al. 1979, Redmond et al. 1982). Spruce grouse on the Kenai Peninsula, Alaska, lose few nests but the mean longevity of adult birds is short: 276 days versus 1,022 days for Alberta birds (calculated from Ellison 1974, Herzog 1978, Boag et al. 1979). Both populations lose few chicks in the summer-28% and 33%, respectively (Table 15.6).
620
A. T. BERGERUD
Females have a higher mortality rate than males in Alberta, where losses occur mostly in the summer, but males die at faster rates than females in Alaska, where mortality occurs mostly in the fall and winter (Ellison 1974, Keppie 1979). The grouse in Alberta live in a closed canopy of dense pine (Pinus contortd) averaging 3,936 stems per ha (Boag et al. 1979). The weakest link in the demography of Alberta grouse is the nesting and brooding seasons, when they must be on the ground and are exposed to mammal predators. In winter, when they live in conifers in the thick forests, they are almost invulnerable to raptor predation. The Alaska grouse, on the other hand, have few mammals to contest with when on the ground, but the more-open forest canopies leave them exposed to goshawk predation (cf. Ellison 1974). I suggest that it was the dense stems of the forests inhabited by the Alberta grouse rather than the lack of predators that accounted for the superior survival rates of the Alberta grouse over the Alaska birds. Two other spruce grouse populations also have low mortality—birds in Montana (Stoneberg 1967) and Minnesota (Anderson 1973, Haas 1974). Both populations inhabited dense forests. Females in the Minnesota population quickly sought dense forest when they lost their broods. Additional evidence that secure environments can make up for high predation pressure is that adjacent to the Alberta spruce grouse are ruffed and blue grouse populations, both living in more open habitats, and both having mortality rates 17%-24% higher than those of spruce grouse (Boag 1966, 1976b). Adult white-tailed ptarmigan in Colorado and Montana have low mortality rates of 29%-33% (Choate 1963a, Braun 1969, May 1975). All other ptarmigan populations studied have died at annual rates of approximately 50% or more (Watson 1965, Bergerud 1970a, Gardarsson 1971, Weeden & Theberge 1972, Myrberget 1975b, 1976b). The mortality rates of white-tailed ptarmigan can be explained on the basis of low predation from raptors. First, they are hunted mostly by prairie falcons (Falco mexicanus) and golden eagles (Aquila chrysaetos), the very effective predators —gyrfalcons and peregrine falcons (Falcoperegrinus) are absent from these southern habitats. Second, white-tailed ptarmigan frequent rock fields of approximately their body size (Schmidt 1969, May 1975). Their strategy is to squat and crouch adjacent to or beneath such rocks. These ptarmigan are inconspicuous and difficult to detect in such situations. Third, the birds are distributed in widely scattered, small patches on high mountains. This dispersal requires raptors to distribute themselves coarsegrained relative to the patches of birds-they must spend time traveling between patches and cannot afford to be as specialized for ptarmigan as are gyrfalcons in more continuous distributions in the arctic. Finally, whitetails live high in the mountains in heavy cloud cover, which hinders raptors. Sometimes they are safe from raptors for days on end. It is this unique, safe mountain habitat exploited by the white-tails that has provided them with their low, annual mortality rate. Rock, white-tailed, and willow ptarmigan are found in the same mountain
POPULATION ECOLOGY OF GROUSE
621
ranges in northern British Columbia (Weeden 1959b), where gyr falcons hunt them. Even here, white-tailed ptarmigan are less vulnerable to raptors. The gyrfalcons at Chilkat Pass, British Columbia, commonly hunt from perches of approximately 1,200 m in elevation (Mossop pers. comm.). These rock ledges are below the elevation used by whitetails and at the same elevations as the rock ptarmigan (Fig. 15.17). Both species are therefore relatively safe from attack from above. Blue grouse on Vancouver Island are another of the long-lived populations. Goshawks and eagles are the common raptors—blue grouse are likely too large for Cooper's hawks. Blue grouse, like spruce grouse, live in conifers in winter. Birds are able to sit high in dense foliage and eat needles without exposure. As noted in Chapter 14, birds are often close to the main stem and surrounded by branches. The birds do not have to go to the ground to snow roost. In contrast, ruffed grouse visit leafless aspen trees early and late in the day to feed when light is least favorable for raptors, and depend on snow-roosting for cover overnight (Fig. 4.7). In some years there is inadequate snow for roosting, and mortality rates increase. The needle/conifer habitat is responsible for the low mortality rates of some blue and spruce grouse populations compared to ruffed grouse.
Fig. 15.17. Gyrfalcons at Chilkat Pass, British Columbia, commonly hunt from perches at about 1,200-m altitude. Perches are below the habitats frequented by white-tailed ptarmigan and at the same elevation as terrain inhabited by rock ptarmigan. White-tailed and rock ptarmigan are relatively safe because gyrfalcons search below them on the valley bottom for willow ptarmigan. Mountain outline and ptarmigan distribution from Weeden (1959a).
622
A. T. BERGERUD
One ruffed grouse population living along the shores of Puget Sound in Washington shows a low mortality rate. Brewer (1980) provides an annual mortality rate of approximately 23% for 31 banded, ruffed grouse males. In this example, the forest did not provide cover, since it was relatively open and logged (Salo 1978), having perhaps only 500 stems per ha, which is well within the hunting tolerance of goshawks (Reynolds et al. 1982). However, the area lies outside the breeding range of goshawks (Cramp et al. 1980). Thus the low mortality rate results from the lack of predators, and not from effective cover. This predator/cover hypothesis predicts that the survival rates of individuals will be improved if the effective raptor populations are substantially reduced. In Scotland, two raptors of red grouse—goshawks and peregrine falcons—no longer hunt red grouse, and red fox populations have also been lowered on many intensely managed moors. Red grouse lay clutches of approximately 7 eggs (Jenkins et al. 1963) and their predicted natural mortality rate should be about 45% (Fig. 15.2). Consistent with this reduction in predation, the mortality rate of the lightly hunted, Kerlock population (1962-65) was only 39%, calculated from Jenkins et al. (1967). The predation/cover hypothesis also predicts that the grouse in North America living south of the goshawk distribution should have lower mortality rates than those farther north. Male lesser and Attwater's prairie chickens can be expected to have mortality rates less than that reported for the greater prairie chicken in Wisconsin (53%, Hamerstrom & Hamerstrom 1973). The return of Attwater's prairie chicken males to their former leks in Texas was 65%; only 35% were missing (n = 20; Jurries 1981).
15.6.4 Mortality rates and reproductive risk An alternative hypothesis to the predator/cover hypothesis to explain the low and high mortality modes is that these differences arise because females with small clutches have lower reproductive risks than females with large clutches. Females have a high mortality rate during the nesting period and in the summer with broods when they frequent short, open cover. The mortality of chicks in broods is about 40%-50% independently of clutch size (Fig. 15.1); therefore, a female that hatches 6 eggs can expect to lose 3 chicks, whereas a female with 12 will lose 6. If these chicks die from predation, the risk to a hen with 12 chicks might be greater than that to one with a smaller brood. One index of female risk is to compare the observed number of females dying in nesting and brood activities, with the expected number of deaths based on prorating the annual survival of the population (both males and females) to the total number of days that females have been monitored nesting and raising chicks. The reproductive-risk index is the observed deaths of hens divided by the expected deaths. All six polygynous grouse species have been radio-monitored during the summer reproductive season, and the deaths of females are known. The sample is un-
POPULATION ECOLOGY OF GROUSE
623
doubtedly biased by the radio-tracking technique, but this bias should allow a comparison of reproductive risk between the low and high clutch modes. Females of the high clutch mode—prairie chicken, ruffed grouse, and sharp-tailed grouse — were monitored in Minnesota, and all suffered significantly greater mortality during the reproductive period than that expected, based on the annual survival statistics (Table 15.7). In contrast, three populations with small clutches — spruce grouse in Alberta, blue grouse on Vancouver Island, and sage grouse in Colorado—had approximately the same number of deaths per time unit as would be expected based on the annual survival rate (Table 15.7). The correlation between reproductive risk and clutch size is r = 0.943 (n = 6). Is this clutch-mortality correlation one of cause and effect, or are clutch and mortality parameters correlated because they are both causal with a third factor? I have called this third factor the predator-cover complex (Fig. 15.3). The three populations within the high clutch mode live in the steppe or in deciduous forests and have high mortality rates (Table 15.6). Blue and spruce grouse, with the low clutch mode, live in conifer forests and have a low mortality rate. Sage grouse females with small clutches also live in open habitats and have low mortality rates, but the mortality rate of the male sage grouse is similar to that of the species with large clutches (Braun 1979). Sexual dimorphism in sage grouse may deflect predators from less conspicuous females to males. The cause-effect dilemma between clutch size and survival may be evaluated Table 15.7. Number of radio-tracked females observed to die from predation during the breeding season compared with the number expected to die prorating the annual mortality rate to the time period radio-tracked.
Clutch mode and species High mode Prairie chicken Sharp-tailed grouse Ruffed grouse Low mode Spruce grouse Blue grouse Sage grouse a
b c d
Clutch size
Days radio Different tracked3 females b
that died Obs.
Exp. c
Risk (0/E)
13.5 12.5 11.0
1,192 1,018 2,420
21 15 17
lld 8 12
2.5 2.3 5.0
4.4 3.5 2.4
4.7 6.4 7.0
1,718 4,661 2,361
29 49 42
2 6 2
2.0 4.6 2.8
1.0 1.3 0.7
Major references for radio tracking: Schiller 1973, Haas 1974, Maxson 1974, Herzog 1977a, 1979, Sopuck 1979, Svedarsky 1979, Petersen 1980. Some radio-tracked females excluded from sample in most studies for a variety of reasons. Expected calculated (total days tracked/ [-0.4343/log. sx] [365]). Figures are conservative because many terminations of radio-tracked females were probably caused by predation but were not considered dead unless radio or carcass was found.
624
A. T. BERGERUD
by looking at the mortality rates of males, which in these populations do not assist females in their reproductive effort. If mortality rates of males are vastly different than those of females, female reproductive risk could be the causal factor in the differences in the low and high modes; but if male mortality parallels that of females, the predator-cover complex, which they both share, can be hypothesized to be the causal factor. The annual mortality rates for the low clutch populations are: (1) spruce grouse in Alberta, males 28% versus females 37% (Keppie 1975b, Boag et al. 1979); (2) blue grouse on Moresby Island, males 32% versus females 26%; at Middle Quinsan, Vancouver Island, males 31% versus females 33%; at Comox Burn, Vancouver Island, males 25% versus females 29% (Chap. 2, Boag et al. 1979, Robinson 1980, Zwickel et al. 1983); and (3) sage grouse in Colorado, males 57% and females 37% (Braun 1979). For the high clutch mode the mortality rates for prairie chickens in Wisconsin are males 53% versus females 56% (Hamerstrom & Hamerstrom 1973). There are no mortality figures for ruffed grouse females; males commonly have annual mortality rates of approximately 50% ± 5 % (Table 3.2, Dorney & Kabat 1960, Gullion & Marshall 1968, Rusch et al. 1984). The adult sex ratio in ruffed grouse is generally 55 % males and 45 % females (Table 15.8), or a mortality rate for females of approximately 56%. There are no natural mortality rates for unhunted, sharp-tailed grouse populations. The proportion of adult males to females in the harvest is commonly 55:45 (Table 15.8). Male and female sharp-tailed grouse banded in South Dakota had similar mortality rates (Robel et al. 1972). The difference in the mortality rates of males between the low and high clutch modes is about 20% (35% versus 55%), but within populations, there is little difference in mortality between males and females. Only in the sage grouse is the difference between male and female mortality rates (20%) similar to the difference between the low and high mortality rates of females (20%). This lack of difference in mortality between males and females supports the view that the greater mortality rate of grouse with large clutches compared to those with smaller clutches is not cause and effect. The greater mortality of birds with larger clutches likely results from the less secure, predator-cover complex of their habitat.
15.6.5 Differential mortality of males and females Parental investment differs between males and females. Males advertise for females and invest little in care of the young, but females nest and defend their young, and direct their movement. Hence, the two sexes have different mortality regimes and there is no theoretical reason for equal proportions of male and female adults in a population (cf. Angelstam 1984). On Moresby Island, British Columbia, the male:female ratio in 1975 of the blue grouse raised on the island
POPULATION ECOLOGY OF GROUSE
625
Table 15.8. Proportion of males in harvested collections and birds trapped in the winter (sample size in parenthesis) % males Grouse groups and species (place) Tundra grouse Willow ptarmigan (Nfld.) Willow ptarmigan (Fin.) Rock ptarmigan (Ice.) White-tailed ptarmigan (Col.) Unweighed means Forest grouse Blue grouse (B.C.) Blue grouse (Wash.) Spruce grouse (Ont.) Ruffed grouse (Mass.) Ruffed grouse (Ohio) Ruffed Ruffed Ruffed Ruffed
grouse grouse grouse grouse
(N.D.) (B.C.) (Alb.) (Alb.)
Unweighed means Steppe grouse Sharp-tailed grouse (Alb.) Sharp-tailed grouse (N.D.) Sharp-tailed grouse (S.D.) Sharp-tailed grouse (Mich.) Prairie chicken (Kan.) Prairie chicken (Wis.)
a
b
Juveniles
Adults
51 (1633) 51 (482) 48 (14943) (66) 45
55 (917) 62 (595) 47 (5009) 65 (233)
49
57
46 47 49 51 51 49 53 51 53 51
(2502) (302) (766) (680) (12971) (2972) (2496) (1753) (1000) (341)
50
47+ (303) 55 (423) 57 (600) 55 (5365) 59 (2562) 55 (1265) 60 (813) 59 (264) 53 (793)
References
Bergerud 1970a Pulliainen 1975 Gardarsson 1971 Braun & Rogers 1971
Zwickel et al. 1975 Zwickel & Brigham 1970 Lumsden & Weeden 1963 Hager 1954 Dorney 1963 Davis & Stoll 1973 Schulz 1981 Chap. 3 Rusch& Keith 1971 b Hilton & Wishart 1981
56
50 49 54 56 55 62
(2623) (28777) (5324) (2018) (306) (950)
47 (1781) 51 (16693) 52 (2263) 60 (889) 55 (298) 63 (740)
Prairie chicken (Okl.) Sage grouse (Wyo.) Sage grouse (Ida.) Sage grouse (Wyo.)a Sage grouse (Colo.)3
53 46 47 41 47
(491) (2693) (67103) (6451) (5775)
47 30 34 25 34
Unweighed meansb
51
(532) (1964) (43028) (4358) (4588)
Hilton & Wishart 1981 Kobriger 1981 Robel et al. 1972 Ammann 1957 Baker 1953 Hamerstrom & Hamerstrom 1973 Lee 1950 Patterson 1952 Autenrieth 1981 Game Dept. Braun 1979, Braun & Hoffman 1979
54
Adults include yearlings in which sex was known: Wyoming yearlings 28% male and adults 20% male, Colorado yearlings 40% male and adults 28% male. Excludes sage grouse.
626
A. T. BERGERUD
was 43:57 (n = 76), when the entire population was classified. In this insular situation, nearly all females successfully raised young (Chap. 2) because of the absence of ground predators, and females had a low mortality of 26%. Males, on the other hand, showed normal mortality rates of 32% (Fig. 2.6)-red-tailed hawks (Buteo jamaicensis) and occasional goshawks apparently took some males when they advertised and occasionally both males and females in the winter. Unbalanced sex ratios can provide insights into the mortality agents at work. In general, males predominate in the harvests of white-tailed and willow ptarmigan, sharp-tailed grouse, prairie chickens, and ruffed grouse (Lee 1950, Baker 1953, Ammann 1957, Dorney & Kabat 1960, Bergerud 1970a, Table 15.8). Males predominate also in large samples of winter-trapped, adult sharp-tailed grouse (Robel et al. 1972) and prairie chickens (Hamerstrom & Hamerstrom 1973). But females are more common in sage grouse, as shown both by band returns and observation of birds in the winter (Dalke et al. 1963, Beck 1977, Braun 1979, Table 15.8). There is no consistent male or female predominance in either spruce or blue grouse (Lumsden & Weeden 1963, Zwickel & Bendell 1967, Zwickel 1972, Ellison 1974, Zwickel & Brigham 1975, Zwickel et al. 1975, Keppie 1979). The adult sex ratio does not necessarily only reflect differences in the mortality rates of males and females resulting from their different investment strategies. The adult sex ratio could also be modified by the sex ratio of the new recruits. In 10-year-cyclic populations of ptarmigan there is some evidence of fewer juvenile females than males during high populations and/or during declines. Gardarsson (1971) reported the sex ratio of rock ptarmigan chicks 4-25 days old as 56% male. There was a male:female ratio of 62:38 in 93 willow ptarmigan chicks collected in Newfoundland (Bergerud 1970a). The proportion of male rock ptarmigan yearlings at Eagle Creek, Alaska, decreased as breeding success improved; the correlation of percent male yearlings and breeding success the prior season was r = -0.574, F < 0 . 1 0 , n = 9 (Weeden pers. comm.). The proportion of males in the breeding population in Scottish rock ptarmigan also decreased as breeding success the prior year improved, r = -0.718, P < 0.01, n = 18 (data from Watson 1965). These combined findings suggest that cyclic ptarmigan populations may have a higher proportion of male than female chicks when there is high summer mortality of young—female chicks may be intrinsically weaker than males (Bergerud 1970a). Several forest grouse populations show little deviation from unity in the proportion of juvenile males and females (Table 15.8), nor were the adult male:female proportions significantly correlated with the percentages of juvenile males (r = 0.554, n = 9, Table 15.8). Thus, here, adult departures from unity will need an explanation apart from the influence of the survival of juveniles. For the prairie grouse there are more juvenile males than females in the harvest in most states (Table 15.8). Samples secured from winter trappings are also
POPULATION ECOLOGY OF GROUSE
627
skewed to males. Males could be more vulnerable than females if they prospected for advertising sites at leks coincident with hunting and traveled widely between leks in small groups. However, the bias would not explain the predominance of males in the winter trapping. The correlation between the percent juvenile males and adult males was r = 0.852, n = 1. The possibility remains that male chicks in prairie grouse may be more viable than females, as postulated for ptarmigan. The sex ratio of both adults and juveniles in sage grouse is weighted to females (Table 15.8). This disparity is also present in yearlings. These differential mortalities of male juveniles and yearlings represent a significant portion of the unbalanced sex ratio seen in adults, but the large-bodied, adult males can also be expected to die at greater rates than females (see Braun 1979). In dimorphic sage grouse, male chicks may suffer a higher mortality than female chicks, owing to differential growth and energy requirements, a sequence noted for dimorphic Capercaillie (Tetrao urogallus) in Europe (Wegge 1980, Moss & Oswald 1985). The sexual imbalances present in grouse are in part a result of different predation mortality. In general, nest predators decrease in abundance and diversity from south to north (Fig. 15.7), whereas effective raptors increase. The most effective grouse raptors are the gyrfalcon and goshawk, and are found in the north; the golden eagle covers the entire North American grouse range except the high arctic and the east. Ground predators of females should be most serious in the south, especially in the steppe, and predators of adults, especially of conspicuous, displaying males, should be least in the steppe and most in the tundra if the natural predators are still present. All three grouse groups fit these predictions and show an increase in the proportion of females with increasing latitude (decline in ground predators) and a decrease with increased nest predation and declining nesting success (Fig. 15.18). The predation hypothesis also explains the variability of the sex ratio between populations of the same species, for example, for spruce grouse (Table 15.9). Male spruce grouse are more common in Alberta than in Alaska; in Alberta, nest predation is high (Keppie 1980) and 29% of the females disappear in summer, versus 11% in winter (Keppie 1979). Males in Alberta are relatively safe from goshawks because of the high density of lodgepole (Pinus contend) stems. In Michigan predation of males and females is more balanced (Table 15.9). The Yellow Dog plains of Robinson's Michigan study area are on the south edge of the goshawk range—grouse use the more open and less secure, Jack pine coverts (Robinson 1969), but also nesting success is high. The sex ratio is predicted to be nearly balanced. In Alaska nearly all the hens rear chicks —nest predation and predation pressure of females are minimal. But males are commonly taken by goshawks (Ellison 1974). The adult sex ratio there favors females (Table 15.9). The north-south cline in nest-adult predation can be documented further with ptarmigan (Table 15.9). Ptarmigan in Colorado have few raptors that take adults, and mortality is low, but nest predation is the highest recorded for ptarmigan
628
A. T. BERGERUD
Fig. 15.18. Differential mortality of females is associated with predation during the nesting and brood-rearing seasons. Top: The proportion of females in populations increases with latitude because there are fewer mammalian predators of nests farther north. In the Arctic, gyrfalcons take more male ptarmigan, further increasing the proportion of females. Middle: There are fewer females in populations in which many nests are destroyed by predators. Bottom: There are more females in populations in which a high proportion of hens are seen with broods in July and August.
POPULATION ECOLOGY OF GROUSE
629
Table 15.9. Comparison of population parameters between grouse populations subject to different intensities of nest and adult predation3
Population parameters Forest grouse
Clutch size Nesting success Females with young Mortality adult females Mortality adult males Adult sex ratio ( : ) Tundra grouse
Clutch size Nesting success Mortality adult females Mortality adult males Adult sex ratio a
b
Many nest predators but few predators of adults
Predation of nests and adults is similar
Few nest predators but many predators of adults
Spruce grouse, Alberta
Spruce grouse, Michigan
Spruce grouse, Alaska
4.8 (64) 30% (64) 29% (171) 0.33 0.22 55:45
5-7 (9) 78% (9) 72% (117) 0.55 0.50 54:46
7.6 (26) 81% (26) 93% (214) 0.57 0.68 45:55
White-tailed ptarmigan, Colorado
Rock ptarmigan, Michigan
Rock ptarmigan, Iceland
5.9 (60) 57% (60) 0.41 0.27 64:36
7.5 (198) 69% (198) 0.56 0.63 49:51
11.0(301) 92 % b 0.46 0.65 47:53
Boagetal. 1979, Keppie 1979, Redmond etal. 1982, Robinson 1980, Ellison 1974, Braun 1969, May 1975, Weeden & Theberge 1972, pers. comm., Gardarsson 1971. Based on eggs rather than nests.
(Giesen et al. 1980). Females in Colorado have a higher mortality rate than males (Braun 1969). Ptarmigan in Alaska face considerable predation by weasels (Weeden 1965), which are probably more interested in eggs than killing females. Some gyrfalcons are present and displaying males should be selected. Adult males have a higher mortality (63 % versus 56%) at Eagle Creek, Alaska, and are less common than adult females in the spring (49% males, n = 857). Ptarmigan in Iceland nested on an island without mammalian predators but where gyrfalcons selectively hunted males (Chap. 9). Thus, the population included more females than males. (Table 15.9). In Svalbard, there are no gyrfalcons, and the chief predator is the Arctic fox (Alopex lagopus) that searches for nests and chicks. The sex ratio there favors males 56 % (« = 151), since males do not defend nests or broods (Unander & Steen 1985).
15.6.6 Mortality and density If the mortality rate of juveniles and adults is density-dependent, this would dampen population change. Lack (1954, 1966) felt that the regulating factor in
630
A. T. BERGERUD
birds was density-dependent starvation during the winter. I evaluated this hypothesis when possible, by comparing the annual mortality rates of banded birds with spring numbers. However, there were very few data. My second technique was to regress fall numbers against spring numbers the next season. If these plots were linear, they indicated density-independence relationships in overwinter mortality. A curvilinear plot in which spring numbers declined more rapidly as a proportion of fall numbers as the population increased would signify that overwinter mortality was density-dependent. Forest grouse show little annual variation in mortality with density. The mortality rate of banded, adult blue grouse was relatively constant at Middle Quinsam from 1959 to 1976, as the population declined, and at Comox Burn, both when the population first colonized this area from 1962-65, and when numbers increased 1969-77 (Zwickel et al. 1983). The mortality of females at Middle Quinsam increased when the population declined. Forest succession may have played a part in this. The spruce grouse population at Gorge Creek shows a linear plot between fall and spring numbers, not suggestive of density-dependence. The correlation between spring numbers and annual mortality was r = 0.154 (n = 5) (Boag et al. 1979). The spruce grouse of the Yellow Dog plains in Michigan showed a correlation of r = 0.644 (n = 4) between spring numbers and annual mortality (Robinson 1980). Ruffed grouse populations in New York, Minnesota, and Alberta showed significant linear relationships between spring numbers and previous fall numbers; winter mortality was density-independent for the most part (Figs. 15.19, 15.20). Of special interest is the mortality of advertising, ruffed grouse males in relation to density. These displaying males are commonly killed by goshawks near their display sites in the spring (Eng & Gullion 1962), but they are seldom killed on the actual display stage (Meslow 1966, Gullion & Marshall 1968, Rusch & Keith 197la). I think there is competition for these safe sites and that males at high densities may opt to be silent as a survival tactic rather than advertise at a dangerous site (Fig. 15.21, see also Lewis & Zwickel 1982). The regression of mortality on total number of males in the Cloquet population of Minnesota was similar to the regression of mortality of advertising males, suggesting that silence tactics did reduce density-dependent mortality (Fig. 15.21). The mortality of advertising males was positively correlated with density in Minnesota (r = 0.549, n = 20; Fig. 15.21), but in Alberta the mortality of advertising males decreased coincident with increasing density (r = - 0.724, n = 8; Rusch et al. 1984). The mortality rate in Minnesota changed in nonrandom runs as the population cycled (Fig. 15.21). I do not believe that these contrasting mortality correlations are causal based on numbers per se, but that they reflect changes in density-dependent selection for aggressiveness during cycles and increased vulnerability in the years of high numbers—because of both behavior and the switch-over of predators from preferred snowshoe hares (Lepus americanus) to grouse (Chap. 3).
POPULATION ECOLOGY OF GROUSE
631
Fig. 15.19. Six noncyclic populations show a linear plot between numbers in the fall and numbers the next spring, indicating a relatively constant winter mortality and a lack of densitydependence in winter mortality rates (qx). (Data from Bump et al. 1947, Allison 1963, Kobriger 1975, 1981, Linde et al. 1978, Boag et al. 1979, Chap. 2).
632
A. T. BERGERUD
Fig. 15.20. Six cyclic populatons show a linear plot between numbers in the fall and numbers the next spring, indicating a relatively nonvariable winter mortality and a lack of densitydependence in winter mortality rates. (Data from King 1937, Watson 1965, Bergerud 1970a, Gardarsson 1971, Weeden & Theberge 1972, Rusch et al. 1984, pers. comm.)
POPULATION ECOLOGY OF GROUSE
633
Fig. 15.21. Top: Total number of males in spring at the Cloquet ruffed grouse study area in Minnesota, 1959 to 1978. The proportion of silent males increased when the population was high. There were annual runs in mortality rates and high mortality occurred in years when goshawks came south. Bottom left: Mortality rates are correlated with the number of males advertising. The correlation does not improve when regressed against total males, which includes silent males. Bottom right: Changes in the number of males in spring between years are correlated with annual mortality rates. (All data adapted from Gullion 1981.)
634
A. T. BERGERUD
Data to evaluate the density dependence of mortality in steppe grouse are few. There are no long-term studies of banded birds, and workers seldom use a studyarea approach to count birds in the fall and spring. The mortality of female sharptailed grouse in South Dakota and males in Montana was not correlated with the number of males displaying in the spring (South Dakota, r = 0.423, n = 4, 1964-67; Robel et al. 1972) (Montana, r = 0.449, n = 4, 1963-67; Brown 1966b, 1967, 1968b). These species should be able to counter increased susceptibility to predation by forming winter flocks (Chap. 14; see also review by Pulliam & Millikan 1982). Of special interest is the possibility of density-dependent mortality of greater prairie chicken hens while nesting. These hens are frequently ambushed on the nest, and there is reduced breeding success at high numbers (Fig. 15.14). The proportion of female prairie chickens in the South Dakota harvest declined as numbers increased (r = 0.579, n = 16; Linde et al. 1978). To the contrary, the proportion of female sharp-tailed grouse in the harvest was not correlated with density in South Dakota, r = 0.129, n = 20; or North Dakota, r = -0.275, n = 17 (Kobriger 1981, Linde et al. 1978). The nest ambush rate of prairie chicken hens is 9% (n = 185) and for sharp-tailed females only 4% (n = 206) (Table 14.6). This difference may partially account for the differences in the densitydependent mortality of hens between the two prairie grouse species. Mortality rates in tundra grouse were not density-dependent. Willow ptarmigan in Newfoundland had relatively constant, annual losses in 7 years (Bergerud 1970a). Rock ptarmigan in Alaska showed a cyclic pattern in adult mortality (Weeden & Theberge 1972) that was not correlated with density (r = 0.432, n = 9). Nor was there a correlation between adult mortality and density for Myrberget's long study of ptarmigan in Norway (Chap. 11), nor a 7-year study of rock ptarmigan in Iceland (Chap. 9, Gardarsson 1971). 15.7 Theories of population change Populations are constantly changing in size, and population ecologists search for the reproductive and mortality factors that drive these changes. Population size is defined as the number of breeding birds in the spring, including young males that may not actually breed, i.e., birds alive in the spring after the population has spaced itself for reproductive activity. Five common, density-dependent hypotheses explain changes in numbers (Fig. 15.22). (Hi) The threshold-of-security hypothesis specifies that winter cover is inadequate to shelter fall populations; numbers above a threshold are vulnerable to predation or dispersal, and spring populations are more constant than fall populations. (Hz) The winter bottleneck hypothesis specifies that the availability of winter food is variable and commonly in short supply. Birds starve and the magnitude of the loss varies with winter severity, and possibly with numbers.
POPULATION ECOLOGY OF GROUSE
635
Fig. 15.22. Major theories that attempt to explain changes between years in the size of a population of breeding birds.
(Hs) The predator s witch-over hypothesis states that mortality rates are lower and populations increase when predators are rare or buffers (alternative prey) common; when buffers decline, predators switch over, winter mortality increases, and populations decline. (H4) The fourth hypothesis I will call the territorial selfregulation hypothesis, which proposes that breeding space is in short supply. Birds establish territories in the fall or spring, and those unable to successfully
636
A. T. BERGERUD
compete for space are surplus and doomed to die. (Hs) The fifth hypothesis states that breeding success drives spring numbers. Unlike the first four hypotheses, a test implication of this hypothesis is a relatively constant, winter mortality rate (Fig. 15.22). The first four hypotheses predict variable winter-mortality rates dependent upon the requisites postulated in short supply: Hi—cover; HI —food; HS—buffers; H4 —space. Population changes in HS are mostly explained by changes in nesting success and chick survival between years, dependent on variations in cover, predators, and weather.
15.7.1 Threshold-of-security The threshold-of-security hypothesis dates to Errington's (1945) study of bobwhite quail at Prairie du Sac, Wisconsin, 1929-44. He noted that spring populations were generally constant and hypothesized that habitats had well-fixed capacities to protect animals; individuals in excess of this capacity were not safe from predators and died or disappeared (for ruffed grouse, cf. Errington 1937b). Romesburg (1981) felt that the hypothesis was nearly impossible to test. One test implication of this hypothesis is that spring populations should be relatively constant in size compared with variable fall numbers. This implication does not hold for grouse because spring numbers parallel fall numbers (Figs. 15.19, 15.20) and often can change tenfold between years (Keith 1963, Schulz 1981, see Fig. 15.30). I believe that the constancy in numbers that Errington saw was a function of the size of the window he used to view the population. If one picks a small study area of optimal habitat, numbers are often relatively constant (cf. Theberge & Gauthier 1982). Birds will always be attracted, regardless of total numbers, as they space themselves to maximize fitness (Fig. 14.27). Self-contained populations do fluctuate between years (Mercer 1967). When large, regional areas are censused, fluctuations are evident (see Fig. 15.30). Radiotelemetry has shown that dispersal for breeding occurs in all grouse and is the means of spacing. This spacing dampens fluctuations in small, optimal study areas.
15.7.2 Winter bottleneck David Lack (1954) argued in favor of the view that birds face a density-dependent shortage of winter food. Many students then took to the field to test the hypothesis, but even the best examples that Lack marshaled in his final evaluation (Lack 1966) are unconvincing (cf. Chitty 1967). It is really difficult to envisage blue or spruce grouse living in a sea of green conifer needles faced with a food problem. Blue grouse apparently gain weight in winter (Redfield 1973b), and spruce grouse show a minimal decline over winter (Ellison & Weeden 1979, Robinson 1980). Also, both species have constant, winter mortality rates (Boag et al. 1979, Zwickel et al. 1983, Hines 1986). They have no food problem. The sage grouse
POPULATION ECOLOGY OF GROUSE
637
also gains weight in winter (Beck & Braun 1978), and it lives in a sea of sagebrush leaves. If snow accumulates, the sage grouse has the ability to make long movements. Nor do the ptarmigan face winter food problems. Gardarsson (Fig. 9.9) and Mossop (Fig. 10.4) have documented ample food remaining when high populations of rock and willow ptarmigan declined. Willow ptarmigan generally maintain or gain weight in winter (Figs. 10.26, 16.11), as do white-tailed ptarmigan (Fig. 16.11). If birds generally face winter shortages of food, they should feed for long periods, and feeding bouts might be correlated with the severity of the shortages. But the winter feeding budget for grouse appears to represent less than 5 % of a 24-hour period. Further evidence against this hypothesis is that biologists do not find starved birds in winter. Lack argued that dead birds would be hard to locate, but even with radiotelemetry workers have not found birds that died of food problems. Biologists are in the field in winter: Mossop found no obviously starved willow ptarmigan in winter, and birds on his study areas gained weight (Chap. 10); the Hamerstroms winter-trapped prairie chickens for some 20 years and they have never found starved birds (pers. comm.). The dead ptarmigan I have found died mostly in the spring at the very moment foods were becoming more abundant (cf. Weeden 1965, Mercer 1967). Ruffed grouse and prairie chickens are discussed in more detail in Chapter 16.
15.7.3 Switch-over of predators The predator switch-over model dates back at least to Lack's (1954) treatise on the natural regulation of numbers (see also Hagen 1952). He suggested that when snowshoe hares crashed, northern predators would cause a severe impact on alternate prey such as grouse. Later, based on the advice of J. J. Hickey and L. B. Keith, he rejected the hypothesis. Keith (1963) showed that hares peaked in numbers before grouse in nine declines; hares and grouse declined together in five examples; and declines of grouse preceded those of hares eleven times. Later Keith (1974) revitalized the hypothesis, suggesting that grouse could be driven down by switch-over even though the local hare population was still high. Hare populations were not in perfect synchrony and predators might, through numerical responses, concentrate where hares and grouse were still common. Rusch et al. (1978) found support for this hypothesis for a ruffed grouse decline in Manitoba, where predators took more ruffed grouse when hares disappeared. Keith and Rusch, and their students, at Rochester, Alberta, counted hares, ruffed grouse, sharp-tailed grouse, coyotes, lynx (Lynx canadensis), red-tailed hawks, horned owls, goshawks, and mice for one 10-year cycle, 1966-75 (Fig. 15.23). Hares peaked in 1971, whereas the local, ruffed grouse population reached a peak in 1968 and a province wide peak of ruffed grouse occurred in 1970
638
A. T. BERGERUD
Fig. 15.23. Annual changes in the abundance of five species of predators and three species of prey during one 10-year cycle of snowshoe hares at Rochester, Alberta. Also shown are mortality rates of advertising males and percent young birds in harvests of ruffed grouse. (Data from Keith et al. 1977, Adamcik et al. 1978, Rusch et al. 1984.)
POPULATION ECOLOGY OF GROUSE
639
(Rusch et al. 1984). In this study grouse were already beginning to decline before predators were numerous and while hares were still increasing (Fig. 15.23). Also, the survival of young grouse was highest when hares were low; predators should not have been switching from hares to grouse nests and chicks. The mortality rates of advertising male ruffed grouse at Rochester, Alberta, were not highly correlated with predator abundance (goshawk: r = 0.101, n = 8; owls: r = 0.752, n = 8; coyotes: r = 0.480, n = 8; lynx, r = 0.573, n = 5). Increased predators may have boosted mortality rates. But the population declined from 1968 to 1969, even though adult survival was the highest in eight years (Fig. 15.23). The population also declined 1970-71 and 1971-72, when breeding success was the lowest in 8 years. It increased when the ratio of percent juveniles to percent adult mortality was greater than 1.4 (e.g., 1969-70, 87/55 = 1.6), and the population declined 1970 to 1973 when the ratio was less than 1.1. Breeding success was certainly involved in the dynamics of this ruffed grouse population (see Fig. 15.26), and it was not highly correlated with the abundance of the important predators (goshawk: r = -0.514, n = 9; horned owls: r = - 0.370, n — 9; coyotes: r = - 0.415, n = 9). Predator switch-over cannot explain the decline in this ruffed grouse population. I think that the predator switch-over model is useful for forecasting prolonged declines, but should be restricted to changes in goshawk predation following changes in hare abundance (Chap. 3). Goshawks are the ultimate predators of grouse (Eng & Gullion 1962). These birds are mobile, and their flights south occur at 10-year intervals coincident with hare declines (Mueller et al. 1977). Ruffed grouse mortality rates increased in all 4 years that goshawks came south in large numbers in northern Minnesota (Fig. 15.21). Goshawk predation appears responsible for major declines in red-phase males at Watch Lake, British Columbia. (Chap. 3). There are large declines of ruffed grouse in Manitoba at 10-year intervals. Ransom (1965) reported the near disappearance of ruffed grouse between November and January 1962-63. A count of drumming males in Rusch's Manitoba study area was 100 birds in 1971 and only three in 1972 (Rusch et al. 1978). The birds simply vanished in both crashes. I believe goshawk predation is the major factor responsible for these fall and winter crashes that occur after hares have declined. The impact of goshawk predation on ruffed grouse, in addition to a functional response to hare declines, should be influenced by the availability of snow for roosting and plunges (Fig. 14.26), and the flushing behavior of grouse (Chap. 3). Grouse should flush at short distances and take greater risks when the birds are at cyclic highs or declining. At such times goshawks may be moving widely, searching for prey other than hares. It appears that greater variation exists in the annual mortality rate for ruffed grouse than for other grouse species (Figs. 15.21,
640
A. T. BERGERUD
15.23, Gullion & Marshall 1968, Little 1978). This variability may relate to the mix of three factors: (1) abundance of goshawks and buffers, (2) cyclic flushing behavior of grouse, and (3) annual differences in the suitability of snow for cover. Predator switching from hares to grouse is not necessary to explain grouse declines or the 10-year cycle of grouse. Grouse populations such as ptarmigan in Iceland show major 10-year cycles (Gudmundsson 1960), although they do not live in ecosystems that contain cyclic hare populations.
15.7.4 Territorial self-regulation A controversy that has lingered since the 1950s is whether territorial behavior actually limits breeding numbers or if it acts primarily as a dispersal mechanism and means of assessing density (Lack 1954, 1966, Wynne-Edwards 1962, Fretwell & Lucas 1969, Watson & Moss 1970, Davies 1978, Patterson 1981, Wittenberger 198la). The most-quoted and best-documented examples used to support the territorial-limitation hypothesis is the 30-year, ongoing study of red grouse in Scotland (Jenkins et al. 1963, Moss et al. 1984, Watson & Moss 1985). Modifications of their hypothesis have also been applied to blue and spruce grouse (Zwickel 1972, 1980, Boag et al. 1979) and willow ptarmigan (Hannon 1983). In the 1972 red grouse model of population regulation, the quality of spring and summer food affects egg quality and breeding success, as well as the intrinsic spacing behavior of males and females contesting for territories in the fall and early winter. Territory size in turn determines stocking and population density (Watson & Moss 1972, 1980). Birds that do not secure territories are surplus and face a short life expectancy. This model implies that populations are always at a socially induced carrying capacity based on behavior as it determines territory size, and surpluses are available every autumn. In the spacing model, food quality affects not only spacing behavior but also the viability of young. Thus, spacing behavior and breeding success covary together; i.e., following a winter and spring that results in good nutrition, birds will be less aggressive and have smaller territories; coincident with this, they will have improved breeding success. The covariance of both breeding success and spacing behavior makes it extremely difficult to distinguish between the influence of breeding success versus that of territorial size on the subsequent population size. Spacing behavior could drive breeding numbers as Watson & Moss propose, or breeding success could be causal to changes in breeding numbers. With the latter hypothesis, spacing behavior and territory size would be the effects of competition between different numbers of competitors, and territory size would be dependent on prior breeding success and would not drive breeding numbers. The two hypotheses can be diagramed as follows:
POPULATION ECOLOGY OF GROUSE
H4:
Health and/or genetics
Breeding success
Spacing behavior
641
No further consequences
Territory size
Breeding stock
Surplus birds H5:
Health and/or genetics
Breeding success
Breeding stock
Territory size
Potential nesting space
Watson and Moss (1972, 1980) rejected the second hypothesis (Hs), arguing that spacing behavior and territorial size were independent of breeding success. Most recently the Scottish workers have rejected the nutritional aspect of their model as well as Chitty's hypothesis that genetic changes are the major cause of population cycles (Moss et al. 1984). However, they still believe that spacing behavior (territorial competition) limits numbers (Watson & Moss 1985) and now suggest that emigration is a key variable in declines (Moss et al. 1984, Watson et al. 1984b). The major research findings that lead to the spacing model in red grouse include: (1) a staircase decline in breeding numbers coincident with territorial behavior; (2) the presence of large numbers of birds —floaters —during the winter, which do not have territories; and (3) the discovery of large numbers of dead floaters in marginal habitats. Finally, when birds were removed from the territories, they were replaced by the floater birds (Jenkins et al. 1963, 1967, Watson & Jenkins 1968, Watson 1985). But the replacement of territorial birds by other birds does not constitute evidence of territorial limitations in an open system. The red grouse populations were open systems. For example, the number of red grouse on the Low study area in Scotland increased after the breeding season in November-December 1957 and again in October-December 1959. The Kerlock population increased in October 1965. The High area population increased in October 1960. The Glen Muick study had an estimated 124 birds in August 1958, even though 147 birds were shot, 106 were present in the spring of 1959. The Corndavon population was estimated at 307 in August 1957; 265 were shot, but 89 birds were present in 1958. Clearly there was considerable movement between moors in Scotland. Documen-
642
A. T. BERGERUD
tation of small dispersal distances of tagged birds, especially when tagged on heavily hunted and productive moors (Jenkins et al. 1967) does not constitute evidence that the populations were self-contained; ingress could easily exceed egress. Replacement of the vacated territories can be explained by birds shifting to optimal breeding densities (Fig. 14.27), and does not mean that these birds would not have bred elsewhere (Brown 1964, Watson & Moss 1970). The mechanics of population change of spring numbers in the red grouse, territorial model (H4) is density-dependent winter mortality. Fall territorial behavior is postulated to determine territory size, and the nonterritorial birds expelled die before the next spring. Predation was the major cause of death in the Kerloch, High and Low study areas (Jenkins et al. 1963, 1964, 1967). But winter predation of red grouse was not strongly density-dependent, contrary to the model (Fig 15.24A). Actually, the dead birds that the Scottish workers located from October to May represented a relatively constant proportion of the posthunting population; for the Low area, the known winter losses were: 1957-58, 39% (233/537); 1958-59, 27% (172/633); 1959-60, 22% (66/302); 1960-61, 25% (115/468). A constant proportion of birds died over winter, yet the breeding population (1957-61) for the Low area changed from a high of 371 birds in 1957 to a low of 157 birds in 1960 (C.V. = 34%, n = 5). These changes in breeding numbers were not correlated with total dead birds found between fall and spring (Fig. 15.24B). Winter mortality in the study was mostly density-independent and noncompensatory, contrary to the test implication of the territorial-limitation model (Fig. 15.22). The question of whether some red grouse females were prevented from breeding should be addressed, since males can be polygynous, and for grouse in general, females are the sex in short supply. The red grouse population declined at Glen Esk between November and May 1957-58; between September and February 1958-59; and between September and October 1960-61 (Jenkins et al. 1963). Because the sex ratio remained relatively constant, females as well as males must have left or died. The loss of males can be explained by territorial expulsion on territories in the fall, but males would not expel females. Thus, the females themselves must force other females to leave (Hannon 1983). The literature suggests that female red grouse are not strongly territorial until well into the winter (Jenkins et al. 1963), yet large numbers of females left as early as September (1961) and October (1958). Females departed in 1958 even after an estimated 43% of the population had been harvested. An alternative hypothesis to explain this dispersal is that many birds left simply to improve their survival and later breeding prospects, which has been documented in spruce (Keppie 1979, see Fig. 14.28) and ruffed grouse (Eng 1959), and were not "forced" out. The detailed counts of red grouse at Kerloch, Scotland (Jenkins et al. 1967), generally show declines in numbers in October-November and increases in February-March. These observations are con-
POPULATION ECOLOGY OF GROUSE
643
Fig. 15.24. Population statistics from early red grouse studies (Jenkins et al. 1963, 1964, 1967) to evaluate the hypothesis of population limitation through territorial self-regulation (spacing). (A) Winter losses from predation were largely density-independent, 1956-61. (E) Changes in the size of the breeding population from year to year were not correlated with proportions of the posthunting population found dead. (Q Proportional changes in territory size from year to year correlated with prior breeding population; (D) The spring population at Kerloch (unhunted) changed as a constant proportion of the prior year's fall population.
sistent with a two-direction dispersal schedule, as reported for blue, spruce, and ruffed grouse (Rusch & Keith 1971b, Herzog & Keppie 1980), and proposed in the polymorphic model in Fig. 2.26. Female red grouse floaters were seen throughout the winter but few were seen in May (Jenkins et al. 1963). Is it reasonable to believe that females suddenly died in May? A more realistic possibility is that with breeding finally upon them, females no longer could delay
644
A. T. BERGERUD
choosing a nesting site while hoping for a vacancy in a preferred habitat, but had to settle for a second-best location. Because some red grouse moors have been managed to provide vigorous heather growth they should be optimal for nesting cover. The difference between the best and the "other" may be more discontinuous in these red grouse areas than in unmanaged grouse habitats where a more gradual continuum exists. Hence hens at high density may delay their decision to the last moment, and ingress-egress could be considerable and confounding. The principal evidence for the territorial hypothesis should be that nonbreeding females are present in late May and June when productive hens are incubating. But the presence of a large proportion of such hens has not been documented for red grouse or, for that matter, any other grouse population (Table 15.1). Large numbers of red grouse hens have never been radio-tracked to document conclusively that many do not nest because of social behavior. A major radiotracking study, such as that conducted by Hines (1986) with blue grouse, would certainly help clarify the demography of red grouse. Changes in territory size in red grouse are correlated with changes in prior breeding success (Fig. 15.24C) and have not been shown to be correlated with the quality or quantity of food available to males in the fall (Henderson 1977, Miller & Watson 1978, Lance 1978b). A more parsimonious explanation than H4 is that territory size is the consequence of breeding success and is inverse to the number of birds that compete for space, as documented for other ptarmigan populations (Pedersen 1984, Bergerud et al. 1985) (i.e., H5). Spring numbers in red grouse are correlated with fall numbers (Fig. 15.24D, see also Watson et al. 1984b), as predicted by the breeding-success hypothesis. There are examples of vacant advertising sites for ruffed grouse, blue grouse, and willow and rock ptarmigan—populations are not always at a "socially induced carrying capacity" (Gardarsson 1971, Chap. 9, Little 1978, Lewis & Zwickel 1982). Bendell and Elliott (1967, p. 78) said, "It follows that even in dense populations of 0.44 males per acre, more males could have joined the breeding population." A growing consensus is that the evolutionary function of territory in birds is to space an individual away from others, as Lack (1966) argued (Davies 1978). Gardarsson (1971, Chap. 9) rejected territorial spacing as a means of control in rock ptarmigan, as did Myrberget and Mossop for willow ptarmigan (Chaps. 10, 11). The birds studied by Gardarsson and Myrberget left their study areas on islands in the fall. Heavy winter mortality occurred in winter flocks in the absence of territorial behavior. Nearly all the ptarmigan that returned in the spring in both areas were able to secure breeding space. Watson (1985) believes that the extremely high densities of red grouse in Scotland are a valuable vehicle to address the question of inherent mechanisms of population limitation. It is my opinion that since these populations are so intensively managed and unnatural, they are not an adequate laboratory to address questions of natural population control and adaptive behavioral strategies. The
POPULATION ECOLOGY OF GROUSE
645
chapters in this book have emphasized the role of predation as a prime mover in the evolution of reproductive tactics to enhance survival. The ground predators in Scotland are controlled and reduced in numbers, and nesting success is unusually high for a noninsular population. Further, the nesting habitat is managed by cultural practices and is discontinuous. Birds without territories must frequent habitats that are less safe (Jenkins et al. 1963), and these birds may be weak from atypical nutritional and disease problems (Jenkins et al. 1963, 1967, Hudson et al. 1985). Thus, it is not surprising that these "waiting birds," of reduced social status, may suffer increased mortality, and this compensatory mortality has been considered the means of population regulation (Watson & Moss 1979). Nonbreeding birds in other, more natural grouse populations may actually improve their longevity by avoiding the risks of advertising and nesting by seeking protective cover (Maxson 1977, Little 1978, Sopuck 1979, Lewis & Zwickel 1982). Again, dense populations restricted to discontinuous and limited habitats should display ingress-egress tendencies atypical of birds with a wider, more continuous range of habitat options. But the red grouse studies do provide insight into the ultimate limits of increase, where numbers are maintained above the natural controls of the extrinsic environments. These studies also provide an experimental laboratory where we can make comparisons with natural populations; but it may be unwise to extrapolate findings from these managed systems to more natural situations (Lack 1965, 1966).
15.7.5 Breeding success drives numbers The earliest grouse biologists believed that breeding success drove numbers. Ralph King pioneered ruffed grouse studies in Minnesota in the 1930s and said (1937, p. 524), "The first essential, of course, is an October population in excess of the desired April population. It must be in excess in order to allow for winter losses." The granddad of all studies, the mammoth, New York ruffed grouse study of 1932-42, resulted in the following conclusion: "The factors of increase include primarily the various components of the reproductive potential for the species" (Bump et al. 1947, p. 512). My primary technique to evaluate whether breeding success is paramount to changes in breeding numbers was to regress population change between years (total birds, or density, in year 2 divided by total birds, or density, in year 1) against the breeding success in year 1 (young birds per adult in August, or young per brood in August). A second method of evaluation was to calculate the coefficient of variation of breeding success versus mortality statistics. The test implication of the breeding-success hypothesis (Fig. 15.22) is that adult mortality is relatively constant compared to variation in breeding success. Variations in population size in the steppe grouse are generally conceded to result from changes in prior production. The standard management techniques for these three grouse species include the counts of summer broods and the determi-
646
A. T. BERGERUD
Fig. 15.25. The proportional change in density between adjacent years is correlated with breeding success the previous year in six populations of steppe grouse; s equals the mean number of juveniles per adults in the autumn required to balance average overwinter losses of juveniles and adults and maintain mean population size. (Data from Hillman & Jackson 1973, Kobriger 1975, 1981, Berg 1977, Lindeet al. 1978, Autenrieth 1981; some years unavailable or excluded for various objective reasons.)
nation of breeding success by assessing age ratios of harvested birds. Changes in breeding numbers are correlated with prior production in prairie chickens in Kansas (data from Horak 1971), Wisconsin, and South Dakota (Fig. 15.25). Annual changes in the density of sharp-tailed grouse males on leks are correlated with production in Montana (Brown 1966b, 1967, 1968b), Saskatchewan, Minnesota, South Dakota, and North Dakota (Fig. 15.25). Similarly, the abundance
POPULATION ECOLOGY OF GROUSE
647
of strutting, male sage grouse in Idaho reflected prior production. The correlation between sage grouse production and subsequent numbers was not significant for statistics from North Dakota or Colorado (Rogers 1964, Braun 1979, Hoffman 1979, government reports). However, the census technique of counting males on leks can be misleading (Beck & Braun 1980). The number of males attending a particular lek shows great variation between days (Hartzler 1972, Jenni & Hartzler 1978, Emmons & Braun 1984), and young males do not arrive until late in the season. Yearling:adult ratios are positively correlated with prior chick production in sage grouse (Dalke et al. 1963), indicating that production does drive numbers and that census methodology is inadequate to determine spring numbers. In general, steppe grouse show large annual variation in juvenile:adult ratios, but mortality rates are relatively constant. The variation in production results primarily from changes in nesting success, because the coefficient of variation (C.V.) in brood size is less than that in juvenile:adult ratios. For example, the C.V. for annual brood size of sharp-tailed grouse during 18 years in North Dakota was 11%, whereas for juvenile: adult ratios it was 38% (n = 32). In South Dakota, the C.V. in brood size of prairie grouse for 28 years was 12 %, the C.V. in the number of juveniles per adult is much greater (prairie chickens: 44%, n = 18; sharp-tailed grouse: 32%, n = 31) (Hillman & Jackson 1973, Kobriger 1975, 1981, Linde et al. 1978). The mean C.V. in annual mortality rates for steppe grouse in two studies was 11% (Brown 1966b, 1967; Robel et al. 1972), whereas the mean C.V. in the juvenile:adult ratios was an average of 46% for 11 studies. In the steppe environment, productivity varies much more than does mortality and is the major force in population change. The annual fluctuation in breeding numbers of forest grouse is also correlated with prior breeding success. Ruffed grouse populations changed in relation to variations in breeding success in New York, Alberta, Wisconsin, New Hampshire, and British Columbia (Fig. 15.26). Spruce grouse numbers changed in response to the percentage of juveniles in the autumn for populations in Alberta and Michigan (Fig. 15.26). Like the steppe grouse, there is greater variation in breeding success (C.V. = 31%, n = 8 studies) than in annual mortality (C.V. = 19%, n = 1 studies). A major problem in evaluating the factors responsible for changes in forest grouse numbers is that many workers rely on counts of advertising males to detect population changes. But there is a large component of silent males in most populations, and this component varies between years and is positively correlated with densities (Gullion 1981, Rusch et al. 1984). Most silent males are yearlings and are overlooked in census techniques such as drumming counts (ruffed grouse) and hooting tallies (blue grouse). Populations that showed changes correlated with production were only those in which researchers included census techniques that accounted for silent males,
648
A. T. BERGERUD
Fig. 15.26. The proportional change in density between adjacent years is correlated with breeding success the previous year for six populations of forest grouse; the s values (see legend for Fig. 15.25) for ruffed grouse in Alberta and Wisconsin are biased by the increased vulnerability of juvenile birds to hunting. (Data from King 1937, Bump et al. 1947, Dorney & Rabat 1960, Allison 1963, Gullion 1970c, Robinson 1980, Rusch et al. 1984; some years unavailable.)
such as the complete counts of spruce grouse in Michigan (Robinson 1980) and the King strip census in Minnesota. Authorities disagree whether breeding success determines breeding numbers
POPULATION ECOLOGY OF GROUSE
649
of blue grouse. Fred Zwickel, from his studies of blue grouse on Vancouver Island, has argued that there are more than enough juveniles alive in the fall to replace annual adult losses, and breeding success does not determine breeding numbers (Zwickel & Bendell 1967, Zwickel, 1982, Zwickel et al. 1983). However, Redfield (1972) also studied blue grouse on Vancouver Island and felt that changes in breeding success did alter spring numbers. His data show a correlation between production and the abundance of yearlings the next year (Fig. 15.27, middle). Fowle (1944), who pioneered blue grouse studies on Vancouver Island, also felt that when losses among chicks were high it could result in stabilized or declining populations. I found that in the closed, blue grouse system on Moresby Island that when the number of chicks per hen in August was greater than 1.3, the population increased the following spring (Fig. 2.10). The most intensively studied population of blue grouse on Vancouver Island is that at Comox Burn (Zwickel & Bendell 1967, Lance 1967, 1970, Mossop 1971, Zwickel 1972, Zwickel et al. 1977, Zwickel 1980). This population has clearly responded to annual changes in production; there is a significant correlation between chick production and yearling females the next year (r = 0.796, 1969-76; Fig. 15.27) and between the number of chicks per hen and yearling females again the next spring (r = 0.728, 1967-74; data from Mossop 1971, Zwickel et al. 1977,1983). Also, overwinter survival of juveniles was not density dependent in 8 years, and thus could not be the means of population regulation (data from Lewis 1979, Zwickel 1983, Zwickel et al. 1983). The survival of adults at Comox Burn has been relatively constant from 1969 to 1978, and cannot explain annual changes in numbers (Zwickel et al. 1983). If productivity drives numbers of blue grouse on Vancouver Island, breeding success should differ significantly between decreasing and increasing populations. The mean number of chicks per brood in August for the Vancouver Island populations that were increasing was 3.8 + 0.14 (n = 14), whereas for stable or declining populations it was 2.9 ± 0.13 (n - 29; t = 5.265, P < 0.01; data from Zwickel & Bendell 1967, Mossop 1971, Redfield 1972, Zwickel et al. 1977, Bendell pers. comm.). The Comox Burn (CB) population increased from 1968 to 1972, at the same time the population at Middle Quinsam (MQ) decreased. During these years, the Comox Burn birds had 61 %-80% of the females with broods and a mean brood size of approximately 3.8 chicks per brood female; birds in the decreasing population at MQ shared a nesting success of 53 % and an annual brood size of about 2.5 (Fig. 15.28). Brood hens are more easily sighted than nonbrood hens and need to be corrected at a 1.9:1 ratio (Mossop 1971). Mortality rates of yearlings and adults were similar for both populations, approximately 42% of the adults and yearlings died each year at MQ from 1970 to 1976, and about 40% of the adults and yearlings from 1969 to 1977 disappeared at CB (Fig. 15.28) (Zwickel
650
A. T. BERGERUD
Fig. 15.27. The percent yearlings or total yearlings is correlated with prior breeding success in three forest grouse populatons in which clutch sizes are small and mortality rates are low. (Data from Mossop 1971, Redfield 1972, Zwickel et al. 1977, 1983, Smyth & Boag 1984.)
POPULATION ECOLOGY OF GROUSE
651
Fig. 15.28. Top: Population changes of blue grouse at Comox Burn and Middle Quinsam, Vancouver Island, British Columbia, in relation to nesting success and chicks per brood statistics (averaged by 2-year intervals). Bottom: The eruption of two blue grouse populations when breeding success was very high and the more typical blue grouse populations that increase following colonization of newly logged areas and that finally decline as the forest canopy closes in. (Data from Redfieldetal. 1970, Mossop 1971, Zwickel etal. 1977, 1983, Frandsen 1980, Bendell pers. comm., and British Columbia game-check data.)
652
A. T. BERGERUD
et al. 1983). Overwinter survival of juveniles was calculated at about 50% in both areas for 1959 to 1964 (Zwickel & Bendell 1972), but only 40% for older juveniles at Comox, 1969-76 (Zwickel 1983). We can use these demographic parameters to evaluate whether production alone is sufficient to drive numbers. At CB, 100 adults and yearlings should provide 138 chicks by fall, of which 55 should live until the next spring and replace the 40 yearlings and adults that died that year. The population should increase from 100 to 115 birds, or at a rate-of-increase of r = 0.14. At MQ, 100 adults and yearlings produced on average 66 chicks until fall, of which 33 should live over winter to replace the 42 older birds that die; this population should change from 100 to 91 individuals, or r = - 0.09. These figures compare favorably with the real-world estimates. The MQ population declined from 1968 to 1972 at a rate-of-increase of r = -0.10, and the CB population increased between 1968 and 1972 at a rate of r = 0.13 (Fig. 15.28). The growth of these populations, 35 km apart, is adequately explained by differences in breeding success. Two population explosions of blue grouse have been documented on Vancouver Island that provide insight into the cause of population fluctuations. At Lower Quinsam (LQ), the number of birds rapidly increased to over 90 males/km2 by 1951 (Fig. 15.28). The second eruption occurred at Copper Canyon (CC), where the population increased from approximately 58 males/km2 in 1964 to 86 males/km2 in 1967. In both instances these eruptions were accompanied by very high breeding success (Fig. 15.28). Thus, it is clearly possible for a blue grouse population to double in 3 years if nesting success is greater than 80% and four or five chicks per brood survive until fall. Both populations plunged quickly from these high densities when breeding success returned to more typical levels. These eruptions also provide further evidence that blue grouse populations cannot be limited by territorial space. Within 3 years they doubled, and reached densities four to five times greater than normal (Fig. 15.28). The two eruptions even occurred in habitats in different stages of reforestation. The common ingredient was unusually high production. Evidence is substantial that breeding success drives numbers of blue grouse on Vancouver Island. Last, Mines (1986a), on Hardwicke Island, British Columbia, placed radios on 293 juvenile blue grouse in the fall and followed them until they died or recruited. The birds were killed by raptors throughout the winter in the absence of spacing behavior; males died at a rate of 72 % -79 %, and females at 65 % -72 %. This overwinter mortality was 74% (1979-80), 76% (1980-81), 67% (1981-82), and 73% (1982-83) (C.V. = only 6%), and was independent of population size. Five of 28 yearling females did not nest, but they were not prevented from holding space in the breeding habitat—a test condition of the territorial limitation model (Watson & Moss 1970). The number of recruits in year 2 was linearly correlated with prior production in year 1; for females r = 0.97, and for males r
POPULATION ECOLOGY OF GROUSE
653
= 0.92 (n = 4 years). Hence variation for breeding success was the mechanism of population change between spring numbers. The spring abundance of tundra grouse also changes in relation to variations in the breeding success the previous year (Fig. 15.29, Chaps. 10, 11, May 1975, Pedersen et al. 1983, Pedersen 1984). But unlike the steppe and forest grouse, in ptarmigan the coefficient of variation for the annual mortality statistics is approximately equal to the C.V. of juvenile:adult ratios —31 % (n = 3 studies) and 39% (n = 5 studies), respectively. Weeden and Theberge (1972) also concluded that vagaries in rock ptarmigan numbers in Alaska were influenced both by changes in production and by mortality. Red grouse populations also respond principally to changes in breeding success (Bergerud et al. 1985). There is a significant correlation between breeding success and the proportional changes in numbers of red grouse in the Glen Esk, Low and High populations (r = 0.883, n = 7; from Jenkins et al. 1963), and for three lightly hunted areas (Kerloch, Forvie, and Locknagar, r = 0.681, n = 26) (Fig. 15.29). But the statistics from two other red grouse populations, on two "rich" Scottish moors—Glen Muick and Corndavon—failed to show positive correlations between breeding success and subsequent densities (Jenkins et al. 1967). It was based on these two exceptions that Jenkins and Watson rejected the hypothesis that production drives numbers of red grouse. Both populations were highly productive and at high densities, and both sustained harvest losses of approximately 45% from 1958 to 1965 (Jenkins et al. 1967). These harvests were sufficiently high to negate the role of production in altering subsequent breeding numbers. The mean estimate of breeding success in August was 144 young per 100 adults (1958-71). In a fall population of 244 birds, 110 would have been harvested and an additional 50 would have died from other causes between fall and spring. I estimated this 50-bird loss (38%) from the overwinter mortality of birds in an isolated red grouse population at Forvie, 1957-65, excluding 1958-59 (from Table 2 in Jenkins et al. 1967). In addition, in an earlier study of red grouse, Jenkins et al. (1963) reported posthunting loss before spring of greater than 25 %, based on dead birds found at their Low study area and the disappearance of banded birds from December to March. This 25 %-plus loss would be conservative because some birds would not be found. If we subtract the 110 birds killed during the hunting season plus the 50 that died from natural causes from a fall population of 244, there would be insufficient birds to maintain 100 breeding birds. To evaluate this in another manner, assume that these red grouse suffer a 65% annual mortality (Jenkins et al. 1967). Then a population of 244 with a survival rate of 35% would have only 85 remaining at the end of the year—again, this is inadequate to maintain numbers. Production could not drive numbers in these two rich moors, because there was an excess of hunting mortality. However, these populations did maintain their numbers from 1957 to 1971
654
A. T. BERGERUD
Fig. 15.29 The proportional changes in density between adjacent years are correlated with breeding success the previous year for six populations of tundra grouse. (Data from Jenkins et al. 1963, 1967, Watson 1965, Bergerud 1970a, Gardarsson 1971, Weeden & Theberge 1972, pers. comm., May 1975.)
(Jenkins et al. 1967, Moss et al. 1975). Additional birds must have come from less productive moors where hunting losses were lower. For example, the Locknagar population maintained a breeding-success estimate of 1.05 juveniles:adult (1961-71) and had a low hunting loss (Jenkins et al. 1967). If these birds sustained a 38% overwinter loss (205-78 = 127), there would still be sufficient
POPULATION ECOLOGY OF GROUSE
655
stock to maintain their own numbers (using an index of 100) and yet have sufficient colonizers (127 - 100 = 27) that could improve their fitness by shifting to the more productive Glen Muick or Corndavon moors. The Scottish, red grouse biologists have done interesting fertilizer experiments (Miller et al. 1970) that provide evidence that breeding success determines numbers in red grouse. They fertilized heather in May 1965. Birds invaded the improved heather cover in 1966, but the density of birds did not increase in that year, as expected if nutrition influences spacing behavior (Watson & Moss 1972). Instead, the population increased in size the second year after fertilization (1967), when the increased production of young from 1966, which resulted from the improved nutrition in 1966, first took territories. The experimental population also remained higher than the control population in the third year (1968) after fertilization, when the heather no longer showed the improved growth or nutritional quality brought about by the fertilizer. This further increase is consistent with the hypothesis that, again, it was the high breeding success the previous year that determined the density of birds the next season. Watson and O'Hare (1979) noted that in 19 populations of red grouse in Ireland with low brood survival, 14 decreased the next year, but in 5 areas where brood survival was high, 4 increased the following year. Production appears to drive numbers in red grouse in Ireland as well as in Scotland. Another example in which hunting probably negates a direct link between production and subsequent changes in numbers is the several small populations of white-tailed ptarmigan in Colorado studied by Braun and Rogers (1971). These populations were harvested at rates ranging from 14% to 63%. These hunted populations seem to have maintained their numbers, but again there is extensive movement between these small populations and ingress from less hunted populations likely explains stability. Heavily hunted populations do not provide a fair test of the breeding-success hypothesis. Recently, Watson and Moss (1985) appear to have rejected their original model (Watson & Moss 1972, 1980) in which breeding success has no further consequence for the subsequent size of spring numbers, and the dynamics of the system, in their view, was determined by compensatory winter mortality (Watson & Moss 1979). They state (Watson & Moss 1985, p. 283): "Changes in spring numbers from one spring to the next could be predicted from the observed chick production ratio each year." This is clearly the breeding-success hypothesis proposed (Bergerud 1970a). But in the same paper, Watson and Moss also say that the density of grouse at Glen Esk and Kerloch is limited by territorial behavior; do they equate density with total numbers? Since territorial behavior occurs after breeding success is determined, and territory size and breeding success are negatively correlated (Bergerud et al. 1985), territorial behavior must be an effect of earlier events, and breeding success lies closer to the core of population regulation.
656
A. T. BERGERUD
The observation that there are more than enough chicks in the fall to replace adult losses is not evidence that production does not drive numbers. These chicks are not surplus. Each population has a unique and relatively constant, overwinter mortality regime dictated by the particular cover characteristics and predators in a given region. In some years, high production of chicks would provide more than enough recruits to compensate for the typical, density-independent adult losses, and the population the next spring would increase. Yet in another season, reduced production or survival of chicks in their first summer will fail to provide sufficient recruits to replace the largely constant loss of yearlings and adults, and next year's population will be lower. Clearly, if production did not provide recruitment that on average balances adult losses, the population would become extinct. In all, I have examined 37 sets of data from substantial studies of grouse. Thirty studies show significant or nearly significant correlations between changes in breeding success and population changes the next year. The exceptions include the white-tailed ptarmigan and red grouse populations that are heavily hunted, and some sage and ruffed grouse populations that probably are inaccurately censused. One blue grouse population in Alberta showed satisfactory breeding success, but numbers continued to decline from 1955 to 1964; high adult mortality was the suspected cause (Boag 1966). My conclusion is that change in breeding success is a sufficient variable to drive numbers. This does not mean that it is a necessary factor. Clearly, if mortality rates do change between years, this can affect breeding numbers, as was true for ruffed grouse in Minnesota (Fig. 15.21). But variations in mortality rates of the majority of grouse populations are only about half those found in breeding success. A generalization from this review is that breeding success is the dominant demographic parameter that alters spring breeding numbers. Similar conclusions have been reached for Hungarian partridge (Perdix perdix) (Potts 1980), black grouse (Anglestam 1983), and pheasants (Linder et al. 1960). 15.8 The 10-year cycle in grouse Since the 1930s, grouse biologists have been intrigued by and have inquired into the underlying causes for the recurring periods of scarcity and abundance of northern grouse. The 10-year cycle in grouse is well recognized (Keith 1963). Leopold (1933, p. 50) spoke of cycles as "periodical associations of more or less fixed length and amplitude." Davis (1957) felt that cycles could be distinguished on the basis that abundance changed in a nonrandom sequence, wherein one could predict the future growth of a population. This discussion is limited to the long, nonrandom fluctuations in abundance of 9- to 11-year intervals. Myrberget (1984 and Chap. 11) has provided a comprehensive explanation for the shorter 3- to 4year cycles (phase-forgetting, quasi-cycles; Niebet & Gurney 1982) in willow ptarmigan, based on cycles in mice, plant cover, and the predator switch-over
POPULATION ECOLOGY OF GROUSE
657
phenomenon. This switch-over model as an explanation for the 3- to 4-year cycle has also been documented for black grouse (Angelstam 1983, Angelstam et al. 1984). The 10-year cycle is most pronounced in northern ruffed grouse (Keith 1963) and arctic ptarmigan populations. Willow and rock ptarmigan in North America, Iceland, Greenland, and Scotland all show 10-year cycles (Salomonsen 1939, Braestrup 1941, Gudmundsson 1960, Watson 1965, Bergerud 1970a, Weeden & Theberge 1972). The cycle has become less pronounced or has disappeared in populations of ruffed grouse in New York, New Brunswick, and Maine, where earlier naturalists once recorded the periodic rises and falls in numbers (cf. Bump etal. 1947, Palmer 1949, Keith 1963). The prairie chicken population in Wisconsin also showed 10-year cycles in abundance from about 1850 to 1940 (Schorger 1944, Grange 1948), and then the pattern was broken (Hamerstrom & Hamerstrom 1973). At present, the cycle of ruffed grouse appears to be damping out in the Cloquet study area in northern Minnesota, where ruffed grouse research began and where field biologists have been using the King census in the spring since 1932 (King 1937, Gullion 1981). Two other generalizations are that cyclic populations are found in large, continuous blocks of habitat (Leopold 1931) and in northern latitudes (Keith 1963). These continuous habitats include aspen parkland, coniferous forests, and arctic tundra. The boundary between the northern, cyclic populations and more southern, noncyclic populations stretches from Alberta to Minnesota, following the southern edge of the aspen parkland (Fig. 15.30). Prairie grouse in the farmlands and grasslands south of the aspen parklands are not cyclic, but sharp-tailed and ruffed grouse in the parkland show regular changes in abundance. One can distinguish further differences in the continuity of habitats used by cyclic ruffed grouse in northern Minnesota and Wisconsin, and those of noncyclic ruffed grouse in the southern portions of these states. The cyclic populations live in more homogeneous, conifer-aspen forests on glacial moraines, whereas the random, fluctuating populations reside in small blocks of relic habitats interspersed with farmland. This forest-farm interface is plainly evident on county forest maps of Wisconsin (see also Curtis 1959), and crosses the state at about 45°N. latitude, just north of the Hamerstroms' prairie chicken study area at Buena Vista Marsh, Wisconsin (Fig. 15.30, see also Curtis 1959).
15.8.1 Frequency- and density-dependent selection In 1970 I proposed that the mechanics of change for 10-year, cyclic populations resulted from regular changes in breeding success (Bergerud 1970a). In this model, breeding success varied in response to density-dependent selection among genotypes that produced chicks of different intrinsic viability, as predicted by Chitty's (1967) hypothesis of density-dependent selection between behavioral morphs. My argument then was that variations in amplitude, periodicity, and syn-
658
A. T. BERGERUD
Fig. 15.30. Comparison of the regular, cyclic fluctuations of grouse populations in the aspen parkland and boreal forest and the irregular fluctuations of populations south of the aspen parkland. All graphs are based on counts of birds in the spring except those from Manitoba and Saskatchewan, which are based on harvest statistics and thus influenced by breeding success. The increasing amplitude in the graphs using the latter four indexes reflects increasing hunting pressure rather than increasing natural amplitude. The cycles are damping out in ruffed grouse in the Turtle Mountains, North Dakota, and in the Cloquet area in Minnesota. (Data from Gullion 1970c, 1981, Hamerstrom & Hamerstrom 1973, Hillman & Jackson 1973, Kobriger 1975, 1981, Boag 1976b, Rusch 1976, Berg 1977, 1979, Keith et al. 1977, Linde et al. 1978, Anonymous 1980, Thompson & Moulton 1981, Schulz 1983, Rusch et al. 1984, pers. comm., R. K. Anderson pers. comm., S. R. Barber pers. comm., Alberta Game Dept. records.)
POPULATION ECOLOGY OF GROUSE
659
chronization in cycles could be explained by superimposing the vagaries of posthatch weather on these smooth alterations in chick viability. Ricker (1954) had suggested that cycles of different amplitude could result from the interaction of smooth reproduction curves and a sequence of random factors. The polymorphic, density-dependent selection hypothesis is consistent with theoretical research into density-dependent selection (Charlesworth 1971, Roughgarden 1971, Giesel 1974, Slatkin 1979). Anderson and Arnold (1983, p. 649) stated that "when resources are limited, the genotypes in a Mendelian population will be forced into competition as the population grows in size. If the selective values of the genotypes respond differentially to population size, then density selection will occur. Experimental studies with a variety of organisms have shown that such density-regulated selection does in fact occur" (see review in Prout 1980). Breeding-success statistics are available for five cyclic grouse populations (Fig. 15.31). Breeding success was highest when these populations increased their numbers and less when the populations were declining. In all cases, breeding success was correlated with subsequent population changes (Figs. 15.20, 15.29). There are no reported examples of cyclic grouse populations that do not display rather smooth annual changes in breeding success (Figs. 15.31, 15.32, 15.33). These relatively nonrandom changes in breeding success are probably a necessary condition to generate the 10-year cycle in grouse. The smooth variation in chick survival in cyclic populations is not directly related to density. Chick survival generally declines before peak populations and remains low for several years after numbers fall. However, this delayed, densitydependent impact can be traced back to the competition faced by parent birds in breeding. Chick survival improves when at low densities there is no longer competition for requisites and the parents that compete at high densities have died. These results are consistent with the theory of density-dependent, polymorphic selection in which birds with less genetic variation (homozygotes) are more successful in competing for nest sites and/or mates at high densities, but in so doing provide less-viable young and the seeds of decline (see Mac Arthur 1968, Bergerud 1970a). In red grouse, chick survival was higher in heterozygous broods than in homozygous broods (Henderson 1977). In Newfoundland, ptarmigan showed differences in calling frequency, vulnerability to hunting (Fig. 15.34), dispersal distances, renesting frequency, autumn weights, and male fidelity to broods, among cohorts (Bergerud 1970a). The ptarmigan on Brunette Island varied in many parameters between cohorts as the population declined from a high in 1962 (Fig. 15.35). Rock ptarmigan in Scotland called more when their numbers were declining (Watson 1965). Captivebreeding studies have now shown that calling frequency is at least partly inherited in red grouse (Moss et al. 1982b). Reduced nesting success is also suggested in ptarmigan when numbers are high
660
A. T. BERGERUD
Fig. 15.31. Juveniles per adult statistics for five cyclic populations (legend, bottom right). Breeding success was greater than the number of juveniles per adult needed to maintain numbers (s) in years before the peak. Just before the peak year (highest breeding population), breeding success declined in six of seven population peaks recorded. After the peak, breeding success was inadequate to maintain numbers (i.e., < s). Breeding numbers again began to increase in four of the populations when breeding success was greater than 5, (Data from Watson 1965, Jenkins et al. 1967, Bergerud 1970a, Gardarsson 1971, Weeden & Theberge 1972, Moss et al. 1975.)
or declining (Jenkins et al. 1963, Watson 1965, Bergerud 1970a). Experimental evidence shows differences in intrinsic viability and behavior of hand-reared chicks between cohorts in red grouse (Jenkins et al. 1965, Moss et al. 1981, Moss et al. 1984), willow ptarmigan (Bergerud 1970a), and rock ptarmigan (Theberge & Bendell 1980). In general, the survival of chicks in aviaries parallels that of those in the wild. Clutch size may also change between cohorts, as there are significant correlations between densities and clutch size in cyclic, Alaskan rock ptarmigan (r = -0.568, n = 9), cyclic ruffed grouse in Alberta (r = -0.538, n = 8), willow ptarmigan in Newfoundland (r = 0.600, n = 9), and rock ptarmigan in Iceland (r = - 0.816, n = 8) (Bergerud 1970a, Gardarsson 1971, Weeden
POPULATION ECOLOGY OF GROUSE
661
Fig. 15.32. Nesting success (estimated by the proportion of females with broods) and chick survival in relation to density for cyclic rock (top) and willow (bottom) ptarmigan populatons. Breeding success changed in smooth patterns, was highest just before peak spring densities, and began to increase again just before the population started to increase. (Data from Watson 1965, Bergerud 1970a.)
662
A. T. BERGERUD
Fig. 15.33. Demography of the cyclic rock ptarmigan population at Eagle Creek, Alaska (39km 2 ), 1960 to 1971. (Data from McGowan 1972, Weeden &Theberge 1972, pers. comm.)
POPULATION ECOLOGY OF GROUSE
663
Fig. 15.34. Vulnerability of willow ptarmigan to hunting in Newfoundland showed a cyclic pattern. Birds were most susceptible to hunting after the peak, in 1962, when they were probably the most aggressive. Vulnerability was also correlated with chick viability. (Data from Bergerud 1970a, 1972.)
& Theberge 1972, Rusch et al. 1984; see Fig. 15.33). These correlations can be explained better by differences in the fecundity of genotypes than from densityrelated or nutritional-physiological stress. The breeding success model builds on the foundation that all grouse populations, whether cyclic or noncyclic, are polymorphic in spacing behavior. I
664
A. T. BERGERUD
Fig. 15.35. In the cyclic population of willow ptarmigan on Brunette Island, Newfoundland, birds showed a high correlation among demographic, behavioral, and morphological parameters as their numbers declined from 1962 to 1965, suggesting a smooth change in intrinsic quality (Data from Mercer 1967.)
hypothesize that there are two suites of tactics, density-tolerant and densityintolerant phenotypes (section 14.9). The relative success of these tactics varies in space and time as populations fluctuate. Space and dispersal are the more important mediators for noncyclic populations, whereas time is the dominant pacer
POPULATION ECOLOGY OF GROUSE
665
in cyclic populations. The evidence is extensive for these two suites of tactics (Chaps. 1,2,3).
15.8.2 The single locus, density-dependent model The genetic background underlying the density-tolerant, intolerant phenotypes is not known. In red grouse, aggressive behavior and dominance had an additive heritability component of 0.6 between fathers and sons (Moss et al. 1982b). The aggressive behavior of blue grouse males on Stuart Island showed no detectable change in 7 years between fathers and sons (Table 2.7), whereas the docile phenotype introduced to nearby Portland Island produced both aggressive and docile progeny in the first generation (Fig. 2.21). In our genetic model of the 10-year cycle (Chap. 12), we equated aggressiveness with homozygosity (aa) and docility (density-tolerance) with heterozygosity (Aa) based on the behavioral outcomes of the island introductions. Thus, the cyclic fluctuations were generated on the basis of a single locus with Mendelian segregations and complete penetrance. The cycle was driven by genetic changes in the viability of chicks and competition for space between three genotypes. Our model in Chapter 12 had several components of realism. Actual brood sizes observed in Newfoundland (Bergerud 1970a) were used in the simulation (high chick survival was attributed to heterosis). The amplitude of the cycles varied in response to stochastic changes in chick survival from spring weather (Bergerud 1970a). At approximately 100-year intervals, cycles were skipped, consistent with the real world. The cycle damped out in grouse populations with low mortality rates, again in tune with actual observations. But was the model too simple? Many researchers feel that complex behavior is best explained by polygenic inheritance—the additive action of many loci (Cade 1984). Possibly the correlations of so many morphological and behavioral parameters in cyclic grouse (see Fig. 15.35) might result from major chromosome differences in inversions or karyotypes. The aggressiveness of house mice varied between 2n = 26 and 2n = 24 karyotypes (Capanna et al. 1984). At this time we know practically nothing about the genetic basis of behavior in birds. Single-locus segregations are a common finding for polymorphisms in morphology for vertebrates (Chap. 3). The substitution of alleles at a single locus can cause changes in rate of wing vibrations in Drosophila (Schicher 1973) and female receptivity in mosquitoes (Aedes atropalpus) (Gwadz 1970). Regardless of the genetic foundation of the densitytolerant, density-intolerant behavior suites, the genetic mechanism would need to be discontinuous in segregations in random mating systems to explain the rapid changes between adjacent cohorts. Another criticism of the single-locus model in Chapter 12 is that the switching mechanism between increase and decrease phases was density-dependent competition for limited space and nonbreeding of the docile genotype. However, the evi-
666
A. T. BERGERUD
dence is that nearly all hens in each of the 9 grouse species generally nest regardless of densities (Table 15.1). Females of the lek species do not appear to have mutually exclusive nesting ranges that could provide a means of densitydependent restraint in nesting.
15.8.3 The female choice, density-dependent model The cycle model might be made more robust by dropping the simplistic restriction of single-locus genetics and the space limitation. Instead, assume that adaptive gene combinations are maintained by linkage disequilibrium or pleiotropy for each of 2 (or 3) density phenotypes. The cycle is still driven by changes in chick survival owing to genetic influences, but the switching mechanism is not mediated by space but by hens pairing nonrandomly at high and low densities, preferring males of a similar phenotype. For ptarmigan in Newfoundland, I noted positive pair assortment within cohorts in 1962 and 1963 when the population was generally aggressive and displayed bimodalism in chick survival (Fig. 15.36, Bergerud 1970a). Pairing was more random in 1964 and 1965 as the population declined and the bimodalism in the viability of chicks became less clear (Fig. 15.36). I have provided a diagram (Fig. 15.37) of how a female-choice paradigm could operate based on data from the cyclic, Scottish rock ptarmigan population studied by Adam Watson (see Fig. 15.32). In 1950-51 the Scottish population was high and the best habitat continuously occupied by males (Watson 1965). Under these conditions females should select males with large territories and avoid those with small territories (Chap. 10, Miller & Watson 1978, Hannon 1983). Pairing monogamously with males with the largest territories would ensure maximum spacing of nests, which would minimize nest predation within the constraints of the relatively close spacing (Fig. 15.37). Males with large territories would be aggressive (Watson 1964, Watson & Miller 1971). Likewise, aggressive females would have the pick of territory sizes by excluding docile females (Hannon 1982). Hence, the social competitions of male vs. male and female vs. female would lead to positive assortment of the density-intolerant phenotype. These aggressive pairs in the best habitats would produce less-viable young (Jenkins et al. 1963, Mercer 1967, Bergerud 1970a, Theberge & Bendell 1980). Docile pairs in secondary habitats might also fare poorly (Bergerud & Mercer 1972). The Scottish population declined after 1952 when the birds were generally aggressive and breeding success was low (< 0.5 chicks per adult, Fig. 15.32). The correlation between spacing behavior (1952, 1954-56, 1960-62; aggressive acts/100 min) and breeding success was r = -0.833, n = 7, P < 0.05) (data from Watson 1965). Not only did the population decline, but the sex ratio of adults shifted heavily to males. The correlation between aggressiveness and the proportion of males the next year was (r = 0.878, n = 7, P < 0.01); the correlation between breeding success and the proportion of males the next year in 18 years was highly significant (r = -0.718) (from Watson 1965).
POPULATION ECOLOGY OF GROUSE
667
Fig. 15.36. The August brood sizes of willow ptarmigan in Newfoundland were often significantly bimodal in local populations in the years after a cyclic peak in 1960-61. (Data from Bergerud 1970a.)
A generalization is that males increase in cyclic ptarmigan populations as numbers decrease (Watson 1965, Bergerud 1970a, Fig. 15.33). In Newfoundland and Iceland more young, female chicks died than males (Bergerud 1970a, Gardarsson 1971). The significance of these deaths imbalance is that they provide changes in the frequency of the phenotypes because the viability of aggressive chicks is less than that of docile chicks (Bergerud 1970a, Theberge & Bendell 1980, Moss et al. 1984). Also, aggressive parents may provide less care (Jenkins et al. 1963, Bergerud 1970a). By 1954-56 the Scottish population had declined by one-third; under these circumstances females could be less particular in the selection of various-sized territories—hence pairs should be a more random assortment between phenotypes. However, there would be lag effects because each new cohort pairs at densities different from that faced by the parents (Bergerud 1970a). The age structure would become progressively older because of reduced cohort contributions. In
SELECTS SPACE (a) AND POLYGYNY
Fig. 15.37. The density-selection model of female choice to explain the 10-year cycle. Data are adapted from Watson (1965). In 1950-51 the population was high; females should have chosen aggressive males (ad) with large territories to increase their space away from other nesting females. An increase in the frequency of the a allele should have resulted in reduced viability and survival of chicks, especially (aa) females, and in a decline in the population. At midcycle, in 1952-56, females could choose either aggressive males or docile males that defended good nesting cover. With no change in a/A frequency, chick survival would have remained low and the population would have continued to decrease. In 1957-59, when the population was small and territories were hardly touching, females would choose males that defended good nesting cover (A). With an increase in frequency of A, chick viability improved and breeding success was greater than needed for stability (s). In 1960-61, as density increased, females could choose either (a) or (A) males, and breeding success remained greater than s. At high densities, 1962-63, females again chose aggressive males (aa), setting into place another decline. Females may choose to be polygamous and select males with relatively large territories when densities are high, and territories are small overall.
POPULATION ECOLOGY OF GROUSE
669
1974, declining populations of willow ptarmigan at Chilkat Pass had the following age structure: > 4 years-15%; 3 years-27%; 2 years-23%; and yearlings —35% (Mossop pers. comm.; n — 26 banded males and females on territory). This array is significantly different from that which would result considering only a constant, annual adult mortality of about 50% (Chap. 10). At the bottom of the cycle in Scotland in 1957-59 the population had declined by two-thirds. By then, females, which I believe were predominantly nonaggressive, had a maximum choice of males since males outnumbered them 3 to 2 (Fig. 15.37). The territories of males should have been generally large and uniform in size (see Figs. 9.6, 10.19), hence females could have chosen males that incorporated the best nesting cover into their defended space. On Moresby Island it was the docile blue grouse males (CC stock) that defended the smallest advertising ranges (Bergerud & Hemus 1975), but in the best cover (Fig. 2.4). If this sequence held for rock ptarmigan, there should have again been positive assortment in pairing, but unlike the nonrandom pairing at the "high," at the "low," nonaggressive hens would have selected nonaggressive (density-tolerant) males. These mating combinations would produce more-viable progeny. After 1958, chick survival improved for the Scottish birds and breeding success exceeded 0.75 young per adult, which was the stabilizing breeding-success figure for that population (Fig. 15.29). The increase phase had been triggered. By 1962-63 the population in Scotland was again dense (Watson 1965), favoring pairing by aggressives, setting in place another decline and cycle. The hypothesis of female choice to explain switching in growth phases of the cycle (Fig. 15.37) requires that there be changes in delayed density-dependence in spacing behavior that can be traced to a genetic basis. Moss et al. (1984) have now documented a regular change in aggressive behavior in cock red grouse during one complete 8-year cycle that had a heritable basis of h2 = 0.6 (Moss et al. 1982b). They state, "The results confirm that genetic changes in aggression take place during a population fluctuation" (Moss & Watson 1985, p. 275). Further, by selectively pairing ptarmigan, these workers developed high and low aggressive lines in a captive population (Hr = 0.9, Moss et al. 1985). They report (p. 258): "The results are consistent with simple quantitative genetic inheritance of the ability to dominate others, with equal contributions from each sex." This simple inheritance system is facilitated if birds mate assortatively, and Gjestal (1977) has documented this for captive ptarmigan. The Scottish workers rejected the genetic-aggression explanation for cycles of red grouse because the proportion of subordinate birds increased before the population started to decline, and the most submissive birds occurred just after the peak in numbers (Moss et al. 1984, Watson & Moss 1985). They reasoned that these sequences suggest changes in behavior as a result of, rather than as the cause of, population change. But the female-choice explanation (Fig. 15.37) predicts delayed sequences. When numbers are high and territories contiguous, assorta-
670
A. T. BERGERUD
tive pairing of aggressive birds should result in their progeny being more homozygous for aggressiveness, and these birds will contest for breeding space in later years (Bergerud 1970a). In Newfoundland, ptarmigan were most approachable and aggressive just after the peak in numbers (Fig. 15.34, Mercer 1967, Bergerud 1970a), as were ruffed grouse at Watch Lake (Fig. 3.16). In Scotland the peak in inherited docility in red grouse occurred in 1973 after the high in numbers in 1972 (Moss & Watson 1985). This sequence is the converse of that in Newfoundland (Fig. 15.34) and Watch Lake (Fig. 3.16), and of the Chitty hypothesis (Moss & Watson 1985). However, in Scotland there was extensive emigration when numbers were high; the birds that left should have been the density-intolerant phenotype (Fig. 2.4). The remaining birds and their progeny would then have been less aggressive. The Scottish results may have resulted because a closed system was not studied. A robust female-choice explanation for the 10-year cycle in grouse should account for the cycles not only in monogamous ptarmigan, but also in polygynous ruffed grouse and sharp-tailed grouse (Fig. 15.30). A major difference in the two mating systems is that with polygyny some males breed several females and males do not defend nesting space or cover that females could choose. Females for these species must generally choose between the males that continue to advertise after male vs. male competition. The most complete data available for polygynous grouse is the ruffed grouse study presented in Chapter 3. Red-phase, ruffed grouse males at Watch Lake, British Columbia, were more aggressive than gray birds, and their frequency in the advertising population increased with density (Fig. 15.38). The frequency of red hens also increased as numbers expanded (Fig. 15.38), and these hens were more aggressive than gray hens (Fig. 3.13) and raised fewer chicks per successful female (Table 3.5). Positive assortment should have occurred in mating for ruffed grouse, on the basis of frequency alone. If birds of the color-behavior phenotype sorted themselves in space as did blue grouse on Moresby Island and Vancouver Island, there would have been positive assortment. The ruffed grouse in Area 1 at Watch Lake were generally more aggressive than those in Area 2 (Figs. 3.10, 3.11, 3.15). Positive pairing assortment between color morphs has been recorded in snow geese (Anserr caerulescens) (Cooke & Cooch 1968) and the Arctic skua (Catharacta skua) (O'Donald 1977). For both species there were differences in behavior between morphs. If we combine assertive matings in grouse, plus genetic linkage combinations involving color-behavior and possibly chick survival, we have the necessary basis for density-dependent selection resulting in cycles. A confounding factor at Watch Lake was the appearance of goshawks in 1972-73 and 1980-81. These raptors probably killed many more red cocks than gray males. This predation should have altered the underlying genetic frequencies in one season and could have affected the switching of the cycle from declining to increasing. Still that population experienced negative recruitment 1971-74 and 1979-82, before
POPULATION ECOLOGY OF GROUSE
671
Fig. 15.38. Top: Proportion of red grouse hens in the ruffed grouse population at Watch Lake, British Columbia (Chap. 3), increased when densities increased. Bottom: Proportion of red cocks advertising at Watch Lake also increased as numbers increased.
and after the goshawk flights. Further there was positive recruitment in both 1974 and 1975 when the frequency of the color phases did not change (Table 3.1). An underlying, intrinsic cyclic mechanism is still suggested. Females of the steppe grouse, like sharptails, generally mate with more centrally located males at leks, but they also choose which leks they will visit (Fig. 14.9). As populations increase, new, small leks of yearlings appear. Males at these leks may be more density-intolerant and aggressive. Such males could improve their breeding success in increasing populations as density intolerant fe-
672
A. T. BERGERUD
males expand into habitats that are less densely occupied, and visit these new, nearby leks. This is strictly speculation; there are no studies to decide if there is a positive assortment in pairing for steppe grouse, and we have only the first suggestion that there are even behavioral phenotypes (Chap. 5).
15.8.4 Relevance of the density-dependent model to characteristics of cyclic populations A satisfactory theory to explain 10-year, cyclic fluctuations should account for the following characteristics of tetraonids. (1) The hypothesis should offer an explanation of why cyclic populations are in northern, continuous biomes. (2) The theory should account both for the damping and disappearance of cyclic fluctuations in southern habitats and, conversely, for the appearance of nonrandom fluctuations such as those that have recently occurred in red grouse populations at Kerloch, Scotland (Moss et al. 1984). (3) A valid model should also provide an explanation for the skips that often occur in a series of cycles (Leopold 1931, Fig. 15.30). Furthermore, a satisfactory account should specify (4) why cyclic behavior is absent in populations in which birds have low, annual mortality rates, such as blue grouse on Vancouver Island, spruce grouse in Alberta, and whitetailed ptarmigan in Colorado. In addition, a hypothesis is needed that (5) can identify the reasons why one population is cyclic, and one immediately adjacent to it is noncyclic. Some of the red grouse populations in Scotland fluctuate irregularly (MacKenzie 1952, Hudson et al. 1985), yet rock ptarmigan living immediately nearby and 200 m higher in elevations are cyclic (Fig. 15.32). Ruffed grouse in the Turtle Mountains of North Dakota are cyclic but sharp-tailed grouse living only a few kilometers away fluctuate in a noncyclic fashion. The densitydependent selection model is sufficient to account for these characteristics in grouse. Cyclic populations differ from noncyclic populations in that they are found in northern biomes where there are few nest predators (Figs. 15.7, 15.30). Nesting success in cyclic populations is invariably in excess of 60% (cf. Watson 1965, Kupa 1966, Bergerud 1970a, Gardarsson 1971, Weeden & Theberge 1972, Rusch et al. 1984). This high nesting success, coupled with typical brood attrition and overwinter loss, provides sufficient recruits for cyclic populations to increase (Table 15.10) so that ultimately density-dependent competition will occur. Females in noncyclic populations generally have a nesting success of less than 60%. This high loss, coupled with chicks lost in summer and overwinter mortality of juveniles and adults, does not provide sufficient recruitment for continuous population growth (Table 15.10). A necessary but not sufficient condition for cyclic populations is high nesting success. Originally, Maine, New York, and New Brunswick were covered with contiguous forests. Nest predation would have been lower than now, and nesting sue-
POPULATION ECOLOGY OF GROUSE
673
Table 15.10. Life equation of a hypothesized, typical, cyclic grouse population compared with that of a noncyclic population Cyclic population Population parameters: 50 breeding males and 50 breeding females; mean clutch size that hatches = 9 eggs (reduced to account for infertility, etc.); 0.70 nesting success (Table 15.2); 0.55 survival of chicks in broods; 0.45 survival of juveniles overwinter; annual mortality of adults 0.50 (Table 15.6). Population change year 1 to year 2: Recruitment (m x ): 50 females x 9 eggs x 0.70 nesting success x 0.55 brood s x x 0.45 overwinter sx = 78 yearlings. Mortality (q x ): 100 adults (year 1) x 0.50 sx = 50 adults 50 adults + 78 yearlings = 128 birds (year 2) (m x > q x ): Population increases Noncyclic population Population parameters: 55 breeding males and 45 breeding females (Table 15.8); mean clutch size that hatches = 9 eggs (reduced to account for infertility, etc.); 0.50 nesting success (Table 15.2); 0.55 survival of chicks in broods; 0.45 survival of juveniles overwinter (Gullion & Marshal 1968); annual mortality of adults 0.50 (Table 15.6). Population change year 1 to year 2: Recruitment (m x ): 45 females x 9 eggs x 0.50 nesting success x 0.55 summer survival of young X 0.45 overwinter sx = 50 yearlings. Mortality (q x ): 100 adults (year 1) x 0.50 sx = 50 adults. 50 adults + 50 yearlings = 100 birds (year 2) (mx = q x ): Population stable
cess of ruffed grouse in these areas would have been comparable to the still-high nesting success of cyclic ruffed grouse in Alberta, about 76% (Rusch et al. 1984). When settlers arrived, small farms appeared; domestic animals accompanied the homesteaders, and habitat diversity increased. The common nest predators found in settled landscapes—red foxes, skunks (Mephitis mephitis), raccoons (Procyon lotor), and corvids—all should have increased. The cycle has disappeared in New York (Bump et al. 1947, Keith 1963), where from 1930 to 1942 the nesting success of ruffed grouse was 61% (Bump et al. 1947). This is borderline nesting success for a cyclic population. Noncyclic ruffed grouse in Massachusetts had a 50% success rate (Banasiak 1951), and in New Brunswick, where the cycle has also disappeared, on average only 47% of the hens were successful from 1965 to 1966 (Neave 1965).
674
A. T. BERGERUD
Nest predation not only reduces potential recruitment, it also brings extrinsic "noise" to what was a more intrinsic, self-contained system. Mice and insect populations fluctuate and foxes, skunks, and weasels increase and switch from buffer species to grouse (Bump et al. 1947; Chap. 11). Random weather effects influence plant cover, hence, the rate of nest predation (Fig. 15.11). Populations cannot show smooth changes in breeding success when predation of nests is high and variable. The cycle of ruffed grouse in northern Minnesota is becoming less pronounced. There was no clear peak for an expected high in 1960-61 at Cloquet (Fig. 15.30). Predation of eggs had been 24% from 1933 to 1934 (n = 477), but increased to 32% by 1959-66 (n = 169) (Kupa 1966). Overall success at Cloquet was 69% (Kupa 1969, Gullion 1970c), whereas in central Minnesota a marginally cyclic population (Fig. 15.30) showed nesting success of approximately 59% (Maxson 1974). Nest predation should increase in the years ahead because the aspen forests of northern Minnesota are becoming fragmented, and the cycle will disappear. In the Turtle Mountains of North Dakota the amplitude of the 10-year cycle of ruffed grouse has moderated in the past 30 years (Fig. 15.39). The forest in the United States portion of the mountains is rapidly being converted to farmland (Schulz 1983). Foxes are replacing coyotes as the dominant canid (A. Sargeant pers. comm.). Foxes are a much more effective nest predator of grouse than are coyotes because of the much higher density of foxes —a home range of 13 km2 versus 62 km2, respectively, in North Dakota (Scott 1982). The percentage of females in the population has declined in each of the peaks of the last three cycles (Fig. 15.39), possibly owing to increased predation. This cycle is probably being damped because predators can destabilize nesting success (< 60%) at progressively lower population peaks. When predation increases, populations can generate positive growth only at lower densities than formerly (Fig. 15.39). Only at low densities can individuals be spaced in cover of good quality and remain inconspicuous in the face of more species and numbers of effective predators. Also, when densities are reduced, competition among individuals for advertising and nest sites will be less intense—space buffers social interaction and competition among genotypes, thereby damping cyclic fluctuations. The prairie chicken population in Wisconsin once showed 10-year cycles of abundance. Peaks were about 1857, 1867, 1878, 1887, 1897 (Schorger 1944), 1909, 1915, 1923 (Leopold 1931), and 1933 (Grange 1948, Anonymous 1976). Minor peaks may have occurred in about 1940 and 1950 (Hamerstrom & Hamerstrom 1973), but the cycle was damping out, and the major peaks in the 1960s and 1970s never occurred (Fig. 15.30). In the mid-1800s this chicken population had rapidly increased following clearing of the land (Leopold 1931, Schorger 1944). The creation of habitat was a necessary but not a sufficient condition for
POPULATION ECOLOGY OF GROUSE
675
Fig. 15.39. Ruffed grouse in the aspen parkland of the Turtle Mountains, North Dakota, are cyclic, whereas sharp-tailed grouse in the grasslands farther south are noncyclic. The grasslands have sufficient kinds and numbers of predators to reduce nesting success so that density-dependent selection seldom occurs. However,the cycle of ruffed grouse is damping out. With each successive peak, there were fewer females, consistent with the hypothesis of increased rates of predation and the disappearance of the cycle in the future. (Data from Hillman & Jackson 1973, Kobriger 1975, 1981, Linde et al. 1978, Schulz 1981.)
population expansion; to increase, a population must also have positive recruitment. When new habitats are created quickly, grouse can spread out and avoid nest concentrations, and thus reduce nest losses. But possibly more important, in this example there was a lag in predation pressure following settlement. Red
676
A. T. BERGERUD
foxes did not enter southwestern Wisconsin until about 1895 (Leopold 1931). Skunks increased after settlement (Scott 1956) and trapped raccoons increased from an estimated 6,000 in 1930 to 200,000 by 1975 (Anonymous 1976). Red foxes increased throughout the Midwest to high levels during the period 1931-45 (Sargeant 1983). By 1940 and 1950 birds were losing the nesting habitat (Westemeier 1971b) needed to counter increasing predator populations. The percentage of female prairie chickens in the population during the low, noncyclic years (1949-65) was 36% (n = 890) (Hamerstrom & Hamerstrom 1973), whereas at the end of the cycling phase, 10 to 15 years earlier, the percentage of females in the population had been 47% (1934-41, n = 528; Hamerstrom & Hamerstrom 1949). The cycle disappeared when nesting success dropped below 60% in the 1930s and 1940s (Gross 1930, Hamerstrom 1939, Grange 1948). Missing were the consecutive years of population increase needed for densitydependent selection. The high population in 1982 (Fig. 16.18) raises the exciting question of whether this increase forecasts the return of density-dependent selection and the return of the cycle to central Wisconsin. Since then the population has again declined, suggesting a cyclic pattern. Predator populations have increased since humans caused erosion of the forests and altered prey diversity. Increased predators include the major nest predators of grouse—red foxes (Sargeant 1984), skunks, and corvids. Raccoon populations moved north. Other predators disappeared with settlement—the wolf (Canis lupus), lynx, and so on—but their disappearance meant little in grouse dynamics, as these species were not serious nest predators (cf. Johnson & Sargeant 1977). As the predators of field and farm expanded, nesting success in cyclic populations should have dropped below 60%. The density-dependent selection model predicts that the cycles will disappear in these situations. In the future, the question of cycles will be even more contentious and more biologists will ask: Grouse cycles: are they real (Gullion 1973)? They become less real with each passing decade, but they were of our past. The density-dependent selection model also offers an explanation for the appearance of cycles where formerly none existed. Harvest statistics for red grouse at Kerloch, Scotland, showed irregular fluctuations from 1921 to approximately 1937 (Jenkins & Watson 1970). Population counts starting in 1962 showed nonrandom fluctuation; there were four runs of increase and decrease in population growth in the 14 years from 1963 to 1977 (Watson et al. 1973, Watson & Moss 1980). Because of the reduction of predators in Scotland, red grouse enjoy nesting success in excess of 75% (Jenkins et al. 1963, 1967), and thus the potential to cycle is present, although there has been extreme variability in the survival of chicks in summer. The research in Scotland documented that the quality of heather affects the hens' nutrition and the viability of chicks at birth—the Siivonen hypothesis (Moss et al. 1975). Disease of chicks may also be a problem (Hudson et al. 1985). The quality of heather, in turn, is affected by random variation in weather, which would add phenotypic "noise" to population growth.
POPULATION ECOLOGY OF GROUSE
677
However, from 1969 to 1971 there was a progression of good breeding seasons (Watson & Moss 1980, Moss et al. 1984) that allowed the population to build up. This high density could then have resulted in competition between densitytolerant and density-intolerant phenotypes, providing the template for cycles. This competition did occur; cocks at Kerloch were less aggressive when the population was increasing than when it was declining (Moss et al. 1984). Also, egg size, chick viability, and chick growth rates were higher when the population was expanding than when it was high and declining. These behavioral and physical parameters have a heritable component (Moss et al. 1982b, 1984), yet the red grouse workers reject a genetic explanation for their findings because results from wild eggs and subsequent progeny did not agree with those from the common lines of birds bred and reared in captivity. However, the genotypic recombinations from aviary crosses would not replicate field crosses because female choice of mates or their territorial requisites would not be duplicated in the aviary. Rock ptarmigan in Scotland live only a few hundred meters higher in elevation than some neighboring, noncyclic red grouse. These rock ptarmigan in Scotland also have the high nesting success needed for true cycles, 76% (Watson 1965). In addition, these birds initiate nesting at the same time that spring green-up occurs (Watson 1965). Thus, compared with red grouse, there should be reduced variability in maternal nutrition and chick viability in rock ptarmigan. Our genetic simulation of the population dynamics of this rock ptarmignan population studied by Watson (1965), gave a cycle slightly longer than 10 years, in close agreement with the actual cycle periodicity (Chap. 12, Fig. 15.32). The potential phenotypic variability between years for laying hens of changing nutrition is generally absent for northern cyclic grouse. For example, the sharptailed grouse in Saskatchewan and Manitoba generally hatch their eggs in the second or third week of June. Females are thus laying eggs in the last week of April. The growing season generally begins about 20 April along the cyclic/noncyclic interface through North Dakota and Minnesota (Fig. 15.30). Grouse north of this line have a reduced potential for nutrition to affect intrinsic chick viability. For grouse south of the line this potential is present, but the sequence has not yet been documented in North American grouse. A more important environmental perturbation between cyclic sharp-tailed grouse in Alberta, Manitoba, and Minnesota, and noncyclic sharp-tailed grouse in North Dakota is nest predation. The Dakota birds live in grasslands where herbaceous nesting cover varies between years depending on rainfall. These birds face an array of prairie and farm predators. In contrast, cyclic sharp-tailed grouse in Manitoba live in the aspen parklands where the moisture regime is more constant and nesting cover more diversified. In addition, these latter populations are hunted by fewer kinds of predators, which exist at lower densities. One of the main predators in the aspen parklands is the coyote, which exists in much lower densities than does the fox in the grasslands to the south. The potential for smooth changes in breeding success resulting from density-
678
A. T. BERGERUD
dependent selection is now absent in the prairie grouse of the grasslands. However, prairie grouse could have cycled in the brief period between the disappearance of the buffalo and the wolves, and the increased arrival of the new contingent of field and farm predators on the coattails of settlement. During this period nest cover should have improved and only few predators were about (Chap. 16). The density-dependent selection hypothesis provides an explanation for skips that occasionally occur in cyclic populations. Our simulation of competition among genotypes with different fitnesses has failed to demonstrate cycles at approximately hundred-year intervals (Fig. 12.3). Skips in cycles will occur whenever breeding populations remain low, so that genotypic sorting would not occur. A prediction of the Chitty hypothesis relevant to this discussion is that population declines can be prevented if competition is reduced by removal experiments (Krebs 1978a). A series of random, inclement springs or disasters could reduce chick survival so that populations failed to increase as predicted using past trends. The Newfoundland willow ptarmigan skipped a cyclic peak in about 1930; there were large forest fires in 1927 and 1929 (Bergerud 1971), and these fires could have killed young birds, destroyed nests, and removed nesting cover. The willow ptarmigan at Chilkat Pass failed to decline in number in 1980-81 (Hannon 1982, pers. observ.), whereas other populations in the Yukon crashed in 1981 (Mossop pers. comm.). The Chilkat population was extremely heavily harvested, and we could not find "waiting" birds in 1980-81, as Mossop had in 1972 (Figs. 10.20, 10.22). This harvest could have delayed selection for density-intolerant birds, the forerunners of decline. Many grouse populations do not cycle—even some in which birds begin nesting before new plant growth appears. Hens in nearly all noncyclic populations have nesting success that on average is less than 60%. For example, white-tailed ptarmigan have a 57% nesting success (Giesen et al. 1980); blue grouse on Vancouver Island, a 55 % success rate (Zwickel 1975, Zwickel & Carveth 1978); and spruce grouse in Alberta hatched only 19 of 66 nests, or 29% (Keppie & Herzog 1978, Keppie 1982). More important, we could not generate cycles in our simulation when populations had low, annual mortality rates less than 40-50%. Birds in these populations that show low mortality rates also have smaller clutches and raise smaller broods (Figs. 15.1, 15.4). The low turnover of adults thus reduces the potential for large variations in brood size and prohibits a rapid, cyclic, polymorphic response between adjacent cohorts. However, these populations had the potential to erupt in our simulation (Chap. 12) if they encountered a run of good breeding seasons. Such eruptions have indeed occurred in blue grouse (Fig. 15.28). Adjacent cyclic populations are often synchronized (Keith 1963), but populations become more out-of-step when one compares populations that are geographically well-spaced (Bergerud 1970a). The interaction of smooth reproductive curves of chicks with different viability and vulnerability to random extrinsic
POPULATION ECOLOGY OF GROUSE
679
weather can partly explain both this synchronization and the absence of it (Leslie 1959, Chitty 1960). The susceptibility of chicks to mortality caused by weather would vary among genotypes. Adjacent populations face the same weather systems and would thus become more uniform in genetic character. Predation by raptors that range widely between populations might also improve the synchronization of adjacent populations if this predation resulted in differential culling of genotypes. In one cycle of ruffed grouse at Watch Lake, British Columbia, redphase birds represented one-third of the population, but in the second peak, they represented only one-fourth (Fig. 3.4). However, in both instances, after the winter crash in which goshawks killed proportionally more red-phase than grayphase birds, the percentage of reds was similar, 2-4% (Figs 3.4, 3.16). In summary, the density-dependent selection hypothesis to explain the 10-year cycle in grouse is based on the belief that all populations are polymorphic in density preference, with some individuals more adapted to living in high densities (density-tolerant) and others to living in low densities (density-intolerant) (Chap. 14). Given that the extrinsic environment is relatively benign, recruitment can generally exceed mortality for several consecutive years. As the population builds, the relative advantages of the density-intolerant birds increase because of the competition among males for advertising sites and because of female choice of successful, aggressive males. In effect, all females breed annually but many males do not; thus fitness paces genotypic variability between genotypes. Aggressive, density-intolerant genotypes that are favored with intense competition are postulated to be more homozygous (Fig. 2.26), produce less-viable young (Bergerud 1970a), and contribute less to population growth than heterozygotes. These young die for a variety of reasons, but the summed result is that recruitment becomes insufficient to replace the more-or-less constant, natural mortality of adults, and the population then declines. The 10-year periodicity results from the relatively constant mortality rates of cyclic grouse (50%-60%); hence, a twogeneration turnover period with its inherent lag effects. Extrinsically induced weather and predator mortality have different impacts on the genotype percentages in populations, thereby reducing genetic variability and adding some generation synchronization to nearby populations exposed to similar extrinsic regimes.
15.9 Limitation of breeding numbers No simple hypothesis of population regulation can be expected to explain the great diversity of numerical responses observed in grouse, but a reasonable, comprehensive body of theory can be fashioned from a few general principles. Any model must include those population parameters that have clearly been documented in the literature to influence breeding numbers. A viable population is maintained only if production to autumn equals overwinter losses (Table 15.6). A population cannot sustain both a high loss of eggs
680
A. T. BERGERUD
and young and a high mortality rate of adults. The first generalization is that there is a south-north continuum for these parameters; loss of nests is greatest in the south, where there is more predation of nests. Adult mortality is probably higher in the north, where there are more effective raptors; however, there are few studies of adult mortality in the south to verify this. The southern distribution of grouse is probably limited by nest predation, and the northern distribution limited by protective cover to avoid predators. A second generalization is that grouse die primarily from predation. There is no substantial evidence that grouse of North America, other than young chicks, starve or succumb to disease in normal populations. The vulnerability of grouse to predation changes seasonally. They are most susceptible when they must compromise their inconspicuousness to advertise and to nest and rear chicks. It should be expected that the mortality rates of males and females will be different, because they differ in their investment in offspring. A third principle is that individuals of each species are preadapted to a specific predator-escape cover complex, i.e., their species-specific habitat. Willow ptarmigan are found in the tall-shrub zones of the tundra. Gyrfalcons also occupy this zone and ptarmigan depend on willow shrubs for cover to escape this raptor (Chap. 10). Each habitat has its own array of coevolved predators and vegetation with unique morphological characteristics used by grouse to escape these predators. Birds of each grouse species, indeed of each population, will be subject to a unique mortality rate that is extrinsically determined by resident predators and cover. Annual mortality rates are more constant than breeding success rates. But mortality can be expected to vary if predator abundance changes, or if the searching pattern of predators varies with changing, alternate-prey abundance. Mortality rates may vary among years if the inconspicuousness of the grouse is altered. Grouse behavior will add another dimension—density-intolerant birds have a suite of behaviors that improves their fitness in some intraspecific competitive situations, but at the same time may increase their vulnerability to predation. If mortality rates change among years, they must alter breeding numbers. But mortality rates are only about half as variable as changes in breeding success. Mortality is induced by predator populations in which individual grouse have evolved their own adaptations to maximize fitness. This coevolution lends stability to the mortality rates of grouse. Breeding success, in contrast to mortality rates, varies more in response to the extrinsic environment, outside biological adaptation, and without lag effects or buffers. I can find little evidence of density-dependence in the mortality rates of adult grouse. A relatively constant proportion of the population dies between breeding seasons. There were exceptions—the death of prairie chicken hens in summer increased with numbers as did the mortality of advertising, male ruffed grouse. A major predator like the goshawk generally hunts ruffed grouse, regardless of
POPULATION ECOLOGY OF GROUSE
681
population size, but the goshawk may change its range in response to encounter rates. It is the breeding-success side of the population equation that most influences changes in breeding numbers of grouse. Nearly all hens nest; clutch size varies among species and populations, but it shows little variation between years that is correlated with population change. Nesting success varies markedly in relation to: (1) changes in nesting cover, (2) predator pressure, (3) age structure (experience of the hen), and (4) female behavior. Nesting success of grouse is the most variable parameter in the dynamics of their populations, and through its influence on breeding success, contributes more to changes in population size between years that any other parameter. The survival of chicks also has a fairly large coefficient of variation (Table 15.6). As in other bird species, very young grouse are the weakest link in the lifehistory of a cohort; chicks die from many causes before they can thermoregulate. Apparently no populations do not lose at least 25% of the chicks that hatch (Fig. 15.1). The major variables that likely alter chick survival are spring temperature, insect abundance, intrinsic viability, predation, concealing cover, and maternal condition. Also, the parental solicitousness may change between cohorts in cyclic species and affect chick survival. We can recognize at least five extrinsic patterns and one intrinsic pattern to breeding success. (1) Chick survival can be influenced by the viability of young at hatch, which in turn is affected by maternal nutrition or parasitism mediated through weather and food in the prelaying period. Red grouse are an example (Fig. 15.40). (2) Breeding success can be influenced by nesting cover, which is influenced by rainfall. Sharp-tailed grouse in South Dakota are an example. (3) Breeding success can vary with stages of plant succession, as for blue grouse on Vancouver Island. When birds first colonize a newly logged area, nesting success and breeding survival are low (Zwickel & Bendell 1967). When after a period of years the forest grows too dense and homogeneous relative to the food for chicks, recruitment declines. (4) Breeding success can vary in relation to cycles in mouse populations and cover and predators switching to nesting grouse. The willow ptarmigan in Norway is an excellent example. (5) Breeding success can vary with June temperatures that affect insect abundance and the need to thermoregulate; the ruffed grouse at Watch Lake (Chap. 3) is an example. (6) The one intrinsic pattern is that breeding success can vary with density-dependent selection between genotypes. Rock ptarmigan in Scotland may be an example. These annual changes in productivity are the driving forces of population changes (Fig. 15.41). The density of grouse, in turn, is a function of the species-specific habitat (space) prorated to the total number of birds alive (Fig. 15.41). It is an effect of last year's population demography; however, the density of birds is the arena in which density-tolerant and density-intolerant morphs compete for forthcoming
682
A. T. BERGERUD
Fig. 15.40. Summary of six scenarios that can affect breeding success and drive numbers: (1) the scenario of June temperature from Chap. 3; (2) soil moisture based on Fig. 15.11; (3) maternal food or parasitism based on Moss et al. 1975 or Hudson et al. 1985; (4) predator switch-over from Chap. 11; (5) forest succession based on Zwickel & Bendell (1967), Redfield (1970), Zwickel et al. (1977, 1983), and Frandsen (1980); (6) the 10-year cycle scenario based on Fig. 15.32 —data from Watson (1965).
POPULATION ECOLOGY OF GROUSE
683
Fig. 15.41. Summary model of population limitation in grouse.
fitness. For noncyclic populations with high nesting losses, the sorting takes place primarily by dispersal of these morphs to seek those densities that suit their evolved suite of breeding characteristics and will result in maximum fitness (Chap. 2). For the cyclic populations living in more homogeneous and stable environments, with fewer extrinsic restraints to breeding success, dispersal is less important. For these birds, density-dependent competition for advertising and nesting sites is more intense. The ultimate question—the lead sentence in the introductory lecture on population ecology—What prevents the unlimited increase of a population; why don't natural populations continue to increase as does man?—can be partly answered with respect to grouse. Many populations fluctuate free of density-dependent restraint; they may increase when food is favorable for chicks or again decrease when nesting cover is reduced in dry cycles. These populations fluctuate in re-
684
A. T. BERGERUD
sponse to a shortage of time, as explained in the random-walk theory of Andrewartha and Birch (1954). To study these populations is to never secure an answer to the question of the ultimate check of population growth. But there are a few insights. The ultimate limit of natural, noncydic populations coexisting with undisturbed predators is density-dependent nesting success. If the environment is benign and natural populations continue to increase, the first requisite in short supply is adequate space so that hens can nest at sufficient distances from each other such that predation is reduced and sufficient progeny are hatched to equal natural, adult losses. Potts (1980) and Angelstam (1983) have reached a similar conclusion—that nesting success is the density-dependent damping mechanism in Hungarian partridge and black grouse. The inverse relationship between population growth and density in grouse comes down, in the end, to nesting success as it is influenced by space, cover, and predators. In those cases where there are insufficient predators in the environment to cause negative, density-dependent nesting success—the cyclic populations — density-dependent selection between genotypes at high densities can result in chicks of insufficient viability to maintain the population, and again population growth is limited.
15.10 Summary The number of breeding grouse in a population constantly changes between years. Documentation of the mechanics of this change is the primary aim of this chapter. Parameters that could influence annual changes in the number of grouse are: percentage of hens that nest, variations in clutch size and nesting success, survival of chicks in summer, survival of juveniles in winter, and variations in annual mortality rates of adults. Nearly all hens attempt to nest. Clutch size is relatively constant between years and is also insufficient to account for the large, annual changes in the number of grouse. There are, however, large differences in clutch size between populations of the same species. The characteristic clutch size of each population is hypothesized to have arisen from selection with respect to lifetime fitness as influenced by the characteristic longevity of individuals in specific populations. Clutch size may have been further modified in some populations from that predicted by the expected longevity of an individual female because of directional selection from nest predators against females that lay large clutches. Nesting success of grouse is generally low; as a group only about 58% of the nests hatch. Predation accounts for an average of 79% of nest failures. Mortality of chicks is also high; on average 44% of chicks die before fall. But chick mortality is not correlated with clutch size and occurs regardless of the presence or absence of predators. Abundant insect food appears to be a necessary precondition for high survival of chicks. Mortality rates of juveniles and adults show large differences among populations, ranging from 18 % to 81 %. The differences can be attributed to the unique, predator-cover complex in which each
POPULATION ECOLOGY OF GROUSE
685
population lives. Birds in some populations are quite secure from predators. In populations that show this low mortality mode (< 45% annual qx), selection favors females that lay small clutches. In noncyclic populations with a preponderance of males, mortality of females is high during the nesting period. The greater proportion of males in cyclic ptarmigan populations is possibly explained by the increased mortality of female chicks during population declines. The model of the 10-year cycle proposed by Bergerud (1970a) is further refined by the hypothesis that the switching mechanism between density-tolerant and density-intolerant phenotypes and population increases and declines is mediated by female choice of aggressive males with large territories when the density of birds is high, and by choice of docile males that control high-quality nesting cover when populations are low and birds are spaced far apart. The major conclusion of this chapter is that variations in breeding success drive population changes between years; overwinter mortality is relatively constant. The ultimate damping mechanism to the growth of noncyclic populations is postulated to be density-dependent nest predation; that to the growth of cyclic populations, density-dependent changes in mate choice between genotypes that provide chicks with differing intrinsic viabilities.
16
Increasing the Numbers of Grouse A. T. Bergerud
16.1 Introduction A tendency of management biologists is to proceed on the basis that if there is more habitat, it follows that there will be more grouse. It is true that habitat is the template for population growth, but it is only a necessary, and not a sufficient, precondition. Mortality rates are generally fixed by the predator-cover complex in which grouse live. For a population to increase, breeding success must be greater than that needed to stabilize numbers. The strategy of management should be to improve reproductive success. The format of this chapter generally follows the topic outline discussed by Aldo Leopold in Game Management (1933). He proposed five major management schemes: control of hunting, control of predators, control of food and water, control of cover, and control of disease. Disease and water are not major limiting factors of grouse in North America (but see Potts et al. 1984 for red grouse [Lagopus lagopus scoticus] in England), and are not included. In lieu of those factors I have added control of space, which represents manipulation of the interaction of birds with their predator-cover complex, to enhance survival and breeding opportunities. Before addressing these limiting factors, however, I will discuss the determinants of density.
16.2 Mechanisms of density 16.2.1 Nesting success and density In general we count males to obtain an index of density, but males space themselves to be near females (Bradbury 1981, Chap. 14). Thus, it is of paramount interest that the locations of females are the basic spacing unit. Females, in turn, 686
INCREASING THE NUMBERS OF GROUSE
687
space their nests to remain inconspicuous (Chap. 14.)- They reduce their contact with males and other females to minimize the risk of predation. The major determinant of density in grouse is, I believe, the space needed for successful nesting so that mx (recruitment) = qx (mortality). Evidence now indicates that the prelaying ranges of female blue (Dendragapus obscurus), spruce (Dendragapus canadensis), and possibly ruffed grouse (Bonasa umbellus) are generally mutually exclusive (Rusch & Keith 1971a, Zwickel 1972, Herzog & Boag 1978, Bergerud & Butler 1985). The prelaying ranges of females of the three ptarmigan species (Lagopus spp.) are also mutually exclusive (see Chap. 13). Anglestam (1983) reported that the ranges of female black grouse (Tetrao tetrix), a lek species, also were mutually exclusive, but this has not been shown in the literature for females of the three North American lek species. However, M. Gratson (pers. comm.) indicates that the nests of sharptailed grouse (Tympanuchus phasianellus) are mostly spaced. Also, a recent increase in prairie chickens (Tympanuchus cupido) in Minnesota has been accompanied by an increase in the number of leks and not in the number of displaying males per lek (Fig. 16.1). This association suggests that spacing occurs in prairie chickens. The prelaying ranges of the lek species are large (Fig. 16.2), and, because such large ranges could not be easily defended without a serious trade-off in energy and loss of inconspicuousness, the spacing mechanism that probably occurs is mutual avoidance rather than mutual exclusion. Females of each grouse group (forest, tundra, and steppe) travel in prelaying ranges of different sizes (Bradbury 1981), and these ranges are larger where nest predation is serious (Fig. 16.2). My interpretation is that females faced with more predation pressure search larger prelaying ranges before investing in a specific nest site (Chap. 14). The difference in the size of prelaying ranges, coupled with the fact that males space themselves to maximize contact with females, suggests that males of lek species will always be at lower densities than forest grouse males, when the entire habitat is considered. Also, the larger the prelaying ranges of the females, the greater the number of males that will be clustered at advertisement sites (Bradbury 1981). Similarly, tundra grouse should, and do, generally occur at higher densities than forest grouse because prelaying ranges of ptarmigan are smaller than those of forest grouse (Fig. 16.2). The low densities reported for ptarmigan in the high Arctic (Weeden 1963) resulted from the inclusion of large blocks of unsuitable habitat in the tabulations rather than from actual, low densities. A more meaningful index to density would be to measure the mean distances between displaying males, or, if there were data, between nesting females. Lance (1970) gave the distance between five blue grouse nests at Comox Burn, British Columbia, as 274 + 101 m. The distance between four adjacent, willow ptarmigan (Lagopus lagopus} nests at Chilkat Pass, British Columbia, was 120 + 26 m (Fig. 13.5). The mean size of the prelaying ranges for five adult, blue grouse
688
A. T. BERGERUD
Fig. 16.1. The prairie chicken population in Minnesota recently increased. This increase was accompanied by an increase in the number of leks and not by an increase in the number of males per lek. (Data from Svedarsky pers. comm.)
hens at Comox Burn was 6.3 + 3.19 ha (Hannon 1978), and for 18 willow ptarmigan hens at Chilkat Pass, approximately 2.1 ha (Hannon 1982). Using either measurement (distances between nests or sizes of prelaying ranges), ptarmigan were spaced 2.3 times closer together than were blue grouse hens. How does this compare with male densities? The mean density of advertising, blue grouse males at Comox Burn from 1960 to 1967 was approximately 17 males/km2 (Zwickel et al. 1983), whereas willow ptarmigan males at Chilkat Pass from 1958 to 1976 occurred at an average density of 36 males/km2 (Fig.
INCREASING THE NUMBERS OF GROUSE
689
Fig. 16.2. Sizes of prelaying ranges of females are larger where there is more nest predation. Nesting statistics from Table 15.2; prelaying ranges from Table 14.2.
10.2). Thus, the ratio of blue grouse density to that of ptarmigan is similar to the ratio obtained from nest spacing and prelaying ranges. The hypothesis that the size of prelaying ranges generally determines density predicts that the density of populations faced with few nest predators will be higher than that of populations coping with many predators and high losses. In the former populations, females can nest closer together—reduce their space—and still have sufficient nesting success such that mx = qx. The fairly distinct, prelaying ranges of each grouse species may reflect an evolutionary stable strategy (ESS) (sensu Maynard-Smith & Price 1973). Selection may have favored females that invested the appropriate search effort in their specific predator-cover complex, which maximized their lifetime breeding success. An alternative hypothesis to explain densities is that the spacing of grouse in a specific habitat is determined by food supplies (Jenkins et al. 1967). Red-grouse biologists noted early in their work that territories were smaller and densities of grouse were greater when heather (Calluna vulgaris) was abundant (Miller et al. 1966). But heather serves as both the primary nest cover and food for red grouse. Watson (1964, 1970) and Watson and Moss (1970) proposed that the quantity of heather would influence the visibility of males, hence their interactions, and that the quality of heather as food would further mediate the intrinsic aggressiveness
690
A. T. BERGERUD
of males, and hence territory size. However, they have been unable to document that food resources working through the behavior of males determine fall territory size and thus densities (Lance 1978b, Miller & Watson 1978). An alternative explanation is that the correlation between heather abundance and red grouse densities is a function of the cover value of heather for nesting hens. Males can afford to advertise and defend smaller territories and still attract hens when their territories include heather of high quality for nesting. Territory sizes are smaller after fertilization (review Watson et al. 1984) because of the improved growth of nesting cover for females. Densities of ptarmigan are positively correlated with nesting success (Fig. 16.3). Three populations that showed extremely high numbers lived on islands relatively free of mammal predators: Brunette Island, Newfoundland (Fig. 15.35); Tran0y Island, Norway (Fig. 11.1); and Hrisey Island, Iceland (Fig. 9.1). The other high-density populations are red grouse in Scotland, where predators are controlled, and willow ptarmigan at Chilkat Pass, British Columbia, where nest predation is relatively light (Weeden 1959b, Hannon 1982, Hannon & Smith 1984). The ptarmigan hens in six populations nested farther apart and had a reduced nesting success (Fig. 16.3); had the males in these populations
Fig. 16.3. Spring density of ptarmigan is higher for populations in which there is little nest predation. (Data from Choate 1963a, Jenkins et al. 1963, 1967, Watson 1965, Bergerud 1970a, Braun & Rogers 1971, Gardarsson 1971, Myrberget Chap. 11, Weeden & Theberge 1972, Watson & O'Hare 1979, Giesen et al. 1980, Hannon 1982.)
INCREASING THE NUMBERS OF GROUSE
691
defended much smaller territories, they would have reduced their fitness by not attracting females or by an increased loss of nests. By avoiding males with small territories (Fig. 10.19), females may set the lower limit of territory size. Forest grouse also space their nests, and densities should increase when nests can be spaced close together. The highest mean density reported for ruffed grouse occurred at Rochester, Alberta—32 males and females/km2 in the spring. Nesting success there was 76% (Rusch et al. 1984). A lower density of ruffed grouse occurred in Kentucky—6 males/km2, and nesting success was low, approximately 25% (Hardy 1950). It has been argued that the abundance of aspen (Populus tremuloides) is the essential factor determining the density of ruffed grouse (Gullion 1977a, 1977b, 1982), but the demographic aspects of why densities could be higher in aspen forests than in other forest types have not been specified. The annual mortality rates of ruffed grouse are similar for several populations, from Wisconsin to Alberta (Dorney & Kabat 1960, Gullion & Marshall 1968, Boag 1976b, Rusch et al. 1984). All these populations live in aspen forests. But the ruffed grouse population with the lowest reported mortality rate lives in the Pacific Northwest, in an alder-maple (Alnus spp., Acer spp.) forest (Brewer 1980). The correlation presented by Gullion between aspen and ruffed grouse density may be in fact a nesting success-density relationship, since nesting improves with latitude (Fig. 15.7).
16.2.2 The concept of a stabilizing density An older argument in the literature is whether a population's mean density averaged over many years represents a biological balance (Nicholson 1933) or is merely a statistical average of random, varying fluctuations (Andrewartha & Birch 1954). If nesting success mediates density, the mean density of a population represents the average density that just provides sufficient recruitment to equal mortality-i.e., the stabilizing density that provides stabilizing breeding successes (Fig. 16.4). Prairie chicken and sharp-tailed grouse (Tympanuchus phasianellus) in South Dakota provide statistics showing fluctuations around different mean densities (Fig. 16.4). The populations are partly sympatric and the areas are censused by similar methods. These populations are probably affected by randomly varying soil moisture that influences plant cover, hence nesting success (see Fig. 15.11). They show similar juvenile:adult ratios in the autumn, but the stabilizing recruitment (prairie chicken, s = 2.20; sharp-tailed grouse, s = 2.60) occurs at a much lower density for prairie chickens than for sharp-tailed grouse (Figs. 15.25, 16.4). Prairie chickens nest mostly in grasslands in open landscapes; sharp-tailed grouse nest mostly in heterogeneous, grass-shrub cover (Chap. 5), as well as enjoy slightly less predation pressure on nests (Table 15.2). It is possible that chickens compromise plant cover more often than sharp-tailed grouse and use
692
A. T. BERGERUD
Fig. 16.4. Sharp-tailed grouse maintained a mean breeding density higher than that of prairie chickens in South Dakota (top). The stabilizing breeding success (s) is similar for both sharp-tailed grouse and prairie chickens, but the decline in breeding success with increased density is much more pronounced for prairie chickens than for sharp-tailed grouse, resulting in a lower stabilizing density for prairie chickens (bottom) (see Figs. 15.13, 15.14; data from Linde et al. 1978).
INCREASING THE NUMBERS OF GROUSE
693
space more often as a nesting tactic. Their lower stabilizing density supports this hypothesis. Pheasants (Phasianus colchicus) in South Dakota occupy much of the prairie grouse range, nest in similar cover, and face similar predators. Wagner et al. (1965) discussed in depth their observation that the mean density of pheasants was greater in South Dakota than in Wisconsin or Indiana. They recognized that these differences in mean densities resulted from compensatory intraspecific competition, but they did not specify the mechanism. Their hypothesis that females died more in warm, dry springs from physiological stress had no density-dependent connotation. However, such mortality could be explained by the lack of nesting cover in dry cycles and increased predation. Later, Gates and Hale (1974) documented spacing of nesting pheasants, and Dumke and Pils (1973) and Gates and Hale (1975) showed that nest predation and summer predation of hens were density-dependent. The highest density of pheasants reported in the literature was on Pelee Island (Stokes 1954). Pheasants there had a consistently high rate of increase, irrespective of densities (Wagner et al. 1965), and nesting success was higher than elsewhere in the pheasant range (Gates & Hale 1975). With few predators, females could nest near each other and still be successful. The combined findings of these pheasant investigations are consistent with the hypothesis that stabilizing densities have biological reality in that they provide on average stabilizing breeding success. There should be a unique stabilizing density for each predator-cover complex. The density of grouse is driven primarily by nesting success, as it contributes to breeding success; density, in turn, feeds back through inversity to influence nesting success (Fig. 16.5). But these relationships are not a closed circuit. If adult mortality is density-dependent it would moderate nesting numbers and nesting success. Except for the mortality both of nesting, prairie chicken hens and of advertising, ruffed grouse males, I have found little evidence of such densitydependent mortality. Grouse have adopted breeding strategies to minimize such mortality. The major short-circuit in the nesting success-density interaction is brought about by changes in nesting success in response to random changes in the extrinsic environment. Nesting cover and the abundance of buffers and predators change between years, independent of grouse density, and thereby importantly moderate the role of density in influencing nesting success (Fig. 16.5). In Figure 16.6 I have attempted to diagram how breeding success influences mean breeding densities. Consider again prairie chickens and sharp-tailed grouse in South Dakota. Both populations have high nesting success when densities are low (Fig. 15.13, 15.14). The reproductive success of females in the sharp-tailed grouse population declined at an increasing rate as numbers increased. The rate of successful, nesting female prairie chickens also declined as numbers rose, but the decline was more rapid and less curvilinear (Fig. 15.13). Consider that the
694
A. T. BERGERUD
Fig. 16.5. Nesting success is the principal parameter that drives density, but density feedsback and influences nesting success. Nesting success is modified from year to year by extrinsic factors (cover, predators, and buffer species) that further change the influence of density (competition for nesting sites) on nesting success.
two mx lines in Figure 16.6 (top) represent the prairie chicken (I) and sharp-tailed grouse (II) populations—the prairie chicken requiring a more favorable mixture of cover and space than sharp-tailed grouse at similar densities for similar recruitment. Adults in both populations probably have similar natural mortality rates. The stabilizing densities of these two populations should result when recruitment (mx) and mortality (qx) equal each other. Prairie chickens will stabilize at a lower density than will sharptails. The favorableness of the nesting habitat will override other factors in determining the mean breeding densities of these populations. The variation in the number of grouse seeking space each spring within a population will be a major factor in nesting success and will ultimately influence the density next year (Fig. 16.6, bottom). This is a space-cover trade-off. As numbers
INCREASING THE NUMBERS OF GROUSE
695
Fig. 16.6. Top, A: Difference in the stabilizing density of two populations depends on how cover and predation affect the balance between recruitment and mortality; recruitment varies more between populations than does adult mortality. Bottom, B: The difference in nesting success between years that drives density in a population (Fig. 16.5) is greatly influenced by differences between favorable and unfavorable nesting cover in the spacing of birds in spring; this success in turn depends on prior recruitment and the toal numbers competing for the space.
696
A. T. BERGERUD
increase, the best cover for avoiding predators is compromised; some birds (density-intolerant morphs) will seek habitats that have more marginal cover, but that are less populated and thus will have more space per bird. Space will partly compensate for cover in fitness. Still, breeding success will be reduced in such marginal habitat, thereby dampening population fluctuations and annual variations in density.
16.3 Control of hunting The history of hunting has passed through several stages. There was a time in pioneer days when there was no thought of conservation; hunters killed what they desired. However, even as early as 1708, New York State had a closed season on ruffed grouse in the summer to protect young and females (Schorger 1945). The attitude of game preservation increased and was discussed in some detail by Leopold (1933), who realized himself that populations could be overhunted. He remarked, "The virtual disappearance of both quail hunting and bird dogs from some shot-out quail states is a case in point" (Leopold 1933, p. 211). However, wildlife biologists argued that surpluses of game birds were available for harvest. For example, Bump et al. (1947, p. 370) concluded that "the general effect of man's hunting on grouse as currently practiced is not detrimental," in referring to 17% of the preseason ruffed grouse population that was harvested. In the mid 1940s a cyclic low in ruffed grouse approached in the Midwest; Minnesota and Wisconsin closed their season, but Michigan "hung tough." When the season opened in 1948 in Wisconsin and Minnesota, these states appeared to have no more grouse than did Michigan (Ammann 1950). The hypothesis was supported that cycles were natural ecological events not caused by hunting, and the belief was in place that you could not stockpile game.
16.3.1 Compensation principle The compensation principle was ushered in during the 1950s. Allen (1954, p. 131) said, "If we fail to take a hunting harvest, Nature does it for us. It is quite possible, and usual, for the hunter to get in ahead of natural mortality factors and convert the annual surplus of game to his own use merely by taking it before something else happens to it." DeStefano and Rusch, (1982, p. 31) recently said, in questioning the compensatory principle, that "generations of students digested the principle and many biologists came to accept the idea that most game animals present in summer and fall would succumb to late fall and overwinter mortality and that fall hunting would mainly harvest surplus animals." The primary evidence that biologists use to argue that compensatory natural mortality occurs is that hunted populations are commonly the same as unhunted populations when spring counts are taken (Palmer 1956); however, this by itself
INCREASING THE NUMBERS OF GROUSE
697
is not evidence that natural mortality is compensatory. We could expect birds to fill voids created by hunting when they seek advertising and nesting sites in the spring. By moving to habitats of lower densities, birds should be able to improve their fitness, as predicted by one hypothesis of the Fretwell-Lucas habitat model (see Fig. 14.27). Hunters commonly seek grouse in "optimum" habitats, in which sometime after harvest, yearling birds should prospect and settle in their attempt to optimize the space-cover trade-off. Despite this, examples of populations that are being reduced by hunting pressure are now filtering into the literature. In these cases, the size of the buffer (unhunted) zones appears inadequate to provide sufficient recruitment to give the resemblance of compensatory natural mortality (cf. Gullion 1982, Kubisiak 1982).
16.3.2 Is hunting mortality additive to overwinter mortality? One irrefutable test of the question of whether hunting is additive is to compare the annual mortality rates of banded adults between areas with different exploitation levels. For the test to be valid, birds on these areas should have taken part in breeding activities, and thus could be expected to show philopatry. Fred Zwickel banded blue grouse at Comox Burn, on Vancouver Island, British Columbia, from 1962 to 1965 and again from 1969 to 1977 (Zwickel & Bendell 1967, Zwickel et al. 1983). During the first period hunting was light, but hunting pressure increased in the later period (Zwickel 1982). Of particular interest is that mortality rates of banded females were significantly greater during the second period (Fig. 16.7). Zwickel (1982, p. 1,060) said, "This selective removal of females has apparently increased the mortality rate of females." Fortunately, this study had a built-in control: most blue grouse males migrate before the hunting season. Mortality rates of males did not significantly increase between the periods 1962-65 and 1969-77. This indicates that hunting was the principal cause of the changes in mortality of blue grouse between the two periods. Are there other examples? Clait Braun banded white-tailed ptarmigan (Lagopus leucurus) in the high mountains of Colorado from 1966 to 1969 (Braun 1969, Braun & Rogers 1971). Birds in two populations, Crown Point and Mt. Evans, were hunted each fall, whereas ptarmigan in Rocky Mountain National Park were not hunted. The lifetable data Braun (1969) presented indicate that birds in the two hunted populations had mortality rates approximately double those of birds in the unhunted populations (Fig. 16.7). In this example, hunting appears completely additive. At Mt. Evans 51 % of the population was harvested. If we assume that the remaining 49% alive at the end of the season died at the natural mortality rate (46% for females at Rocky Mountain National Park), then 49% x 46% = 23 %; and 23 % plus 51 % harvested equals 74%, compared to the observed total mortality of 76% (Fig. 16.7). Braun (1969, p.86) felt that hunting increased the annual mortality rate by 15 % at Crown Point and by approximately 27% at Mt. Evans. He remarked that
698
A. T. BERGERUD
F\g. 16.7. Mortality rate of banded yearlings plus adults in populations that were subjected to different intensities of hunting pressure. (Data from Braun 1969, Fischer & Keith 1974, Zwickel et al. 1983.)
INCREASING THE NUMBERS OF GROUSE
699
"perhaps hunting mortality was not entirely replacive and was partially additive." These populations in Colorado were maintained by immigration from surrounding areas (Braun & Rogers 1971). Hunting of banded ruffed grouse at Rochester, Alberta, was also additive (Fig. 16.7). The harvest of banded birds within 201 m of roads was 23 %, and the total annual mortality was 77% (Fischer & Keith 1974). Birds harvested at distances greater than 201 m from roads had only a 1 % mortality rate caused by hunting, and their total annual mortality was 64%. Again we can look at unshot birds to confirm the estimates. After the hunting season 77% of the birds should still be alive near the road, and they should have a natural mortality rate of 63%; 0.77 x 0.63 = 0.485, thus a 48.5% natural mortality rate plus the 23% hunting mortality rate equals 71.5% total mortality. This is reasonably close to the observed mortality of 77%. Fischer and Keith (1974, p. 593) concluded that "fall hunting increased total annual mortality of males less than 201 meters from roads, and was therefore in part additive." Frederick and Fran Hamerstrom banded prairie chickens in central Wisconsin from 1949 to 1965 (Hamerstrom & Hamerstrom 1973). The hunting season was closed during these years except in 1951. The two cohorts exposed to this hunting harvest had mortality rates higher than later cohorts that were not hunted (Fig. 16.8). The Hamerstroms concluded (p. 36), "It would seem that the hunting season increased mortality by about 25%." Gordon Gullion has been banding ruffed grouse males for the past 30 years in central Minnesota, 1956-1985. Terry Little (1978) also banded males in central Minnesota at Crow Wing County, 1970-74. The hunting pressures for three study areas in Minnesota lined up as follows: Mille Lacs > Crow Wing > Cloquet. Cloquet is a refuge, but birds are harvested when they move outside (Gullion & Marshall 1968). Grouse at both Crow Wing and Mille Lacs had significantly greater mortality rates than Cloquet (Fig. 16.8). The mortality figures from Cloquet are similar to the losses reported for other ruffed grouse populations living in lightly hunted areas in North America, i.e., 50-60% (Dorney & Kabat 1960, Davies 1973, Boag 1976). Gullion (1982, p.21) concluded for the Mille Lacs region: "This population depression appears to be the result of excessive hunter harvest each fall. Evidently too many potential breeders are being removed by hunting." It is important to note that Mille Lacs is being intensively managed, whereas the surrounding areas have lower densities, thus reducing the potential of ingress by birds to fill vacancies created by hunting (see also Kubisiak 1982). David Jenkins and Adam Watson began studies of red grouse in 1956 and the work continues under the direction of Watson and Robert Moss. In an early paper they reported a mortality rate of 71 % for banded adults that were hunted on their study areas (Jenkins et al. 1963). The overall mortality rate of 828 banded adults on hunted moors in Scotland was 67% (Jenkins et al. 1963). It was concluded that
700
A. T. BERGERUD
Fig. 16.8. Mortality rates of banded yearlings plus adults in populations that were subjected to different intensities of hunting pressure. The red grouse in Ireland and Kerloch were not banded. (Data from Jenkins et al. 1963, 1967, Hamerstrom & Hamerstrom 1973, Little 1978, Watson & O'Hare 1979, Gullion 1981, 1982.)
INCREASING THE NUMBERS OF GROUSE
701
natural mortality was compensatory, "shooting exploited part of the surplus, and natural mortality mostly eliminated other birds that could not obtain or hold territories" (p. 375). Two largely unhunted populations of red grouse are reported in the literature; one is at Kerloch, Scotland, and the other in Ireland. The overwinter mortality figures for these populations, computed by comparing fall populations with spring numbers (no banded birds available), were 39% for Kerloch (1962-65) and 42% for Ireland (1969-71) (Fig. 16.8 data from Jenkins et al. 1967, Watson & O'Hare 1979). The Kerloch birds also had an overwinter loss of approximately 41 % in the years 1966-71 (data from Moss et al. 1975). Red grouse have a high survival rate over the summer, hence losses between fall and spring should approximate losses for the entire year. A comparison of the overwinter mortality rate of approximately 41 % for the two unhunted populations with the annual mortality rate of 67-71% for the heavily hunted populations suggests that, contrary to the view of Moss et al. (1982a), hunting is additive in red grouse. A second index that can be used to decide if hunting is additive is to compare the rate of population change over winter across an array of densities. If hunting is additive, spring numbers should be linearly correlated with fall numbers. As fall numbers increase, there would be no increase in compensatory, overwinter natural loss. If natural overwinter mortality is compensatory, the rate of overwinter loss should increase with increased numbers, and the plot of spring numbers on fall numbers should be curvilinear. Spring numbers were regressed on the previous fall numbers for 12 populations (Figs. 15.19, 15.20). Ten populations show a linear plot. Neither the sharptailed grouse population in South Dakota nor the rock ptarmigan (Lagopus mutus) population in Scotland showed a completely linear relationship. In the case of rock ptarmigan, the curvilinear relation resulted from 2 years with maximum fall numbers. A third index I used to evaluate whether hunting was additive to overwinter mortality was to compare the natural mortality of birds that were heavily hunted with the expected natural mortality based on clutch size (Fig. 15.1). Three populations that received considerable harvests and for which small clutch sizes are the rule are those of red grouse in Scotland, blue grouse on Comox Burn, Vancouver Island, and spruce grouse in Alaska (Jenkins et al. 1963, 1967, Ellison 1974, Zwickel 1977). A regression of mortality rates on clutch sizes for the seven populations with the lowest clutch sizes in Fig. 15.2 (all lightly hunted) is Y = - 6.76 + 7.112X. Substituting the representative clutch size for red grouse (mean of 7.2 eggs), the expected natural mortality is 45 %; for blue grouse (6.4 eggs) it is 39% and for spruce grouse (7.6 eggs) it is 47%. However, the annual mortality rates reported for these hunted populations are 71% red grouse, 50% female blue grouse, and 61 % spruce grouse (Jenkins et al. 1963, Zwickel et al. 1983, Ellison 1974). The observed natural and hunting mortality rates of birds in these popula-
702
A. T. BERGERUD
tions are 26%, 11%, and 14%, greater, respectively, than the predicted values based on clutch size. Hunting mortality is additive in all three populations, and red grouse show the largest additive component.
16.3.3 Compensatory aspects of hunting Although hunting is clearly additive to overwinter mortality, it is probably not additive to the mortality that occurs in the breeding season. If densities are reduced by hunting, the density-dependent effect of competition for optimum sites will be reduced when birds space themselves in the spring to advertise and search for nesting sites. We know that birds enjoy a higher breeding success when densities are low (Fig. 15.14). I propose the generalization that hunting mortality will be additive to natural mortality whenever the natural mortality occurs at the time birds are not spaced, such as in winter when they seek safe feeding sites in conifers, as do spruce and blue grouse; or travel in winter flocks, as do the tundra and steppe grouse. The hypothesis of why natural overwinter mortality is additive to hunting losses is that natural mortality is caused mainly by predators (Table 15.5). The major predators of grouse—the great horned owl (Bubo virginianus), the red fox (Vulpes vulpes), the goshawk (Accipiter gentilis) and the gyrfalcon (Falco rusticolus)—a\l have winter territories (Errington et al. 1940, Sargeant 1972, Pils & Martin 1978, Petersen 1979, Newton 1979). The spacing of these predators prevents marked numerical responses to grouse abundance. The number of raptors observed overwinter increased as the posthunting, red grouse population increased (r = 0.706, n = 9) at the Low and High study areas in Scotland (Jenkins et al. 1964). Further, the number of red grouse found killed by raptors increased linearly as the population increased (r = 0.746, n — 9, Jenkins et al. 1964). Predators probably change the size of their territories as grouse numbers change, but the spacing of predators should still result in relatively constant overwinter natural mortality regardless of the size of posthunting populations. Possibly at extremely high harvest rates with few grouse remaining, predators may move to other areas or switch to other food sources, and then the natural winter mortality could be partly compensatory. The control of hunting should involve removing a sufficient number of birds to ensure that hunting losses, plus a constant and predictable natural loss over winter, will provide optimum spring spacing (distances between birds) for high breeding success. Also, maximum spacing of advertising males and nesting females in the safest sites will decrease density-dependent mortality from predation during the breeding season. The optimum hunting pressure for maximum compensatory reproduction and survival when birds are spaced will be specific for each population, depending upon the kinds and number of winter predators, as well as the abundance of nest predators. Grouse populations with large numbers
INCREASING THE NUMBERS OF GROUSE
703
of nest predators, especially canids, should show the greatest positive response to heavy hunting removals. As the proportion of the habitat that is hunted increases in the future, one can predict more and more examples of hunting reducing the size of subsequent spring populations. 16.4 Control of food One of Leopold's (1933) decimating factors was starvation, and one of his welfare factors was food. Poor food reduced productivity indirectly by decreasing the breeding rate and weakening defenses against starvation. Lack (1954) felt that birds were limited by winter food because (1) predators and disease appeared to be ruled out as limiting factors, (2) birds were more numerous where food was abundant, (3) related species ate different foods, and (4) there was fighting among birds for food in winter. Lack was not deterred that no one had found starving birds. It must seem obvious to biologists that more food translates into more birds. Since Leopold wrote Game Management, more research has been conducted on food than on any other ecological factor. Nor is much letup in sight: 13 of 69 (19%) articles on grouse published between 1964-72 in the Journal of Wildlife Management were on food, and in the past 13 years 12% of the articles were food related (Table 16.1). The sophistication of analysis has become more complex; measurements of food quantity have been replaced by measurements of food quality (Linden 1984, Remington & Braun 1985). Trace elements were once in vogue, but recently we have embarked on chemical-warfare analysis, secondary compounds, and digestive limits (see review by Bryant & Kuropat 1980, Bryant et al. 1983). Few seem concerned that 30 years after Lack's assessment we still have not documented density-dependent starvation, even with the advent of radiotracking. Johnsgard (1973, p. 309) said, "Dependable and nutritious winter food sources are critical to the survival of all grouse." True, but it does not necessarily follow that food is in short supply. The two species that are the best candidates for food shortage are ruffed grouse and prairie chickens. If these species are not limited by food, the other species, which have more abundant food, should not be limited.
16.4.1 Ruffed grouse and aspen buds Two recent cyclic declines of ruffed grouse in northern Minnesota were associated with increased mortality of banded males (Fig. 14.26, Fig. 15.21, Gullion 1967, Gullion & Marshall 1968, and Little 1978). In both declines winter mortality of males was correlated with changes in spring counts of birds (1959-65, r = -0.868; 1971-74, r = -0.718). Little (1978, p. 106) speculated: "Svoboda and Gullion (1972) documented substantial variations in winter
A. T. BERGERUD
704
Table 16.1. Subject content of articles published on grouse in the Journal of Wildlife Management, 1937 to 1985 % total articles
19371945
19461954
19551962
1963
19641972
19731981
19821985
Techniques Determining sex and age Capture methods Census methods Other
12 0 0 6
8 8 0 0
15 19 7 0
11 3 8 0
13 7 3 3
12 2 6 12
11 0 4 7
12 6 5 5
Habitat studies Drumming sites Food and diet Other
0 47 0
8 25 0
0 7 0
3 8 11
6 19 13
8 12 12
19 11 11
6 16 9
0 0
0 0
0 0
0 13
3 4
4 4
15 4
3 4
6
17
0
3
1
4
0
3
0 18 6 6 0
0 17 0 8 8
7 11 15 4 15
3 3 8 16 13
4 4 1 1 16
8 0 4 2 10
7 11 0 0 0
5 6 4 5 11
17
12
27
38
69
50
27
240
Subject content
Population dynamics Nesting studies Hunting Sex and age composition Dispersal and movement Disease Predation Status General Total articles
Total
food utilization for these grouse, so fluctuations in food production combined with deteriorating snow conditions may determine the severity of individual winters for grouse." Gullion (1977a) has argued that a shortage of aspen buds can cause a decline in ruffed grouse—the large decline in 1971-72 and 1972-73 he attributed to shortages of aspen buds. Aspen buds have also been counted from 1971 to 1974 at Cedar Creek, 120 km south of the area studied by Little and Gullion during the general decline of grouse in Minnesota. There was no shortage of aspen (Chap. 4, Table 4.6) even though aspen was much less abundant at Cedar Creek than farther north. Also, secondary foods were both available and acceptable (Chap. 4). The southern distribution of ruffed grouse does not coincide with aspen. There are viable populations of grouse living south of aspen in Oregon, Washington, Kentucky, and Virginia. I believe nesting success, rather than food, is the limiting factor, contrary
INCREASING THE NUMBERS OF GROUSE
705
to Gullion (1984). Huempfner radio-tracked grouse and found a gradual decline in the number of birds through the winter, from predation, not starvation (Huempfner pers. comm.). Svoboda and Gullion (1972) counted overwinter aspen buds for 8 years, but these counts were not correlated with mortality rates nor changes in population size (estimated from Gullion 1981;r = -0.558,r = 0.166, respectively). That ruffed grouse spend less than 5% of their time feeding (Table 4.2) implies that food per se is not limiting. We have suggested that aspen is used not because it is necessary for nutrition, but because its characteristics allow optimal foraging with a reduced predation risk (Fig. 14.24). Gullion (1982) proposed that aspen had a defense mechanism that resulted in less use in some years. But another explanation is that aspen may be a risky place to feed under some snow conditions. Ruffed grouse may need suitable snow cover near aspen to escape goshawks by last-minute snow plunges (Fig. 14.24). Grouse fare well in some winters when they feed infrequently in aspen (cf. annual food supplies in Svoboda & Gullion 1972 vs. mortality rates in Gullion 1981); these data demonstrate that aspen buds are not a prerequisite for a healthy ruffed grouse population.
16.4.2 Prairie chickens and grain Of all the grouse, prairie chickens are the most granivorous (Hamerstrom 1950). Snow should seriously limit the ability of chickens to locate food. Hamerstrom (1941) documented that captive birds lost weight on diets of buds alone. However, weight loss in winter may be typical and by itself is not evidence that food limits numbers of grouse. The major evidence that food limits prairie chickens is that numbers increased with early settlement and the planting of cereal grains. Leopold (1931, p. 165) noted the correlation: "Here was an increase [in chickens] caused by, or at least associated with the introduction of settlements and grain feed." From that point on the argument snowballed. One can hardly find a thesis on prairie chickens that does not restate this association, accepted in the repeating as cause and effect. Chickens increased with more food. They "followed the plow"; but the hypothesis can be questioned. First, the argument is asymmetrical—the birds increased by "following the plow" but they decreased because of the "cow and the plow." More food resulted in more birds, but even more food but less cover resulted in declines. Second, a population can increase only when recruitment (mx) is greater than mortality (qx). Adult birds would show philopatry to the nests and lek ranges they selected as yearlings. For range expansion, breeding success would have had to exceed adult losses. The additional yearlings would push the frontier of the range forward (see Fig. 16.1). No one has explained why recruitment would have exceeded mortality because of more winter food. Winter starvation of prairie chickens has not been documented. The two research teams that walked the winter landscapes for decades have not found starv-
706
A. T. BERGERUD
ing birds (Fred and Fran Hamerstrom, Leo Kirsch, pers. comm.). When the birds were abundant in the 1870-90s, they also migrated from the northern states (Cooke 1888, Gross 1930, Schmidt 1936). This suggests that even when they were abundant, there was insufficient winter food on the range to hold them north. But we should not equate insufficient food to remain north with insufficient food to maintain numbers. Sharp-tailed grouse also increased at approximately the same time chickens reached high numbers (Fig. 16.9). Sharp-tailed grouse do not require grain, and high populations have developed in the past, north of settlements and cultivated grains (Snyder 1935). The most telling counter argument to the plow hypothesis is that the supposed increase in prairie chickens is not correlated with settlement and farming. Peak numbers were reached in Illinois and Indiana from 1860 to 1870 (Sparling 1979), before the native prairie was drained and tilled (Weaver 1954). The chickens must have erupted in numbers between 1870 and 1890. As many as 300,000 were shipped out of Nebraska in 1874 (Kobriger 1964 quoting Aughey 1878). At this time settlement hardly had a foothold in the more western ranges in the Dakotas and Nebraska. The fait accompli that Leo Kirsch has documented, using interviews and historical literature, including references by Teddy Roosevelt, James Audubon, and the Lewis and Clark expedition, is that birds were in the western prairies before settlement and that in fact they increased in areas like his home at Woodworth, North Dakota, about 1880-90 (Fig. 16.9), before the region was settled (Kirsch & Kruse 1973). The first homesteaders found the prairie chicken already common when they arrived. The native prairie was still not under plow and some buffalo (Bison bison) still roamed until approximately 1883 (Roe 1951). Yet the population erupted. This response could not have been to new food resources. An alternative to the hypothesis of increased food is that prairie chickens followed the grass (Kirsch & Kruse 1973, Kirsch 1983). They did not dog the heels of the homesteaders, but instead they followed the buffalo hunter! The original herds of buffalo were almost beyond imagination. Seton (1909) estimated 40 million on the plains. At their peak it is thought that they could have numbered 60-70 million; probably no other continent, not even Africa, has ever produced wild game animals in such great numbers (Anonymous 1965). If grazing by domestic cows can nowadays adversely affect prairie grouse breeding success, what must have been the effect from this horde? There are many statements of the impact of buffalo on the native herbage. "Hundreds of thousands of acres literally eaten to the turf by the immense herds of buffalo. . . Scarcely a blade of grass standing after the herd passed" (from Marlin 1967). "Prairie so completely trodden by buffalo next to impossible to walk" (Audubon 1960). Buffalo had eaten the grass until it was very short, making food scarce, and the lakes were polluted by their wastes (see Roe 1951, p. 362). "The United States Army found it difficult to find sufficient grass for their horses
INCREASING THE NUMBERS OF GROUSE
707
Fig. 16.9. Original range of the prairie chicken and its range after expansion once the buffalo declined. Also shown are the reported dates of historical peak populations of sharptailed grouse and prairie chickens. (Data from Leopold 1931, Roe 1951, Sharpe 1968, Johnsgard & Wood 1968, Sparling 1979.)
708
A. T. BERGERUD
and mules" (Roe 1951, p. 361). Reid and Gannon (1928) quote Alexander Henry, who on 15 February 1802 was in North Dakota: "Buffalo have destroyed the grass and our horses are starving" (other references in Kirsch & Kruse 1973). Now imagine, incomprehensible as it may be, that the southern herd of buffalo (living south of the Platte River) was still 3 to 4 million in 1871 and that by the close of the hunting season in 1874—just 4 years later—this herd had ceased to exist. It had been utterly annihilated (Hornaday in Roe 1951). Hunters then moved to the northern herd of about 1.5 million animals. Hornaday laments that "the hunting season which began in October 1882 and ended in February, 1883, finished the annihilation of the great northern herd and left but a few small bands of straggling numbers, only a very few thousand individuals all told" (Hornaday in Roe 1951, p. 458). The grazers were gone in one decade. The wet years then occurred (Kiel et al. 1972, see also Will 1946), and the grasses flourished. There were as yet no serious predators; the wolf (Canis lupus} had been eliminated and the red fox and the skunk (Mephitis spp.) had hardly arrived (Seton 1953, Johnson & Sargeant 1977). Prairie chickens and sharp-tailed grouse must have erupted because of phenomenal nesting success, but they did so when most of the range was still prairie. Prairie chickens followed the grass, not the plow, and the high populations in the 1870-90s are not evidence that they were then or are now limited by food. 16.4.3 Food is not limiting Food is generally not limiting for grouse. The major evidence is that grouse have not been reported starving in natural systems where predators are present. Grouse in several populations gained weight in the winter: sage grouse (Centrocercus urophasianellus} (Beck & Braun 1978); blue grouse (Redfield 1973b); whitetailed ptarmigan (May 1975); and willow ptarmigan (Fig. 10.26). Rock ptarmigan in Norway (65°40'N) maintain their weights in winter (Mortensen et al. 1985), and rock ptarmigan in Svalbard (78°N) put on fat in the autumn to tide them over 4 months without sunlight (Mortensen & Blix 1985, Steen & Unander 1985). Prairie chickens lost weight in several winters in Wisconsin (Fig. 16.10), but dead birds were not found and the declines in weight were not correlated with changes in numbers of birds (see Fig. 16.18). The spring weights of ruffed grouse were not correlated with prior snow conditions in Minnesota (Fig. 14.26) or British Columbia (Table 3.2). Grouse reach their lowest annual weight levels in the breeding season and not in the winter (Fig. 16.11). Another line of evidence that food is not limiting is that grouse spend a very small fraction of their time feeding, whether they live in the arctic (Fig. 10.9) or in Minnesota and feed on aspen buds. Feeding is most vigorous during the short, morning and evening foraging bouts, when light intensity is low (Braun & Schmidt 1971; Svoboda & Gullion 1972; Figs. 4.7, 10.9). Vigilance and preda-
INCREASING THE NUMBERS OF GROUSE
709
Fig. 16.10. Monthly mean weights of male (top) and female (bottom) prairie chickens in Wisconsin in relation to snow depth. The percent change in the number of total males counted on booming grounds from the previous spring is shown at bottom of figure. The large decrease in 1952 was a result of the hunting season in 1951. (Data from Hamerstrom & Hamerstrom 1973 and pers. comm.)
tion risk seem to be of major concern in the length of feeding, food abundance does not. The actual amounts of food were measured in three studies in this volume when populations were high. In all cases food supplies remained, yet populations declined for other reasons (Table 4.6, Figs. 9.10, 10.4). Future studies of foraging and foods will need to consider predation risk as a major variable in feeding schedules and food preferences. Rock ptarmigan in Scotland, where there are no gyrfalcons, will eat heather, but in Iceland they take birch (Betula spp.) and willow (Salix spp.) (Gardarsson & Moss 1970). Birch and willow may well provide better cover from gyrfalcons than the shorter heather. Ptarmigan take birch in
710
A. T. BERGERUD
Fig. 16.11. Annual weight cycle for eight grouse populations. (Data from Bump et al. 1947, Keith 1962, West & Meng 1968, Hillman & Jackson 1973, May 1975, Ellison & Weeden 1979, Rusch et al. 1984, pers. comm. See also Szuba & Bendell 1984.)
INCREASING THE NUMBERS OF GROUSE
711
preference to other species in Norway, but willow is preferred in North America (see Bryant & Kuropat 1980). Possibly these differences relate to resin levels (Bryant & Kuropat 1980), but I suspect that the ultimate value in species selection will be associated with characteristics of plant species that permit rapid ingestion of adequate food while providing suitable cover to escape predators.
16.5 Control of cover Cover means vegetative or other shelter for grouse (Leopold 1933). But here I shall use it in a more restrictive sense—i.e., to enhance inconspicuousness or escape from predators. This definition excludes cover that is used to protect birds from weather. Substrates that can be considered as cover include: herbaceous vegetation, woody stems and logs, rocks, and snow which is used for snowroosting and in which grouse may plunge. In addition, cover is provided by low clouds and fog and crepuscular light conditions. Gardarsson (Chap. 9) observed rock ptarmigan take cover from gyrfalcons in Iceland by flying to a fence and frequenting a human settlement. A ruffed grouse in flight from a Cooper's hawk (Accipter cooperii) approached Prawdzik (1963) and burrowed into the snow nearby. Steppe (Chap. 5) and tundra (Chap. 10) grouse find cover by traveling in flocks. Some grouse populations have sufficient cover that birds can remain inconspicuous when they feed. Birds in conifer habitats have the highest annual survival rates (Fig. 15.4). I believe the reason is the protective "cover value" of these conifers. Cover is commonly separated into that for nesting, brood rearing, roosting, and escape. A more appropriate classification might be breeding season and nonbreeding season (winter) cover. Both males and females are prepared to compromise their own inconspicuousness to breed and to produce progeny. In winter their own inconspicuousness and survival are their primary concern. The argument has been made (Chap. 14) that fall movements are primarily the shift from cover that favors reproductive activities to that which provides maximum concealment in the absence of leaf cover and the presence of the new contrasting substrate, snow.
16.5.1 Cover during the breeding season A primary management effort to increase the numbers of grouse should be to improve nesting success by enhancing nesting cover and reducing nest predation. With the exception of prairie grouse biologists, few workers have tried to manage grouse by improving nesting cover. It is difficult to improve the nesting cover for ptarmigan in the Arctic, but there may be some opportunities. The order of density of ptarmigan in Newfoundland was: Brunette Island > the St. Shott's barrens > the Portugal Cove barrens (Mercer 1967, Bergerud & Huxter 1969b, Bergerud 1970a). There were no
712
A. T. BERGERUD
mammal predators on Brunette Island and grouse hens could successfully nest near each other. There was also more prostrate Krummholz, conifer cover at Brunette Island than at St. Shotts and Portugal Cove so that nesting females could remain concealed from corvids (Corvus spp.) and gulls (Lams spp.) (Fig. 16.12). The birds on the St. Shott's barrens were spaced closer together and had a higher breeding success than did the birds on the Portugal Cove barrens; pairs were closely associated with the presence of preferred nesting cover (Fig. 16.13; Bergerud & Huxter 1969a,b). Peters (1958) recommended burning these Newfoundland barrens to increase food supplies, but burning also destroyed the slowgrowing, Krummholz nesting cover. The lower density of birds at Portugal Cove than at St. Shott's is in fact the result of prior fires that destroyed climax vegetation that serves as permanent nesting cover. This example illustrates that managers should not reduce nesting habitat to increase food supplies. The forest grouse of North America illustrates a different problem for cover management. Hens use more open canopies with herbaceous growth to conceal nests, but males generally advertise in more protected, shade-tolerant cover to escape raptors. At present, cover management for these species takes the form of improving forest composition for displaying males (Kubisiak et al. 1980, Gullion 1982, Schulz 1982). This may result in shifting males about and decreasing the proportion of silent males, but it cannot result in a substantial increase in the population; breeding success of females drives numbers (Chap. 15, Fig. 15.41). In fact, such management to improve advertising locations for males could increase mortality of females. Females should seek the best available nest cover. If males are at great distances, females will need to compromise their inconspicuousness to move farther to breed. Here is where edge and interspersion of cover play a role. By increasing the interface between the cover type needs of males and females, mobility can be reduced along with predation risk. Cover management for forest and steppe grouse must be directed at improving the concealment of nests and hens. One male can breed several females, and a population with a sex ratio weighted to females would have a higher potential rate-of-increase than a population with a balanced proportion of males and females. Males and females have different cover needs and this distinction must be clearly recognized in habitat manipulation. Nest-brood cover has always been considered the crux of factors limiting prairie grouse (Hamerstrom et al. 1957, Kirsch 1974). Biologists have long noted that prairie grouse, especially prairie chickens, declined during periods of drought, especially in the 1930s and 1950s and when nesting cover was reduced by overgrazing. There have been cycles of wet and dry years for the past 500 years (Will 1946). The breeding success of several widely spaced populations of prairie grouse on the great plains is significantly correlated (Fig. 16.14). This broad synchronization is probably a result of annual variations in soil moisture being in phase over large areas of the grassland. When the soil bank program was in-
INCREASING THE NUMBERS OF GROUSE
713
Fig. 16.12. Comparison of nesting and brood-rearing utilization and cover availability among three ptarmigan barrens (Mercer 1967, Bergerud & Huxter 1969a).
714
A. T. BERGERUD
Fig. 16.13. Locations of ptarmigan pairs on the St. Shott's Barrens, Newfoundland, 1959, in relation to the abundance of nesting cover— prostrate Abies-Picea (Bergerud unpubl. data).
itiated, prairie grouse increased (Kirsch et al. 1973). However, there was also evidence of declines in populations when grasslands were idled for long periods (Kirsch 1974). Kirsch argued that the quality of grasslands for prairie chickens was not well understood, but that residual cover should average 52 cm in height and be sufficiently dense to hide nesting hens. Nesting success should be the key variable to evaluate the quality of grasslands. In prairie grouse, there is more annual variability in nesting success than in brood size. The understanding of the quality of grassland cover will be unraveled when we know more about searching techniques of predators for nests, as affected by cover characteristics. Nesting hens generally need dense cover, whereas young chicks need open-
INCREASING THE NUMBERS OF GROUSE
715
Fig. 16.14. Breeding success of birds in widely scattered prairie grouse populations on the Great Plains is correlated, suggesting wet and dry cycles of moisture occurring over large geographical areas. (Data from Mevel 1973, Robertson 1979, Kobriger 1975, 1981, Sisson 1976, Berg 1977, 1979, Linde et al. 1978, Hilton & Wishart 1981, Manitoba Game Records, S. R. Barber pers. comm.)
716
A. T. BERGERUD
Fig. 16.15. Visual obstruction readings (VOR) from 100 sites along transects in 1977 in each of two fields in North Dakota that were burned in 1975 and planted with the same seed mixture. The nests are those of dabbling ducks, and the broods are sharp-tailed grouse. (Data from Leo Kirsch pers. comm.)
ness, insects, and warmth—these are different requirements (Fig. 16.15). The syntax "nest-brood cover" is inappropriate. Prairie chickens in Minnesota had high nesting success, but low chick survival (Chap. 6). Admittedly, some habitats may provide both cover and openness. Bunchgrasses in Illinois provide tall
INCREASING THE NUMBERS OF GROUSE
717
clumps to hide eggs and hens, yet they also provide interspaces for chicks seeking insects without hinderance of vegetation (Yeatter 1963). The juxtaposition of nest cover and brood foraging habitat may be one component of high-quality grassland cover. However, they are not one and the same at the microhabitat level. Bunchgrasses may eventually be seen as another key to grassland quality. The greater prairie chicken now nests in many areas in sod grass, such as Poapratensis. But hens are, and were, commonly killed on the nest in Minnesota (Svedarsky 1979), Michigan (Ammann 1957), Wisconsin (Gross 1930), and Kansas (Bowen 1971), most frequently by canids which rely on vision (cf. Wells & Lehner 1978). In contrast there are no reports of lesser prairie chickens or Attwater's prairie chickens being killed on the nest (Horkel et al. 1978, Riley 1978, pers. comm., Sell 1979, Lehmann 1941). These two races still nest in bunchgrasses, and use these with considerable preference to sod grasses (Table 16.2) (Bent 1932, Jones 1963, Copelin 1963, Horkel etal. 1978, Riley 1978, Davis etal. 1979, Sell 1979, Wilson 1982). Bunchgrasses in Texas did not significantly differ from sod grasses in cover density (VOR readings) or height (Table 16.2). By nesting in bunchgrasses, hens may have secure cover yet sufficient visibility to detect the approach of predators in time to decide escape tactics. That sage grouse hens are rarely killed on the nest and that the growth form of sagebrush is most similar to that of bunchgrasses is consistent with this hypothesis. Greater prairie chickens that nest in sod grasses often do so within 1 m of openings created by mowing (Westemeier 1973). Prairie chickens do well in tame, sod grasses (Hamerstrom et al. 1957, Westemeier 1973, Hamerstrom & Hamerstrom 1973), but to be "lost" in a field of exotic bluegrass (cf. photo in Hamerstrom et al. 1957, p. 39) may Table 16.2. Attwater's prairie chicken use of bunch and sod grasses, visual obstruction reading, and height of these grass growthforms (data from Cogar et al. 1977)
Parameters % area Visual obstruction readings (VOR) Spring Summer
Clumped, bunch grasses 34
2.22 2.47
Unclumped, sod grasses 18
2.02 2.21
Difference, bunch vs. sod 16
0.20 0.26
Height (cm) Spring Summer
49.2 42.9
52.0 48.8
% nests (n = 19)
68
32
36
% observations (n = 633)
76
16
60
2.8 5.9
718
A. T. BERGERUD
not be the same as to be lost in a sea of clumps of big bluestem (Andropogon gerardi), and may cost some hens their lives. When the dry years come, as they must, or the buffalo—or nowadays the cow-not only are the height and density of grasses reduced but the species composition changes. Bunchgrasses are generally the warm-season species and are less resistant to grazing and drought. The sod grasses are the cool-season grasses and are more resistant to grazing and drought (Weaver 1954). Prairie chickens need bunchgrasses that grow in the summer and provide residual cover for early nesting the next spring, to give that added margin of time to renest. Bunchgrasses also have space between clumps, which allows chick movement, and are associated more often with less litter than are sod grasses. They thus may support fewer mice, possibly fewer insects, and hence probably fewer foxes and skunks. Bunchgrasses are needed for high-quality nesting habitats.
16.5.2 Cover during winter Grouse generally leave the breeding range when they become conspicuous with leaf fall and snow cover. Blue and spruce grouse seek the safe conifer habitats, and the ruffed grouse may snow roost to escape predator detection. The ptarmigan and prairie grouse use shrubs for winter cover, also snow burrow, and use the advantages of large flocks, as do sage grouse. These tundra and steppe grouse also hide by using the crepuscular periods. The primary management aim for manipulation of winter habitat should be to prevent predators from obtaining advantages in particular predator-cover complexes. Managers should try to understand how grouse use the available cover to successfully cope with resident predators. Biologists must avoid altering habitat in such a fashion that the advantages are passed to predators. For example, in the past, prairie chickens made long migrations that may have been particularly advantageous if they were able to shift south of wintering goshawks. Today, however, with cultivated grains the close interspersion of winter and breeding cover in many prairie chicken areas has resulted in reduced movement and increased mortality of chickens when goshawks move south after hare (Lepus americanus) declines. 16.6 Predator control The controversy about increasing game stocks through predator reduction is forever debated by naturalists and hunters. In earlier times raptors and nest predators were considered vermin, and control programs were generally aimed to reduce these predators. Then the Errington view began to prevail that predators took the surplus and that control would have little effect on prey numbers. Now predator control, which we call predator management for a ring of respectability, is again being suggested, but this time the control is directed at nest predators, primarily mammals, and not at raptors which principally hunt adult birds. Few advocates
INCREASING THE NUMBERS OF GROUSE
719
in these times are prepared to decimate the hawks and owls, even if it should mean more grouse.
16.6.1 Predator-control arguments The evidence is generally without exception that if nest predators are substantially reduced the number of nests that hatch will increase. One confounding influence, however, in evaluating these data, is that there may be predator exchange—if one species is removed, another may, in the absence of the first, increase its predation. Nesting success of grouse increased following the reductions of predators in New York and Texas (Edminster 1939, Lawrence & Silvy 1981). Waterfowl enjoy increased hatching success following predator control or when they nest on islands, as have members of the phasianinae (Stoddard 1931, Chesness et al. 1968, Schranck 1972, Duebbert & Kantrud 1974, Trautman et al. 1974, Potts 1980, Duebbert 1982, Lokemoenetal. 1984, Sargeantetal. 1984). If we are prepared to reduce mammalian nest predators, we can have more grouse, both in the fall and the next spring, because breeding success is the primary force in population change. Still, some biologists feel that it is unethical to reduce predators of nests. But ethics are an individual matter. Aldo Leopold, a philosopher in conservation ethics, did not view predator reduction as unethical and said (Leopold 1931, p. 225), "The fox question is not so much of whether foxes do more harm than good, but rather a question of what density of fox population affords the best balance between harm and good." Some would view as unethical a food pyramid too topheavy in predators with a reduced herbivore stratum and a largely underutilized plant base. Another argument is that there is a balance of nature and that predator reduction is an unjustified intrusion. Charles Elton, the father of ecology, would not agree, having said (1930, pp. 16-17) that "it is assumed that an undisturbed animal community lives in a certain harmony, the 'balance of nature.' The picture has the advantage of being an intelligible and apparently logical result of natural selection in producing the best possible world for each species. It has the disadvantage of being untrue." Connell and Sousa (1983) in their extensive review of the stability and persistence of a wide variety of animal populations, from protozoans to rodents, concluded that the evidence in the past 50 years upholds Elton's description. It is only our limited perspective of evolutionary time that persuades us to believe that the whole is greater than the parts and that stability exists. The composition of predators of the steppe and forest is greatly changed from that of 100 years ago. Wolves have been eliminated and coyote (Canis latrans) populations have expanded north and east. The red fox was introduced and expanded south and west. Skunks, raccoons, (Procyon lotof), and opossum (Didelphis virginiand) moved north. Crows expanded across the prairies with increased nesting sites. Kalmback (1938) reported a 69% nesting success of waterfowl in
720
A. T. BERGERUD
North Dakota 50 years ago. But now nesting success for many dabbling ducks is often less than 50%. The dynamics of the predator-prey adaptive race are always in a state of flux. The past is not a guide for the future. In addition, humans have completely upset the interactions between predators and grouse in the steppe and forest. Populations of some predators have shifted, and others have increased. Road-kills and domestic animals provide alternative foods for foxes and skunks (Fig. 16.16). Cover for nesting grouse has been reduced or altered. Ambush sites and vantage lookouts for raptors have been provided. All is changed and changing, and the balance-of-nature concept has no special relevance to the real world of dynamic populations. The proportion of males and females in several grouse populations is unbalanced because of differential rates of predation. Prairie chickens and ruffed grouse generally have more males than females (Table 15.8, Fig. 16.17), and this is partly a result of increased mortality of females during nesting and raising broods. The differential mortality of males and females has occurred because of an increased abundance and success of ground predators since settlement. Southern ptarmigan populations also have more males than females. These males may have a higher survival rate than females because of fewer raptors in the south (cf. Chap. 9). Populations with an excess of males will have lower potential rates-ofincrease compared with populations, such as sage grouse, where females predominate.
Fig. 16.16. Some examples of how human activities have upset predator-prey interactions.
INCREASING THE NUMBERS OF GROUSE
721
Fig. 16.17. In steppe and tundra populations, there are commonly more males than females. In the steppe, human disturbance has reduced nesting space and cover, reduced raptor numbers, and increased mammalian nest predators. These disturbances are selective in favoring the differential survival of males. In the tundra, raptors have been reduced, again favoring male rather than female survival. (Data from Table 15.8.)
722
A. T. BERGERUD
The view that a strong habitat is the best defense against excessive predation has merit, especially if "habitat" includes space as well as cover. Still, predator populations can be so abundant that habitat alone will not suffice to permit a population to increase. The story of the Attwater's prairie chicken is a case in point. This race of prairie chicken is now reduced to fewer than 2,500 birds and restricted to about 120,000 ha along the Texas coast (Lawrence & Silvy 1980). Nesting success for Attwater's hens is terrible. Commonly, less than 25% are successful and most hens lose their clutches to predators (Brownlee 1971-74, Horkel et al. 1978). The sex ratio is heavily biased to males (Horkel & Silvy 1980). The population faces extinction because of high levels of predation. Large sums have been spent to purchase land. Additional land has been donated to the Aransas Refuge, but it is in small blocks which act as predator magnets. Money should now be spent to reduce or at least experiment with reducing nest predation by fencing predators out of key nesting habitat, and on predator reduction programs. This grouse could be saved. Why buy more land and invest in artificial rearing—a "hanging-on" strategy—when the key to recovery is to reduce skunks, armadillos (Dasypus novemcinctus), raccoons, and opossums? Habitat alone does not assure that grouse populations can cope with today's predator populations in today's space.
16.6.2 The return of prairie chickens in Wisconsin A success story in which habitat management efforts can be partly evaluated against changes in predator abundance is the prairie chicken population in central Wisconsin. Prairie chickens in Wisconsin are probably the most intensively managed grouse in North America. In the late 1970s this population staged a spectacular recovery (Fig. 16.18). Frederick and Fran Hamerstrom began watching these chickens in the 1930s and still do so today. They have devoted many years of their lives to saving this population (Hamerstrom 1939, Hamerstrom & Hamerstrom 1949, 1955, 1973, Hamerstrom et al. 1957); one might say their work was "strictly for the chickens" (Hamerstrom 1980). Bird buffs-"the boomers"-came from the cities to watch the dawn rise over the displaying males performing their ancient rituals. The Hamerstroms coaxed and cajoled the "concerned" for funds to purchase land. They saved many acres of the central Wisconsin grasslands, the necessary template, and the day did not come, as it had on Martha's Vineyard, Massachusetts, when a solitary heath hen (T. c. cupido) male boomed alone in the springs of 1929 and 1930; he was the last of his race. But the fluctuations of prairie chickens in Wisconsin are not totally explained by alterations in the extent of their grassland habitat. The population at Buena Vista generally declined from 1950 to 1967-69, and recovered after 1970 (Fig. 16.18). The decline between 1953 and 1969 was 70%, yet through management efforts of the Hamerstroms and others the grassland acreages were held constant,
INCREASING THE NUMBERS OF GROUSE
723
Fig. 16.18. Number of male prairie chickens counted on leks in central Wisconsin, 1950-82. Also shown are two indexes to the abundance of red foxes, and the extent of grasslands in 1953, 1969, and 1979. (Data from Hamerstrom & Hamerstrom 1973, Anonymous 1976, Burke 1979, Pils et al. 1981, R. K. Anderson pers. comm.)
88 km 2 in 1953 and 87 km2 in 1969 (Hamerstrom & Hamerstrom 1973). The population increased to approximately 550 males by 1981 (a 400% increase), yet grassland acres were only 52 km2 in 1980, a 40% decline (Burke 1979). Nor did the dynamics of this population correspond to an increase in food. The sowing of corn in food patches began in the 1950s and was expanded in the 1960s, yet the population continued to decline. The numbers of birds on the CarsonSherry marsh increased in 1975, similar to those at Buena Vista, but little or no corn was planted at Carson-Sherry. The marsh with the most corn, Leola, is the one area that showed the least recovery (B. Gruthoft pers. comm. to M. Gratson,
724
A. T. BERGERUD
Fig. 16.18). It is apparent that food was not a problem for these populations. Changes in the number of birds were not correlated with snow statistics or winter weights (Figs. 16.10, 16.18). In the mid-1960s DDT was sprayed on the marshes (Fig. 16.18) and may have affected insect abundance and thus directed predation at nests. However, the population was declining long before this spraying program was under way. In summer 1972 a new program of burning and mowing these grasslands began upon the urging of J. Toepfer and R. Anderson. The goal was to open up rank grasslands and improve brood habitat. These practices were directed to Buena Vista and Leola marshes. Some mowing occurred at Carson-Sherry but the marsh was not burned (B. Gruthoft pers. comm. to M. Gratson). However, the populations that increased after 1975 were those at Buena Vista and Carson-Sherry (Fig. 16.18). These management practices may have improved breeding success, especially in view of the dwindling acreage in grasslands. Although burning was initiated to enhance brood survival, it may have had a greater influence on nesting success. Variations in brood size between prairie grouse populations show less variation than does nesting success (Chap. 15). In both North Dakota and Minnesota, nesting success of prairie grouse was greater in burned or disturbed habitats than in those that were undisturbed (Kirsch & Kruse 1973, Chap. 6). Westemeier (1973), in Illinois, also showed an increase in the density of prairie chickens 2-4 years after burning. Disturbing grassland sods could have several effects that would reduce predation. Early in secondary succession new grasses in a burned area may be dense enough to hide nesting hens, yet also in some areas sparse enough for feeding by young chicks—the ultimate interspersion (Fig. 16.19). Burning also provides vigorous, upright culms that resist snow loads better than older stems, improving residual nest cover (Ramharter 1976). Burning and mowing may reduce the diversity of plant species, reduce litter, and reduce alternative prey (Tester & Marshall 1961, Halvorsen 1981). With less litter to attract insects and mice, and fewer other alternative prey (small birds), predators may seek more diversified habitats elsewhere. Young, succulent plant stages and less litter could provide less scent for predators. Some burning may also make grasslands more homogeneous, providing fewer clues to predators on where to search. Only by radio-tracking predators hunting in burned, mowed, and undisturbed habitats can these hypotheses be tested. An alternative to the explanation that habitat-manipulation practices have directly resulted in the recovery of chickens in Wisconsin is that nest predation has varied in response to fox numbers. The red fox population increased and spread north in Wisconsin in the 1940-50s (Richards & Hine 1953). The number of prairie chicken males on the Buena Vista and Leola marshes declined at the same time. The correlation between the number of booming males and the num-
INCREASING THE NUMBERS OF GROUSE
725
Fig. 16.19. Grass succession following burning (top). Nesting cover for the lek species should be halfway between leks (bottom), where it may be fenced to keep out nest predators. Hens need space and cover for nesting that are adjacent to more open habitat used by broods. (Adapted from Gratson 1983.)
726
A. T. BERGERUD
her of red foxes trapped in Wisconsin between 1950 and 1963 is significant (r = -0.848, P < 0.01; Fig. 16.18). The fox population remained high during the low years for chickens, the 1960s, then started down in about 1968 (Pils et al. 1981), possibly because coyotes were increasing (O. Rongstad pers. comm.). Chickens began to increase in 1970, and the correlation between total males (Buena Vista, Leola, and Carson-Sherry) and fox numbers between 1962 and 1978, using Pil's index, was again significant (r = -0.782, n = 17, P < 0.01; Fig. 16.18). Nesting success in 1984 was extremely high (78%, n = 9 nests, pers. files). The independent variation in fox numbers may have been more instrumental in the dynamics of this prairie chicken population than were the many years of manipulation of habitat. Prairie chickens in Minnesota also increased in the late 1970s (Fig. 16.1) and this increase was correlated with the price of fox furs (Chap. 6, Fig. 6.18). Sargeant et al. (1984) reported the density of fox in North Dakota as one family/5.7 km2 in 1964, decreasing to one family/20 km2 in 1969, then increasing to one family/7.0 km2 in 1973. The population then declines 1973 to 1977, then again increases. The expansion of the prairie chicken population in Minnesota into surrounding habitats is consistent with the view that the increase resulted from improved breeding success, and that the habitat had not been limiting. The point is that birds can utilize habitats less intrinsically secure when there are fewer predators. The predator-cover complex is what managers should strive to understand, and the requirement of a "strong habitat" should be viewed with the predatorgrouse interaction in mind.
16.7 Control of space The "control of space" in the discussion that follows means the use and manipulation of area (space), to mitigate unfavorable predator-cover complexes so that nesting success can improve. This space concept is not synonymous with the concept of a "space-factor", as outlined by Hamerstrom et al. (1957). These biologists have shown that prairie chickens select habitats with wide, unobstructed views of the horizon. Prairie chickens are granivorous and feed on the ground in the open. To escape raptors they often need a long lead time, and obstructed views would reduce the effectiveness of their breeding and wintering tactics. Kupa (1966), on the frontispiece of his thesis, wrote: "The female ruffed grouse on her nest—key to population." But his advice was not heeded. Instead, grouse biologists, and especially students of the ruffed grouse, have concentrated on improving the habitat for advertising males. Between 1937 and 1985, there have been 15 articles in the Journal of Wildlife Management on the vegetative characteristics of where ruffed grouse males drum, and only eight on nesting studies, for all nine grouse species (8 of 240 articles, or 3%) (Table 16.1). The principal means of improving grouse numbers is to increase nesting success, not
INCREASING THE NUMBERS OF GROUSE
727
the survival of males (Chap. 15). Space, in which grouse may disperse their nests, is the key concept of successful management.
16.7.1 Size and uniformity of space Spatial considerations in the literature have been specified only for the steppe grouse. Kirsch (1974) recommended blocks of 160 acres (65 ha) of nesting cover to maintain prairie chicken populations. Arthaud (1968) thought 60 acres (24 ha) were needed in Missouri. Hamerstrom et al. (1957) recommended a patterning of 80-acre (32-ha) parcels of nest-brood cover interspersed with other habitat types. But there seems to be nothing magic about these sizes. There should be no fixed space; what is desired is an area of sufficient size that hens can maximize nesting success in the face of the predator-cover complex they must contend with. The size of the space will vary with the quality of the cover and with the predators. All hens could nest near each other on an island, such as Pelee Island, without predators, and management of cover would hardly be needed. A large space would not suffice for sage grouse if the individual sagebrush plants (Artemisia spp.) that hens nest under were sufficiently rare that coyotes could economically search in a plant-to-plant pattern for nests. The inconspicuousness in space of dummy clutches in sage grouse habitat and the nests of prairie chickens were compromised when workers marked the nests with conspicuous poles that attracted corvids (Bowen 1971, Autenrieth 1981). Uniformity is needed between selected and nonselected nest sites. There should be hundreds of tree bases from which to choose a nest site for spruce grouse, countless scattered stumps and logs for blue grouse hens to evaluate, and numerous rock fields where white-tailed females can make their choices. Such "likeness" means that predators cannot take advantage of the heterogeneity of cover to narrow their search. The hens are lost in a sea of potential sites. Wildlife managers can actually create nesting "traps" by improving habitat heterogeneity (see Sugden & Beyersbergen 1986). The best and a large-scale example is the creation of refuges for waterfowl. The refuges serve as a respite for ducks from hunters but also as a magnet for nest predators. The Horicon Marsh in Wisconsin is one such trap; nesting waterfowl concentrate there but have an extremely low nesting success (see Livezey 1981, McCabe 1985). If managers create and maintain cover at the expense of space, they should consider fencing the cover to exclude predators. If grouse are given no choice between space and restricted cover, they will compromise space for cover, therein facilitating the searching effort of predators.
16.7.2 Females space away from males Females of the lekking grouse do not nest near the display grounds of males as the older literature suggests; in fact they do the converse, they nest away from
728
A. T. BERGERUD
males. Nesting females should avoid males that are interested in breeding and whose unwanted attention can compromise inconspicuousness in space. When nests are located by radiotelemetry rather than by a biased search near leks, the locations of nests are normally distributed between leks, with the modal distance equal to one-half the distance between leks (Fig. 13.10; see also Davis et al. 1979, Autenrieth 1981). Most females also avoid nesting in the areas regularly visited by males when the latter are feeding or roosting away from the lek (Chap. 5, see Rothenmaier 1979). Nesting cover for steppe species should be provided midway between leks (Fig. 16.19). If nesting cover is planted near leks, it will receive little use. If it is utilized because of a cover-space trade-off—i.e., there is no other cover—birds may show reduced nesting success. The mean distance of successful prairie chicken nests in Minnesota from nearest leks was 1,171 + 103 m (« = 20), whereas unsuccessful nests were closer (975 + 149 m, n = 10) (Table 6.4). Christenson (1970) found successful, sharp-tailed grouse nests 1,210 + 180 m (n = 11) from leks and unsuccessful nests 740 ± 100 m (n = 10, P < 0.05) from leks. Sharp-tailed grouse in Saskatchewan were also more successful when they nested long distances from displaying males (Pepper 1972). Mammalian predators may frequently visit leks. Prospecting females should improve their chances by avoiding areas where predator routes converge. Females leave their nests during crepuscular periods (Table 14.5)—the same times when males move to leks and are most interested in locating females. Hens can thus avoid males at these predictable locations. Nesting cover should be created in places where the inconspicuousness of females will not be compromised by conspicuous males.
16.7.3 The big new space If a large block of homogenous new habitat is suddenly created, colonizing birds may enjoy a high, initial nesting success because they can nest far apart. In addition, the new space may often lack alternative prey and seldom be frequented by predators. Large-scale examples of this big space should have been the midgrass prairies between the 1870s and 1890s, when the buffalo were suddenly gone. Nesting cover must have appeared in large blocks during wet cycles, and there were still few foxes, skunks, and crows on the plains during this period. Large spaces have occurred on Vancouver Island when fires created new habitats for blue grouse (Redfield et al. 1970, Bendell 1974). The expansions that occurred in these blue grouse populations were measured by counting advertising males, but the growth of these populations would have arisen because females initially could nest far apart from each other. Their progeny then recruited to the habitat. Males were then attracted to these areas because of the presence of females. Initially there was little vegetative cover for males or females in these new burns (see Zwickel & Bendell 1967, Plate I), and space must have compensated for this lack of cover.
INCREASING THE NUMBERS OF GROUSE
729
Another example is the response of ruffed grouse to logging. Populations were high in both Minnesota and Wisconsin in the early 1900s (Leopold 1931, King 1937, Grange 1948, Keith 1963). Large nesting areas were created by forest fires and the clearing of coniferous forests, releasing herbaceous and shrub growth. Females could reach high nesting densities in these new spaces before densitydependent predation dampened population increase.
16.7.4 Space and grouse introductions At the 15th Prairie Grouse Technical Conference there was interest in reestablishing new populations of prairie grouse. A common reintroduction technique for steppe grouse has been to identify a likely lek location in some general habitat— low cover with wide horizons—and release males and females at the location in the presence of "dummy" males and taped vocalizations of advertising males. These procedures should generally fail because nesting females are the basic spacing unit. Males do not visit leks to be with other males; nor do females select nesting sites to be near males. The priority in these introductions should be first to locate or create prime nesting cover. Then prospective lek locations can be created away from the nesting habitat—at a distance of at least 1 km. Females should have their best cover onehalf the distance between proposed leks (Fig. 16.19). One could place yearling females in cages in April at the proposed lek location. Yearling males should then be released nearby. The males should display to the penned females (Toepfer 1979). Adults should not be used because they would scatter, attempting to return to where they had nested and advertised previously. After the males are firmly established, displaying near females, the females should be released. They should scatter to the nesting cover provided but return to breed. Once copulations have occurred, the reintroduction should be in place. Another way to increase the numbers or distribution of grouse is to transplant birds to a new space where they cannot leave, such as an island. If the island is without predators the population should quickly expand. The best examples of this technique are the introductions of pheasants on Protection Island and Peele Island, Washington (Einarsen 1945, Stokes 1954). The highest density of sharptailed grouse reported in the literature was on Drummond Island, Michigan (Ammann 1957). Ammann (1957) documented another factor in introductions of sharp-tailed grouse that directly relates to space. Introductions were more successful when the Michigan transplanted stock came from populations that were increasing. More of these founders from increasing populations would be the density-tolerant phenotypes that should produce young that are more viable than if founders were taken from declining populations, in which the density-intolerant morphs would be frequent and their offspring less viable.
730
A. T. BERGERUD
16.8 Successful management The highest densities of grouse over a considerable area are the red grouse populations on managed moors in the British Isles. Numbers in the spring commonly exceed 100 birds/km2. These high densities result from intensive management; favorable nesting success has resulted from improved nesting cover owing to burning practices and the control of nest predators, particularly crows, foxes (Hewson 1984), and stoats (Mustela spp.). This system has now been in place for over 130 years (Lovat 1911, Leopold & Ball 1931), and was well established before intense research began in 1956 (Jenkins et al. 1963). Unfortunately, many biologists have not used the red grouse story to understand how to increase numbers of grouse, but instead have used the Scottish studies as an example of how natural populations are regulated. These Scottish populations have been intensively managed; food and nest cover have been increased, predators have been reduced, and hunting mortality is heavy and additive; thus current research findings are generally inappropriate for systems that are more natural (Lack 1965, 1966, Angelstam 1983, Bergerud et al. 1985). The unmanaged red grouse populations in Ireland provide a control such that we can evaluate the management success in Scotland. The nesting success of Irish grouse is about 60%, whereas success in Scotland is over 80% (Table 16.3). Birds in both populations have similar clutch sizes, but mortality is much higher in Scotland, indicating additive hunting mortality (Fig. 16.8). The Irish grouse exist at low densities in which the summer rate-of-increase declines with increased numbers (Table 16.3). They are not living in marginal habitat as argued by Watson and O'Hare (1979); they live in unmanaged habitats where their more Table 16.3. Comparison of the demographic parameters of unmanaged population of red grouse in Ireland and managed population in Scotland Parameters Spring breeding density (birds/km2) Clutch size Nesting success (%) Chicks per adult in August Adult annual mortality rate (%) Regression breeding success on spring densities Regression proportional change Spring numbers between years on prior breeding success a b
c
Unmanaged 3
Managed13
5.0 7.4 (13) c 69 (13) 0.8 = 45-50
84.4 7.0 (395) 83 (395) 1.6 67-71
r = 0.989 (4)
r = -0.125 (27)
r = 0.900 (3)
r =
0.594 (22)
Data from Ireland (Watson and O'Hare 1979). Data from Scotland —Glen Esk, Glen Muick, and Corndavon moors—all heavily hunted (Jenkins et al. 1963, 1967). Sample size in parentheses.
INCREASING THE NUMBERS OF GROUSE
731
typical nesting success has resulted in nesting hens being well spaced, and hence at low densities. The management of red grouse developed by landowners clearly shows how to increase the numbers of grouse. Improve nesting habitat and reduce predators and nesting success will improve. This should result in a stabilizing density, in which recruitment (mx) = mortality (qx), that is higher than in unmanaged populations. This is the oldest, managed grouse system, and the most successful. Fossils of grouse date to the Miocene (review in Johnsgard 1973). The oldest remains have been located in North America. If grouse arose in the Americas, they had ancestors in common with the phasianinae. The phasianinae occupied the south and came to depend on seeds. Grouse expanded north and adapted to steppe-forest interfaces (Short 1967); possibly the prairie chicken secondarily invaded the south and adopted a granivorous diet. The greatest advance for grouse may have occurred when they evolved their bill and the digestive physiology that permitted them to eat the needles, leaves, and buds of the dominant cover species. Their home became their food. They could remain in their cover to feed, and could stay north in the winter. They adapted long ago to eat northern plant species that are superabundant. Grouse have no food problems. Yet, biologists forever expect to find that grouse are food-limited. That most major plant species act as both food and cover is confounding, but plant species are selected more often for their cover value than for their food value. We must turn our emphasis away from food; not only does food not limit, it does not even set breeding densities. To the contrary, grouse could not perfectly adapt to predators as they had to food because predators were coevolving at equal rates with grouse. The key to increasing the numbers of grouse is to increase their nesting success —more space and cover, and/or fewer nest predators. There is no other practical way to have more grouse.
16.9 Summary This chapter opens with a discussion of the mechanisms that determine the density of grouse. Evidence indicates that grouse females space their nests to reduce predation risk and that the prelaying ranges they search increase in size with increased nest losses. Males space themselves to be near nesting females. Sufficient space and cover for successful nesting are factors that influence the densities of populations. Each population has its own stabilizing density, determined by the predator-cover complex it inhabits. Hunting mortality is additive to natural mortality over winter, but may be partly compensatory to natural mortality during breeding periods. Also, reduced spring densities, brought about by hunting mortality, may provide inverse rate-of-gains if nests are far apart and difficult for predators to locate. Food supplies of adults neither limit population fluctuations nor determine mean densities. Control of space and cover for nesting and control of nest predators are the major management tools available to increase the nesting success of hens and thus increase the numbers of grouse.
This page intentionally left blank
References
This page intentionally left blank
References Adamcik, R. S., A. W. Todd, and L. B. Keith. 1978. Demographic and dietary responses of great horned owls during a snowshoe hare cycle. Can. Field Nat. 92:156-66. Alison, R. M. 1976. Female ptarmigan reoccupying nest site. Auk 73:657. Alice, W. C., N. E. Collias, and C. Z. Lutherman. 1939. Modification of the social order in flocks of hens by the injection of testosterone propionate. Physiol. Zool. 12:412-40. Allen, A. A. 1934. Breeding season behaviour of the ruffed grouse. Trans. Am. Game Conf. 20:311-22. Allen, D. L. 1954. Our wildlife legacy. Funk and Wagnalls, New York. 422 pp. Allen, H. M. and H. Parker. 1977. Willow ptarmigan remove broken eggs from the nest. Auk 94:612-13. Allison, D. G. 1963. Basic features of the New Hampshire ruffed grouse census. J. Wildl. Manage. 27:614-16. Allred, W. J. 1942. Predation and the sage grouse. Wyo. Wildl. 7:3-4. Alway, J. H. and D. A. Boag. 1979. Behaviour of captive spruce grouse at the time broods break up and juveniles disperse. Can. J. Zool. 57:1311-17. Ammann, G. A. 1944. Determining age of pinnated and sharp-tailed grouse. J. Wildl. Manage. 8:170-71. Ammann, G. A. 1950. A ten year comparison of the ruffed grouse kill in Michigan, Minnesota and Wisconsin. Mich. Dep. Conserv. Game Div. Rep. 1096. 2 pp. Ammann, G. A. 1957. The prairie grouse of Michigan. Mich. Dep. Conserv. Game Div. Fed. Aid in Wildl. Restoration Proj. 5-R, 37-R, 70-R. 200 pp. Anderson, L. J. 1973. Habitat use, behavior, territoriality, and movements of the male spruce grouse of northern Minnesota. M.S. thesis. Univ. Minnesota, St. Paul. Anderson, R. K. 1969. Prairie chicken responses to changing booming-ground cover type and height. J. Wildl. Manage. 33:636-43. Anderson, W. W. and J. Arnold. 1983. Density-regulated selection with genotypic interactions. Am. Nat. 121:649-55. Andersson, M. and C. G. Wiklund. 1978. Clumping versus spacing out: experiments on nest predation in Fieldfares (Turdus pilaris). Anim. Behav. 26:1207-12. Andersson, M., C. G. Wiklund, and H. Rundgren. 1980. Parental defense of offspring: a model and an example. Anim. Behav. 28:536-42. Andrewartha, H. G. and L. C. Birch. 1954. The distribution and abundance of animals. Univ. Chicago Press, Chicago. 782 pp. Angelstam, P. 1983. Population dynamics of Tetraonids, especially the black grouse Tetrao tetrix L., in boreal forests. Ph.D. thesis. Uppsala Univ., Sweden. Angelstam, P. 1984. Sexual and seasoned differences in mortality of black grouse Tetrao tetrix in boreal Sweden. Ornis Scand. 15:123-34. 735
736
REFERENCES
Angelstam, P., E. Lindstrom, and P. Widen. 1984. Roleofpredation in short-term population fluctuations of some birds and mammals in Fennoscandia. Oecologia 63:199-208. Angelstam, P., E. Lindstrom, and P. Widen. 1985. Synchronous short-term population fluctuations of some birds and mammals in Fennoscandia —occurrence and distribution. Holarctic Ecol. 8:285-98. Anonymous. 1965. American buffalo. U.S. Dep. Conserv. Note 12. U.S. Dep. Inter. Anonymous. 1976. Wisconsin game and fur harvests. Wis. Dep. Nat. Resour. 25 pp. Anonymous. 1980. Saskatchewan game bird management objectives and strategies for the 80's. Sask. Dep. Tourism and Renewable Resour. 78 pp. Archibald, H. L. 1973. Spring drumming activity and space use of ruffed grouse. Ph.D. thesis. Univ. Minnesota, St. Paul. Archibald, H. L. 1974. Directional differences in the sound intensity of ruffed grouse drumming. Auk 91:517-21. Archibald, H. L. 1975. Temporal patterns of spring space use by ruffed grouse. J. Wildl. Manage. 39:472-81. Archibald, H. L. 1976a. Spatial relationships of neighboring male ruffed grouse in spring. J. Wildl. Manage. 40:750-60. Archibald, H. L. 1976b. Spring drumming patterns of ruffed grouse. Auk 93:808-29. Armleder, H. M. 1980. Habitat use and movements of hens and broods in relation to the demography of blue grouse. M.Sc. Thesis. Univ. Toronto, Toronto. Arnold, S. J. 1983. Sexual selection: the interface of theory and empiricism. Pp. 67-107 in P. Bateson, ed. Mate choice. Cambridge Univ. Press, Cambridge. Arthaud, F. L. 1968. Populations and movements of the prairie chicken related to land use in southwestern Missouri. M.A. thesis. Univ. Missouri, Columbia. Artmann, J. W. 1970. Spring and summer ecology of the sharp-tailed grouse. Ph.D. thesis. Univ. Minnesota, St. Paul. Aschoff, J. 1966. Circadian activity pattern with two peaks. Ecology 47:657-62. Ash, A. N. 1979. The effect of urea fertilizer on the habitat, population dynamics, and local distribution of blue grouse. Ph.D. thesis. Univ. Toronto, Toronto. Aubin, A. E. 1970. Territory and territorial behavior of male ruffed grouse in southwestern Alberta. M.Sc. thesis. Univ. Alberta, Edmonton. Aubin, A. E. 1972. Aural communication in ruffed grouse. Can. J. Zool. 50:1225-29. Audubon, M. R. 1960. Audubon and his journals. Dover Press, New York, Vol. 1, 532 pp. Vol. 2, 554 pp. Aughey, S. 1878. Some facts and considerations concerning the beneficial work of birds. U.S. Entomol. Comm. Annu. Rep. 1877:338-50. Austad, S. N. 1984. A classification of attentive reproductive behaviours and methods of field-testing ESS models. Am. Zool. 24:309-19. Autenrieth, R. E. 1969. Impact of strip spray on vegetation and sage grouse use on summer habitat. Proc. West. States Sage Grouse Workshop 6:147-57. Autenrieth, R. E. 1981. Sage grouse management in Idaho. Idaho Dep. Fish and Game, Wildl. Bull. 9. 238 pp. Bailey, W. J., W. M. Sharp, R. B. Hazel, and G. Davis. 1955. Food habit trends of ruffed grouse in the Centre County "Barrens." Pennsylvania State Univ., Agric. Exp. Sta. Bull. 604. 18 pp. Baker, M. F. 1952. The ecology and management of prairie chickens in Kansas. Ph.D. thesis. Univ. Kansas, Lawrence. Baker, M. F. 1953. Prairie chickens of Kansas. State Biol. Survey, Univ. Kans. Misc. Publ. 5. 68pp. Baker, R. R. and G. A. Parker. 1979. The evolution of bird coloration. Philos. Trans, of the Royal Society of London, Ser. B. 287:63-120. Bakke, E. L. 1980. Movements and habitat use of ruffed grouse in the Turtle Mountains, North Dakota. M.S. thesis. Univ. North Dakota, Grand Forks.
REFERENCES
737
Balser, D. S., H. H. Dill, and H. K. Nelson. 1968. Effect of predator reduction on waterfowl nesting success. J. Wildl. Manage. 32:669-82. Banasiak, C. F. 1951. Ruffed grouse, Bonasa umbellus umbellus (Linne), populations and cover use on two areas in central Massachusetts. M.S. thesis. Univ. Massachusetts, Amherst. Barichello, N. 1983. Gyrfalcon nest site selection. M.Sc. thesis. Univ. British Columbia, Vancouver. Barichello, N. and D. Mossop. 1983. Productivity of gyrfalcons as a response to changes in ptarmigan abundance. 34th. Alaska Sci. Conf., Whitehorse, Yukon Territ. (abstract) Barrett, R. W. 1970. Behavior of ruffed grouse during the breeding and early brood rearing periods. Ph.D. thesis. Univ. Minnesota, St. Paul. Bateman, A. J. 1948. Intra-sexual selection in Dwsophila. Heredity 2: 349-68. Bateson, P. (ed.). 1983. Mate choice. Cambridge Univ. Press, Cambridge. 462 pp. Batterson, W. M. and W. B. Morse. 1948. Oregon state grouse. Oreg. State Game Comm. Fauna Ser. 1. 29pp. Beasom, S. L. 1974. Intensive short-term predator removal as a game management tool. Trans. N. Am. Wildl. Nat. Resour. Conf. 39:230-40. Beck, A. M. and R. J. Vogl. 1972. The effects of spring burning on rodent populations in a brush prairie savanna. J. Mammal. 53:336-45. Beck, T. D. I. 1975. Attributes of a wintering population of sage grouse, North Park, Colorado. M.S. thesis. Colorado State Univ., Fort Collins. Beck, T. D. I. 1977. Sage grouse flock characteristics and habitat selection in winter. J. Wildl. Manage. 41:18-26. Beck, T. D. I. and C. E. Braun. 1978. Weights of Colorado sage grouse. Condor 80:241-43. Beck, T. D. I. and C. E. Braun. 1980. The strutting ground count: variation, traditionalism, management needs. Proc. West. Assoc. Fish and Wildl. Agencies 60:558-66. Beckerton, P. R. and A. L. A. Middleton. 1982. Effects of dietary protein levels on ruffed grouse reproduction. J. Wildl. Manage. 46:569-79. Beil, C. E. 1969. The plant associations of the Cariboo Aspen-Lodgepole pine—Douglas fir Parkland Zone. Ph.D. thesis. Univ. British Columbia, Vancouver. Bendell, J. F. 1954. A study of the life history and population dynamics of the sooty grouse, Dendragapus obscurus fuliginosus (Ridgway). Ph.D. thesis. Univ. British Columbia, Vancouver. Bendell, J. F. 1955a. Age, breeding behavior and migration of sooty grouse, Dendragapus obscurus fuliginosus (Ridgway). Trans. N. Am. Wildl. Conf. 20:367-81. Bendell, J. F. 1955b. Disease as a control of a population of blue grouse, Dendragapus obscurus fuliginosus (Ridgway). Can. J. Zool. 33:195-223. Bendell, J. F. 1972. Population dynamics and ecology of the Tetraonidae. Proc. Int. Ornithol. Congr. 15:81-89. Bendell, J. F. 1974. Effects of fire on birds and mammals. Pp. 73-138 in T. T. Kozlowski and C. E. Ahlgren, eds. Fire and ecosystems. Academic Press, New York. Bendell, J. F. and P. W. Elliott. 1966. Habitat selection in blue grouse. Condor 68:431-46. Bendell, J. F. and P. W. Elliott. 1967. Behaviour and the regulation of numbers in blue grouse. Can. Wildl. Serv. Rep. Ser. 4. 76 pp. Bendell, J. F., D. G. King, and D. H. Mossop. 1972. Removal and repopulation of blue grouse in a declining population. J. Wildl. Manage. 36:1153-65. Bendell, J. F. and F. C. Zwickel. 1979. Problems in the abundance and distribution of blue, spruce, and ruffed grouse in North America. Pp. 48-63 in T. W. I. Lovel, ed. Woodland Grouse Symp. Culloden House, Inverness, Scotland. Bengtson, S. A. 1971. Hunting methods and choice of prey of gyrfalcons Falco rusticolus at Myvatn in northeast Iceland. Ibis 113:468-76. Bent, A. C. 1932. Life histories of North American Gallinaceous birds. U.S. Natl. Mus. Bull. 162. Wash., D. C. 490pp.
738
REFERENCES
Berg, W. E. 1975. Sharp-tailed grouse hunter check, northwestern Minnesota, 1974. Minn. Wildl. Res. Q. 35:94-100. Berg, W. E. 1977. Sharp-tailed grouse hunter check, northwestern Minnesota, 1976. Minn. Wildl. Res. Q. 37:108-20. Berg, W. E. 1979. Ruffed grouse drumming count, spring 1979. Minn. Wildl. Res. Job Progress Rep. 18 pp. Berger, D. D., F. Hamerstrom, and F. N. Hamerstrom, Jr. 1963. The effect of raptors on prairie chickens on booming grounds. J. Wildl. Manage. 27:778-91. Bergerud, A. T. 1970a. Population dynamics of the willow ptarmigan Lagopus lagopus alleni L. in Newfoundland 1955-1965. Oikos 21:299-325. Bergerud, A. T. 1970b. Vulnerability of willow ptarmigan to hunting. J. Wildl. Manage. 34:282-85. Bergerud, A. T. 1971. The past abundance of willow ptarmigan on the Avalon Peninsula of Newfoundland. Can. Field-Nat. 85:21-23. Bergerud, A. T. 1972. Changes in the vulnerability of ptarmigan to hunting in Newfoundland. J. Wildl. Manage. 36:104-9. Bergerud, A. T. and H. E. Butler. 1985. Aggressive and spacing behaviour of female blue grouse. Auk 102:313-22. Bergerud, A. T. and H. D. Hemus. 1975. An experimental study of the behavior of blue grouse (Dendragapus obscurus). I. Differences between the founders from three populations. Can. J. Zool. 53:1222-37. Bergerud, A. T. and D. S. Huxter. 1969a. Breeding season habitat utilization and movement of Newfoundland willow ptarmigan. J. Wildl. Manage. 33:967-74. Bergerud, A. T. and D. S. Huxter. 1969b. Effects of hunting on willow ptarmigan in Newfoundland. J. Wildl. Manage. 33:866-70. Bergerud, A. T. and W. E. Mercer. 1966. Census of the willow ptarmigan in Newfoundland. J. Wildl. Manage. 30:101-13. Bergerud, A. T. and W. E. Mercer. 1972. Spring foods of willow ptarmigan Lagopus lagopus alleni in southeastern Newfoundland. Oikos 23:213-17. Bergerud, A. T. and D. H. Mossop. 1985. The pair bond in ptarmigan. Can. J. Zool. 62:2129-41. Bergerud, A. T., D. H. Mossop, and S. Myrberget. 1985. A critique of the mechanics of annual changes in ptarmigan numbers. Can. J. Zool. 63: 2240-48. Bergerud, A. T., S. S. Peters, and R. McGrath. 1963. Determining sex and age of willow ptarmigan in Newfoundland. J. Wildl. Manage. 27:700-711. Berner, A. and L. W. Gysel. 1969. Habitat analysis and management considerations for ruffed grouse for a multiple use area in Michigan. J. Wildl. Manage. 33:769-78. Bernhoft, L. S. 1969. Reproductive ecology of female sharp-tailed grouse and food habits of broods in southwestern North Dakota. M.S. thesis. Univ. North Dakota, Grand Forks. Berry, J. D. and R. L. Eng. 1985. Interseasonal movements and fidelity to seasonal use areas by female sage grouse. J. Wildl. Manage. 49:237-40. Berry, R. J. and P. E. Davis. 1970. Polymorphism and behaviour of arctic skua, Stercorarius parasiticus. Proc. Royal Society London, Ser. B 175:225-67. Beyer, C., K. Larsson, and M. L. Cruz. 1979. Neural mechanisms probably related to the effects of steroids on sexual behavior, in C. Beyer, ed. Endocrine control of sexual behaviour. Raven Press, New York. Blackford, J. L. 1963. Further observations on the breeding behavior of a blue grouse population in Montana. Condor 65:485-513. Blake, C. S. 1970. The response of sage grouse populations to precipitation trends and habitat quality in south central Idaho. Proc. West. Assoc. State Game and Fish Comm. 50:452-62. Blom, R. 1980. Spring foods of the willow grouse Lagopus I. lagopus on Tranoy, northern Norway. Cand. Real, thesis. Univ. Bergen, Norway.
REFERENCES
739
Blom, R. and S. Myrberget. 1978. Experiments on shooting territorial willow grouse Lagopus lagopus. Cinclus 1:29-33. Blus, L. J. and J. A. Walker. 1966. Progress report on the prairie grouse nesting study in the Nebraska sandhills. Nebr. Bird Rev. 34:23-30. Boag, D. A. 1965. Indicators of sex, age, and breeding phenology in blue grouse. J. Wildl. Manage. 29:103-8. Boag, D. A. 1966. Population attributes of blue grouse in southwestern Alberta. Can. J. Zool. 44:799-814. Boag, D. A. 1976a. The effect of shrub removal on occupancy of ruffed grouse drumming sites. J. Wildl. Manage. 40:105-10. Boag, D. A. 1976b. Influence of changing grouse density and forest attributes on the occupancy of a series of potential territories by male ruffed grouse. Can. J. Zool. 54:1727-36. Boag, D. A. and J. H. Alway. 1980. Effect of social environment within the brood on dominance rank in gallinaceous birds (Tetraonidae and Phasianidae). Can. J. Zool. 58:44-49. Boag, D. A. and J. H. Alway. 1981. Heritability of dominance status among Japanese quail: a preliminary report. Can. J. Zool. 59:441-44. Boag, D. A., K. H. McCourt, P. W. Herzog, and J. H. Alway. 1979. Population regulation in spruce grouse: a working hypothesis. Can. J. Zool. 57: 2275-84. Boag, D. A., S. G. Reebs, and M. A. Schroeder. 1983. Egg loss among spruce grouse inhabiting lodgepole pine forests. Can. J. Zool. 62:1034-37. Boag, D.A. and K. M. Sumanik. 1969. Characteristics of drumming sites selected by ruffed grouse in Alberta. J. Wildl. Manage. 33:621-28. Boggs, C. E., E. Norris, andJ. B. Steen. 1977. Behavioural and physiological temperature regulation in young chicks of the willow grouse (Lagopus lagopus). Comp. Biochem. Physiol. 58(A):371-72. Borgia, G. 1979. Sexual selection and the evolution of mating systems. Pp. 19-80 in M. S. Blum and N. A. Blum, eds. Sexual selection and reproductive competition in insects. Academic Press, New York. Bowen, D. E., Jr. 1971. A study of dummy nests and greater prairie chicken (Tympanuchus cupido pinnatus) nests in northeastern Kansas with notes on female nesting behavior. M.S. thesis. Kansas State Univ., Manhattan. Bowman, T. J. and R. J. Robel. 1977. Brood break-up, dispersal, mobility, and mortality of juvenile prairie chickens. J. Wildl. Manage. 41:27-34. Boyd, H. 1981. Prairie dabbling ducks. Can. Wildl. Serv. Prog. Note 119. 9 pp. Bradbury, J. W. 1981. The evolution of leks. Pp. 138-169 in R. D. Alexander and D. W. Tinkle, eds. Natural selection and social behavior. Chiron Press, New York. Bradbury, J. W. and R. M. Gibson. 1983. Leks and mate choice. Pp. 109-38 in P. Bateson, ed. Mate choice. Cambridge Univ. Press, Cambridge. Bradbury, W. C. 1915. Notes on the nesting of the white-tailed ptarmigan in Colorado. Condor 17:214-22. Braestrup, F. W. 1941. A study on the arctic fox in Greenland. Immigrations, fluctuations in numbers based mainly on trading statistics. Medd. om Graland, 13:1-101. Brander, R. B. 1965. Factors affecting dispersion of ruffed grouse during late winter and spring on the Cloquet Forest Research Center, Minnesota. Ph.D. thesis. Univ. Minnesota, St. Paul. Brander, R. B. 1967. Movements of female ruffed grouse during the mating season. Wilson Bull. 79:28-36. Braun, C.E. 1969. Population dynamics, habitat, and movements of white-tailed ptarmigan in Colorado. Ph.D. thesis. Colorado State Univ., Fort Collins. Braun, C. E. 1979. Evolution of the effects of changes in hunting regulations on sage grouse populations. Colo. Div. Wildl., Game Res. Rep. Proj. W-37-R-32, Job 9a. Pp. 11-35.
740
REFERENCES
Braun, C. E., T. Britt, and R. O. Wallestad. 1977. Guidelines for maintenance of sage grouse habitats. Wildl. Soc. Bull. 5:99-106. Braun, C. E. and D. M. Hoffman. 1979. Vulnerability and population characteristics of sage grouse in Moffat County. Colo. Div. Wildl., Game Res. Rep. Proj. W-37-R-32, Job 11. Pp. 163-99. Braun, C. E., R. W. Hoffman, and G. E. Rogers. 1976. Wintering areas and winter ecology of whitetailed ptarmigan in Colorado. Colo. Div. Game, Fish and Parks, Spec. Rep. 38. 38 pp. Braun, C. E. and G. E. Rogers. 1971. The white-tailed ptarmigan in Colorado. Colo. Div. Game, Fish and Parks, Tech. Publ. 27. 80 pp. Braun, C. E. and R. K. Schmidt. 1971. Effects of snow and wind on wintering populations of whitetailed ptarmigan in Colorado. Pp. 238-250 in A. O. Haugen, ed. Proc. Snow and Ice Symp. Iowa Coop. Wildl. Res. Unit, Iowa State Univ., Ames. Braun, C. E., R. K. Schmidt, and G. E. Rogers. 1973. Census of Colorado white-tailed ptarmigan with tape recorded calls. J. Wildl. Manage. 37:90-93. Brewer, L. W. 1980. The ruffed grouse in western Washington. Wash. State Dep. Game, Biol. Bull. 16. 102 pp. Brewer, R. and K. L. Harrison. 1975. The time of habitat selection in birds. Ibis. 117:521-22. Brittas, R. and V. Marcstrom. 1982. Studies in willow grouse Lagopus lagopus of some possible measures of condition in birds. Ornis Fenn. 59:157-69. Brown, C. P. 1946. Food of Maine ruffed grouse by seasons and cover types. J. Wildl. Manage. 10:17-28. Brown, J. L. 1964. The evolution of diversity in avian territorial systems. Wilson Bull. 76:160-69. Brown, J. L. 1969. Territorial behavior and population regulation in birds: a review and reevaluation. Wilson Bull. 81:293-329. Brown, R. L. 1966a. Response of sharptail breeding populations to annual changes in residual grassland cover. Proc. Annu. Conf. Western Assoc. State Game and Fish Comm. 46:219-22. Brown, R. L. 1966b. Sharptail grouse population study. Mont. Dep. Fish and Game. Proj. W-91-R-6 and W-91-R-7. 23 pp. Brown, R. L. 1967. Sharptail grouse population study. Mont. Dep. Fish and Game. Proj. 91-R-8, Job II-E. 16 pp. Brown, R. L. 1968a. Effects of land-use practices on sharptail grouse. Mont. Dep. Fish and Game. Proj. W-91-R-9, Job II-F. 11 pp. Brown, R. L. 1968b. Sharptail grouse population study. Mont. Dep. Fish and Game. Proj. W-91-R9. Job II-E. 18 pp. Brownlee, W. C. 1971-74. Seasonal territorial ranges of Attwater's prairie chicken. Texas Parks and Wildl. Dep. Proj. W-100-R-Z, 3, 4, and 5. 13 pp. Bryant, J. P. and P. J. Kuropat. 1980. Selection of winter forage by subarctic browsing vertebrates: the role of plant chemistry. Annu. Rev. Ecol. Syst. 11:261-85. Buehler, D. A. and L. B. Keith. 1982. Snowshoe hare distribution and habitat use in Wisconsin. Can. Field-Nat. 96:19-29. Bump, G., R. W. Darrow, F. C. Edminster, and W. F. Crissey. 1947. The ruffed grouse: life history, propagation, management. N.Y. State Conserv. Dep., Holling Press, Buffalo. 915 pp. Burke, C. J. 1979. Effect of prey and land use on mating systems of harriers .M.S. thesis. Uni v. Wisconsin, Stevens Point. Byrkjedal, I. 1980. Nest predation in relation to snow-cover—a possible factor influencing the start of breeding in shorebirds. Ornis Scand. 11:249-52. Cade, T. J. 1960. Ecology of the peregrine and gyrfalcon populations in Alaska. Univ. Calif. Publ. Zool. 63:151-290. Cade, W. H. 1981. Alternative mating strategies: Genetic difference in crickets. Science 212:563-64. Cade, W. H. 1984. Genetic variation underlying sexual behavior and reproduction. Am. Zool. 24:355-66.
REFERENCES
741
Caldwell, J. P. 1982. Disruptive selection: a tail color polymorphism in Acris tadpoles in response to differential predation. Can. J. Zool. 60:2818-27. Caldwell, P. J. 1976. Energetic and population considerations of the sharp-tailed grouse in the aspen parkland of Canada. Ph.D. thesis. Kansas State Univ., Manhattan. Campbell, H. 1972. A population study of lesser prairie chickens in New Mexico. J. Wildl. Manage. 36:689-99. Cancelado, R. and T. R. Yonke. 1970. Effect of prairie burning on insect populations. J. Kans. Society 43:274-81. Cannon, R. W. and F. L. Knopf. 1979. Lesser prairie chicken responses to range fires at the booming ground. Wildl. Society Bull. 7:44-46. Cannon, R. W. and F. L. Knopf. 1981. Lek numbers as a trend index to prairie grouse populations. J. Wildl. Manage. 45:776-78. Capanna, E., M. Cord, D. Mainardi, S. Parmigiani, and P. F. Brain. 1984. Karyotype and intermale aggression in wild house mice: ecology and speciation. Beh. Genet. 14:195-208. Caraco, T., S. Martindale, and H. R. Pullion. 1980. Avian flocking in the presence of a predator. Nature (London) 285:400-401. Carr, H. D. 1967. Effects of sage brush spraying on abundance, distribution and movements of sage grouse. M.S. thesis. Colorado State Univ., Fort Collins. Carrick, R. 1963. Ecological significance of territory in the Australian magpie, Gymnorhina tibicen. Proc. Int. Ornithol. Congr. 13:740-53. Carrick, R. and M. D. Murray. 1964. Social factors in population regulation of the silver gull Larus novaehollandiae. C.S.I.R.O. Wildl. Res. 9:189-99. Cartwright, B. W. 1944. The "crash" decline in sharp-tailed grouse and hungarian partridge in western Canada and the role of the predator. Trans. N. Am. Wildl. Conf. 9:324-30. Cebula, J. J. 1966. Radio-telemetry as a technique used in greater prairie chicken (Tympanuchus cupidopinnatus) mobility studies. M.S. thesis. Kansas State Univ., Manhattan. Charlesworth, B. 1971. Selection in density-regulated populations. Ecology 52:469-74. Charlesworth, B. and J. T. Giesel. 1972. Selection in populations with overlapping generations. II. Relationship between gene frequency and demographic variables. Am. Nat. 106:388-401. Charnov, E. L. and J. P. Finerty. 1980. Vole population cycles: a case for kin-selection? Oecologia 45:1-2. Charnov, E. L. and J. R. Krebs. 1974. On clutch size and fitness. Ibis 116:217-19. Chesness, R. A., M. M. Nelson, and W. H. Longley. 1968. The effect of predator removal on pheasant reproductive success. J. Wildl. Manage. 32:683-97. Chitty, D. 1958. Self-regulation of numbers through changes in viability. Cold Spring Harbor Symp. Quant. Biol. 22:277-80. Chitty, D. 1960. Population processes in the vole and their relevance to general theory. Can. J. Zool. 38:99-113. Chitty, D. 1964. Animal numbers and behavior. Pp. 41-53 in J. R. Dymond, ed. Fish and wildlife. A memorial to W. J. K. Harkness. Honyman Press, Canada. Chitty, D. 1967. The natural selection of self-regulatory behaviour in animal populations. Proc. Ecol. Society Aust. 2:51-78. Chitty, D. 1970. Variation and population density. Symp. Zool. Society London 26:327-33. Chitty, D. and E. Phipps. 1966. Seasonal changes in survival in mixed populations of two species of vole. J. Anim. Ecol. 35:313-31. Choate, T. S. 1960. Observations on the reproductive activities of white-tailed ptarmigan (Lagopus leucurus) in Glacier Park, Montana. M.S. thesis. Montana State Univ., Missoula. Choate, T. S. 1963a. Ecology and population dynamics of white-tailed ptarmigan (Lagopus leucurus) in Glacier National Park, Montana. Ph.D. thesis. Montana State Univ., Missoula. Choate, T. S. 1963b. Habitat and population dynamics of white-tailed ptarmigan in Montana. J. Wildl. Manage. 27:684-99.
742
REFERENCES
Christensen, K. and H. Pedersen. 1982. Variation in chromosome number in the blue fox (Alopex lagopus) and its effect on fertility. Hereditas 97:211-15. Christenson, C. D. 1970. Nesting and brooding characteristics of sharp-tailed grouse (Pedioecetes phasianellus jamesi Lincoln) in southwestern North Dakota. M.S. thesis. Univ. North Dakota, Grand Forks. Christisen, D. M. 1969. National status and management of the greater prairie chicken. Trans. N. Am. Wildl. and Nat. Resour. Conf. 34:207-17. Cody, M. L. 1966. A general theory of clutch size. Evolution 20:174-84. Cogar, V. F., J. D. Horkel, and N. J. Silvy. 1977. Vegetation type preferences of Attwater's prairie chicken in coastal prairie. Proc. Annu. Conf. Southeast. Assoc. Fish and Wildl. Agencies 31:41-50. Cole, L. C. 1954. Some features of random population cycles. J. Wildl. Manage. 18:2-24. Collias, N. E. 1950. Hormones and behavior with special reference to birds and mammals and the mechanism of hormone action, in E. S. Gordon, ed. A symposium of steroid hormones. Univ. Wisconsin Press, Madison. Comins, H. N., W. D. Hamilton, and R. M. May. 1980. Evolutionarily stable dispersal strategies. J. Theor. Biol. 82:205-30. Connell, J. H. andW. P. Sousa. 1983. On the evidence needed to judge ecological stability or persistence. Am. Nat. 121:789-824. Connelly, J. W., Jr. 1982. An ecological study of sage grouse in southeastern Idaho. Ph.D. thesis. Washington State Univ., Pullman. Cooke, F. and F. G. Cooch. 1968. The genetics of polymorphism in the snow goose Anser caoerulescens. Evolution 22:289-300. Cooke, F. and P. J. Mirsky. 1972. A genetic analysis of lesser snow goose families. Auk 89:863-71. Cooke, W. W. 1888. Report on bird migration in the Mississippi Valley in the years 1884 and 1885. U.S. Dep. Agric. Div. Econ. Ornithol. Bull. 2. Cooper, C. R. 1977. Differences in behaviour between populations of captive blue grouse. M.Sc. thesis. Univ. Toronto, Toronto. Copelin, F. F. 1963. The lesser prairie chicken in Oklahoma. Okla. Wildl. Conserv. Dep. Tech. Bull. 6. 58 pp. Cowardin, L. M., D. S. Gilmer, and C. W. Shaiffer, 1985. Mallard recruitment in the agricultural environment of North Dakota. Wildl. Monog. 92. 37 pp. Craig, J. V., L. C. Ortman, and A. M. Gull. 1965. Genetic selection for social dominance ability in chickens. Anim. Behav. 13:114-31. Cramp, S., et al. 1980. Handbook of the birds of Europe, the Middle East and North Africa. The birds of the western Palearctic. Vol. 2. Hawks to bustards. Oxford Univ. Press, Oxford. 695 pp. Cringan, A. T. 1970. Reproductive biology of ruffed grouse in southern Ontario, 1964-1969. J. Wildl. Manage. 34:756-61. Crook, J. H. 1965. The adaptive significance of avian social organisations. Symp. Zool. Society London 14:181-218. Croze, H. 1970. Searching image in carron crows. Z. Tierpsychol. 5:1-35. Curio, E. 1975. The functional organization of antipredator behaviour in the pied flycatcher: a study of avian visual perceptions. Anim. Behav. 23:1-115. Curio, E. 1976. The ethology of predation. Springer-Verlag, Berlin. 250 pp. Curtis, J. T. 1959. The vegetation of Wisconsin. Univ. Wisconsin Press, Madison. 657 pp. Dalke, P. D., D. B. Pyrah, D. C. Stanton, J. E. Crawford, and E. F. Schlatter. 1960. Seasonal movements and breeding behavior of sage grouse in Idaho. Trans. N. Am. Wildl. Conf. 25:396-407. Dalke, P. D., D. B. Pyrah, D. C. Stanton, J. E. Crawford, and E. F. Schlatter. 1963. Ecology, productivity, and management of sage grouse in Idaho. J. Wildl. Manage. 27:810-41. Dapkus, D. C. 1976. Differential survival involving the Burnsi phenotype in the northern leopard frog Ranapipiens. Herpetology 32:325-27.
REFERENCES
743
Dargan, L. and R. J. Keller. 1940. Survey of 1940: North Park, Jackson County, Colo. Colo. Dept. Game and Fish. Sage grouse survey, Vol. 2. 28 pp. Darwin, C. 1871. The descent of man and selection in relation to sex. John Murray Press, London. 490 pp. Davies, N. B. 1978. Ecological questions about territorial behaviour. Pp. 317-50 in J. R. Krebs and N. B. Davies, eds. Behavioural ecology. Blackwell Sci. Publ., London. Davies, N. B. and T. R. Halliday. 1979. Competitive mate searching in male common toads, Bufo bufo. Anim. Behav. 27:1235-67. Davies, R. G. 1973. A study of demography and behaviour of ruffed grouse in British Columbia. M.S. thesis, Univ. Victoria, Victoria. Davis, C. A., T. Z. Riley, R. A. Smith, H. R. Suminski, andM. J. Wisdom. 1979. Habitat evaluation of lesser prairie chickens in eastern Chaves County, New Mexico. Dep. Fish, and Wildl. Sci., New Mexico State Univ., Agric. Exp. Station, Las Cruces. 141 pp. Davis, D. E. 1957. The existence of cycles. Ecology 38:163-64. Davis, J. A. and R. J. Stoll, Jr. 1973. Ruffed grouse age and sex ratios in Ohio. J. Wildl. Manage. 37:133-41. Dawkins, R. 1980. Good strategy or evolutionary stable strategy? Pp. 337-70 in G. W. Barlow and J. Silvererg, eds. Sociobiology: beyond nature/nurture. Westville Press, Boulder, Colo. Dawkins, R. and T. R. Carlisle. 1976. Parental investment, male desertion and a fallacy. Nature 262:131-33. DeStefano, S. and D. H. Rusch. 1982. Some historical aspects of ruffed grouse harvests and hunting regulations in Wisconsin. Wis. Acad. Trans. 70:27-35. De Vos, G. J. 1979. Adaptedness of arena behaviour in black grouse (Tetrao tetrix) and other grouse species (Tetraoninae). Behaviour 68:277-314. De Vos, G. J. 1983. Social behaviour of black grouse, an observational and experimental field study. Ardea 71:1-103. Dixon, J. 1927. Contribution to the life history of the Alaska willow ptarmigan. Condor 29:213-23. Dixson, A. F. 1980. Androgens and aggressive behavior in primates: a review. J. Aggress. Behav. 6:37-67. Doerr, P. D., L. B. Keith, andD. H. Rusch. 1971. Effects of fire on a ruffed grouse population. Proc. Annu. Tall Timbers Fire Ecol. Conf. 10:25-46. Doerr, P. D., L. B. Keith, D. H. Rusch, and C. A. Fischer. 1974. Characteristics of winter feeding aggregations of ruffed grouse in Alberta. J. Wildl. Manage. 38:601-15. Dominey, W. J. 1984. Alternative mating tactics and evolutionary stable strategies. Am. Zool. 24:385-96. Donaldson, J. L. and A. T. Bergerud. 1974. Behaviour and habitat selection of an insular population of blue grouse. Syesis 7:115-27. Dorney, R. S. 1959. Relationship of ruffed grouse to forest cover types in Wisconsin. Wis. Conserv. Dep. Tech. Bull. 18. 32 pp. Dorney, R. S. 1963. Sex and age structure of Wisconsin ruffed grouse populations. J. Wildl. Manage. 27:599-603. Dorney, R. S. 1966. A new method for sexing ruffed grouse in late summer. J. Wildl. Manage. 30:623-25. Dorney, R. S. and C. Kabat. 1960. Relation of weather, parasitic disease, and hunting to Wisconsin ruffed grouse populations. Wis. Conserv. Dep. Tech. Bull. 20. 64 pp. Drobney, R. D. and R. D. Sparrowe. 1977. Land use relationships and movements of greater prairie chickens in Missouri. Trans. Mo. Acad. Sci. 10,11:146-60. Duebbert, H. F. 1982. Nesting of waterfowl on islands in Lake Aububon, North Dakota. Wildl. Soc. Bull. 10:232-37. Duebbert, H. F. and H. A. Kantrud. 1974. Upland duck nesting related to land use and predator reduction. J. Wildl. Manage. 38:257-65.
744
REFERENCES
Dumke, R. T. and C. M. Pils. 1973. Mortality of radio-tagged pheasants on the Waterloo Wildlife Area. Wis. Dep. Nat. Resour. Tech. Bull. 72. 52 pp. Dumke, R. T. and C. M. Pils. 1979. Renesting and dynamics of nest site selection by Wisconsin pheasants. J. Wildl. Manage. 43:705-16. Dunn, P. O. and C. E. Braun. 1985. Natal dispersal and lek fidelity of sage grouse. Auk 102:621-27. Edminster, F. C. 1939. The effect of predator control on ruffed grouse populations in New York. J. Wildl. Manage. 3:345-52. Edminster, F. C. 1947. The ruffed grouse, its life story, ecology and management. Macmillan, New York. 385 pp. Edminster, F. C. 1954. American game birds of fields and forest; their habits, ecology and management. Charles Scribner's Sons, New York. 490 pp. Edmonds, M. 1974. Defence in animals. Longman, N.Y. 357 pp. Einarsen, A. S. 1945. Some factors affecting ring-necked pheasant population density. The Murrelet 26:29,39-44. Elliot, C. A. 1979. Insect availability as a food source for spruce grouse chicks. B.Sc. thesis. Univ. New Brunswick, Fredericton. Elliott, P. W. 1965. Factors affecting the local distribution of blue grouse on a breeding range. M.Sc. thesis. Univ. British Columbia, Vancouver. Ellison, L. N. 1967. Spring movements and behavior of territorial and non-territorial male Alaskan spruce grouse. Abst. Wildl. Society Northwest Sect. Annu. Conf. 18 pp. Ellison, L. N. 1971. Territoriality in Alaskan spruce grouse. Auk 88:652-64. Ellison, L. N. 1972. Role of winter food in regulating numbers of Alaskan spruce grouse. Ph.D. thesis. Univ. California, Berkeley. Ellison, L. N. 1973. Seasonal social organization and movements of spruce grouse. Condor 75:375-85. Ellison, L. N. 1974. Population characteristics of Alaskan spruce grouse. J. Wildl. Manage. 38:383-95. Ellison, L. N. and R. B. Weeden. 1966. Small game and furbearers investigations. Alaska Dep. Fish and Game Proj. W-13-R-1, Work Plan B. (mimeogr.). Ellison, L. N. and R. B. Weeden. 1979. Seasonal and local weights of Alaska spruce grouse. J. Wildl. Manage. 43:176-83. Elton, C. 1930. Animal ecology and evolution. Clarendon Press, Oxford. 96 pp. Elton, C. S. 1942. Voles, mice and lemmings. Clarendon Press, Oxford. 496 pp. Emlen, S. T. and L. W. Oring. 1977. Ecology, sexual selection and the evolution of mating systems. Science 197:215-23. Emmons, S. R. 1980. Lek attendance of male sage grouse in North Park, Colorado. M.S. thesis. Colorado State Univ., Fort Collins. Emmons, S. R. and C. E. Braun. 1984. Lek attendance of male sage grouse. J. Wildl. Manage. 48:1023-28. Eng, R. L. 1959. A study of the ecology of male ruffed grouse (Bonasa umbellus L.), on the Cloquet Forest Research Center, Minnesota. Ph.D. thesis. Univ. Minnesota, St. Paul. Eng, R. L. 1963. Observations on the breeding biology of male sage grouse. J. Wildl. Manage. 27:841-46. Eng, R. L. and G. W. Gullion. 1962. The predation of goshawks upon ruffed grouse on the Cloquet Forest Research Center, Minnesota. Wilson Bull. 74:227-42. Eng, R. L., E. J. Pitcher, S. J. Scott, and R. J. Greene. 1979. Minimizing the effect of surface coal mining on a sage grouse population by a directed shift of breeding activities. Pp. 464-68 in G. A. Swanson, ed. The mitigation symposium. Gen. Tech. Rep. RM-65, Rocky Mt. For. Range Exp. Stn., Ft. Collins, Colo. Eng, R. L. and P. Schladweiler. 1972. Sage grouse winter movements and habitat use in central Montana. J. Wildl. Manage. 36:141-46.
REFERENCES
745
Erikstad, K. E. 1978. Relations between survival of chicks, brood movements and habitat utilization in willow grouse. Cand. Real, thesis. Univ. Bergen, Norway. Erikstad, K. E. 1979. Effects of radio packages on reproductive success of willow grouse. J. Wildl. Manage. 43:170-75. Erikstad, K. E. 1982. Effekten ar kaldt og fuktig vaer pa lirypekyllingenes naering sadferd og overlerelse. Pp. 207-25 in E. 0stbye and I. Mysterud, eds. H0yfjell0kologisk Forskningsstasjon, Finse. Univ. Bergen and Oslo. Erikstad, K. E. 1985a. Growth and survival of willow grouse chicks in relation to home range size, brood movements and habitat selection. Ornis Scand. 16:181-90. Erikstad, K. E. 1985b. Territorial breakdown and brood movements in willow grouse Lagopus I. lagopus. Ornis Scand. 16:95-98. Erikstad, K. E. 1985c. Clutch size and egg size variation in willow grouse Lagopus I. lagopus. Ornis Scand. 16:88-94. Erikstad, K. E. and R. Andersen. 1983. The effect of weather on survival, growth rate and feeding time in different sized willow grouse broods. Ornis Scand. 14:249-52. Erikstad, K. E., R. Blom, and S. Myrberget. 1981. Territorial hooded crows as predators of willow ptarmigan nests. J. Wildl. Manage. 46:109-14. Erikstad, K. E. and T. K. Spids0. 1982. The influence of weather on food intake, insect prey selection and feeding behaviour in willow grouse chicks in northern Norway. Ornis Scand. 13:176-82. Errington, P. L. 1935. The 1934 drought and southern Iowa bobwhite. Iowa Bird Life 5:18-21. Errington, P. L. 1937a. Food habits of Iowa red foxes during a drought summer. Ecology 18:53-62. Errington, P. L. 1937b. Winter carrying capacity of marginal ruffed grouse environment in northcentral United States. Can. Field-Nat. 51:31-34. Errington, P. L. 1945. Some contributions of a fifteen-year local study of the northern bobwhite to a knowledge of population phenomena. Ecol. Monogr. 15:1-34. Errington, P. L. 1951. Concerning fluctuations in populations of the prolific and widely distributed muskrat. Am. Nat. 85:273-92. Errington, P. L., F. Hamerstrom, and F. N. Hamerstrom, Jr. 1940. The great horned owl and its prey in north central United States. Iowa State College. Agri. Exp. Stn. Bull. 277:757-850. Evans, R. M. 1961. Courtship and mating behavior of sharp-tailed grouse (Pedioecetes phasianellus jamesi, Lincoln). M.S. thesis. Univ. Alberta, Edmonton. Evans, R. M. 1969. Territorial stability in sharp-tailed grouse. Wilson Bull. 81:75-78. Ewing, J. 1924. Plant successions of the brush-prairie in northwestern Minnesota. J. Ecol. 12:238-66. Fairchild, L. 1984. Male reproductive tactics in an explosive breeding toad population. Am. Zool. 24:407-18. Falconer, D. S. 1960. Introduction to quantitative genetics. Oliver and Boyd Press, Edinburgh. 365 pp. Fallis, A. M. and C. E. Hope. 1950. Observations of ruffed grouse in southern Ontario with a discussion on cycles. Can. Field-Nat. 64:82-85. Falls, J. B. and M. K. McNicholl. 1979. Neighbor-stranger discrimination by song in male blue grouse. Can. J. Zool. 57:457-62. Farmes, R. E. and G. H. Maertens. 1973. The status of the prairie chicken in Minnesota: What now? Pp. 92-94 in W. D. Svedarsky and T. Wolfe, eds. Proc. The prairie chicken in Minnesota. Univ. Minn., Crookston. Feyerherm, A. M., L. D. Bark, and W. C. Burrows. 1966. Probabilities of sequences of wet and dry days in Minnesota. Kans. Tech. Bull. 139j. Manhattan. 55 pp. Ficken, R. W., M. S. Ficken, andJ. P. Mailman. 1978. Differential aggression in genetically different morphs of the white-throated sparrow (Zonotrichia albicollis). Z. Tierpsychol. 46:43-57. Fischer, C. A. and L. B. Keith. 1974. Population responses of central Alberta ruffed grouse to hunting. J. Wildl. Manage. 38:585-600.
746
REFERENCES
Fisher, L. W. 1939. Studies of the eastern ruffed grouse in Michigan (Bonasa umbellus umbellus). Mich. Agric. Exp. Station, East Lansing. Tech. Bull. 166. 46 pp. Fisher, R. A. 1930. The genetical theory of natural selection. Clarendon Press, Oxford. Fisher, R. A. 1958. The genetical theory of natural selection. 2nd ed. Dover Press, New York. Fleming, T. H. 1973. Numbers of mammal species in North and Central American forest communities. Ecology 55:555-63. Fogleman, J. C., P. S. Corn, and D. Pettus. 1980. The genetic basis of a dorsal color polymorphism in Rana pipiens. J. Hered. 71:439-40. Foster, M. S. 1983. Disruption, dispersion, and dominance in lek-breeding birds. Am. Nat. 122:53-72. Fowle, C. D. 1944. The sooty grouse, Dendragapus fuliginous fuliginous (Ridgway), on its summer range. M.Sc. thesis. Univ. British Columbia, Vancouver. Fowle, C. D. 1960. A study of the blue grouse (Dendragapus obscurus), on Vancouver Island, British Columbia. Can. J. Zool. 38:701-13. Frandsen, D. H. 1980. Density, habitat, and behaviour of blue grouse. M.Sc. thesis. Univ. Toronto, Toronto. French, N. R., T. Y. Tagami and P. Hayden. 1968. Dispersal in a population of desert rodents. J. Mammal. 49:272-80. Fretwell, S. D. 1972. Populations in a seasonal environment. Princeton Univ. Press, Princeton, N.J. 217 pp. Fretwell, S. D. and H. L. Lucas, Jr. 1969. On territorial behavior and other factors influencing habitat distribution in birds. I. Theoretical development. Acta Biotheoretica 19:16-36. Gadgil, M. 1971. Dispersal: population consequences and evolution. Ecology 52:253-61. Gadgil, M. and W. H. Bossert. 1970. Life history consequences of natural selection. Am. Nat. 104:1-24. Gandelman, R. and B. Svare. 1974. Mice: pregnancy termination, lactation and aggression. Horm. Behav. 5:397-405. Gardarsson, A. 1971. Food ecology and spacing behavior of rock ptarmigan (Lagopus mutus) in Iceland. Ph.D. thesis. Univ. California, Berkeley. Gardarsson, A. 1979. Waterfowl populations of Lake Myvatn and recent changes in numbers and food habits. Oikos 32:250-70. Gardarsson, A. and R. Moss. 1970. Selection of food by Icelandic ptarmigan in relation to its availability and nutritive value. Pp. 47-71 in A. Watson, ed. Animal populations in relation to their food resources. Blackwell Sci. Publ., Oxford and Edinburgh. Gates,!. M. 1971. The ecology of a Wisconsin pheasant population. Ph.D. thesis. Univ. Wisconsin, Madison. Gates, J. M. and J. B. Hale. 1974. Seasonal movement, winter habitat use and population distribution of an east central Wisconsin pheasant population. Wis. Dep. Nat. Resour. Tech. Bull. 76. 55 pp. Gates, J. M. and J. B. Hale. 1975. Reproduction of an east central Wisconsin pheasant population. Wis. Dep. Nat. Resour. Tech. Bull. 85. 70 pp. Gates, R. J. 1985. Observations on the formation of a sage grouse lek. Wilson Bull. 97:219-21. Gershenson, S. 1945. Evolutionary studies on the distribution and dynamics of melanism in the hamster (Cricetus cricetus L.). Genetics 30:207-51. Gibson, R. M. and J. W. Bradbury. 1985. Sexual selection in lekking sage grouse: phenotypic correlates of male mating success. Behav. Ecol. Sociobiol. 18:117-23. Giesel, J. T. 1974. Fitness and polymorphism for net fecundity distribution in iteroparous populations. Am. Nat. 108:321-31. Giesen, K. M. 1977. Mortality and dispersal of juvenile white-tailed ptarmigan. M.S. thesis. Colorado State Univ., Fort Collins. Giesen, K. M. and C. E. Braun. 1976. Renesting of white-tailed ptarmigan in Colorado. ColoradoWyoming Acad. Sci. J. 8:71-72.
REFERENCES
747
Giesen, K. M. and C. E. Braun. 1979a. Nesting behavior of female white-tailed ptarmigan in Colorado. Condor 81:215-17. Giesen, K. M. and C. E. Braun. 1979b. Renesting of white-tailed ptarmigan in Colorado. Condor 81:217-18. Giesen, K. M., C. E. Braun, and T. A. May. 1980. Reproduction and nest-site selection by whitetailed ptarmigan in Colorado. Wilson Bull. 92:188-99. Gill, R. B. 1965. Distribution and abundance of a population of sage grouse in North Park, Colorado. M.S. thesis. Colorado State Univ., Fort Collins. Gillespie, J. 1974. Polymorphism in patchy environments. Am. Nat. 108:145-51. Gjesdal, A. 1977. Social rank, mating and egg fertilization in willow ptarmigan (Lagopus lagopus lagopus). Poultry Science 56:41-44. Gladfelter, H. L. and R. S. McBurney. 1971. Mating activity of ruffed grouse. Auk 88:176-77. Gleason, H. A. and A. Cronquist. 1963. Manual of vascular plants of northeastern United States and adjacent Canada. Van Nostrand Reinhold, New York. 810 pp. Godfrey, G. A. 1967. Summer and fall movements and behavior of immature ruffed grouse (Bonasa umbellus [L]). M.S. thesis. Univ. Minnesota, St. Paul. Godfrey, G. A. 1970. Snow roosting behavior of immature ruffed grouse. Auk 87:578-79. Godfrey, G. A. 1975. Home range characteristics of ruffed grouse broods in Minnesota. J. Wildl. Manage 39:287-98. Godfrey, G. A. and W. H. Marshall. 1969. Brood break-up and dispersal of ruffed grouse. J. Wildl. Manage. 33:609-20. Gower, W. C. 1939. The use of the bursa of Fabricius as an indication of age in game birds. Trans. N. Am. Wildl. Conf. 4:426-30. Grange, W. B. 1948. Wisconsin grouse problems. Wis. Conserv. Dep. Publ. 328. 318 pp. Gratson, M. W. 1981. Goshawks prey on radio-tagged sharp-tailed grouse. J. Field Ornith. 53:55-56. Gratson, M. W. 1983. Habitat, mobility, and social patterns of sharp-tailed grouse in Wisconsin. M.S. thesis. Univ. Wisconsin, Stevens Point. Grondahl, C. R. 1956. Predation studies using artificial pheasant nests. N.D. Game and Fish Dep. Fed. Aid Wildl. Restoration Proj. W-35-R-3. 14 pp. Gross, A. O. 1928. The heath hen. Mem. Boston Society Nat. Hist. 6:491-588. Gross, A. O. 1930. Progress report of the Wisconsin prairie chicken investigation. Wise. Conserv. Comm., Madison. 112 pp. Gudmundsson, F. 1960. Some reflections on ptarmigan cycles in Iceland. Proc. Int. Ornithol. Congr. 12:259-65. Gullion, G. W. 1963. Populations of grouse. Minn. Div. Game and Fish Quart. Rept. 23:77-113. Gullion, G. W. 1964. Evaluation of food, cover and other grouse management practices. Minn. Div. Game and Fish. Proj. C-35-R-8. 71 pp. (mimeogr.). Gullion, G. W. 1965. Improvements in methods for trapping and marking ruffed grouse. J. Wildl. Manage. 29:109-16. Gullion, G. W. 1966a. A viewpoint concerning the significance of studies of game bird food habits. Condor 68:372-76. Gullion, G. W. 1966b. The use of drumming behavior in ruffed grouse population studies. J. Wildl. Manage. 30:717-29. Gullion, G. W. 1967. Selection and use of drumming sites by male ruffed grouse. Auk. 84:87-112. Gullion, G. W. 1970a. Factors affecting ruffed grouse populations in the boreal forests of northern Minnesota, USA. Finnish Game Res. 30:103-17. Gullion, G. W. 1970b. Factors influencing ruffed grouse populations. Trans. N. Am. Wildl. Conf. 35:93-105. Gullion, G. W. 1970c. The ruffed grouse in northern Minnesota. Univ. Minnesota For. Wildl. Relations Proj. Revisal FW67. No. 49. 37 pp.
748
REFERENCES
Gullion, G. W. 1973. Grouse cycles: are they real? Minn. Sci. 29:6-7,10-11. Gullion, G. W. 1977a. Forest manipulation for ruffed grouse. Trans. N. Am. Wildl. and Nat. Resour. Conf. 42:449-58. Gullion, G. W. 1977b. Maintenance of the aspen ecosystem as a primary wildlife habitat. Proc. Int. Congr. Game Biol. 13:256-65. Gullion, G. W. 1981. Non-drumming males in ruffed grouse population. Wilson Bull. 93:372-82. Gullion, G. W. 1982. Rejuvenation and maintenance of forest habitats for the American ruffed grouse. Pp. 11-29 in T. W. I. Lovel, ed. Proc. 2nd Internal!. Symp. on Grouse. World Pheasant Assoc., Exning, England. Gullion, G. W. 1984a. Managing northern forests for wildlife. The Ruffed Grouse Society, Coraopolis, Pa. 72 pp. Gullion, G. W. 1984b. Grouse of the North Shore. Willow Creek Press, Oshkosh, Wis. 136 pp. Gullion, G. W., R. T. King, andW. H. Marshall. 1962. Male ruffed grouse and thirty years of forest management on the Cloquet Forest Research Center, Minnesota. J. For. 60:617-22. Gullion, G. W. and W. H. Marshall. 1968. Survival of ruffed grouse in a boreal forest. Living Bird 7:117-67. Gwadz, R. W. 1970. Monofactorial inheritance of early sexual receptivity in the mosquito, Aedes atropelus. Anim. Behav. 18:358-61. Gyllensten, U. 1984. Biochemical genetic variation, intraspecific genetic structure and postglacial evolution of select North European vertebrate species. Ph.D. thesis. Stockholm, Sweden. Gyllensten, U. 1985. Temporal allozyme frequency changes in density fluctuating populations of willow grouse (Lagopus lagopus L.). Evolution 39:115-21. Haas, G. F. 1974. Habitat selection, reproduction, and movements in female spruce grouse. Ph.D. thesis. Univ. Minnesota, St. Paul. Hagen, Y. 1935. Ringmerking av lirype i Rauland og Tinn. Nytt. Mag. Naturv. 75:243-88. Hagen, Y. 1952a. The gyrfalcon (Falco r. rusticolus) in Dovre, Norway. Some breeding records and food studies. Del Norske Videnskaps — Akademi i Oslo Mat. Naturv. Klassa 4:1-37. Hagen, Y. 1952b. Rovfuglene og Viltpleien. Oslo (Gyldendal). 603 pp. Hager, D. C., Jr. 1954. Ruffed grouse, Bonasa umbellus umbellus (L.), populations on two areas in central Massachusetts. M.S. thesis. Univ. Massachusetts, Amherst. Hale, J. B. and R. S. Dorney. 1963. Seasonal movements of ruffed grouse in Wisconsin. J. Wildl. Manage. 27:648-56. Hale, J. B. and R. F. Wendt. 1951. Ruffed grouse hatching dates in Wisconsin. J. Wildl. Manage. 15:195-99. Halliday, W. E. D. 1937. A forest classification of Canada. Can. Dep. Mines, Resour., For. Serv. Bull. 89 pp. Halvorsen, H. H. 1981. An evaluation of grassland management practices for wildlife in central Wisconsin. M.S. thesis. Univ. Wisconsin, Stevens Point. Hamerstrom, F. 1950. Range, habits and food requirements of prairie chicken. Wis. Conserv. Bull. 15:9-11. Hamerstrom, F. 1980. Strictly for the chickens. Iowa State Univ. Press, Ames. 174 pp. Hamerstrom, F., D. D. Berger, and F. N. Hamerstrom, Jr. 1965. The effect of mammals on prairie chickens on booming grounds. J. Wildl. Manage. 29:536-42. Hamerstrom, F. N., Jr. 1936. A study of the nesting habitats of the ring-necked pheasant in northern Iowa. Iowa State Coll. J. Sci. 10:173-203. Hamerstrom, F. N., Jr. 1939. A study of Wisconsin prairie chicken and sharp-tailed grouse. Wilson Bull. 51:105-20. Hamerstrom, F. N., Jr. 1941. A study of Wisconsin prairie grouse: breeding habits, winter foods, endoparasites, and movements. Ph.D. thesis. Univ. Wisconsin, Madison. Hamerstrom, F. N., Jr. 1963. Sharptail brood habitat in Wisconsin's northern pine barrens. J. Wildl. Manage. 27:793-802.
REFERENCES
749
Hamerstrom, F.N., Jr. 1981. Do prairie chicken cocks move from exterior to interior territories as they grow older? Proc. Prairie Grouse Tech. Counc. 14:16. Hamerstrom, F. N. Jr. and F. Hamerstrom. 1949. Daily and seasonal movements of Wisconsin prairie chickens. Auk 66:312-37. Hamerstrom, F. N., Jr. and F. Hamerstrom. 1951. Mobility of the sharp-tailed grouse in relation to its ecology and distribution. Am. Midi. Nat. 46:174-226. Hamerstrom, F. N., Jr. and F. Hamerstrom. 1955. Population density and behavior in Wisconsin prairie chickens (Tympanuchus cupido pinnatus). Proc. Int. Ornithol. Congr. 11:459-66. Hamerstrom, F. N., Jr. and F. Hamerstrom. 1961. Status and problems of North American grouse. Wilson Bull. 73:284-94. Hamerstrom, F. N., Jr. and F. Hamerstrom. 1973. The prairie chicken in Wisconsin. Wis. Dep. Nat. Resour. Tech. Bull. 64. 52 pp. Hamerstrom, F. N., Jr., F. Hopkins, and A. J. Rinzel. 1941. An experimental study of browse as a winter diet for prairie chickens. Wilson Bull. 53:185-95. Hamerstrom, F. N., Jr., O. E. Mattson, and F. Hamerstrom. 1957. A guide to prairie chicken management. Wis. Conserv. Dep. Tech. Wildl. Bull. 15. 128 pp. Hamilton, W. D. 1971. Geometry for the selfish herd. J. Theor. Biol. 31:295-311. Hannon, S. J. 1978. The reproductive cycle, movements, and pre-nesting behavior of adult and yearling females in a population of blue grouse. M.Sc. thesis. Univ. Alberta, Edmonton. Hannon, S. J. 1980. The cackle call of female blue grouse; does it have a mating or aggressive function? Auk 97:404-7. Hannon, S. J. 1982. Female aggressiveness, breeding density, and monogamy in willow ptarmigan. Ph.D. thesis. Univ. British Columbia, Vancouver. Hannon, S. J. 1983. Spacing and breeding density of willow ptarmigan in response to an experimental alteration of sex ratio. J. Anim. Ecol. 52:807-20. Hannon, S. J. 1984. Factors limiting polygyny in the willow ptarmigan. Anim. Behav. 32:153-61. Hannon, S. J., B. R. Smard, F. C. Zwickel, andJ. F. Bendell. 1979. Differences in the gonadal cycles of adult and yearling blue grouse. Can. J. Zool. 57:1283-89. Hannon, S. J. and J. N. M. Smith. 1984. Factors influencing age-related reproductive success in willow ptarmigan. Auk 101:848-54. Hannon, S. J., L. G. Sopuck, and F. C. Zwickel. 1982. Spring movements of female blue grouse; evidence for socially induced delayed breeding in yearlings. Auk 99:687-94. Hannon, S. J. and F. C. Zwickel. 1979. Probable non-breeders among female blue grouse. Condor 81:78-82. Hanson, H. C. 1953. Muskeg as sharp-tailed grouse habitat. Wilson Bull. 65:235-41. Hanssen, I., H. J. Grav, J. B. Steen, and H. Lysnes. 1979. Vitamin C deficiency in growing willow ptarmigan (Lagopus lagopus lagopus). J. Nutr. 109:2260-76. Hanssen, I. and J. Ness. 1982. Chick nutrition and mortality in captive willow ptarmigan (Lagopus I. lagopus). Acta Vet. Scand. 23:456-65. Hanssen, I., J. Ness, and J. B. Steen. 1982. Parental nutrition and chick production in captive willow ptarmigan (Lagopus 1. lagopus). Acta Vet. Scand. 23:528-38. Hanssen, I. and F. Utne. 1985. Spring phenology, egg quality and chick production in willow ptarmigan Lagopus 1. lagopus in northern Norway. Fauna norv. Ser. C., Cinclus 8:77-81. Hansson, L. 1982. Food as a limiting factor for small rodent numbers. Oecologia 37:297-314. Hardy, F. C. 1950. Ruffed grouse studies in eastern Kentucky. Ky. Div. Fish and Game, Fed. Aid Unit. Proj. 18-R. 25 pp. Harpending, H. C. 1979. The population genetics of interactions. Am. Nat. 113:622-30. Harper, J. L. 1977. Population biology of plants. Academic Press, London. Hart, C. M., O. S. Lee, and J. B. Low. 1950. The sharp-tailed grouse in Utah: its life history, status and management. Utah State Dep. Fish and Game. Fed. Aid Div. Publ. 3. 79 pp. Hartzler, J. E. 1972. An analysis of sage grouse lek behavior. Ph.D. thesis. Univ. Montana, Missoula.
750
REFERENCES
Hartzler, J. E. 1974. Predation and the daily timing of sage grouse leks. Auk 91:532-36. Harvey, M. F. and R. W. Barbour. 1965. Home range of Microtus ochrogaster as determined by a modified minimum area method. J. Mammal. 46:398-402. Haukioja, E. and T. Hakala. 1975. Herbivore cycles and periodic outbreaks. Formulation of a general hypothesis. Rep. Kevo Subarctic Res. Station 12:1-9. Haukioja, E., K. Kapiainen, P. Niemela, and J. Tuomi. 1983. Plant availability hypothesis and other explanations of herbivore cycles: complementary or exclusive alternatives? Oikos 40:419-32. Healy, M. C. 1967. Aggression and self-regulation of population size in deer mice. Ecology 48:377-92. Hedberg, J. 1980. Habitat selection by spruce grouse in eastern Maine. M.S. thesis. Univ. Maine, Orono. Hedrick, P. W., M. E. Ginevan, and E. P. Ewing. 1976. Genetic polymorphism in heterogeneous environments. Ann. Rev. Ecol. Syst. 7:1-32. Heisler, I. L. 1985. Quantitative genetic models of female choice based on "arbitrary" male characters. Heredity 55:187-98. Henderson, B. A. 1977. The genetics and demography of a high and low density of red grouse Lagopus I. scoticus. J. Anirn. Ecol. 46:581-92. Henderson, F. R., R. W. Brooks, R. E. Wood, and R. B. Dahlgren. 1967. Sexing of prairie grouse by crown feather patterns. J. Wildl. Manage. 31:764-69. Henderson, F. R. and W. W. Jackson. 1967. History of selected dancing grounds in South Dakota. Proc. 7th Prairie Grouse Tech. Counc. (abstract.) Herzog, P. W. 1977a. Dispersion and mobility in a local population of spruce grouse (Canachites canadensis Franklinif). M.Sc. thesis. Univ. Alberta, Edmonton. Herzog, P. W. 1977b. Summer habitat use by white-tailed ptarmigan in southwestern Alberta. Can. Field-Nat. 91:367-71. Herzog, P. W. 1978. Food selection by female spruce grouse during incubation. J. Wildl. Manage. 42:632-35. Herzog, P. W. 1979. Effects of radio-marking on behavior, movements, and survival of spruce grouse. J. Wildl. Manage. 43:316-23. Herzog, P. W. 1980. Winter habitat use by white-tailed ptarmigan in southwestern Alberta. Can. Field-Nat. 94:159-62. Herzog, P. W. and D. A. Boag. 1977. Seasonal changes in aggressive behavior of female spruce grouse. Can. J. Zool. 55:1734-39. Herzog, P. W. and D. A. Boag. 1978. Dispersion and mobility in a local population of spruce grouse. J. Wildl. Manage. 42:853-65. Herzog, P. W. and D. M. Keppie. 1980. Migration in a local population of spruce grouse. Condor 82:366-72. Hess, E. H. 1972. Imprinting in a natural laboratory. Sci. Am. 227:24-31. Hewson, R. 1984. Changes in the numbers of foxes (Vulpes vulpes) in Scotland. J. Zool. Lond. 204:561-69. Hickey, J. J. 1955. Some American population research on gallinaceous birds. Pp. 326-396 in A. Wolfson, ed. Recent studies in avian biology. Univ. Illinois Press, Urbana. Higgins, K. F., L. M. Kirsch, H. F. Duebbert, A. T. Klett, J. T. Lokemoen, H. W. Miller, and A. D. Kruse. 1977. Construction and operation of cable-chain drag for nest searches. Wildl. Leafl. 512. U.S. Fish and Wildl. Serv., Wash., D.C. 14 pp. Hillman, C. N. and W. W. Jackson. 1973. The sharp-tailed grouse in South Dakota. S.D. Dep. Game, Fish and Parks. Tech. Bull. 3. 64 pp. Hilton, M. and W. Wishart. 1981. A summary of the Camp Wainwright grouse season September 5, 7-12, 1981. Alberta Energy and Natl. Resour. Fish and Wildl. Div. 9 pp. Hines, J. E. 1986a. Recruitment of young in a declining population of blue grouse. Ph.D. thesis. Univ. of Alberta, Edmonton.
REFERENCES
751
Mines, J. E. 1986b. Survival and reproduction of dispersing blue grouse. Condor 88:43-49. Hjorth, I. 1967. Fortplantninsbetweenden inom honsfagelfamilien Tetraonidae. Var. Fagelvarld 26:193-241. Hjorth, I. 1968. Significance of light in the initiation of morning display of the black grouse (Lyrurus tetrix L.). Viltrevy 5:39-94. Hjorth, I. 1970. Reproductive behaviour in Tetraonidae with special reference to males. Viltrevy 7:183-596. Hoffman, R. S. 1961. The quality of the winter food of blue grouse. J. Wildl. Manage. 25:209-10. Hoffman, R. W. 1979. Population dynamics and habitat relationships of blue grouse. Colo. Div. Wildl., Game Res. Rep. Proj. W-37-R-32, Job 5. Pp. 201-44. Hoffman, R. W. and C. E. Braun. 1975. Migration of a wintering population of white-tailed ptarmigan in Colorado. J. Wildl. Manage. 39:485-90. Hofslund, P. B. 1973. An invasion of goshawks. Raptor Res. 7:107-8. Hogen-Warburg, A. J. 1966. Social behaviour of the ruff(Philomachus pugnax). Ardea 55:109-299. Hoglund, N. H. 1964. Fright moulting in Tetraonids. Viltrevy 2:419-25. Hoglund, N. H. 1980. Studies on the winter ecology of the willow grouse (Lagopus lagopus lagopus L.). Viltrevy 11:249-70. Horak, G. J. 1974. Population dynamics of prairie chicken in grassland and cropland areas. Kans. For., Fish and Game Comm., Final Rep. Proj. W-23-R, Job D-3. 98 pp. Horkel, J. D. 1979. Cover and space requirements for Attwater's prairie chicken (Tympanuchus cupido attwateri) in Refugio County, Texas. Ph.D. thesis. Texas A&M Univ., College Station. Horkel, J. D., R. S. Lutz, and N. J. Silvy. 1978. The influence of environmental parameters on nesting success of upland game birds. Proc. Annu. Conf. Southeast. Assoc. Fish and Wildl. Agencies 32:234-41. Horkel, J. D. and N. J. Silvy. 1980. Evolutionary considerations in creating artificial leks for Attwater's prairie chicken. Pp. 42-47 in P. A. Vohs, Jr. and F. L. Knopf, eds. Proc. Prairie Grouse Symp., Oklahoma State Univ., Stillwater. Horn, H. S. 1968. The adaptive significance of colonial nesting in the Brewer's blackbird Euphagus cyanocephalus. Ecology 49:682-94. Hornocker, M. C. 1969. Winter territoriality in mountain lions. J. Wildl. Manage. 33:457-64. Howard, R. D. 1984. Alternative mating behaviors of young male bullfrogs. Amer. Zool. 24:397-406. Hudson, P. J. 1986. The effect of a parasitic nematode on the breeding production of red grouse. J. Anim. Ecol. 55:85-92. Hudson, P. J., A. P. Dobson, and D. Newborn. 1985. Cyclic and noncyclic populations of red grouse: a role for parasitism? Pp. 79-89 in D. Rollinson and R. M. Anderson, eds. Ecology and genetics of host-parasite interactions. Academic Press, London. Huempfner, R. A. 1981. Winter arboreal feeding behavior of ruffed grouse in east-central Minnesota. M.S. thesis. Univ. Minnesota, St. Paul. Huempfner, R. A. and G. J. Erickson. 1975. A designation of plant communities on part of the Cedar Creek Natural History Area, Benthel, Minnesota. In-house plant community key and maps available at Cedar Creek, Univ. Minnesota. 25 pp. Huempfner, R. A., S. J. Maxson, G. J. Erickson, and R. J. Schuster. 1975. Recapturing radio-tagged ruffed grouse by nightlighting and snow-burrow netting. J. Wildl. Manage. 39:821-23. Huff, D. E. 1970. A study of selected nutrients in browse available to the ruffed grouse. M.S. thesis. Univ. Minnesota, St. Paul. Huff, D. E. 1973. A preliminary study of ruffed grouse-aspen nutrient relationships. Ph.D. thesis. Univ. Minnesota, St. Paul. Hungerford, K. E. 1951. Ruffed grouse populations and cover use in northern Idaho. Trans. N. A. Wildl. Conf. 16:216-24.
752
REFERENCES
Hunt, F. 1982. Regulation of population cycles by genetic feedback: existence of periodic solutions of a mathematical model. J. Math. Biol. 13:271-82. Hurst, G. A. 1970. The effects of controlled burning on arthropod denstiy and biomass in relation to bobwhite quail (Colinus virginianus) brood habitat. Ph.D. thesis. Mississippi State Univ. Hussell, D. J. T. 1972. Factors affecting clutch size in Arctic passerines. Ecol. Monogr. 42:317-64. Huxley, J. S. 1938. Threat and warning colouration in birds with a general discussion of the biological functions of colour. Zool. Society, London. Pp. 430-55. Innes, D. J. and L. E. Haley. 1977. Inheritance of a shell-color polymorphism in the mussel. J. Hered. 68:203-4. Irving, L., G. C. West, and L. J. Peyton. 1967a. Winter feeding program of Alaska willow ptarmigan shown by crop contents. Condor 69:69-11. Irving, L., G. C. West, L. J. Peyton, and S. Paneak. 1967b. Migration of willow ptarmigan in arctic Alaska. Arctic 20:77-85. Jacobs, J. 1981. How heritable is innate behaviour? Z. Tierpsychol. 55:1-18. Jamieson, I. G. 1982. Spatial patterns and behavior of yearling blue grouse and their relation to recruitment into the breeding population. M.Sc. thesis. Univ. Alberta, Edmonton. Jamieson, I. G. 1983. Seasonal changes in spatial patterns and behavior of yearling blue grouse on the breeding range. Can. J. Zool. 61:2777-80. Jamieson, I. G. and F. C. Zwickel. 1983. Spatial patterns of yearling male blue grouse and their relation to recruitment into the breeding population. Auk 100:653-57. Janson, R. G. 1951. Prairie grouse population trends in South Dakota. Midwest Wildl. Conf. Pp. 52-59. Janson, R. G. 1955. Prairie grouse brood studies. 1953-1954. S.D. Dep. Game, Fish and Parks, Fed. Aid in Wildl. Restoration. Proj. W-17-R-8. 8 pp. Jenkins, D. and A. Watson. 1962. Fluctuations in a red grouse (Lagopus scoticus) population 1956-59. Pp. 96-117 in E. D. LeCren and M. Holdgate, eds. The exploitation of natural animal populations. Blackwell Sci, Publ., Oxford. 399 pp. Jenkins, D. and A. Watson. 1970. Population studies on red grouse, Lagopus lagopus scoticus. Trans Int. Congr. Union Game Biol. 7:121-41. Jenkins, D., A. Watson, and G. R. Miller. 1963. Population studies on red grouse, Lagopus lagopus scoticus (Lath.) in north-east Scotland. J. Anim. Ecol. 32:317-76. Jenkins, D., A. Watson, andG. R. Miller. 1964. Predation and red grouse populations. J. Appl. Ecol. 1:183-95. Jenkins, D., A. Watson, and G. R. Miller. 1967. Population fluctuations in the red grouse Lagopus lagopus scoticus. J. Anim. Ecol. 36:97-122. Jenkins, D., A. Watson, and N. Picozzi. 1965. Red grouse chick survival in captivity and in the wild. Trans. Int. Congr. Game Biol. 6:63-70. Jenni, D. A. and J. E. Hartzler. 1978. Attendance at a sage grouse lek: implications for spring censuses. J. Wildl. Manage. 42:46-52. Jensen, L. and R. A. Ryder. 1965. Breeding behavior of the white-tailed ptarmigan near Aspen, Colorado. Colorado-Wyoming Acad. Sci. J. 5:52-53. Johns, J. E. and E. W. Pfeiffer. 1963. Testosterone induced incubation patches of phalaopes. Science 140:1225-26. Johnsgard, P. A. 1973. Grouse and quails of North America. Univ. Nebraska Press, Lincoln. 553 pp. Johnsgard, P. A. 1983. The grouse of the world. Univ. Nebraska Press, Lincoln 413 pp. Johnsgard, P. A. and R. E. Wood. 1968. Distributional changes and interaction between prairie chickens and sharp-tailed grouse in the midwest. Wilson Bull. 80:173-88. Johnson, D. H. and A. B. Sargeant. 1977. Impact of red fox predation on the sex ratio of prairie mallards. U.S. Dep. Inter. Fish and Wildl. Serv. Wild. Res. Rep. 6. 56 pp. Johnson, M. D. 1964. Feathers from the prairie. N.D. Dep. Game and Fish, Proj. W-67-R-5. 240 pp.
REFERENCES
753
Jones, D. M. and J. B. Theberge. 1982. Summer home range and habitat utilisation of the red fox (Vulpes vulpes) in a tundra habitat, northwest British Columbia. Can. J. Zool. 60:807-12. Jones, R. E. 1963. Identification and analysis of lesser and greater prairie chicken habitat. J. Wildl. Manage. 27:757-78. J0rgensen, E. and A. S. Blix, 1985. Effects of climate and nutrition on growth and survival of willow ptarmigan chicks. Ornis Scand. 16:99-107. J0rgenson, J. P. 1977. Pinnated grouse (Tympanuchus cupido pinnatus) movements and habitat utilization in the northern Great Plains. M.S. thesis Univ. North Dakota, Grand Forks. Jurries, R. W. 1981. Requirements of Attwater's prairie chicken. Texas Fed. Aid in Wildl. Restoration Act, Final Rep. Proj. W-108-R-4, Job 13. 3 pp. Kalela, O. 1962. On the fluctuations in the numbers of arctic and boreal small rodents as a problem of production biology. Ann. Acad. Sci. Fenn. Ser. A, 4 Biol. 66:1-38. Kalmbach, E. R. 1938. A comparative study of nesting waterfowl on the Lower Souris Refuge. 1936-1937. Trans. N. Am. Wildl. Conf. 3:610-23. Kaufman, J. H. 1983. On the definitions and functions of dominance and territoriality. Biol. Rev. Cambridge Philosophical Soc. 58:1-20. Keith, L. B. 1962. Fall and winter weights of Hungarian partridges and sharp-tailed grouse from Alberta. J. Wildl. Manage. 26:336-37. Keith, L. B. 1963. Wildlife's ten-year cycle. Univ. Wisconsin Press, Madison. 201 pp. Keith, L. B. 1974. Some features of population dynamics in mammals. Trans. Int. Congr. Game Biol. 11:17-58. Keith, L. B., A. W. Todd, C. J. Brand, R. S. Adamcik, andD. H. Rusch. 1977. An analysis of predation during a cyclic fluctuation of snowshoe hares. Trans Int. Congr. Game Biol. 13:151-75. Keith, L. B. and L. A. Windberg. 1978. A demographic analysis of the snowshoe hare cycle. Wildl. Monogr. 58. 70 pp. Keller, B. and C. J. Krebs. 1970. Microtus population biology. III. Reproductive changes in fluctuating populations of M. ochrogaster and M. pennsylvanicus in southern Indiana, 1965-67. Ecol. Monogr. 40:263-94. Keller, R. J., H. R. Sheperd, and R. N. Randall. 1941. Survey of 1941: North Park, Jackson County, MofTat County, including comparative data of previous seasons. Colo. Dep. Game and Fish, Sage Grouse Survey. Vol. 3. 31 pp. Kelsall, J. P., E. S. W. Telfer, and R. D. Wright. 1977. The effects of fire on the ecology of the boreal forest with particular reference to the Canadian North: a review and selected bibliography. Can Wildl. Serv. Occas. Pap. No. 32, 58 pp. Keppie, D. M. 1975a. Clutch size of the spruce grouse, Canachites canadensis franklinii, in southwest Alberta. Condor 77:91-92. Keppie, D. M. 1975b. Dispersal, overwinter mortality and population size of spruce grouse (Canachites canadensis franklinii). Ph.D. thesis. Univ. Alberta, Edmonton. Keppie, D. M. 1977a. Inter-brood movements of juvenile spruce grouse. Wilson Bull. 89:67-72. Keppie, D. M. 1977b. Snow cover and the use of trees by spruce grouse in autumn. Condor 79:382-84. Keppie, D. M. 1979. Dispersal, overwinter mortality, and recruitment of spruce grouse. J. Wildl. Manage. 43:717-27. Keppie, D. M. 1980. Similarity of dispersal among siblings male spruce grouse. Can. J. Zool. 58:2102-4. Keppie, D. M. 1982. A difference in production and associated events in two races of spruce grouse. Can. J. Zool. 60:2116-23. Keppie, D. M. and P. W. Herzog. 1978. Nest site characteristics and nest success of spruce grouse. J. Wildl. Manage. 42:628-32. Kermott, L. H., III. 1982. Breeding behavior in the sharp-tailed grouse. Ph.D. thesis. Univ. Minnesota, St. Paul.
754
REFERENCES
Kermott, L. H. and L. W. Oring. 1975. Acoustical communication of male sharp-tailed grouse (Pedioecetes phasianellus) on a North Dakota dancing ground. Anim. Behav. 23:315-86. Kessler, W. B. 1977. Availability and use of Attwater's greater prairie chicken habitat. Proc. Prairie Grouse Tech. Counc. 12:13-14. Kettlewell, H. B. D. 1973. The evolution of melanism. Clarendon Press, Oxford. 423 pp. Kiel, W. H., A. S. Hawkins, and N. G. Perret. 1972. Waterfowl habitat trends in the aspen parkland of Manitoba. Can. Wildl. Serv. Rep. Ser. 18. 63 pp. King, D. G. 1971. The ecology and population dynamics of blue grouse in the sub-alpine. M.S. thesis. Univ. British Columbia, Vancouver. King, J. A. 1955. Social behavior, social organization and population dynamics of a black-tailed prairie dog town in the Black Hills of South Dakota. Univ. Michigan, Contrib. Lab. Vertebr. Biol. 67:1-123. King, R. T. 1937. Ruffed grouse management. J. For. 35:523-32. Kirkpatrick, M. 1982. Sexual selection and the evolution of female choice. Evolution 36:1-12. Kirkpatrick, M. 1985. Evolution of female choice and male parental investment in polygynous species: the demise of the "sexy son." Am. Nat. 125:788-810. Kirkpatrick, M. 1986. The handicap mechanism of sexual selection does not work. Am. Nat. 127:222-40. Kirsch, L. M. 1956. Spring prairie grouse census and habitat inventory methods. Wildl. Leafl. 13. U.S. Fish and Wildl. Serv., Wash., D.C. 7 pp. Kirsch, L. M. 1969. Waterfowl production in relation to grazing. J. Wildl. Manage. 33:821-28. Kirsch, L. M. 1974. Habitat management considerations for prairie chickens. Wildl. Soc. Bull. 2:124-29. Kirsch, L. M. 1983. Historical ecological records of great plains grasslands workshop: management of public lands in the northern great plains, Bismarck, North Dakota. 3 pp. Kirsch, L. M., A. T. Klett, and H. W. Miller. 1973. Land use and prairie grouse population relationships in North Dakota. J. Wildl. Manage. 37:449-53. Kirsch, L. M. and A. D. Kruse. 1973. Prairie fires and wildlife. Proc. Annu. Tall Timbers Fire Ecol. Conf. 12:289-303. Klebenow, D. A. 1968. Nesting and brood habitat of sage grouse. Ph.D. thesis. Univ. Idaho, Moscow. 61 pp. Klebenow, D. A. 1969. Sage grouse nesting and brood habitat in Idaho. J. Wildl. Manage. 33:649-62. Klett, A. T. and D. H. Johnson. 1982. Variability in nest survival rates and implications to nesting studies. Auk 99:77-87. Klimstra, W. D. and J. L. Roseberry. 1975. Nesting ecology of the bobwhite in southern Illinois. Wildl. Mongr. 41. 37 pp. Klomp, H. 1970. The determination of clutch-size in birds: a review. Ardea 58:1-124. Kluyver, H. M. and L. Tinbergen. 1953. Territory and regulation of density in titmice. Arch. Neerl. Zool. 10:265-86. Knapp, A. K. 1984. Water relations and growth of three grasses during wet and drought years in a tallgrass prairie. Oecologia 65:35-43. Kobriger, G. D. 1965. Status, movements, habitats, and foods of prairie grouse on a sandhills refuge. J. Wildl. Manage. 29:788-800. Kobriger, G. D. 1975. Correlation of sharp-tailed grouse population parameters. N.D. Outdoors 38(5): 10-13. Kobriger, G. D. 1981. Prairie grouse population data. N.D. State Game and Fish Dep. Rep. B-318, Proj. W-67-R-21, Jobs B-V-1-5. 30 pp. Koertge, N. 1979. The problem of appraising scientific theories. Pp. 228-251 in P. D. Asquith and
REFERENCES
755
H. E. Kyburg, Jr., eds. Current research in the philosophy of science. Philos. Sci. Assoc., East Lansing, Mich. 533 pp. Kohn, S. C. 1976. Sharp-tailed grouse nesting and brooding habitat in southwestern North Dakota. M.S. thesis. South Dakota State Univ., Brookings. Korschgen, L. J. 1962. Food habits of greater prairie chicken in Missouri. Am. Midi. Nat. 68:307-18. Korschgen, L. J. 1966. Foods and nutrition of ruffed grouse in Missouri. J. Wildl. Manage. 30:86-100. Koskimies, J. 1957. Flocking behaviour in capercaillie, Tetrao urogallus (L.), and blackgame, Lyrurus tetrix (L.). Finnish Game Res. 18:1-32. Krajina, V. J., ed. 1965. Biogeoclimatic zones and biogeocoenoses of British Columbia. Ecology of western North America. Dep. of Bot. Univ. British Columbia, Vancouver. 112 pp. Krebs, C. J. 1966. Demographic changes in fluctuating populations of Microtus californicus. Ecol. monogr. 36:239-73. Krebs, C. J. 1978a. A review of the Chitty hypothesis of population regulation. Can. J. Zool. 56:2463-80. Krebs, C. J. 1978b. The experimental analysis of distribution and abundance. 2nd ed. Harper and Row, New York. Krebs, C. J. 1985. Do changes in spacing behaviour drive population cycles in small mammals? Pp. 295-312 in R. M. Sibley and R. H. Smith, eds. Behavioural ecology. Blackwell Sci Publ., Oxford. Krebs, C. J., M. S. Gaines, B. L. Keller, J. H. Myers, and R. H. Tamarin. 1973. Population cycles in small rodents. Science 179:35-41. Krebs, C. J. and J. H. Myers. 1974. Population cycles in small mammals. Adv. Ecol. Res. 8:267-399. Krebs, C. J., I. Wingate, J. Leduc, J. A. Redfield, M. Taitt, and R. Hilborn. 1976. Microtus population biology: dispersal in fluctuating populations of M. townsendii. Can. J. Zool. 54:79-95. Krebs, J. R. 1971. Territory and breeding density in the great tit, Parus major L. Ecology 52:2-22. Krebs, J. R. 1973. Behavioural aspects of predation. Pp. 73-111 in P. P. G. Bateson and P. H. Klopfer eds. Perspectives in ethology. Plenum Press, New York. Krebs, J. R. 1974. Colonial nesting and social feeding as strategies for exploiting food resources in the Great Blue Heron (Ardea herodias). Behaviour 52:99-131. Krebs, J. R., M. H. Roberts, and J. M. Culen. 1972. Flocking and feeding in the great tit, Parus major L.-an experimental study. Ibis 114:507-30. Kristensen, J. 1973. Distraction display in female blue grouse. B.Sc. honours thesis, Univ. Alberta, Edmonton. Kristoffersen, S. 1937. Undersokelser over lirypens forplantinings forhold. Tromso Museums rypenndersokelser, Nytt Mag. Naturv. 77:141-94. Kruijt, J. P., G. J. De Vos and I. Bossema. 1972. The arena system of black grouse. Proc. Int. Ornithol. Congr. 15:399-423. Kruijt, J. P. and H. A. Hogan. 1967. Social behaviour on the lek in the black grouse Lyrurus tetrix tetrix (L.). Ardea 55:203-20. Kubisiak, J. F. 1982. The impact of hunting on ruffed grouse populations. Paper presented at Ruffed Grouse Workshop, Oct. 7, Traverse City, Mich. 18 pp. Kubisiak, J. F., J. C. Moulton, and K. R. McCaffery. 1980. Ruffed grouse density and habitat relationships in Wisconsin. Wis. Dep. Nat. Resour. Tech. Bull. 118. 16 pp. Kuhn, T. S. 1962. The structure of scientific revolutions. Univ. Chicago Press, Chicago. 172 pp. Kuhn, T. S. 1965. Logic of discovery or psychology of research? Pp. 231-78 in I. Lakatos and A. Musgrave, eds. Criticism and the growth of knowledge. Cambridge Univ. Press, Cambridge. Kupa, J. I. 1966. Ecological studies of the female ruffed grouse (Bonasa umbellus L.) at the Cloquet Forest Research Center, Minnesota. Ph.D. thesis. Univ. Minnesota, St. Paul.
756
REFERENCES
Lack D. 1939. The display of the black cock. British Birds 32:290-303. Lack, D. 1947. The significance of clutch sizes. Ibis 89:302-52. Lack, D. 1954. The natural regulation of animal numbers. Clarendon Press, Oxford. 343 pp. Lack, D. 1965. Evolutionary ecology. J. Anim. Ecol. 34:223-31. Lack, D. 1966. Population studies of birds. Clarendon Press, Oxford. 341 pp. Lack, D. 1968. Ecological adaptations for breeding in birds. Methuen Press, London. 409 pp. Laine, K. and H. Henttonen. 1983. The role of plant production in microtine cycles in northern Fennoscandia. Oikos 40:407-18. Lakatos, I. 1965. Falsification and the methodology of scientific research programmes. Pp. 91-196 in I. Lakatos and A. Musgrave, eds. Criticism and the growth of knowledge. Cambridge Univ. Press, Cambridge. Lakatos, I. and A. Musgrave. 1965. Criticism and the growth of knowledge. Cambridge Univ. Press, Cambridge. 282 pp. Lake, P. E. 1975. Gamete production and the fertile period with particular reference to domestic birds. Symp. Zool. Lond. No. 35:225-44. Lance, A. N. 1967. A telemetry study of dispersion and breeding biology in blue grouse. M.Sc. thesis. Univ. British Columbia, Vancouver. Lance, A. N. 1970. Movements of blue grouse on the summer range. Condor 72:437-44. Lance, A. N. 1978a. Survival and recruitment success of individual young cock red grouse Lagopus L. scoticus tracked by radio-telemetry. Ibis 120:369-78. Lance, A. N. 1978b. Territories and the food plant of individual red grouse. II. Territory size compared with an index of nutrient supply in heather. J. Anim. Ecol. 47:307-13. Lance, A. N. 1983. Selection of feeding sites by hen red grouse Lagopus lagopus scoticus during breeding. Ornis Scand. 14:78-80. Lande, R. 1981. Models of speciation by sexual selection on polygenic traits. Proc. Nat. Acad. Sci., USA 78:3721-25. Langvatn, R. 1977. Characteristics and relative occurrence of remnants of prey found at nesting places of gyrfalcon Falco rusticolus. Ornis Scand. 8:113-25. Larsen, J. A. and J. F. Lahey. 1958. Influence of weather upon a ruffed grouse population. J. Wildl. Manage. 22:63-70. Latham, R. M. 1947. Differential ability of male and female game birds to withstand starvation and climatic extremes. J. Wildl. Manage. 11:139-49. Lawrence, J. S. and N. J. Silvy. 1980. Status of the Attwater's prairie chicken-an update. Pp. 29-33 in P. A. Vohs, Jr. and F. L. Knopf, eds. Proc. Prairie Grouse Symp., Oklahoma State Univ., Stillwater. Lawrence, J. S. and N. J. Silvy. 1981. Effect of predator reduction on reproductive success of Attwater's prairie chicken. Proc. Prairie Grouse Tech. Counc. 14:8-9. Lazarus, J. 1972. Natural selection and the functions of flocking in birds: a reply to Murton. Ibis 111:556-58. Lee, L. 1950. Kill analyses for the lesser prairie chicken in New Mexico, 1949. J. Wildl. Manage. 14:475_77. Lehmann, V. W. 1941. Attwater's prairie chicken: its life history and management. U.S. Fish and Wildl. Serv., N. Am. Fauna Ser. 57. 63 pp. Leslie, P. H. 1959. The properties of certain lag type of population growth and the influence of an external random factor on a number of such populations. Physiol. Zool. 32:151-59. Leopold, A. 1931. Report on a game survey of the north central states. Sporting Arms and Ammunition Manufacturers' Inst., Madison, Wis. 299 pp. Leopold, A. 1933. Game management. Charles Scribner's Sons, New York and London. 481 pp. Leopold, A. and J. N. Ball. 1931. British and American grouse management. Am. Game Mag. July-August. Sept.-Oct. Pp. 1-5.
REFERENCES
757
Levins, R. 1968. Evolution in changing environments. Princeton Univ. Press, Princeton, N.J. 120 pp. Lewis, E. 1904. The nesting habitats of the white-tailed ptarmigan in Colorado. Bird Lore 6:117-21. Lewis, R. A. 1979. Suitability and selection of territorial sites used by male blue grouse. M.Sc. thesis. Univ. Alberta, Edmonton. Lewis, R. A. 1981. Characteristics of persistent and transient territorial sites of male blue grouse. J. Wildl. Manage. 45:1048-51. Lewis, R. A. 1984. Density movements and breeding success of female blue grouse in an area of reduced male density. Can. J. Zool. 62:1556-60. Lewis, R. A. 1985. Do blue grouse form leks? Auk 102:180-83. Lewis, R. A. and F. C. Zwickel. 1980. Removal and replacement of male blue grouse on persistent and transient territorial sites. Can. J. Zool. 58:1417-23. Lewis, R. A. and F. C. Zwickel. 1981. Differential use of territorial sites by male blue grouse. Condor 83:171-76. Lewis, R. A. and F. C. Zwickel. 1982. Survival and delayed breeding in male blue grouse. Can. J. Zool. 60:1881-84. Linde, D. A., M. Muck, H. H. Pietz, and L. Roth. 1978. Grouse management surveys, 1978. S.D. Dep. Game, Fish and Parks. Game Rep. 80-6, Proj. W-95-R-13, Jobs 1-7. 24 pp. Linder, R. L., L. David, L. Lyon, and C. P. Agee. 1960. An analysis of pheasant nesting in south central Nebraska. Trans. N. Am. Wildl. Conf. 25:214-30. Lindroth, H. and L. Lindgren. 1950. On the significance for forestry of the capercaillie, Tetrao urogallus L., feeding on pine-needles. Suomen Rusta 4:60-81. Lismann, H. W. 1932. Die Umwelt dws kampffishes (Betta splendens Regen). Cited in: P. R. Marler and W. J. Hamilton, 1966. Mechanisms of animal behavior. John C. Wiley & Sons, New York. Little, T. W. 1978. Populations, distributions and habitat selection by drumming male ruffed grouse in central Minnesota prior to clearcutting. Ph.D. thesis. Univ. Minnesota, St. Paul. Livezey, B. C. 1981. Duck nesting in retired croplands at Horicon National Wildlife Refuge, Wisconsin. J. Wildl. Manage. 45:27-37. Lokemoen, J. T., H. F. Duebbert, and D. E. Sharp. 1984. Nest space, habitat selection and behavior of waterfowl on Miller Island, North Dakota. J. Wildl. Manage. 48:309-21. Lomnicki, A. 1978. Individual differences between animals and the natural regulation of their numbers. J. Anim. Ecol. 47:461-75. Lovat, Lord (ed.). 1911. The grouse in health and disease. 2 vols. Smith and Elder & Co., London. Low, D. J. 1975. Qualities of individuals in a replacement stock of blue grouse. M.Sc. thesis. Univ. British Columbia, Vancouver. Lumsden, H. G. 1965. Displays of the sharptail grouse. Ont. Dep. Lands and For., Res. Rep. 66. 68pp. Lumsden, H. G. 1968. The displays of the sage grouse. Ont. Dep. Lands and For., Res. Rep. 83. 94pp. Lumsden, H. G. 1970. The shoulder-spot display of grouse. Living Bird 9:65-74. Lumsden, H. G. and R. B. Weeden. 1963. Notes on the harvest of spruce grouse. J. Wildl. Manage. 27:587-91. Lutz, R. S. 1979. The response of Attwater's prairie chicken to petroleum development. M.Sc. thesis. Texas A&M Univ., College Station. MacArthur, R. H. 1968. Selection for life tables in periodic environments. Am. Nat. 103:381-83. MacDonald, S. D. 1968. The courtship and territorial behavior of Franklin's race of the spruce grouse. Living Bird 7:5-25. MacDonald, S. D. 1970. The breeding behavior of the rock ptarmigan. Living Bird 9:195-238. MacKenzie, J. M. D. 1952. Fluctuations in the numbers of British Tetraonids. J. Anim. Ecol. 21:128-53.
758
REFERENCES
Marcstrom, V. and N. H. Hoglund. 1981. Factors affecting reproduction of willow grouse (Lagopus lagopus) in two highland areas of Sweden. Viltrevy 11:285-314. Marler, P. 1955. Studies of fighting in chaffinches (2), the effect of dominance relations of disguising females as males. Br. J. Anim. Behav. 3:137-46. Marlin, J. C. 1967. The grasslands of North America: prolegomena to its history with addenda and postscripts. Peter Smith, Gloucester, Mass. Marr, J. W. 1961. Ecosystems of the east slope of the Front Range in Colorado. Univ. Colo. Serv. in Biol. 8. 132 pp. Marshall, W. H. 1954. Ruffed grouse and snowshoe hare populations on the Cloquet Experimental Forest, Minnesota. J. Wildl. Manage. 18:109-12. Marshall, W. H. 1965. Ruffed grouse behavior. Bioscience 15:92-94. Marshall, W. H. and G. W. Gullion. 1963. A discussion of ruffed grouse populations. Cloquet Forest Research Center, Minnesota. Trans. Int. Union Game Biol. 6:93-100. Marshall, W. H. and M. S. Jensen. 1937. Winter and spring studies of the sharp-tailed grouse in Utah. J. Wildl. Manage. 1:87-99. Marshall, W. H. and J. J. Kupa. 1963. Development of radio-telemetry techniques for ruffed grouse studies. Trans. N. Am. Wildl. Conf. 28:443-56. Martinka, R. R. 1972. Structural characteristics of blue grouse territories in southwestern Montana. J. Wildl. Manage. 36:498-510. Martinson, R. K. 1963. The relationship of weather to productivity in upland game birds. N.D. State Game and Fish Dep. Proj. W-67-R-4, Job 17. 24 pp. Martinson, R. K. and C. R. Grondahl. 1966. Weather and pheasant populations in southwestern North Dakota. J. Wildl. Manage. 30:74-81. Maxson, S. J. 1974. Activity, home range, and habitat usage of female ruffed grouse during the egglaying, incubation, and early brood periods as determined by radio telemetry. M.S. thesis. Univ. Minnesota, St. Paul. Maxson, S. J. 1977. Activity patterns of female ruffed grouse during the breeding season. Wilson Bull. 89:439-55. Maxson, S. J. 1978. Spring home range and habitat use by female ruffed grouse. J. Wildl. Manage. 42:61-71. May, T. A. 1970. Effects of sagebrush control on distribution and abundance of sage grouse. Colo. Game Bird Survey. Proj. W-37-R-23, Job 8a. Pp. 115-38. May, T. A. 1975. Physiological ecology of white-tailed ptarmigan in Colorado. Ph.D. thesis. Univ. Colorado, Boulder. May, T. A. and C. E. Braun. 1972. Seasonal foods of adult white-tailed ptarmigan in Colorado. J. Wildl. Manage. 36:1180-86. Maynard-Smith, J. 1958. The theory of evolution. Penguin Press, Harmondsworth. 320 pp. Maynard-Smith, J. and C. R. Price. 1973. The logic of animal conflict. Nature 246:15-18. McBride, G. 1958. The measurement of aggressiveness in the domestic hen. Anim. Behav. 6:87-91. McCollom, R. E., P. B. Siegel, and H. P. Van Krey. 1974. Responses to androgens in lines of chickens selected for mating behavior. Horm. Behav. 2:31-42. McCourt, K. H. 1969. Dispersion and dispersal of female and juvenile Franklin's grouse in southwestern Alberta. M.S. thesis. Univ. Alberta, Edmonton. McCourt, K. H., D. A. Boag, and D. M. Keppie. 1973. Female spruce grouse activities during laying and incubation. Auk 90:619-23. McCoy, E. D. and E. F. Connor. 1980. Latitudinal gradients in the species diversity of North American mammals. Evolution 34:193-203. McEwen, L. C., D. B. Knapp, and E. A. Milliard. 1969. Propagation of prairie grouse in captivity. J. Wildl. Manage. 33:276-83. McGowan, J. D. 1972. Population characteristics of rock ptarmigan. Alaska Dep. Fish and Game, Final Rep. Fed. Aid in Wildl. Restoration. 14 pp.
REFERENCES
759
McLachlin, R. A. 1970. The spring and summer dispersion of male Franklin's grouse in lodgepole pine forest in southwestern Alberta. M.Sc. thesis. Univ. Alberta, Edmonton. McNicholl, M. K. 1978. Behaviour and social organization in a population of blue grouse on Vancouver Island. Ph.D. thesis. Univ. Alberta, Edmonton. McNicholl, M. K. 1979. Individual variation in behaviour among blue grouse. N. Am. Bird Bander 4:16-18. Mercer, W. E. 1967. Ecology of an island population of Newfoundland willow ptarmigan. Newfoundland and Labrador Wildl. Sci. Tech. Bull. 2. 97 pp. Merrell, D. J. and C. F. Rodell. 1968. Seasonal selection in the leopard frog Ranapipiens. Evolution 22:284-88. Meslow, E. C. 1966. The drumming log and drumming log activity of male ruffed grouse. M.S. thesis. Univ. Minnesota, St. Paul. Mevel, D. J. 1973. Sex and age ratios of sharp-tailed grouse and hungarian partridge. Sask. Dep. Tourism and Renewable Resour. 15 pp. Miller, G. R., D. Jenkins, and A. Watson. 1966. Heather performance and red grouse populations. I. Visual estimates of heather performance. J. Appl. Ecol. 3:313-26. Miller, G. R. and A. Watson. 1978. Territories and the food plant of individual red grouse. I. Territory size, number of mates, and brood size compared with the abundance, production and diversity of heather. J. Anim. Ecol. 47:293-305. Miller, G. R., A. Watson, andD. Jenkins. 1970. Responses of red grouse populations to experimental improvement of their food. Pp. 323-35 in A. Watson, ed. Animal populations in relation to their food resources. Blackwell Sci. Publ., Oxford and Edinburgh. Miller, H. W. and D. H. Johnson. 1978. Interpreting the results of nesting studies. J. Wildl. Manage. 42:471-76. Mitchell, D. L. 1957. Dummy nest study. N.D. Dep. Game and Fish Fed. Aid Wildl. Rest. Proj. W-35-R-3. 9 pp. Mitchell, J. and B. Vodehnal. 1982. Surveys and management of Prairie grouse. S.D. Fed. Aid Wildl. Rest. Proj W-15-R-38. 15 pp. Mohr, C. O. 1947. Table of equivalent populations of North American small mammals. Am. Midi. Nat. 37:233-49. Morse, D. H. 1970. Ecological aspects of some mixed species foraging flocks of birds. Ecol. Mongr. 40:119-68. Mortensen, A. and A. S. Blix. 1985. Seasonal changes in the effects of starvation on metabolic rate and regulation of body weight in Svalbard ptarmigan. Ornis Scand. 16:20-24. Mortensen, A., E. S. Nord0y, and A. S. Blix. 1985. Seasonal changes in the body composition of the Norwegian rock ptarmigan Lagopus mutus. Ornis Scand. 16:25-28. Moss, R. 1972. Social organization of willow ptarmigan on their breeding grounds in interior Alaska. Condor 74:144-51. Moss, R. 1973. The digestion and intake of winter foods by wild ptarmigan in Alaska. Condor 75:293-300. Moss, R. 1983. Gut size, body weight, and digestion of winter foods by grouse and ptarmigan. Condor 85:185-93. Moss, R. and I. Hanssen. 1980. Grouse nutrition. Nutrition abstracts and reviews, Ser. B. 50:555-67. Moss, R. and J. Oswald. 1985. Population dynamics of Capercaillie in a north-east Scottish glen. Ornis Scand. 16:229-38. Moss, R., P. Rothery, and I. B. Trenholm. 1985. The inheritance of social dominance rank in red grouse (Lagopus lagopus scoticus). Aggress. Behav. 11:253-59. Moss, R. and A. Watson. 1980. Inherent changes in the aggressive behaviour of a fluctuating red grouse Lagopus lagopus scoticus population. Ardea 68:113-19.
760
REFERENCES
Moss, R. and A. Watson. 1982. Heritability of egg size, hatch weight, body weight, and viability in red grouse (Lagopus lagopus scoticus). Auk 99:683-86. Moss, R. and A. Watson. 1985. Adaptive value of spacing behaviour in population cycles of red grouse and other animals. Pp. 275-94 in R. M. Sibly and R. H. Smith, eds. Behavioural ecology. Blackwell, Oxford. Moss, R., A. Watson, and J. Ollason. 1982a. Animal population dynamics. Chapman and Hall, London. Moss, R., A. Watson, and R. Parr. 1974. A role of nutrition in the population dynamics of some game birds (Tetraonidae). Trans. Int. Congr. Game Biol. 11:193-201. Moss, R., A. Watson, and R. Parr. 1975. Maternal nutrition and breeding success in red grouse (Lagopus lagopus scoticus). J. Anim. Ecol. 44:233-44. Moss, R., A. Watson, R. A. Parr, D. Watt, and W. Glennie. 1978. Egg size and other factors in relation to early mortality in red grouse (Lagopus lagopus scoticus) chicks. Ibis 120:407. Moss, R., A. Watson, and P. Rothery. 1984. Inherent changes in body size, viability and behaviour of a fluctuating red grouse (Lagopus lagopus scoticus) population. J. Anim. Ecol. 53:171-89. Moss, R., A. Watson, P. Rothery, and W. W. Glennie. 1981. Clutch size, egg size, hatch weight and laying date in relation to early mortality in red grouse Lagopus lagopus scoticus chicks. Ibis 123:450-62. Moss, R., A. Watson, P. Rothery, and W. Glennie. 1982b. Inheritance of dominance and aggressiveness in captive red grouse Lagopus lagopus scoticus. Aggress. Behav. 8:1-18. Mossop, D. H. 1971. A relation between aggressive behavior and population dynamics in blue grouse. M.Sc. thesis. Univ. British Columbia, Vancouver. Moyles, D. L. J. 1977. A study of territory establishment by and movements of male sharp-tailed grouse (Pedioecetesphasianellus) relative to the arena. M.Sc. thesis. Univ. Alberta, Edmonton. Moyles, D. L. J. 1981. Seasonal and daily use of plant communities by sharp-tailed grouse (Pedioecetes phasianellus) in the parklands of Alberta. Can. Field-Nat. 95:287-91. Moyles, D. L. J. and D. A. Boag. 1981. Where, when, and how male sharp-tailed grouse establish territories on arenas. Can. J. Zool. 59:1576-81. Muehrcke, J. P. and C. M. Kirkpatrick. 1970. Observations of ecology and behavior of Indiana ruffed grouse. Proc. Indiana Acad. Sci. 79:177-86. Mueller, H. C., D. D. Berger, and G. Allez. 1977. The periodic invasions of goshawks. Auk 94:652-63. Murphy, G. I. 1968. Pattern in life history and the environment. Am. Nat. 102:391-403. Mussehl, T. W. 1963. Blue grouse brood cover selection and land-use implications. J. Wildl. Manage. 27:546-55. Mussehl, T. W. and P. Schladweiler. 1969. Forest grouse and experimental spruce budworm insecticide studies. Mont. Fish and Game Dep. Tech. Bull 4. 53 pp. Myers, J. H. and C. J. Krebs. 1971. Genetic, behavioral, and reproductive attributes of dispersing field voles Microtus pennsylvanicus and Microtus ochrogaster. Ecol. Monogr. 41:53-78. Myhre, K., M. Cabanac, and C. Myhre. 1975. Thermoregulatory behavior and body temperature in chicks of willow grouse (Lagopus lagopus lagopus). Poultry Sci. 54:1174-79. Myrberget, S. 1970a. Fra et h0nsehaukrede i Troms. Sterna 9:5-8. Myrberget, S. 1970b. On the part played by predation in short-term variations in the population of willow grouse, Lagopus lagopus, in Norway. Trans. Int. Congr. Game Biol. 9:458-64. Myrberget, S. 1970c. Reproductive success of young and old willow ptarmigan, Lagopus lagopus lagopus, in north Norway. Finnish Game Res. 30:169-72. Myrberget, S. 1972. Fluctuations in a north Norwegian population of willow grouse. Proc. Int. Ornithol. Congr. 15:107-20. Myrberget, S. 1974a. Vaeret som synkroniserende faktor for smagnagersvingninger. Medd. Stat. Viltunders. 2:1-16.
REFERENCES
761
Myrberget, S. 1974b. Variations in the production of the willow grouse Lagopus lagopus (L.) in Norway, 1963-1972. Ornis Scand. 5:163-72. Myrberget, S. 1975a. Vegetasjonskartlegging av Tran0y i Senja. Trondheim (DVF). Myrberget, S. 1975b. Age distribution, mortality and migration of willow grouse, Lagopus lagopus, on Senja, north Norway. Astarte 8:29-35. Myrberget, S. 1975c. Norske studier omkring 1930 over lirypas reproduksjon. Sterna 14:51-64. Myrberget, S. 1975d. Gjenfangstrater ved to merkemetododer for lirype. Sterna 14:29-32. Myrberget, S. 1976a. En censusmetode for hekkende ryper. Fauna 29:79-85. Myrberget, S. 1976b. Hunting mortality, migration, and age composition of Norwegian willow grouse (Lagopus lagopus). Norwegian J. Zool. 24:47-52. Myrberget, S. 1976c. Lirypas reirhabitat. Medd. norsk Viltforsk. 3:1-30. Myrberget, S. 1977. Size and shape of eggs of willow grouse Lagopus lagopus. Ornis Scand. 8:39-46. Myrberget, S. 1980. Eggpredasjon hos lirype pa ei nord-norsk 0y. Viltrapport 10:108-13. Myrberget, S. 1981. Diet of willow grouse Lagopus lagopus chicks on a coastal island. Fauna norv. Ser. C, Cinclus 4:58-63. Myrberget, S. 1982a. Bestandsvariasjoner hos lirype i Norge 1932-1971. Medd. norsk Viltforsk. 3:1-31. Myrberget, S. 1982b. Fluctuations in microtine populations in an island area in northern Norway 1958-1981. Fauna norv. Ser. A 3:7-11. Myrberget, S. 1982c. Krakebestanden pa ei nordnorsk 0y 1960-1981. Fauna 35:8-10. Myrberget, S. 1983a. Vacant habitats during a decline in a population of willow grouse Lagopus lagopus. Fauna norv. Ser. C, Cinclus 6:1-7. Myrberget S. 1983b. Desertion of nests by willow grouse Lagopus lagopus. Fauna norv. Ser. C, Cinclus 6:109-13. Myrberget, S. 1984a. Population cycles of willow grouse Lagopus lagopus on an island in northern Norway. Fauna norv. Ser. C, Cinclus 7:46-56. Myrberget, S. 1984b. Population dynamics of willow grouse Lagopus lagopus on an island in Norway. Fauna norv. Ser. C, Cinclus 7:95-105. Myrberget, S. 1985. Egg predation in an island population of willow grouse Lagopus lagopus. Fauna norv. Ser. C, Cinclus 8:82-87. Myrberget, S. (in press). Cycles in an island population of willow grouse Lagopus lagopus 1960-1980. Myrberget, S. and A. Aandahl. 1976. Jordugle og haukugle sompredatorer pa lirype. Fauna 29:93-94. Myrberget, S., K. E. Erikstad, R. Blom, and T. K. Spids0. 1985. Estimation of nesting success and frequency of re-laying in willow grouse Lagopus lagopus. Ornis Fenn. 62:9-12. Myrberget, S., K. E. Erikstad, and T. K. Spids0. 1977. Variations from year to year in growth rates of willow grouse chicks. Astarte 10:9-14. Myrberget, S., K. E. Erikstad, and T. K. Spids0. 1982. Demography of a North Norwegian population of willow grouse. Trans. Int. Congr. Game Biol. 14:47-52. Myrberget, S., P. Krigsvoll, and K. E. Erikstad. 1976. Et fors0k pa a desimere en Krakebestand. Sterna 15:127-32. Myrberget, S., 0. Olsvik, and T. Saether. 1981. On a "crash" in a population of willow grouse Lagopus lagopus. Fauna norv. Ser. C, Cinclus 4:64-68. Myrberget, S. and H-J Skar. 1976. Fat and calorific content of willow grouse in autumn and winter. Norwegian J. Zool. 24:41-45. National Academy of Sciences - National Research Council. 1982. Range research: basic problems and techniques. Publ. No. 890 NAS-NRC. Washington, D.C. 341 pp. Neave, D. J. 1967. The population dynamics of ruffed grouse (Bonasa umbellus L.) in central New Brunswick. M.S. thesis. Univ. New Brunswick, Fredericton.
762
REFERENCES
Neave, D. J. and B. S. Wright. 1969. The effects of weather and DDT spraying on a ruifed grouse population. J. Wildl. Manage. 33:1015-20. Nelson, O. C. 1955. A field study of sage grouse in southeastern Oregon with special reference to reproduction and survival. M.S. thesis. Oregon State Coll., Corvallis. Neu, C. W., C. R. Byers and J. M. Peek. 1974. A technique for analysis of utilization-availability data. J. Wildl. Manage. 38:541-45. Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, S.D. 399 pp. Nice, M. M. 1942. Analysis of losses in the nesting of birds (review). Bird Banding 13:90. Nicholls, T. H. and D. W. Warner. 1972. Barred owl habitat use as determined by radio-telemetry. J. Wildl. Manage. 36:213-24. Nicholson, A. J. 1933. The balance of animal populations. J. Anim. Ecol. 2:132-78. Niebet, R. M. and W. S. C. Gurney. 1982. Modelling fluctuating populations. John Wiley and Sons, New York. Nielsen, G. F. 1973. The status of prairie chicken in Minnesota: land acquisition programs. Pp. 86-88 in W. D. Svedarsky and T. Wolfe, eds. Proc. of the prairie chicken in Minnesota. Univ. Minnesota, Crookston. Nielsen, L. S. and C. A. Yde. 1982. The effects of rest-rotation grazing on the distribution of sharptailed grouse. Pp. 147-65 in J. M. Peek and P. D. Dalke ed. Wildlife-Live stock relationships. Dept. Wildl. Res., College of Forestry, Wildlife and Range science, Univ. Idaho, Moscow, Idaho. Noble, G. K. and W. Vogt. 1935. An experimental study of sex recognition in birds. Auk 52:278-86. Nolan, V. 1963. Reproductive success of birds in a deciduous shrub habitat. Ecology 44:305-13. Nugent, D. P. and D. A. Boag. 1982. Communication among territorial female spruce grouse. Can. J. Zool. 60:2624-32. O'Donald, P. 1983. Sexual selection by female choice. Pp. 53-66 in P. Bateson, ed. Mate choice. Cambridge Univ. Press, Cambridge. O'Donald, P. 1977. Mating preferences and sexual selection in the arctic skua. II. Behavioral mechanisms of the mating preferences. Heredity 39:111-29. Olpinski, S. C. 1980. The significance of age differential courtship and aggression behaviour among male Canada spruce grouse Canachites canadensis canace (L.). B.Sc. honours thesis. Univ. New Brunswick, Fredericton. Olstad, O. 1932. Unders0kelser over Lirypens Fortplantings for hold. Nytt Mag. Naturv. 51:1-71. Olsvik, 0. andO. Olsen. 1980. Ryper, bestandstetthet og eggtap. Jakt-Fiske-Friluftsliv 109:18-19. O'Malley, B. W. 1977. Hormonal control of gene expression in reproductive tissue, in R. O. Greep and M. A. Koblinsky, eds. Frontiers in reproductive and fertility control. Part 2. MIT Press, Cambridge, Mass. Orians, G. H. 1969. On the evolution of mating systems in birds and mammals. Am. Nat. 103:589-603. Oring, L. W. 1982. Avian mating systems. Pp. 1-92 in D. S. Farner, J. R. King and K. C. Parkes, eds. Avian biology. Vol. 6. Chiron Press, New York. Owen, R. E. and R. C. Plowright. 1980. Abdominal pile color dimorphism in the bumble bee, Bombus melanopygus. J. Hered. 71:241-47. Page, G. W., L. E. Stenzei, D. W. Winkler, and C. W. Swarth. 1983. Spacing out at Mono Lake: breeding success, nest density, and predation in the snowy plover. Auk 100:13-24. Palmer, R. S. 1949. Maine birds. Vol. 102. Mus. Comp. Biol. Bull. Cambridge, Mass. 656 pp. Palmer, W. L. 1956. Ruffed grouse population studies on hunted and unhunted areas. Trans. N. Am. Wildl. Conf. 21:338-44. Palmer, W. L. 1963. Ruffed grouse drumming sites in northern Michigan. J. Wildl. Manage. 37:656-63. Parker, H. 1981a. Duration of fertility in willow ptarmigan hens after separation from the cock. Ornis Scand. 12:186-87.
REFERENCES
763
Parker, H. 1981b. Renesting biology of Norwegian willow ptarmigan. J. Wildl. Manage. 45:858-64. Parsons, P. A. 1983. The evolutionary biology of colonizing species. Cambridge University Press, Cambridge. 262 pp. Partridge, L. 1980. Mate choice increases a component of offspring fitness in fruit flies. Nature 283:290-91. Patterson, I. J. 1981. Territorial behaviour and the limitation of population density. Pp. 53-62 in H. Klomp and J. W. Woldendorp, eds. The integrated study of bird populations. North-Holland Publ., Amsterdam. Patterson, R. L. 1952. The sage grouse in Wyoming. Sage Books, Denver. 341 pp. Pearson, O. P. 1966. The prey of carnivores during one cycle of mouse abundance. J. Anim. Ecol. 35:217-33. Pedersen, H. C. 1984. Territory size, mating status, and individual survival of males in a fluctuating population of willow ptarmigan. Ornis Scand. 15:197-203. Pedersen, H. C. and J. B. Steen. 1979. Behavioural thermoregulation in willow ptarmigan chicks Lagopus lagopus. Ornis Scand. 10:17-21. Pedersen, H. C. and J. B. Steen. 1985. Parental care and chick production in a fluctuating population of willow ptarmigan. Ornis Scand. 16:270-76. Pedersen, H. C., J. B. Steen, and R. Andersen. 1983. Social organization and territorial behavior in a willow ptarmigan population. Ornis Scand. 14:263-72. Pepper, G. W. 1972. The ecology of sharp-tailed grouse during spring and summer in the aspen parklands of Saskatchewan. Sask. Dep. Nat. Resour. Wildl. Rep. 1. 56 pp. Perrins, C. M. 1977. The role of predation in the evolution of clutch size. Pp. 181-91 in B. Stonehouse and C. Perrins, eds. Evolutionary ecology. Univ. Park Press, Baltimore, Md. Peters, S. S. 1958. Food habits of the Newfoundland willow ptarmigan. J. Wildl. Manage. 22:384-94. Petersen, B. E. 1980. Breeding and nesting ecology of female sage grouse on North Park, Colorado. M.S. thesis. Colorado State Univ., Fort Collins. Petersen, L. 1979. Ecology of great horned owls and red-tailed hawks in southeastern Wisconsin. Wis. Dep. Nat. Resour. Tech. Bull. 111. 63 pp. Petraborg, W. H., E. G. Wellein, and V. E. Gunvalson. 1953. Roadside drumming counts—a spring census method for ruffed grouse. J. Wildl. Manage. 17:292-95. Petrides, G. A. 1949. Viewpoints on the analysis of open season and age ratios. Trans. N. Am. Wildl. Conf. 14:391-410. Phillips, J. and F. Aalerud. 1980. Rypas territoriale atferd. Jakt-Fiske-Friluftsliv 109:48-50. Peilou, E. C. 1977. Mathematical ecology. John Wiley and Sons, New York. 385 pp. Pietz, P. J. and J. R. Tester. 1979. Utilization of Minnesota peatland habitats by snowshoe hare, white-tailed deer, spruce grouse, and ruffed grouse. Final Tech. Rep., Minn. Dep. Nat. Resour. 80 pp. Pietz, P. J. and J. R. Tester. 1982. Habitat selection by sympatric spruce and ruffed grouse in north central Minnesota. J. Wildl. Manage. 46:391-403. Pils, C. M. and M. A. Martin. 1978. Population dynamics, predator-prey relationships and management of the red fox in Wisconsin. Wis. Dep. Nat. Resour. Tech. Bull. 105. 56 pp. Pils, C. M., M. A. Martin, andE. L. Lange. 1981. Harvest, age structure, survivorship, and productivity of red foxes in Wisconsin, 1975-78. Wis. Dep. Nat. Resour. Tech. Bull. 125. 21 pp. Pimentel, D. 1963. Natural population regulation and inter-species evolution. Proc. Int. Congr. Zool. 16:329-36. Pitelka, F. A. 1958. Some aspects of population structure in the short-term cycle of the brown lemming in northern Alaska. Cold Spring Harbour Symp. Quant. Biol. 22:237-51. Pitelka, F. A. and A. M. Schultz. 1964. The nutrient recovery hypothesis for arctic microtine cycles. Pp. 55-68 in D. Crisp, ed. Grazing in terrestrial and marine environments. Blackwell Sci. Publ., Oxford.
764
REFERENCES
Platt, J. B. 1976. Winter activity of gyrfalcons at their nest sites in arctic Canada. Can. Field-Nat. 90:338-45. Platt, J. B. and E. lull. 1976. A study of wintering and nesting gyrfalcons on the Yukon north slope during 1975 with emphasis on their behavior during experimental overflights by helicopter. Chapter 2 in W. W. H. Gunn, R. E. Schweinsburg, C. E. lull, and T. D. Wright, eds. Ornithological studies conducted in the area of the proposed gas pipeline route. Northwest Territory, Yukon and Alaska Arctic Gas Biol. Rep. Ser. Vol. 30. Popper, K. R. 1959. The logic of scientific discovery. Basic Books, New York. 479 pp. Popper, K. R. 1965. Normal science and its dangers. Pp. 51-58 in I. Lakatos and A. Musgrave, eds. Criticism and the growth of knowledge. Cambridge Univ. Press, Cambridge. Porath, W. R. and P. A. Vohs, Jr. 1972. Population ecology of ruffed grouse in northeastern Iowa. J. Wildl. Manage. 36:793-802. Porter, W. F. 1977. Home range dynamics of wild turkeys in southeastern Minnesota. J. Wildl. Manage. 41:434-37. Pot, W., W. Van Delden, and J. P. Kruijt. 1980. Genetic differences in mating success and the maintainance of the alcohol dehydrogenase polymorphism in Drosophila melanogaster. Behav. Genet. 10:43-58. Potts, G. R. 1980. The effects of modern agriculture, nest predation and game management on the population ecology of partridges (Perdix perdix andAlectoris rufa). Pp. 1-79 in A. MacFayden, ed. Advances in ecological research. Vol. 2. Academic Press, London. Potts, G. R., S. C. Tapper, and P. J. Hudson. 1984. Population fluctuations in red grouse: analysis of bag records and a simulation model. J. Anim. Ecol. 53:21-36. Poulsen, E. T. 1979. Model for population regulation with density and frequency dependent selection. J. Math. Biol. 13:305-24. Powell, G. V. N. 1974. Experimental analysis of the social value of flockings by starlings (Sturnus vulgaris) in relation to predation and foraging. Anim. Behav. 22:501-5. Powell, R. A. 1982. Evolution of black-tipped tails in weasels: predator confusion. Am. Nat. 119:126-31. Prawdzik, T. R. 1963. Ruffed grouse escaping a cooper's hawk. J. Wildl. Manage. 27:639-42. Prout, T. 1980. Some relationships between density-independent selection and density-dependent population growth. Pp. 1-68 in M. K. Hecht, W. C. Steele and B. Wallace, eds. Evolutionary biology. Vol. 13. Plenum, New York. Pulliainen, E. 1975a. Choice of prey by a pair of gyrfalcons Falco rusticolus during the nesting period in Forest-Lapland. Ornis Fenn. 52:19-22. Pulliainen, E. 1975b. Structure of two willow grouse (Lagopus lagopus) populations in Finnish Fjeld Lapland in the winters of 1972-1974. Ann. Zool. Fenn. 12:263-67. Pulliainen, E. 1978. Behaviour of a willow grouse Lagopus lagopus lagopus at the nest. Ornis Fenn. 55:141-48. Pulliam, H. R. and G. C. Millikan. 1982. Social organization in the nonreproductive season. Pp. 169-97 in D. S. Farner, J. R. King, and K. C. Parkes, eds. Avian biology. Vol. 6. Academic Press, New York. Pyke, G. H., H. R. Pulliam, and E. L. Charnov. 1977. Optimal foraging: a selective review of theory and tests. Q. Rev. Biol. 52:137-54. Quinn, N. W. S. andD. M. Keppie. 1981. Factors influencing growth of juvenile spruce grouse. Can. J. Zool. 59:1790-95. Ramharter, B. G. 1976. Habitat selection and movements of sharp-tailed grouse (Pedioecetes phasianellus) hens during the nesting and brood rearing periods in a fire maintained brush prairie. Ph.D. thesis. Univ. Minnesota, St. Paul. Rand, A. L. 1947. Clutch size in the spruce grouse and theoretical considerations of some factors affecting clutch size. Can. Field-Nat. 61:127-30.
REFERENCES
765
Ransom, A. B. 1965. Observations of a ruffed grouse decline. Can. Field-Nat. 79:128-30. Rasmussen, D. I. and L. A. Griner. 1938. Life history and management studies of the sage grouse in Utah, with special reference to nesting and feeding habits. Trans N. Am. Wildl. Conf. 3:852-64. Redfield, J. A. 1972. Demography and genetics in colonizing populations of blue grouse (Dendragapus obscurus). Ph.D. thesis. Univ. Alberta, Edmonton. Redfield, J. A. 1973a. Demography and genetics in colonizing populations of blue grouse (Dendragapus obscurus). Evolution 27:576-92. Redfield, J. A. 1973b. Variations in weight of blue grouse (Dendragapus obscurus). Condor 75:312-21. Redfield, J. A. 1974. Genetics and selection at the Ng locus in blue grouse (Dendragapus obscurus). Heredity 33:69-78. Redfield, J. A. 1975. Comparative demography of increasing and stable populations of blue grouse (Dendragapus obscurus). Can. J. Zool. 53:1-11. Redfield, J. A. 1978. Growth of juvenile blue grouse, Dendragapus obscurus. Ibis 120:55-61. Redfield, J. A., F. C. Zwickel, and J. F. Bendell. 1970. Eifects of fire on numbers of blue grouse. Proc. Annu. Tall Timbers Fire Ecol. Conf. 10:63-83. Redfield, J. A., F. C. Zwickel, J. F. Bendell, and A. T. Bergerud. 1972. Temporal and spatial patterns of allele and genotype frequencies at the Ng locus in blue grouse (Dendragapus obscurus). Can. J. Zool. 50:1657-62. Redmond, G. W., D. M. Keppie, and P. W. Herzog. 1982. Vegetative structure, concealment, and success at nests of two races of spruce grouse. Can. J. Zool. 60:670-75. Reid, R. and C. G. Gannon. 1928. Natural history notes on the journals of Alexander Henry. N.D. Hist. Q. 2:168-201. Reynolds, R. T., E. C. Meslow, and H. M. Wight. 1982. Nesting habitat of coexisting Accipiter in Oregon. J. Wildl. Manage. 46:124-38. Rice, L. and A. Carter. 1975. Evaluation of South Dakota grassland management practices as they affect prairie chicken populations, 1973-74 and 1974-75. S.D. Dep. Game, Fish and Parks. Proj. W-75-R-16, 17, Jobs I-V. 16 pp. Rice, L. and A. Carter. 1976. Evaluation of South Dakota grassland management practices as they aifect prairie chicken populations, 1975-1976. S.D. Dep. Game, Fish and Parks. Proj. W-75-R18. Jobs I-V. 17 pp. Rice, L. and A. Carter. 1977. Evaluation of South Dakota grassland management practices as they affect prairie chicken populations 1976-77. S.D. Dep. Game, Fish and Parks. Proj. W-75-R-19, Jobs I-V. 16 pp. Richards, S. H. and R. L. Hine. 1953. Wisconsin fox populations. Wis. Conserv. Dep. Final Rep., Tech. Wildl. Bull. 6. 78 pp. Ricker, W. E. 1954. Effects of compensatory mortality upon population abundance. J. Wildl. Manage. 18:45-51. Ricklefs, R. E. 1969. An analysis of nesting mortality in birds. Smithsonian Contrib. Zool. 9:1-48. Ricklefs, R. E. 1970. Clutch size in birds: outcome of opposing predator and prey adaptations. Science 168:599-600. Ricklefs, R. E. 1972. Breeding biology of birds. Pp. 366-435 in D. S. Farner, ed. Symposium on breeding behavior and reproductive physiology in birds. Natl. Acad. Sci., Washington, D.C. Ricklefs, R. E. 1977. On the evolution of reproductive strategies in birds: reproductive effort. Am. Nat. 111:453-61. Riewe, J. W. 1975. Mammalian carnivores. Pp. 493-501 in L. C. Bliss, ed. Truelove Lowland, Devon Island, Canada: a high arctic ecosystem. Univ. Alberta Press, Edmonton. Riley, T. Z. 1978. Nesting and brood-rearing habitat of lesser prairie chickens. M.S. thesis. New Mexico State Univ., Las Cruces.
766
REFERENCES
Rippin, A. B. and D. A. Boag. 1974a. Recruitment to populations of male sharp-tailed grouse. J. Wildl. Manage. 38:616-21. Rippin, A. B. and D. A. Boag. 1974b. Spatial organization among male sharp-tailed grouse on arenas. Can. J. Zool. 52:591-97. Ritcey, R. W. and R. Y. Edwards. 1963. Grouse abundance and June temperatures in Wells Gray Park, British Columbia. J. Wildl. Manage. 27:604-6. Robel, R. J. 1964. Quantitative indices to activity and territoriality of booming Tympanuchus cupido pinnatus in Kansas. Trans. Kans. Acad. Sci. 67:702-12. Robel, R. J. 1966. Booming territory size and mating success of the greater prairie chicken (Tympanuchus cupido pinnatus). Anim. Behav. 14:328-31. Robel, R. J. 1970. Possible role of behavior in regulating greater prairie chicken populations. J. Wildl. Manage. 34:306-12. Robel, R. J. 1972. Possible function of the lek in regulating tetraonid populations. Proc. Int. Ornithol. Congr. 15:121-33. Robel, R. J. and W. B. Ballard, Jr. 1974. Lek social organization and reproductive success in the greater prairie chicken. Am. Zool. 14:121-28. Robel, R. J., J. N. Briggs, J. J. Cebula, N. J. Silvy, C. E. Viers, and P. G. Watt. 1970a. Greater prairie chicken ranges, movements, and habitat usage in Kansas. J. Wildl. Manage. 34:286-306. Robel, R. J., J. N. Briggs, A. D. Dayton, and L. C. Hulbert. 1970b. Relationships between visual obstruction measurements and weight of grassland vegetation. J. Range Manage. 23:295-97. Robel, R. J., F. R. Henderson, and W. W. Jackson. 1972. Some sharp-tailed grouse population statistics from South Dakota. J. Wildl. Manage. 36:87-98. Robertson, K. 1979. Surveys and management of prairie grouse. Nebr. Game and Parks Comm. Proj. W-15-R-35, Work Plan E-78. 53 pp. Robinson, W. L. 1969. Habitat selection by spruce grouse in northern Michigan. J. Wildl. Manage. 33:113-20. Robinson, W. L. 1980. Fool hen: the spruce grouse on the yellow dog plains. Univ. Wisconsin Press, Madison. 221 pp. Rodgers, R. D. 1980. Ecological relationships of ruffed grouse in southwestern Wisconsin. Wis. Acad. Sci., Arts, and Letters. 68:97-105. Roe, F. G. 1951. The North American buffalo. Univ. Toronto Press, Toronto. 957 pp. Roff, D. A. 1974. Spatial heterogeneity and the persistence of populations. Oecologia 15:245-58. Roff, D. A. 1975. Population stability and the evolution of dispersal in a heterogeneous environment. Oecologia 15:259-75. Rogers, G. E. 1964. Sage grouse investigations in Colorado. Colo. Game, Fish and Parks Dep. Tech. Publ. 16. 132 pp. Rohwer, S., S. D. Fretwell, and D. M. Niles. 1980. Delayed maturation in passerine plumages and the deceptive acquisition of resources. Am. Nat. 115:400-437. Romesburg, H. C. 1981. Wildlife science: gaining reliable knowledge. J. Wildl. Manage. 45:293-313. Rose, M. R., M. L. Dorey, A. M. Coye, and P. M. Service. 1984. The morphology of postponed senescence in Drosophila melanogaster. Can. J. Zool. 62:1576-80. Rothenmaier, D. 1979. Sage grouse reproductive ecology: breeding season movements, strutting ground attendance and site characteristics, and nesting. M.S. thesis. Univ. Wyoming, Laramie. Roughgarden, J. 1971. Density-dependent natural selection. Ecology 52:453-68. Rubenstein, D. I. 1980. On the evolution of alternative mating strategies. Pp. 65-100. in J. E. R. Staddon, ed. Limits to action: the allocation of individual behavior. Academic Press, New York. Rusch, D. H. 1976. The wildlife cycle in Manitoba. Manit. Dep. Mines, Resour. and Wildl. Manage. Wildl. Branch Inf. Ser. 11. 14 pp. Rusch, D. H., M. M. Gillespie, and D. I. McKay. 1978. Decline of a ruffed grouse population in Manitoba. Can. Field-Nat. 92:123-27.
REFERENCES
767
Rusch, D. H. and L. B. Keith. 197la. Ruffed grouse-vegetation relationships in central Alberta. J. Wildl. Manage. 35:417-29. Rusch, D. H. and L. B. Keith. 1971b. Seasonal and annual trends in numbers of Alberta ruffed grouse. J. Wildl. Manage. 35:803-22. Rusch, D. H., L. B. Keith, P. D. Doerr, and C. A. Fischer. 1984. Demography of a ruffed grouse population in central Alberta. Unpubl. report, Dept. of Wildlife Ecology, Univ. Wisconsin, Madison. Rusch, D. H., E. C. Meslow, P. D. Doerr, and L. B. Keith. 1972. Response of great horned owl populations to changing prey densities. J. Wildl. Manage. 36:282-296. Sadlier, R. M. F. S. 1965. The relationship between agonistic behavior and population changes in the deer mouse, Peromyscus maniculatus (Wagner). J. Anim. Ecol. 34:331-52. Safriel, U. N. 1975. On the significance of clutch size in nidifugous birds. Ecology 56:703-8. Salmon, W. C. 1979. Informal analytical approaches to the philosophy of science. Pp. 3-15 in Current research in philosophy of science. Edwards Bros., Ann Arbor, Mich. Salo, L. J. 1978. Characteristics of ruffed grouse drumming sites in western Washington and their relevance to management. Ann. Zool. Fenn. 15:261-78. Salomonsen, F. 1939. Moults and sequence of plumages in the rock ptarmigan (Lagopus mutus [Montin]). Videnskabelige Meddelelser Fra Dansk Naturhistorisk Forening 103:1-491. Salomonsen, F. 1972. Presidential address: zoogeographical and ecological problems in arctic birds. Proc. Int. Ornithol. Congr. 15:25-77. Sargeant, A. B. 1972. Red fox spatial characteristics in relation to waterfowl predation. J. Wildl. Manage. 36:225-36. Sargeant, A. B. 1983. A case history of a dynamic resource—the red fox. Pp. 127-37 inG. C. Sanderson, ed. Midwest Furbearer Manage. Symp. The Wildl. Society. Sargeant, A. B., S. H. Allen, and R. T. Eberhardt. 1984. Red fox predation of breeding ducks in midcontinent, North America. Wildl. Mongr. 89. 41 pp. Sargeant, A. B. and L. E. Eberhardt. 1975. Death feigning by ducks in response to red foxes (Vulpes fulva) Am. Midi. Nat. 94:108-19. Sassaman, C. and R. Garthwaite. 1980. Genetics of a pigment polymorphism in the isopod Porcellio dilatatus. J. Hered. 71:158-60. Savory, C. J. 1975. Seasonal variation in the food intake of captive red grouse. Br. Poult. Sci. 16:471-79. Savory, C. J. 1977. The food of red grouse chicks Lagopus I. scoticus. Ibis 119:1-9. Savory, C. J. 1978. Food consumption of red grouse in relation to age and productivity of heather. J. Anim. Ecol. 47:269-82. Schemnitz, S. D. 1970. Fall and winter feeding activities and behavior of ruffed grouse in Maine. Proc. Northeast Sect. Wildl. Society Annu. Conf. 27:127-40. Schilcher, F. V. 1977. A mutation which changes courtship song in Drosophila melanogaster. Behav. Genet. 7:251-59. Schiller, R. J. 1973. Reproductive ecology of female sharp-tailed grouse (Pedioecetes phasianellus) and its relationship to early plant succession in northwestern Minnesota. Ph.D. thesis. Univ. Minnesota, St. Paul. Schladweiler, P. 1965. Movements and activities of ruffed grouse, Bonasa umbellus (L.), during the summer period. Ph.D. thesis. Univ. Minnesota, St. Paul. Schladweiler, P. 1968. Feeding behavior of incubating ruffed grouse females. J. Wildl. Manage. 32:426-28. Schmidt, F. J. W. 1936. Winter food of the sharp-tailed grouse and pinnated grouse in Wisconsin. Wilson Bull. 48:186-230. Schmidt, R. K., Jr. 1969. Behavior of white-tailed ptarmigan in Colorado. M.S. thesis. Colorado State Univ., Fort Collins.
768
REFERENCES
Schofield, R. D. 1960. A thousand miles of fox trails in Michigan's ruffed grouse range. J. Wildl. Manage. 24:432-34. Schorger, A. W. 1944. The prairie chicken and sharp-tailed grouse in early Wisconsin. Trans. Wis. Acad. Sci., Arts and Letters 35:1-59. Schorger, A. W. 1945. The ruffed grouse in early Wisconsin. Trans. Wis. Acad. Sci., Arts and Letters 37:35-90. Schranck, B. W. 1972. Waterfowl nest cover and some predation relationships. J. Wildl. Manage. 36:182-86. Schroeder, M. A. 1985. Behavioral differences of female spruce grouse undertaking short and long migrations. Condor 87:281-86. Schultz, A. M. 1964. The nutrient-recovery hypothesis for arctic microtine cycles. II. Pp. 57-68 in D. Crisp, ed. Grazing in terrestrial and marine environments. Blackwell Sci. Publ., Oxford. Schulz, J. W. 1981. Ruffed grouse population data. N.D. State Game and Fish Dep. Proj. W-67-R21, JobsB-XI-1, 2, 3. 15pp. Schulz, J. W. 1982. Examination of ruffed grouse populations and habitat requirements on the Turtle Mountains. N.D. State Game and Fish Dep. Proj. W-67-R-22, Jobs B-XIV-2, 3. 22 pp. Schulz, J. W. 1983. Ruffed grouse population data. N.D. State Game and Fish Dep. Proj. W-67-R23, JobsB-XI-1, 2, 3, 4. 19 pp. Schwartz, C. W. 1944. The prairie chicken in Missouri. Mo. Conserv. Comm. 179 pp. Schwartz, C. W. 1945. The ecology of the prairie chicken in Missouri. Univ. Mo. Studies 20:1-99. Scott, J. W. 1942. Mating behavior of sage grouse. Auk 59:477-98. Scott, T. G. ed. 1981. Fisheries and wildlife research 191. U.S. Fish and Wildl. Serv. Denver, Colo. 191 pp. Scott, W. E. 1956. The wildlife resources of Wisconsin. The Natural Resources of Wisconsin, December. Pp. 32-34. Searcy, W. A. and K. Yasukawa. 1981. Does the "sexy son" hypothesis apply to mate choice in redwinged blackbirds? Am. Nat. 117:343-48. Seiskari, P. 1962. On the winter ecology of the capercaillie, Tetrao urogallus, and the black grouse, Lyrurus tetrix, in Finland. Pap. on Game Res. 22:1-119. Selander, R. K. 1965. On mating systems and sexual selection. Am. Nat. 99:129-41. Selander, R. K. 1972. Sexual selection and dimorphism in birds. Pp. 180-230 in B. Campbell, ed. Sexual selection and the descent of man. 1871-1971. Aldine Press, Chicago. Sell, D. L. 1979. Spring and summer movements and habitat use by lesser prairie chicken females in Yoakum County, Texas. M.S. thesis. Texas Tech. Univ., Lubbock. Seton, E. T. 1953. Lives of game animals. Vol. 3. Charles T. Branford, Boston (first published in 1904). 780 pp. Severson, R. 1982. Ninety years of weather reporting in northwest Minnesota. Minn. Rep. 1892-1982. Agric. Exp. Station, Univ. Minn., St. Paul. 46 pp. Sexton, D. A. and M. M. Gillespie. 1979. Effects of fire on the location of a sharp-tailed grouse arena. Can. Field-Nat. 93:74-76. Sharpe, R. S. 1968. The evolutionary relationships and comparative behavior of prairie chickens. Ph.D. thesis. Univ. Nebraska, Lincoln. Short, L. L., Jr. 1967. A review of the genera of grouse (Aves, Tetraoninae). Am. Mus. Novitates 2289. 39 pp. Siivonen, L. 1957. The problem of the short-term fluctuations in numbers of tetraonids in Europe. Pap. on Game Res. 19:1-44. Silver, R., M. O'Connell, and R. Saad. 1979. Effect of androgens on the behavior of birds. Pp. 223-78 in C. Beyer, ed. Endocrine control of sexual behavior. Raven Press, New York. Silvy, N. J. 1968. Movements, monthly ranges, reproductive behavior, and mortality of radio-tagged greater prairie chickens (Tympanuchus cupidopinnatus). M.S. thesis. Kansas State Univ., Manhattan.
REFERENCES
769
Silvy, N. J. and R. S. Lutz. 1978. Attwater's prairie chicken response to petroleum development. Caesar Kleberg Res. Program in Wildl. Ecol., Coll. Agric., Texas A&M Univ., College Station. 69pp. Simmons, K. E. L. 1955. The nature of the predator-reaction of waders towards humans with special reference to the role of the aggressive-escape and brooding drives. Behaviour 8:130-73. Sisson, L. 1976. The sharp-tailed grouse in Nebraska. Nebr. Game and Parks Comm., Lincoln. 88 pp. Skinner, R. M. 1977. A comparison of grassland structure and prairie chicken use in Missouri. Proc. Prairie Grouse Tech. Counc. 12:12. Skutch, A. F. 1949. Do tropical birds rear as many young as they can nourish? Ibis 91:430-55. Slagsvold, T. 1975. Production of young by the willow grouse (Lagopus lagopus (L.) in Norway in relation to temperature. Norwegian J. Zool. 23:269-75. Slagsvold, T. 1984. Clutch size variation of birds in relation to nest predation and the cost of reproduction. J. Anim. Ecol. 53:945-53. Slatkin, M. 1979. Frequency and density dependent selection in a quantitative character. Genetics 93:755-71. Smith, H. M., C. T. Garten, Jr., and P. R. Ramsey. 1975. Genetic heterozygosity and population dynamics in small mammals. Pp. 85-102 in C. L. Markert, ed. Isozymes IV. Genetics and evolution. Academic Press, New York. Smith, S. M. 1967. Seasonal changes in the survival of the black capped chickadee. Condor 69:344-59. Smouse, P. E. 1976. The implications of density-dependent population growth for frequency- and density-dependent selection. Am. Nat. 110:849-60. Smyth, K. E. and D. A. Boag. 1984. Production in spruce grouse and its relationship to environmental factors and population parameters. Can. J. Zool. 62:2250-57. Snyder, L. L. 1935. A study of sharp-tailed grouse. Univ. Toronto Studies Biol. Ser. 40. 66 pp. Soikiro, T., S. Nobuo, and S. Chiba. 1969. On territory of Japanese ptarmigan (Lagopus mutus japonicus Clark) in the Murodo area, Taeyama, Japan Alps, in 1967 and 1968. Misc. Rep. of the Nat'l Park for Nature Study. 1:14-18. Sokal, R. R. and F. J. Rohlf. 1969. Biometry. W. H. Freeman and C., San Francisco. 776 pp. Sonerud, G. A. 1985. Brood movements in grouse and waders as defense against win-stay search in their predators. Oikos 44:287-300. Sopuck, L. G. 1979. Movements and breeding biology of blue grouse in relation to recruitment, reproductive success, and migration. M.Sc. thesis. Univ. Alberta, Edmonton. Sopuck, L. G. and F. C. Zwickel. 1983. Renesting in adult and yearling blue grouse. Can. J. Zool. 61:289-91. Southwick, C. H. 1955. Regulatory mechanisms of house mouse populations: social behavior affecting litter survival. Ecology 36:627-34. Southwood, T. R. E. 1977. Habitat, the templet for ecological strategies? J. Anim. Ecol. 46:337-65. Southwood, T. R. E. and D. J. Cross. 1969. The ecology of the partridge. III. Breeding success and the abundance of insects in natural habitats. J. Anim. Ecol. 38:497-509. Sparling, D. W., Jr. 1979. Reproductive isolating mechanisms and communication in greater prairie chickens (Tympanuchus cupido) and sharp-tailed grouse (Pedioecetesphasianellus). Ph.D. thesis. Univ. North Dakota, Grand Forks. Sparling, D. W. 1981. Communication in prairie grouse. I. information content and intraspecific functions of principal vocalizations. Behav. Neurol. Biol. 32:463-86. Sparling, D. W. 1983. Quantitative analysis of prairie grouse vocalizations. Condor 85:30-42. Sparling, D. W., Jr. and W. D. Svedarsky. 1978. Responses of prairie grouse to avian and mammalian visitors on display grounds in northwestern Minnesota. Prairie Nat. 10:17-22. Spids0, T. K. 1980. Food selection by willow grouse Lagopus lagopus chicks in northern Norway. Ornis Scand. 11:99-105.
770
REFERENCES
Stanton, D. C. 1958. A study of breeding and reproduction in a sage grouse population in southeastern Idaho. M.S. thesis. Univ. Idaho, Moscow. Steen, J. B. 1978. Rypeboka. Oslo (Gyldendal). 231 pp. Steen, J. B., H. C. Pedersen, K. E. Erikstad, K. B. Hansen, K. H0ydal, and A. St0rdal. 1985. The significance of cock territories in willow ptarmigan. Ornis Scand. 16:277-82. Steen, J. B. and S. Unander. 1985. Breeding biology of the Svalbard rock ptarmigan Lagopus mutus hyperboreus. Ornis Scand. 16:191-97. Stenseth, N. C. 1978. Demographic strategies in fluctuating populations of small rodents. Oecologia 33:149-72. Stewart, P. A. 1967. Hooting of sitka blue grouse in relation to weather, season, and time of day. J. Wildl. Manage. 31:28-34. Stewart, R. E. 1975. Breeding birds of North Dakota. Trinity Coll. Center for Environ. Studies. Fargo, N.D. 295 pp. Stirling, I. G. 1965. Studies of the holding, behaviour and nutrition of captive blue grouse. M.Sc. thesis. Univ. British Columbia, Vancouver. Stirling, I. 1968. Aggressive behavior and the dispersion of female blue grouse. Can. J. Zool. 46:405-8. Stirling, I. and J. F. Bendell. 1970. The reproductive behaviour of blue grouse. Syesis 3:161-71. Stoddard, H. L. 1931. The bobwhite quail: its habits, preservation and increase. Charles Scribner's Sons, New York. 559 pp. Stokes, A. W. 1954. Population studies on the ring-necked pheasant on Pelee Island, Ontario. Ont. Dep. Lands and For. Tech. Bull. Wildl. Ser. 4. 154 pp. Stokkan, K-A. 1979a. The effect of permanent short days and castration on plumage and comb growth in male willow ptarmigan (Lagopus lagopus). Auk 96:682-87. Stokkan, K-A. 1979b. Testosterone and daylength-dependent development of comb size and breeding plumage of male willow ptarmigan (Lagopus lagopus lagopus). Auk 96:106-15. Stoneberg, R. P. 1967. A preliminary study of the breeding biology of the spruce grouse in northwestern Montana. M.S. thesis. Montana State Univ., Missoula. Sugden, L. G. and G. W. Beyersbergen. 1986. Effect of density and concealment on American crow predation of simulated duck nests. J. Wildl. Manage. 50:9-14. Svedarsky, W. D. 1979. Spring and summer ecology of female greater prairie chickens in northwestern Minnesota. Ph.D. thesis. Univ. North Dakota, Grand Forks. Svedarsky, W. D. 1980. Juvenile prairie chicken predation by marsh hawk. Raptor Research 14:31-32. Svedarsky, W. D., R. J. Oehlenschlager, and T. D. Tonsager. 1982. A remnant flock of greater prairie chickens in north central Minnesota. Agr. Exp. Station, Univ. of Minnesota, No. 1802, 13 pp. Svedarsky, W. D. and T. D. Tonsager. 1981. A remnant flock of greater prairie chickens in north central Minnesota. Univ. of Minnesota, Crookston. Svoboda, F. J. 1972. Instructions for site selection, plot establishment and catkin production measurements of preferred ruffed grouse winter foods. Univ. Minnesota Forestry Center, Cloquet. Misc. Note 9 pp. Svoboda, F. J. and G. W. Gullion. 1972. Preferential use of aspen by ruffed grouse in northern Minnesota. J. Wildl. Manage. 36:1166-80. Swanson, G. A. and M. I. Meyer. 1973. The role of invertebrates in the feeding ecology of Anatinae during the breeding season. Pp. 143-85 in Proc. Waterfowl Habitat Management Symposium. 1973. Moncton, New Brunswick. Swenson, J. E. 1985. Seasonal habitat use by sharp-tailed grouse, Tympanuchus phasianellus, on a mixed-grass prairie in Montana. Can. Field-Nat. 99:40-46. Szuba, K. J. and J. F. Bendell. 1982. Population densities and habitats of spruce grouse in Ontario. Pp. 199-213 in R. W. Wein, R. R. Riewe, I. R. Methuen, eds. Resources and dynamics of the boreal forest. Association of Canadian Universities for Northern Studies, Ottawa.
REFERENCES
771
Szuba, K. J. and J. F. Bendell. 1984. Weights of Ontario spruce grouse by sex, age, and breeding success. Can. J. Zool. 62:788-92. Tanner, W. D. 1948. A population study of the ruffed grouse (Bonasa umbellus monticola) in Pennsylvania at the low of the cycle. M.S. thesis. Pennsylvania State Univ., State College. Tast, J. and O. Kalela. 1971. Comparison between rodent cycles and plant production in Finnish Lapland. Ann. Acad. Sci. Fenn., Ser. A, 4 Biol. 186:1-14. Tate, J., Jr., M. S. Boyce, and T. R. Smith. 1979. Response of sage grouse to artificially created display ground. Pp. 459-63 in The Mitigation Symposium. Gen. Tech. Rep. RM-65, Rocky Mt. For. Range Exp. Stn., Fort Collins, Colo. Taylor, D. A. 1976. An analysis of some physical characteristics of ruffed grouse (Bonasa umbellus) drumming sites and logs in middle and eastern Tennessee. Tenn. Wildl. Resour. Agency Tech. Rept. 15-25. 72 pp. Taylor, J. 1976. The advantage of spacing out. J. Theor. Biol. 59:485-90. Taylor, L. R. and J. Taylor. 1977. Aggregation, migration and population mechanics. Nature 365:415-20. Taylor, M. A. 1979. Lesser prairie chicken use of man-made leks. Southwest. Nat. 24:706-7. Taylor, M. A. and F. S. Guthery. 1980. Status, ecology, and management of the lesser prairie chicken. U.S. Dep. Agric. For. Serv. Gen. Tech. Rep. RM-77. Rocky Mountain For. and Range Exp. Station, Fort Collins, Colo. 15 pp. Taylor, P. D. and G. C. Williams. 1982. The lek paradox is not resolved. Theor. Pop. Biol. 22:392-409. Tester, J. R. and W. H. Marshall. 1961. A study of certain plant and animal interrelations on a native prairie in northwestern Minnesota. Minn. Mus. Nat. Hist., Univ. Minn. Occasional Pap. 8:1-51. Tester, J. R. and W. H. Marshall. 1962. Minnesota prairie management techniques and their wildlife implications. Trans. N. Am. Wildl. Nat. Resour. Conf. 27:267-87. Theberge, J. B. 1971. Population fluctuation and changes in the quality of rock ptarmigan in Alaska. Ph.D. thesis. Univ. British Columbia, Vancouver. Theberge, J. B. and J. F. Bendell. 1980. Differences in survival and behaviour of rock ptarmigan (Lagopus mutus) chicks among years in Alaska. Can. J. Zool. 58:1638-42. Theberge, J. B. and D. A. Gauthier. 1982. Factors influencing densities of territorial male ruffed grouse, Algonquin Park, Ontario. J. Wildl. Manage. 46:263-68. Theberge, J. B. and G. C. West. 1973. Significance of brooding to the energy demands of Alaskan rock ptarmigan chicks. Arctic 26:138-48. Thomas, V. G. 1982. Energetic reserves of Hudson Bay willow ptarmigan during winter and spring. Can. J. Zool. 60:1618-23. Thompson, D. R. and J. C. Moulton. 1981. An evaluation of Wisconsin ruffed grouse surveys. Wis. Dep. Nat. Resour. Tech. Bull. 123. 13 pp. Thorneycroft, H. B. 1975. A cytogenetic study of the white-throated sparrow, Zonotrichia albicollis (Gimelin). Evolution 29:611-21. Tinbergen, N. 1951. The study of instinct. Oxford Univ. Press, London. 228 pp. Tinbergen, N. 1959. Comparative studies of the behavior of gulls (Laridae): a progress report. Behaviour 15:1-70. Tinbergen, N., M. Imperkoven and D. Franck. 1967. An experiment on spacing out as defence against predators. Behavior 28:307-21. Titus, K. and J. A. Mosher. 1981. Nest-site habitat selected by woodland hawks in the Central Appalachians. Auk. 98:270-81. Todd, A. W., L. B. Keith, and C. A. Fischer. 1981. Population ecology of coyotes during a fluctuation of snowshoe hares. J. Wildl. Manage. 45:629-40. Toepfer, J. E. 1973. Movements and habitat use of greater prairie chickens in central Wisconsin. College of Nat. Resour., Univ. Wisconsin, Stevens Point. 2 pp. (mimeogr.)
772
REFERENCES
Toepfer, J. E. 1976. Reintroduction of prairie chickens into the Crex Meadows Wildlife Area, Burnett County, Wisconsin. 5 pp. (mimeogr.) Toepfer, J. E. 1979. Reintroduction of prairie chickens into the Crex Meadows Wildlife Area Burnett County, Wisconsin, Proc. Prairie Grouse Tech. Counc. 13, 12-14. Tompa, F. S. 1964. Factors determining the numbers of song sparrows, Melospiza melodia (Wilson) on Mandarte Island, B. C., Canada. Acta Zool. Fenn. 109:1-73. Trautman, L. G., L. F. Fredrickson, and A. V. Carter. 1974. Relationships of red foxes and other predators to populations of ring-necked pheasants and other prey. Trans. N. Am. Wildl. Nat. Resour. Conf. 39:24-52. Treisman, M. 1975. Predation and the evolution of gregariousness. I. Models for concealment and evasion. Anim. Behav. 23:779-800. Triver, R. L. 1972. Parental imvestment and sexual selection. Pp. 136-79 in B. G. Campbell, ed. Sexual selection and the descent of man. 1871-1971. Aldine Press, Chicago. Triver, R. L. 1974. Parent-offspring conflict. Am. Zool. 14:249-64. Trobec, R. J. and L. W. Oring. 1972. Effect of testosterone proprionate implantation on lek behavior of sharp-tailed grouse. Am. Midi. Nat. 87:531-36. Unander, S. and J. B. Steen. 1985. Behaviour and social structure in Svalbard rock ptarmigan Lagopus mutus hyperboreus. Ornis Scand. 16:198-204. Vakkari, P. 1980. Polymorphism in Simyra albovenosa (Lepidoptera, Nocturidae). I. Genetic control of the melanic forms. Hereditas 93:181-84. Van Amburg, G. L., J. A. Swaby, and R. H. Pemble. 1981. Response of arthropods to a spring burn of tall grass prairie in northwestern Minnesota. Pp. 240-43 in R. L. Stuckey and K. J. Reese, eds. The prairie peninsula—in the "shadow" of Transeau. Proc. 6th N. Am. Prairie Conf. Ohio State Univ., Columbus. Vance, D. R. and R. L. Westemeier. 1979. Interactions of pheasants and prairie chickens in Illinois. Wildl. Soc. Bull. 7:221-25. Vanderschaegen, P. V. 1970. Food habits of ruffed grouse at the Cloquet Forest Research Center, Minnesota. M.S. thesis Univ. Minnesota, St. Paul. Varley, G. C. and G. R. Gradwell. 1970. Recent advances in insect population dynamics. Annu. Rev. Entomol. 15:1-24. Veen, J. 1977. Functional and causal aspects of nest distribution in colonies of the sandwich tern (Sternas sandvicensis Lath.). Behav. Supplement. 193 pp. Viers, C. E., Jr. 1967. Home range and movements of the greater prairie chicken(Tympanuchus cupidopinnatus) with note on activities. M.S. thesis. Kansas State Univ., Manhattan. Vine, I. 1973. Detection of prey flocks by predators. J. Theor. Biol. 40:207-10. Vogl, R. J. 1964. Vegetational history of Crex Meadows, a prairie savanna in northwestern Wisconsin. Am. Midi. Nat. 72:157-75. Voronin, R. N. 1978. Belaya kuropatka Bol'shezemerskoy tundry. Leningrad (NAUKA). Wagner, F. H. 1957. Late-summer mortality in the pheasant hen. Trans. N. Am. Wildl. Conf. 22:301-15. Wagner, F. H., C. D. Besadny, andC. Kabat. 1965. Population ecology and management of Wisconsin pheasants. Wis. Conserv. Dep. Tech. Bull. 34. 168 pp. Wagner, G. C., L. J. Beuving, and R. R. Hutchinson. 1979. Androgen dependency of aggressive target-biting and paired fighting in male mice. Physiol. Behav. 22:43-46. Wallestad, R. O. 1971. Summer movements and habitat use by sage grouse broods in central Montana. J. Wildl. Manage. 35:129-36. Wallestad, R. O. 1975a. Male sage grouse responses to sagebrush treatment. J. Wildl. Manage. 39:482-84. Wallestad, R. O. 1975b. Life history and habitat requirements of sage grouse in central Montana. Mont. Dep. Fish and Game. 66 pp.
REFERENCES
773
Wallestad, R. O. and D. Pyrah. 1974. Movement and nesting of sage grouse hens in central Montana. J. Wildl. Manage. 38:630-33. Wallestad, R. O. and P. Schladweiler. 1974. Breeding season movements and habitat selection of male sage grouse. J. Wildl. Manage. 38:634-37. Wallestad, R. O. and R. C. Watts. 1972. Factors effecting annual sage grouse productivity in central Montana. Mont. Fish and Game Dep. Proj. 120-R-3, Job GB-2.01. 23 pp. Ward, P. and A. Zahavi. 1973. The importance of certain assemblages of birds as information centers for finding food. Ibis 115:517-34. Waser, P. M. and R. H. Wiley. 1979. Mechanisms and evolution of spacing in animals. Pp. 159-223 in P. Marler and J. G. Vandenbergh, eds. Handbook of behavioral neurobiology. Vol. 3. Social Behavior and Communication. Plenum Press, New York and London. Watson, A. 1964. Aggression and population regulation in red grouse. Nature 202:506-7. Watson, A. 1965. A population study of ptarmigan (Lagopus mutus in Scotland. J. Anim. Ecol. 34:135-72. Watson, A. 1967a. Population control by territorial behaviour in red grouse. Nature 215:1274-75. Watson, A. 1967b. Social status and population regulation in the red grouse (Lagopus lagopus scoticus). Proc. Royal Society Pop. Study Group 2:22-30. Watson, A. 1970. Territorial and reproductive behaviour of red grouse. J. Reprod. Pert., Suppl. 11:3-14. Watson, A. 1972. The behaviour of the ptarmigans. Br. Birds 65:43-117. Watson, A. 1973. Moults of wild Scottish ptarmigan, Lagopus mutus, in relation to sex, climate and status. J. Zool. 171:207-23. Watson, A. 1985. Social class, socially-induced loss, recruitment and breeding of red grouse. Oecologia 67:493-98. Watson, A., R. Hewson, D. Jenkins, and P. Parr. 1973. Population densities of mountain hares compared with red grouse on Scottish moors. Oikos 24:225-30. Watson, A. and D. Jenkins. 1964. Notes on the behaviour of the red grouse. Br. Birds 57:137-70. Watson, A. and D. Jenkins. 1968. Experiments on population control by territorial behavior in red grouse. J. Anim. Ecol. 37:595-614. Watson, A. and G. R. Miller. 1971. Territory size and aggression in a fluctuating red grouse population. J. Anim. Ecol. 40:367-83. Watson, A. and R. Moss. 1970. Dominance, spacing behaviour and aggression in relation to population limitation in vertebrates. Pp. 167-220. in A. Watson, ed. Animal populations in relation to their food resources. Blackwell Sci. Publ., Oxford and Edinburgh. Watson, A. and R. Moss. 1972. A current model of population dynamics in red grouse. Proc. Int. Ornithol. Congr. 15:134-49. Watson, A. and R. Moss. 1979. Population cycles in the Tetraonidae. Ornis Fenn. 56:87-109. Watson, A. and R. Moss. 1980. Advances in our understanding of the population dynamics of red grouse from a recent fluctuation in numbers. Ardea 68:103-11. Watson, A. and R. Moss. 1982. Heritability of egg size, hatch weight, body weight, and viability in red grouse (Lagopus lagopus scoticus). Auk 99:683-86. Watson, A., R. Moss, and R. Parr. 1984a. Effects of food enrichment on numbers and spacing behaviour of red grouse. J. Anim. Ecol. 53:663-78. Watson, A., R. Moss, P. Rothery, and R. Parr. 1984b. Demographic causes and predictive models of population fluctuations in red grouse. J. Anim. Ecol. 53:639-62. Watson, A. and P. J. O'Hare. 1979. Red grouse populations on experimentally treated and untreated Irish bog. J. Appl. Ecol. 16:433-52. Watson, A. and R. Parr. 1981. Hormone implants affecting territory size and aggressive and sexual behavior in red grouse. Ornis Scand. 12:55-61. Watt, P. G. 1969. Mobility patterns, habitat relationships and reproductive success of greater prairie
774
REFERENCES
chickens (Tympanuchus cupido pinnatus) in northeastern Kansas. M.S. thesis. Kansas State Univ., Manhattan. Wayne, P. and G. F. Jolly. 1958. Notes on the breeding of the Iceland gyrfalcon. Br. Birds 51:285-290. Weatherhead, P. J. and R. J. Robertson. 1979. Offspring quality and the polygyny threshold: "the sexy son hypothesis." Am. Nat. 113:201-8. Weatherhead, P. J. and R. J. Robertson. 1981. In defence of the "sexy son" hypothesis. Am. Nat. 117:349-56. Weaver, J. E. 1954. North American prairie. Johnsen Publ. Co., Lincoln, Nebr. 348 pp. Weber, D. A. 1975. Blue grouse ecology, habitat requirements and response to habitat manipulation in north central Utah. Utah Coop. Wildl. Res. Unit. Special Rep. 33. 66pp. Weeden, R. B. 1959a. Ptarmigan ecology and distribution of ptarmigan in western North America. Ph.D. thesis. Univ. British Columbia, Vancouver. Weeden, R. B. 1959b. Ptarmigan research report. Arctic Inst. N. Am. Final Rep. 176 pp. Weeden, R. B. 1963. Management of ptarmigan in North America. J. Wildl. Manage. 27:672-83. Weeden, R. B. 1964. Spatial separation of sexes in rock and willow ptarmigan in winter. Auk 81:534-41. Weeden, R. B. 1965a. Breeding density, reproductive success, and mortality of rock ptarmigan at Eagle Creek, Central Alaska, from 1960 to 1964. Trans. N. Am. Wildl. and Nat. Res. Conf. 30:336-48. Weeden, R. B. 1965b. Grouse and ptarmigan in Alaska. P.R. project W-6-R-5 Vol. 5. Alaska Dept. of Fish and Game, Juneau, Alaska. 110 pp. Weeden, R. B. 1969. Foods of rock and willow ptarmigan in central Alaska with comments on interspecific competition. Auk 86:271-81. Weeden, R. B. and J. B. Theberge. 1972. The dynamics of a fluctuating population of ptarmigan in Alaska. Proc. Int. Ornithol. Congr. 15:90-106. Weeden, R. B. and A. Watson. 1967. Determining the age of rock ptarmigan in Alaska and Scotland. J. Wildl. Manage. 31:825-26. Wegge, P. 1980. Distorted sex ratio among small broods in a declining Capercaillie population. Ornis Scand. 11:106-9. Wegge, P. and B. B. Larsen. 1987. Spacing of adult and juvenile capercaillie, Tetrao urogallus L., cocks during the breeding season. Auk (in press). Weigand, J. P. 1977. The biology and ecology of Hungarian (European gray) partridge (Perdixperdix L.) in north central Montana. Mont. Dep. Fish and Game. Proj. W-91-R-12, W-120-R-1 through 8, Study GB-16.1, Job 1, Final Rep. 338 pp. Weise, C. M. and J. R. Meyer. 1979. Juvenile dispersal and development of site-fidelity in the blackcapped chickadee. Auk 96:40-55. Weller, M. W. 1979. Density and habitat relationships of blue-winged teal nesting in northwestern Iowa. J. Wildl. Manage. 43:367-74. Wells, D. G. 1981. Utilization of a northern Minnesota peatland by sharp-tailed grouse. M.S. Plan B Paper. Univ. Minnesota. Wells, K. D. 1977. The social behaviour of anuran amphibians. Anim. Behav. 25:666-93. Wells, M. C. and P. N. Lehner. 1978. The relative importance of the distance senses in coyote predatory behaviour. Anim. Behav. 26:257-58. West, G. C. 1968. Bioenergertics of captive willow ptarmigan under natural conditions. Ecology 49:1035-45. West, G. C. and M. S. Meng. 1966. Nutrition of willow ptarmigan in northern Alaska. Auk. 83:603-15. West, G. C. and M. S. Meng. 1968. Seasonal changes in body weight and fat and the relation of fatty acid composition to diet in the willow ptarmigan. Wilson Bull. 80:426-41.
REFERENCES
775
Westemeier, R. L. 1971a. Fourth annual report of the prairie grouse committee. Illinois Chapter The Nature Conservancy. 111. Nat. Hist. Survey, Urbana. 25 pp. Westemeier, R. L. 1971b. The history and ecology of prairie chickens in central Wisconsin. Univ. of Wis., Col. of Agric. and Life Sciences. Res. Bull. 281. 61 pp. Westemeier, R. L. 1973. Prescribed burning in grassland management for prairie chickens in Illinois. Proc. Tall Timbers Fire Ecol. Conf. 12:317-38. Westemeier, R. L. and D. R. Vance. 1975. Eighth annual report of the prairie grouse committee. Illinois Chapter-The Nature Conservancy. 111. Nat. Hist. Survey, Urbana. 27 pp. Westerskov, K. 1956. Age determination and dating nesting events in the willow ptarmigan. J. Wildl. Manage. 20:274-79. White, T. C. R. 1978. The importance of a relative shortage of food in animal ecology. Oecologia 33:71-86. Wiens, J. A. 1974. Climatic instability and the "ecological saturation" of bird communities in North American grasslands. Condor 76:385-400. Wiens, J. A. 1976. Population responses to patchy environments. Annu. Rev. Ecol. Syst. 7:81-120. Wiger, R. 1982. Role of self-regulation mechanisms in cyclic populations of Clethrionomys with special reference to C. glareolus: a hypothesis. Oikos 38:60-71. Wiley, R. H. 1973a. The strut display of male sage grouse: a "fixed" action pattern. Behaviour 47:129-52. Wiley, R. H. 1973b. Territoriality and non-random mating in sage grouse Centrocercus urophasianus. Anim. Behav. Monogr. 6:85-169. Wiley, R. H. 1974. Evolution of social organization and life history patterns among grouse (Aves:Tetraonidae). Q. Rev. Biol. 49:201-27. Wiley, R. H. 1978. The lek mating system of the sage grouse. Sci. Am. 238:114-25. Will, G. F. 1946. Tree ring studies in North Dakota. Agric. Exp. Station, N.D. Agric. Coll. Fargo, N.D. Bull. 338. 24 pp. Williams, G. C. 1966a. Adaptation and natural selection: a critique of some current evolutionary thought. Princeton Univ. Press, Princeton, N.J. 307 pp. Williams, G. C. 1966b. Natural selection, the costs of reproduction, and a refinement of Lack's principle. Am. Nat. 100:687-90. Williams, G. C. 1975. Sex and evolution. Princeton Univ. Press. Princeton, N.J. 200 pp. Williams, G. D. V. and G. W. Robertson. 1965. Estimating most probable prairie wheat production from precipitation data. Can. J. Plant Sci. 45:34-47. Williams, J. B., D. Best, and C. Warford. 1980. Foraging ecology of ptarmigan at Meade River, Alaska. Wilson Bull. 92:341-51. Williams, J. C. 1974. Mathematical analysis of red grouse populations. B.Sc. thesis. Univ. York, Toronto, Ontario. Wilson, D. L. 1982. Nesting habitat of lesser prairie chickens in Roosevelt and Lea Counties, New Mexico. M.S. thesis. New Mexico State Univ., Las Cruces. Wilson, E. O. 1975. Sociobiology: the new synthesis. Belknap Press, Cambridge, Mass., and London. 697 pp. Wilson, J. W. III. 1974. Analytical zoogeography of North American mammals. Evolution 28:124-40. Wise, D. R. 1982. Nutrition of wild red grouse (Lagopus lagopus scoticus). J. World Pheasant Assoc. 7:36-41. Wittenberger, J. F. 1978. The evolution of mating systems in grouse. Condor 80:126-37. Wittenberger, J. F. 1979. The evolution of mating systems in birds and mammals. Pp. 271-349 in P. Marler and J. G. Vandenbergh, eds. Handbook of behavioral neurobiology. Vol. 3. Social behavior and communication. Plenum Press, New York and London. Wittenberger, J. F. 198la. Animal social behavior. Duxbury Press, Boston, Mass. 722 pp.
776
REFERENCES
Wittenberger, J. F. 1981b. Male quality and polygyny: "the sexy son" hypothesis revisited. Am. Nat. 117:329-42. Wittenberger, J. F. and R. L. Tilson. 1980. The evolution of monogamy: hypotheses and evidence. Annu. Rev. Ecol. Syst. 11:197-232. Woehr, J. R. 1974. Winter and spring shelter and food selection by ruffed grouse in central New York. M.S. thesis. State Univ. of New York. Wolff, J. O. 1980. The role of habitat patchiness in the population dynamics of snowshoe hares. Ecol. Monogr. 50:111-30. Wrangham, R. W. 1980. Female choice of least costly males; a possible factor in the evolution of leks. Z. Tierpsychol. 54:357-67. Wright, S. 1968. The theory of gene frequencies. Vol. 2. Univ. Chicago Press, Chicago. 511 pp. Wydoski, R. S. 1964. Seasonal changes in the color of starling bills. Auk 81:542-50. Wynne-Edwards, V. C. 1962. Animal dispersion in relation to social behaviour. Oliver and Boyd, Edinburgh. 653 pp. Wynne-Edwards, V. C. 1966. Regulation of numbers as a homeostatic process involving social behavior. Proc. Royal Society Pop. Study Group 1967. 2:7-12. Yeatter, R. E. 1943. The prairie chicken in Illinois. 111. Nat. Hist. Survey Bull. 22:377-416. Yeatter, R. E. 1963. Population responses of prairie chickens to land-use changes in Illinois. J. Wildl. Manage. 27:739-57. Zedja, J. 1966. Litter size in Clethrionomys glareolus. Zool. Listy 16:193-206. Zweifel, R. G. 1981. Genetics of color pattern polymorphism in the California kingsnake. J. Hered. 72:238-44. Zwickel, F. C. 1965. Early mortality and the numbers of blue grouse. Ph.D. thesis. Univ. British Columbia, Vancouver. Zwickel, F. C. 1967. Some observations of weather and brood behavior in blue grouse. J. Wildl. Manage. 31:563-68. Zwickel, F. C. 1972. Removal and repopulation of blue grouse in an increasing population. J. Wildl. Manage. 36:1141-52. Zwickel, F. C. 1973. Dispersion of female blue grouse during the brood season. Condor 75:114-19. Zwickel, F. C. 1975. Nesting parameters of blue grouse and their relevance to populations. Condor 77:423-30. Zwickel, F. C. 1977. Local variations in the time of breeding of female blue grouse. Condor 79:185-91. Zwickel, F. C. 1980. Surplus yearlings and the regulation of breeding density in blue grouse. Can. J. Zool. 58:896-905. Zwickel, F. C. 1982. Demographic composition of hunter-harvested blue grouse in east central Vancouver Island, British Columbia. J. Wildl. Manage. 46:1057-61. Zwickel, F. C. 1983. Factors affecting the return of young blue grouse to breeding range. Can. J. Zool. 61:1128-32. Zwickel, F. C. and J. F. Bendell. 1967. Early mortality and the regulation of numbers in blue grouse. Can. J. Zool. 45:817-51. Zwickel, F. C. and J. F. Bendell. 1972. Blue grouse, habitat, and populations. Proc. Int. Ornithol. Congr. 15:150-69. Zwickel, F. C., J. F. Bendell, and A. N. Ash. 1983. Population regulation in blue grouse. Pp. 212-25 in F. L. Bunnell, D. S. Eastman, and J. M. Peek, eds. Proc. 14th Natural Regulation of Wildlife Populations Symp. For., Wildl. and Range Exp. Station, Univ. Idaho, Moscow. Zwickel, F. C. and J. H. Brigham. 1970. Autumn sex and age ratios of spruce grouse in north-central Washington. J. Wildl. Manage. 34:218-19. Zwickel, F. C., J. H. Brigham, and I. O. Buss. 1975. Autumn structure of blue grouse populations in north-central Washington. J. Wildl. Manage. 39:461-67.
REFERENCES
777
Zwickel, F. C., I. O. Buss, and J. H. Brigham. 1968. Autumn movements of blue grouse and their relevance to populations and management. J. Wildl. Manage. 32:456-68. Zwickel, F. C. and R. G. Carveth. 1978. Desertion of nests by blue grouse. Condor 80:109-11. Zwickel, F. C. and A. N. Lance. 1965. Renesting in blue grouse. J. Wildl. Manage. 29:402-4. Zwickel, F. C. and A. N. Lance. 1966. Determining the age of young blue grouse. J. Wildl. Manage. 30:712-17. Zwickel, F. C., J. A. Redfield, and J. Kristensen. 1977. Demography, behavior, and genetics of a colonizing population of blue grouse. Can. J. Zool. 55:1948-57.
This page intentionally left blank
Index
This page intentionally left blank
Index breeding success, 6, 9, 28, 39, 41, 43, 46, 47, 65-67, 578, 579, 602-615, 618, 626-629, 635, 645-656, 679-686, 693, 694, 715 brood, 9, 10, 32, 41, 43, 46, 47, 73, 516, 517, 537-550, 578-581, 584-586, 602-615, 618, 619, 627-629, 645-653, 656, 681-685, 730, 731 (see also chicks) brood break-up, 548, 550 brood-rearing behavior, 14, 17, 18, 24, 28, 33, 49, 474, 505, 508, 509, 538, 584-586, 622, 712-716 buffer species, 506, 601, 630, 635-640, 681, 682, 693, 694 (see also microtines) burrowing, 564-567 calls, 474, 475 "cackle," call 32, 56, 61, 75 carrying capacity, 636 chicks, 6, 9, 17-20, 28, 32, 33, 41, 43, 46, 47, 51, 66, 67, 70, 75, 516, 517, 537-548, 575-578, 584-586, 602-615, 618, 619, 626-629, 640, 645-653, 656, 681-685, 714, 730, 731 (see also brood) Chitty hypothesis, 4, 24, 25, 28, 29, 70, 71, 75, 434, 435, 641 "cluck" call, 15 clutch, 6, 516, 517, 575, 578, 581-592, 605, 618, 619, 622-624, 629, 681, 684, 701, 702, 730 (see also eggs) communication, 474, 475 (see also advertisement, display) competition, 72, 73, 76, 475, 477-482, 499, 555, 567, 579, 683-685, 693 conifer, 556-558, 565, 711, 718 courtship, 21, 24, 534 (see also advertisement, display) cover, 463, 505, 550-561, 564-567, 576, 587, 588, 599, 602-605, 619-622, 636,
Black Grouse (Chap. 14), 128, 259, 264, 481, 536, 556, 557, 561, 601, 684, 687 Blue Grouse (Chaps. 1, 2, 12, 13, 14, 15, 16) abundance, 5-8, 27, 28, 32, 43-46, 65, 66, 71, 434, 435, 578-731 (see also density, numbers, population size) advertisement, 32, 43, 73, 76, 461, 463, 464, 474-504, 567, 575, 576, (see also) display) age, 6, 8-10, 19, 35, 39-43, 46-53, 59-62, 66, 67, 464, 465, 477-479, 493-505, 528, 429, 534, 537-539, 546, 547, 555, 575, 578-581, 588-590, 599-601, 604-609, 614, 617, 626-629, 645-653, 656, 681, 683-685, 704 aggressive behavior, 4, 15, 21-28, 33-37, 42, 47-53, 55, 56, 61, 65, 68-77, 430, 431, 498, 523, 630, 640, 641, 689, 690 alleles, 35, 434, 435 (see also genetics) ambush, 524-527, 576, 602 antipredator behavior, 15-18, 20, 24, 28, 440, 463, 464, 474, 481, 482, 489-494, 499, 505-512, 517-533, 540, 542-546, 550-552, 555-567, 572, 575-577, 581, 584-592, 686, 687, 709, 717, 728, 729 approachability, 7, 10, 12, 15, 17, 20, 24, 28, 32, 48-56, 58-64, 68, 69, 73, 75, 77 assortment, 74, 75, 574, 685 body condition, 581-584, 640, 641, 681 (see also weights) breeding behavior, 439, 440, 473-504, 533-537, 575-583, 595, 640, 641, 644-656, 702, 711-718 (see also advertisement, aggressive behavior, display, female choice, "hooting," mate choice, mating system, sexual selection, spacing behavior, territorial behavior) 781
782
Blue Grouse (cont'd.) 656, 680-686, 693, 694, 710-718, 727-731 (see also habitat) cover-seeking behavior, 562-567, 576, 595 cover-space tradeoff, 726-729 "crouch-and-run," 15, 16, 24, 28 "crouch-or-fly," 562, 563 cycle, 71, 435, 575, 626, 632, 637-639, 683-685 decision-making behavior, 497, 503, 520, 530 demography, 5-9, 27-29, 35-47, 65-67, 71-77, 434, 435, 578-685 density, 4-8, 27, 28, 32, 57, 58, 65, 68, 70-77, 434, 435, 461, 531, 532, 567-577, 605-609, 614, 615, 629-656, 680-696, 730, 731 density-dependent selection, 681-685 density intolerant/density tolerant, 12-11, 434, 435, 500, 567-577, 680-685, 729 (see also phenotypes) desertion, 505, 506, 527, 530-533, 576, 592 diet (see food) dimorphism, 494-496, 543 disease, 615, 681, 682, 686, 704 dispersal, 31, 35-42, 45, 47, 57, 64, 67, 68, 70-77, 500, 548, 550, 571, 579, 634-636, 640-642 dispersion, 45, 57, 67, 462-464, 482, 517-524, 570-577, 640, 687-689, 704 (see also distribution, spacing behavior) display, 13, 15, 24, 32, 55, 56, 68, 474-475, 477-482, 490, 491, 496, 503 (see also advertisement, distraction behavior) distraction behavior, 7, 16, 24, 33, 49, 58-62, 517-520, 542, 546-549, 576 distribution, 45, 57, 67, 462-464, 640, 687-689 (see also dispersion, latitude, spacing behavior) docile behavior, 53, 55, 70, 71, 77, 430, 431 eggs, 33, 60, 515-517, 524, 581-592, 629, 640, 679 escape behavior, 10, 13, 15, 17, 20, 24, 32, 33, 48-51, 54, 55 experience, 505 "feather-spread," 14-16, 24, 28 female choice, 472, 533-537, 683, 685, 691 (see also mate choice)
INDEX fidelity, 476, 477, 505, 517, 526, 527, 553, 555, 575 (see also philopatry) fire effects, 71, 73, 434, 435, 569, 570, 592, 728, 730 fitness, 29, 70-76, 439, 440, 473, 489, 494, 495, 505, 533, 567-577, 584, 680 flocking, 545, 561-564, 576, 711 "flutter-flight," 14, 24, 28 food, 67, 128, 473, 510, 514-517, 537, 539-541, 550-552, 555-561, 581, 610, 612, 614, 615, 634-637, 640, 644, 681, 683-686, 689, 690, 703, 704, 708-711, 730, 731 foraging, 537, 539-541, 550-552, 555-561, 576, 708 genetics/genotypes, 29, 35, 39, 70, 73-76, 430, 431, 434, 435, 567-571, 641, 681-685 (see also heterozygous/ homozygous) "growl," 15 growth rates, 494-496, 539, 612, 614 habitat, 6, 31, 33-35, 67, 70-76, 463, 473, 483-486, 490, 491, 505, 507, 511, 512, 525, 526, 537, 542-545, 550-561, 565, 567-577, 587, 588, 595, 599, 602-605, 614, 619-622, 680-686, 693, 694, 711-718, 727-731 (see also cover) hatching, 67, 73, 513-515, 539, 578, 592-609, 629, 719 "head-dipping," 15, 24 heterozygous/homozygous, 70, 73-75, 571 home range, 73, 440-443, 461-463, 472, 499, 505-509, 520-523, 531, 538, 542, 545, 546, 687-691, 731 "hooting," 7, 9-13, 24, 28, 31, 32, 43, 47-49, 52, 53, 55, 56, 58, 61, 63, 477 hunting, 642, 686, 696-704, 731 incubation, 518, 519, 524, 576 inversity, 605-609, 693, 694 latitude, 443, 496, 582-584, 592, 594, 628, 680 laying, 515-519, 524, 576, 588, 589, 640 "least-costly male" hypothesis, 472, 627, 628 limiting factors, 679-686, 696-731 longevity, 532, 574, 575, 586-592 (see also mortality) management, 644, 686-731 mate, 683-685 mate choice, 472, 474, 533-537, 576, 683-685, 691
INDEX Blue Grouse (cont'd.) mating system, 439, 440, 461-465, 472 (see also polygamy/polygyny, promiscuity) microtines,681, 682 migration, 550-555, 576, 641-643 mortality, 9, 32, 35-43, 46, 65-67, 463 464, 473, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-636, 642, 645, 647, 649, 650, 652, 656, 679, 680, 683-686, 691, 693-703, 720, 721, 730-731 movements, 64, 462, 499, 528, 529, 537-540, 545, 546, 550-556, 614, 641-643, 704 (see also dispersal, migration) natural selection, 473, 495 "neck-stretching," 15 nest, 9, 441, 443, 463, 464, 503-533, 538, 578, 588-609, 617, 618, 627-629, 647, 649, 651, 652, 679-696, 719, 722, 730, 731 nesting behavior, 9, 34, 36, 39, 41, 43, 65, 66, 76, 441-443, 463, 464, 472, 474, 483-488, 503-533, 538, 555, 567, 575, 576, 578-583, 614, 622-624, 687, 688, 704, 711-718, 726-729 nonbreeding, 35, 41, 66, 475, 534, 578-583, 635, 640-644, 652 (see also surplus birds) numbers, 5, 32, 43, 44-46, 65, 66, 578-731 (see also abundance, density, population size) nutrition, 514-517, 581-584, 614, 640, 641, 644, 681, 682 (see also food) parental investment, 439, 440, 474, 533, 596, 624 permanent sites, 45, 487-489 phenotypes, 29, 53, 64, 65, 71, 73, 76, 77, 499, 567-577, 683-685 philopatry, 505, 526, 527, 553, 555, 575 (see also fidelity) plumage, 464, 465, 548 polygamy/polygyny, 439, 440, 461-465, 642 (see also promiscuity, mating system) polymorphism, 71, 567-577, 643 population dynamics, 3-7, 27, 28, 35-46, 65, 66, 71, 434, 435, 578-685 population size, 5-8, 27, 28, 32, 43, 44-46, 65, 66, 71, 434, 435, 578-696, 730, 731 (see also abundance, density, numbers)
783
predation, 41, 440-443, 463, 464, 472, 527, 530-533, 540, 546-549, 557-560, 576, 581, 584-615, 626-630, 634-640, 642, 652, 680-686, 693-695, 704, 711, 720-722, 728, 729, 731 predator control, 718-726, 730, 731 predator-cover complex, 587, 588, 619-624, 630, 633, 656, 680-686, 693, 711-726, 731 "predator switch-over" hypothesis, 630, 633, 635, 637-640, 681, 682 predators, 41, 67, 73, 441-443, 463, 464, 472, 473, 481, 506, 518-520, 525, 527, 540, 546-549, 557-560, 562-565, 576, 587-589, 592, 594-596, 606, 607, 610-616, 619-622, 626-629, 630, 637-640, 652, 656, 680, 681, 683-686, 689, 690, 693-695, 702, 703, 712, 718-726, 730, 731 productivity, 6, 9, 10, 27, 32, 43, 46, 47, 65, 66, 578, 579, 602-615, 618, 629, 635, 645-656, 679-686 (see also breeding success, broods, chicks, nest, reproductive success) promiscuity, 461-465 (see also polygamy/polygyny, mating system) "qua-qua," 15 renesting, 66, 505, 506, 511, 512, 527, 530-533, 576, 595, 597-599 reproduction, 6, 32, 39, 41, 43, 46, 65, 66, 439, 440, 473-555, 575-577, 622, 702 (see also breeding behavior, broodrearing behavior, nesting behavior) reproductive success, 6, 27, 32, 39, 41, 43, 65, 66, 495, 503, 578, 579, 602-615, 618, 627-629, 635, 636, 645-656, 679-686, 730, 731 (see also breeding success, brood, chicks, nest, productivity) roosting, 545, 564-567, 576, 621 safe site, 469, 475, 489-494, 534, 575, 712 self-regulation, 635, 640-645 sex, 39, 439, 440, 552, 556, 576, 611, 624-629, 642, 704, 720-722 sexual selection, 439, 440, 487-489, 494, 495, 498, 533-537 silent males, 14, 35, 43, 44, 49, 55, 496, 497, 534, 630, 647, 648 snow, 512-515, 550, 557-560, 564-567, 576, 584, 621, 711 space, 686, 726-731
784
INDEX
Blue Grouse (cont'd.) space-cover tradeoff, 694, 696, 726-731 spacing behavior, 35, 42, 46, 58, 67-77, 441, 463, 464, 482, 499, 505, 517-524, 542, 567, 577, 607, 614, 635, 636, 640-645, 683-688, 702, 726-731 (see also territorial behavior) stabilizing density, 691-696, 731 succession, 31, 569, 570, 681, 682 surplus birds, 8, 35, 41, 43, 55, 66, 67, 578-583, 535, 640, 641, 652, 656 (see also nonbreeding) survival, 9, 32, 35-42, 43, 46, 65-67, 463, 464, 473, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-636, 642, 645, 647, 649, 650, 652, 656, 679, 680, 683-686, 691, 693-703, 720, 721, 730-731 territorial behavior, 14, 32, 43-45, 52, 55, 57, 462, 476-479, 497-504, 517-524, 567, 579, 635, 640-645, 683-691 territory, 34, 35-45, 52, 55, 57, 73, 462, 497, 498, 567, 640, 642, 644, 686-691 "threat," 15 "threshold-of-security" hypothesis, 634-636 transient sites, 45, 487 weather, 67, 473, 481, 507, 516, 517, 527, 541, 576, 592, 599, 602-605, 610-612, 614, 635-637, 681, 682, 691, 712 weights, 543, 545, 582-584, 708-710 "whinny," 32 "winter bottleneck" hypothesis, 634-637 winter cover, 550-555, 576, 718 Bonasa umbellus (see Ruffed Grouse) Capercaillie (Chap. 14), 128, 479, 557, 627 Centrocercus urophasianus (see Sage Grouse) Dendragopus canadensis (see Spruce Grouse) D. obscurus (see Blue Grouse) Lagopus lagopus (see Willow Ptarmigan) L. I. scoticus (see Red Grouse) L. leucurus (see White-tailed Ptarmigan) L. mutus (see Rock Ptarmigan) Prairie Chicken (Chaps. 6, 13, 14, 15, 16) abundance, 193, 236-239, 578-585,
640-645, 686-731 (see also density, numbers, population size) advertisement, 465, 474-504, 567, 575, 576, 659 age, 208, 477-479, 493-505, 528, 529, 534, 537-539, 546, 547, 555, 575, 578-581, 588, 590, 599-601, 604-609, 614, 617, 626-629, 645-648, 659-665, 667, 669, 681, 683-685, 704 aggressive behavior, 498, 523, 640, 641, 665-671, 677, 679, 689, 690 alleles, 665 (see also genetics) ambush, 493, 634, 524-527, 576, 602, 717, 718 antipredator behavior, 465-472, 474, 481, 482, 489-494, 499, 505-512, 517-533, 540, 542-546, 550-552, 555-567, 572, 575-577, 581, 584-592, 686, 687, 709, 717, 727-729 assortment, 574, 666-672, 685 body condition, 581-584, 640, 641, 681 (see also weights) "booming," 475 breeding behavior, 201, 225, 226, 238, 239, 465-504, 533-537, 575-583, 595, 640, 641, 644-656, 702, 711-718 (see also advertisement, aggressive behavior, display, "booming," lek, female choice, mate choice, mating system, sexual selection, spacing behavior, territorial behavior) breeding success, 203-206, 229, 236-239, 578, 579, 602-615, 618, 626-629, 635, 645-673, 677, 705 brood, 193, 199, 206-210, 232-234, 238, 239, 516, 517, 537-550, 578, 579, 581, 584-586, 602-606, 607-615, 618, 619, 627-629, 645-648, 656, 659-670, 673, 678, 679, 724, 725, 730, 731 (see also chicks) brood break-up, 548, 550 brood-rearing behavior, 211, 215-220, 224, 225, 230-232, 238, 239, 474, 505, 508, 509, 537-548, 550, 575, 576, 584-586, 622, 659-667, 669, 670, 673, 678, 679, 712-716, 724, 725 buffalo hypothesis, 706, 707 buffer species, 506, 601, 635-640, 674, 681, 682, 693, 694, 724 burning (see fire effects)
INDEX Prairie Chicken (cont'd.) burrowing, 545, 564-567, 576, 621 calls, 474, 475, 659 carrying capacity, 636 chicks, 206-210, 212, 232-234, 238, 239, 516, 517, 537-548, 575, 576, 578, 584, 586, 602-615, 618, 619, 626-629, 640, 645-653, 656, 659-667, 669, 670, 673, 678, 679, 681-685, 714, 724, 730, 731 (see also brood) Chitty hypothesis, 641, 657, 658, 670, 678 clutch, 202-204, 228, 238, 239, 515-517, 524, 575, 578, 581-592, 605, 618, 619, 622-624, 629, 640, 660, 662, 673, 679, 681, 684, 701, 702, 730 communication, 474, 475 (see also advertisement) competition, 475, 477-482, 499, 555, 567, 579, 659, 672-677, 683-685, 693 courtship, 225, 226, 534 (see also advertisement) cover, 194-200, 206, 207, 219-225, 227, 228, 230-232, 235-239, 466, 467, 483-486, 493, 505, 550-561, 564-567, 576, 587, 588, 599, 602-605, 619-622, 636, 656, 669, 672-677, 680-686, 693-694, 710-718, 727-729, 730, 731 (see also habitat) cover-seeking behavior, 562-567, 576, 595 cover-space tradeoff, 726-729 "crouch-or-fly," 562, 563 cycles, 575, 626, 632, 637-639, 656-679, 683-685 decision-making behavior, 497, 303, 520, 530 density, 236-239, 531, 532, 567-577, 605-609, 614, 615, 629-696, 730, 731 density-dependent selection, 657-679, 681-685 density intolerant/density tolerant, 664-672, 679-685, 729 (see also phenotypes) desertion, 505, 506, 527, 530-533, 576, 592 diet (see food) dimorphism, 494, 495, 543 disease, 615, 676, 681, 682, 686, 704 dispersal, 548, 550, 571, 579, 634-636, 640-642, 659 dispersion, 201, 465-472, 482, 517-524, 570-577, 640, 670, 687-689, 704
785
display, 465, 474, 475, 477-482, 503, 567, 575, 576, 659 (see also distraction behavior) distraction behavior, 517-520, 542, 546-549, 576, 659 distribution, 201, 465-472, 658, 687-689, 705-708 (see also dispersion, latitude, spacing behavior) docile behavior, 665-670, 679 dominance, 665 eggs, 202, 203, 212, 228, 515-517, 524, 581-592, 629, 640, 679 (see also clutch) experience, 505 female choice, 465-472, 533-537, 683, 685, 691 (see also mate choice) fidelity, 204-206, 476, 477, 505, 517, 526, 527, 553, 555, 575, 705 fire effects, 204, 206, 219-225, 231, 238, 239, 506, 569, 570, 592, 678, 723, 724, 728, 730 fitness, 439, 440, 465, 466, 473, 489, 494, 495, 505, 533, 567-577, 584, 679, 680 flocking, 545, 561-564, 576, 711 food, 466,473,510,514-517,537,539-541, 550-552, 555-561, 581, 610, 612, 615, 634-637, 640, 644, 676, 681, 683-686, 689, 690, 703-711, 723, 724, 730, 731 foraging, 221, 231, 233, 466, 537, 539-541, 550, 552, 555-561, 576, 708 genetics/genotypes, 465, 567-571, 574, 641, 657-659, 663, 665-672, 674, 677-679, 681-683, 685 growth rates, 494, 495, 539, 612, 614, 677 habitat, 194-200, 206, 207, 219-225, 227, 228, 230-232, 235-239, 466, 467, 483-486, 493, 511, 512, 525, 526, 537, 542-545, 550-561, 565, 567-577, 587, 588, 595, 599, 602-605, 614, 619-622, 657, 658, 669, 672-677, 680-686, 693, 694, 711-718, 722-731 (see also cover) hatching, 202, 207-210, 513-515, 539, 578, 592-609, 629, 677, 719 heterozygous/homozygous, 574, 659, 665, 670, 679 home ranges, 197, 198, 214-220, 226, 227, 235, 236, 238, 239, 440-443, 466, 467, 472, 505-509, 520-523, 531, 542, 545, 546, 687-691, 731 hunting, 642, 658, 659, 663, 676, 678, 686, 696-704, 731
786
INDEX
Prairie Chicken (cont'd.) incubation, 202, 212, 228, 518, 519, 524, 576 introductions, 727 inversity, 605-609, 693, 694 latitude, 443, 496, 582-584, 592, 594, 628, 657, 658, 672-676, 680 laying, 202, 215, 228, 515-519, 524, 576, 588-589, 640, 677 "least-costly male" hypothesis, 465-472, 627, 628 lek, 201, 204, 205, 211, 216-220, 226, 227, 465-472, 476, 477, 483-485, 493, 498, 502-504, 515, 574, 671, 687, 688, 727, 728 limiting factors, 236-239, 679-686, 696-731 longevity, 532, 574, 575, 586-592 (see also mortality) management, 225, 236-239, 644, 686-731 mate, 465-472, 666-672, 677, 679, 683-685 mate choice, 465-472, 474, 533-537, 576, 666-672, 677, 679, 683-685, 691 mating system, 226, 227, 439, 440, 465-472, 665-672 microtines, 681, 682 migration, 550-555, 576, 641-643, 670, 718 mortality, 206-209, 211, 232-235, 238, 239, 532, 537-548, 550-560, 574-579, 581, 586-592, 609-636, 642, 645, 647, 649, 650, 652, 656, 660, 662, 665-669, 672-676, 678, 679, 680, 683-686, 691, 693-696, 703, 720, 721, 730, 731 (see also survival) movements, 197, 211-220, 226, 227, 230-232, 235, 236, 238, 239, 528, 529, 537-540, 545, 546, 550-556, 614, 641-643, 670, 704 (see also dispersal, migration) natural selection, 473, 495 nest, 193, 201-207, 211-213, 215-224, 229, 230, 238, 239, 441-443, 465-472, 503-533, 578, 588-609, 617, 618, 627-629, 647, 659, 661, 662, 672-696, 719, 722, 724, 725, 729, 730, 731 nesting behavior, 193, 201-207, 211-213, 215-224, 229, 230, 238, 239, 441-443, 472, 474, 483-488, 503-533, 555, 567,
575, 576, 578-583, 614, 622-624, 666, 687, 688, 704, 711-718, 724-729 nonbreeding, 475, 534, 578-583, 635, 640-642, 644 numbers, 193, 236-239, 531, 532, 567-731 nutrition, 514-517, 581-584, 614, 640, 641, 644, 677, 681, 682 (see also food) parental investment, 439-440, 477, 533, 596, 624 phenotypes, 567-577, 659, 664-672, 677, 683-685 philopatry, 204-206, 476, 477, 505, 517, 526, 527, 553, 555, 575, 705 plumage, 548, 549 polymorphism, 567-577, 643, 659, 664 population dynamics, 236-239, 578-685 population size, 193, 236-239, 531, 532, 567-731 predation, 203, 204, 207-209, 211, 212, 229, 234-239, 440-443, 465, 472, 527, 530-533, 540, 546-549, 557-560, 576, 581, 584-615, 626-630, 634-636, 637-640, 642, 666, 672-686, 693-695, 704, 711, 720-722, 728, 729, 731 predator control, 718-726, 730, 731 predator-cover complex, 587, 588, 619-624, 656, 680-686, 693, 711-726, 731 "predator switch-over" hypothesis, 635, 637-640, 681, 682 predators, 203, 204, 207-209, 211, 229, 234, 235, 441-443, 465-467, 472, 473, 481, 506, 518-520, 525, 527, 540, 546-549, 557-560, 562-565, 576, 587-589, 592-596, 606, 607, 610-616, 619-622, 626-630, 637-640, 656, 670-679, 680, 681, 683-685, 689, 690, 693, 696, 702, 703, 708, 712, 718-726, 730, 731 productivity, 203-209, 229, 236-239, 495, 503, 578, 579, 602-615, 618, 626-629, 635, 636, 645-673, 677, 679-686, 705, 730, 731 (see also brood, chicks, nest) promiscuity, 439, 440, 465-472 (see also mating system) renesting, 201, 203, 204, 215, 229, 230, 238, 239, 505, 506, 511, 512, 527, 530-533, 576, 595, 597-599, 659 reproduction, 193, 225, 226, 439, 440, 465-555, 575-577, 622, 702 (see also breeding behavior, brood-rearing behavior, nesting behavior)
INDEX Prairie Chicken (cont'd.) reproductive success, 203-209, 229, 236-239, 495, 503, 578, 579, 602-615, 618, 626-629, 635, 636, 645- 673, 677, 679-686, 705, 730, 731 (see also brood, chicks, nest) roosting, 545, 564-567, 576, 621 safe site, 475, 489-494, 534, 575, 712 self-regulation, 635, 640-645 sex, 439, 440, 552, 556, 557, 611, 624-629, 634, 642, 666, 667, 674, 676, 704, 720, 721, 722 sexual selection, 439, 440, 465-472, 482-486, 494, 495, 498, 533-537, 666-672, 677, 679 "sexy-son" hypothesis, 465, 535 snow, 515, 705, 709 space, 686, 726-731 space-cover tradeoff, 674, 694, 696, 726-731 spacing behavior, 441, 465-471, 482, 499, 505, 517-524, 542, 567-577, 607, 614, 635, 636, 640-645, 659, 663-667, 674, 679, 683-688, 702, 726-731 (see also territorial behavior) stabilizing density, 691-696, 731 succession, 569, 570, 681, 682 surplus birds, 578-583, 635, 640, 641, 652, 656 (see also non-breeding) survival, 206-209, 232-235, 238, 239, 473, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-635, 642, 647, 656, 660, 662, 665-667, 669, 672, 679, 680, 683-685, 691, 693-703, 720, 721 (see also mortality) ten-year cycle, 575, 626, 632, 637-639, 656-679, 683-685 territorial behavior, 465-472, 476-479, 497-504, 517-524, 567, 579, 635, 640-645, 659, 666, 668, 669, 683-691 territory, 301, 465-472, 497, 498, 567, 640, 642, 644, 666, 668, 669, 686-691 "threshold-of-security" hypothesis, 634-636 "waiting male" hypothesis, 465-472, 536-537, 627-628 weather, 207-210, 231, 235-239, 473, 481, 507, 515-517, 527, 572, 574, 576, 592, 599, 602-605, 610-612, 614, 635-637, 659, 665, 674, 676, 677-679, 681, 682, 691, 712
787
weights, 515, 516, 543, 545, 582-584, 659, 705, 708-710 "winter bottleneck" hypothesis, 634-637 winter cover, 550-555, 576, 718 Red Grouse (Chaps. 13, 15, 16), 4, 27, 361, 369, 370, 373, 412, 415, 430, 454, 457, 473, 516, 574, 590, 613, 622, 640-645, 653-655, 660, 665-670, 672, 676, 677, 682, 689, 690, 699-702, 730, 731 (see also Willow Ptarmigan) Rock Ptarmigan (Chaps. 9, 13, 14, 15, 16) abundance, 300, 302, 304-311, 316-318, 327, 329, 423, 438, 578-731 (see also density) advertisement, 310-318, 321, 329, 474-504, 567, 575, 576, 659 age 302-310, 319, 321, 323-325, 329, 427, 477-479, 493-505, 528-529, 534, 537-539, 546, 547, 555, 575, 578-581, 659-665, 667, 669, 588-590, 599-601, 604-609, 614, 617, 626-629, 645-648, 653-655, 656, 659-665, 667, 669, 681, 683-685, 704 aggressive behavior, 313-318, 321, 329, 423, 425-427, 429, 430-432, 434, 436, 451-454, 494, 498, 523, 640, 641, 665-671, 677, 679, 689, 690 alleles, 424, 427, 431-434, 665 (see also genetics) ambush, 524-527, 576, 602 antipredator behavior, 313, 318-320, 329, 440-461, 471, 472, 474, 481, 482, 489-494, 499, 505-512, 517-533, 540, 542-546, 550-552, 555-567, 572, 575-577, 581, 584-592, 686, 687, 709, 717, 728, 729 approachability, 449, 450 assortment, 574, 666-672, 685 body condition, 581-584, 640, 641, 681 (see also weights) breeding behavior, 310-319, 321, 329, 439, 440, 473-504, 533-537, 567, 575-583, 595, 640, 641, 644-650, 659, 702, 711-718 (see also aggressive behavior, female choice, mate choice, mating system, sexual selection, spacing behavior, territorial behavior) breeding success, 305-310, 327-329,
788
INDEX
Rock Ptarmigan, (cont'd.) 425-427, 432, 434, 436, 438, 459, 473, 578, 579, 602-615, 618, 626-629, 635, 645-673, 677, 679-686, 693, 694, 715 brood, 303, 305-307, 308, 315, 329, 427-430, 432, 434, 450, 516, 517, 537-550, 575-578, 579, 581, 584-586, 602-615, 618, 619, 626-629, 645-656, 659-670, 672, 673, 677-679, 681-685, 714, 730, 731 brood break-up, 548, 550 brood-rearing behavior, 315, 318, 450, 454, 459, 474, 505, 508, 509, 538, 584-586, 622, 712-716 buffer species, 436, 506, 601, 635-640, 674, 681, 682, 693, 694 burrowing, 564-567 calls, 474, 475, 659 carrying capacity, 636 chicks, 303, 305-307, 308, 315, 329, 427-430, 432, 434, 450, 516, 517, 537-550, 575-578, 579, 581, 584-586, 602-615, 618, 619, 626-629, 640, 645-656, 659-670, 672, 673, 677-679, 681-685, 714, 730, 731 Chitty hypothesis, 329, 423-438, 641, 657, 658, 670, 678 clutch, 305, 306, 328, 329, 426-428, 453, 516, 517, 578, 581-592, 605, 618, 619, 622-624, 629, 660, 662, 673, 681-684, 701, 702, 730 (see also eggs) communication, 474, 475 (see also advertisement, display) competition, 425-428, 436, 438, 475, 477-482, 499, 555, 567, 579, 659, 665, 670, 674, 677-679, 683-685, 693 courtship, 534, (see also advertisement, display) cover, 302, 312, 315, 318-321, 441, 442, 459, 471, 505, 550-561, 564-567, 576, 587, 588, 599, 602-605, 619-622, 636, 656, 669, 672-676, 677, 680-686, 693, 694, 710-718, 727, 729-731 (see also habitat) cover-seeking behavior, 562-567, 576, 595 cover-space tradeoff, 726-729 "crouch-or-fly," 562, 563 cycles, 300, 316, 326, 327, 329, 423-438, 575, 626, 632, 637-639, 656-679, 683-685 decision-making behavior, 440, 497, 503,
520, 530 demography, 300, 304-311, 316-318, 329, 578-685 density, 302-306, 316-318, 324, 329, 423-438, 531, 532, 567-577, 605-609, 614, 615, 629-696, 730, 731 density-dependent selection, 657-679, 681-685 density intolerant/density tolerant, 423, 425, 432-438, 567-577, 664-672, 679, 680-685, 729 (see also phenotypes) desertion, 505, 506, 527, 530-533, 576, 592 diet, (see food) dimorphism, 494, 495, 543 disease, 615, 676, 681, 682, 686, 704 dispersal, 425, 426, 436, 548, 550, 571, 579, 634-636, 640-643, 659 dispersion, 312, 322, 323, 328, 482, 517-524, 570-577, 640, 670, 687-689, 704 (see also distribution, spacing behavior) display, 317, 321, 474, 475, 477-482, 503 (see also distraction behavior) distraction behavior, 447, 449-451, 454, 459, 471, 517, 520, 542, 546-549, 576, 659 distribution, 312, 322, 323, 328, 445, 640, 658, 687-689 (see also dispersion, latitude, spacing behavior) docile behavior, 423, 425, 429, 432, 435-437, 665-670, 679 dominance, 665 eggs, 431, 515-517, 524, 581-592, 629, 640, 679 (see also clutch) experience, 505 fecundity, 425-427, 432, 434, 437 female choice, 427, 451-460, 472, 533-537, 683, 685, 691 (see also mate choice) fidelity, 444, 476, 477, 505, 517, 526, 527, 553-555, 575 fire effects, 569, 570, 592, 678, 728, 730 fitness, 424, 425, 439, 440, 444, 451, 473, 489, 494, 495, 505, 533, 567-577, 584, 679, 680 flocking, 315, 318-321, 329, 545, 561-564, 576, 711 food, 300, 303, 318, 323, 325-329, 437, 447, 457, 458, 473, 510, 514-517, 537, 539-541, 550-552, 555-561, 581, 610,
INDEX Rock Ptarmigan (cont'd.) 612, 614, 615, 634-637, 640, 644, 676, 681, 683-685, 686, 689, 690, 703, 704, 708-711, 730, 731 foraging, 310, 315, 318-321, 324-328, 537, 539-541, 550-552, 555-561, 576, 708 genetics/genotypes, 423-438, 451, 567-571, 574, 641, 657-659, 663, 665-672, 674, 677-679, 681-685 growth rates, 494, 495, 539, 612, 614, 677 habitat, 302, 315, 318, 319, 324, 325, 441, 442, 444, 447, 454, 459, 473, 483-486, 505, 507, 511, 512, 525, 526, 537, 542-545, 550-561, 565, 567-577, 587, 588, 595, 599, 602-605, 614, 619-622, 657, 658, 669, 672-677, 680-686, 693, 694, 711-718, 727-731 (see also cover) hatching, 308, 315, 318, 428, 429, 440, 511-516, 539, 578, 592-609, 629, 677, 719 heterozygous/homozygous, 429, 432, 435, 574, 659, 665, 670, 679 home range, 310, 311, 313, 318, 440-444, 471, 472, 505-509, 520-523, 531, 542, 545, 546, 687-691, 731 (see also territory) hormones, 431, 432, 437, 454 hunting, 302, 303, 308-310, 321-324, 642, 653-655, 658, 659, 663, 676, 678, 686, 696-704, 731 incubation, 315, 431, 454, 518, 519, 524, 576 inversity, 605-609, 693, 694 latitude, 443, 496, 582-584, 592, 594, 657, 658, 672-676, 680 laying, 314, 515-519, 524, 576, 588, 589, 640, 677 limiting factors, 579-686, 696-731 longevity, 532, 574, 575, 586-592 (see also mortality) management, 644, 686-731 mate, 427, 451-460, 666-672, 677, 679, 683-685 mate choice, 427, 451-460, 472, 474, 533-537, 576, 666-672, 677, 679, 683-685, 691 mating system, 313, 314, 427, 439-461, 471, 472, 642, 665, 666-672 microtines, 436, 681, 682 migration, 302, 319, 322-325, 329, 550-556, 576, 641-643, 670
789
molt, 314, 320, 445-448, 454, 457 (see also plumage) monogamy, 313, 314, 439-461, 471, 472, 666 (see also pair bond) mortality, 303, 306-311, 314, 329, 425, 427, 428, 432, 434, 436, 437, 440, 451-453, 460, 473, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-636, 642, 645, 647, 653-656, 660, 662, 665-667, 669, 672-676, 678-680, 683-686, 691, 693-703, 720, 721, 730, 731 movements, 315, 318, 319, 322-325, 329, 436, 528, 529, 537-549, 545, 546, 550-556, 614, 641-643, 655, 670, 704 (see also dispersal, migration) natural selection, 473, 495 nest, 303, 306, 312, 315, 317, 318, 329, 440-446, 451-456, 471, 472, 503-533, 578, 588-609, 617, 618, 627-629, 647, 659, 661, 662, 672-696, 719, 722, 730, 731 nesting behavior, 313, 315, 440, 441-446, 451-456, 458, 474, 483-488, 494, 503-533, 555, 567, 575, 576, 578-583, 614, 622-624, 666, 687, 688, 711-718, 726, 729 nonbreeding, 313, 316-318, 329, 428, 436, 475, 534, 578-583, 635, 640-642, 644 numbers, 300, 302, 304-311, 316-318, 327, 329, 423-438, 578-731 (see also density) nutrition, 303, 320, 323, 328, 329, 514-517, 581-584, 614, 640, 641, 644, 677, 681, 682 (see also food) pair bond, 439, 440, 471 (see also mating system) parental investment, 438-440, 454, 474, 533, 595, 624, 667 phenotypes, 423, 451, 567-577, 624, 659, 664-672, 677 philopatry, 444, 505, 526, 527, 553, 555, 575 plumage, 314, 320, 440, 445-448, 454, 457, 548, 549 poly gamy/polygyny, 313, 314, 459, 460, 642, 670 (see also mating system) polymorphism, 423, 567-577, 643, 659, 664, 679 population dynamics, 300, 304-310, 316-318, 327, 329, 423-438, 578-685
790
INDEX
Rock Ptarmigan (cont'd.) population size, 300, 302, 304-311, 316-318, 327, 329, 423-438, 578-731 (see also density) predation, 304, 306,313-316,318-321,329, 440-443,445-447, 451-456, 461, 472, 527, 530-533, 540, 546-549,557-560, 576, 581, 584-615, 626-630, 634-640, 642, 666, 672-679, 680-686, 693-695, 704,711,720-722, 728, 729, 731 predator control, 718-726, 730, 731 predator-cover complex, 587, 588, 619-624, 656, 680-686, 693, 711-726, 731 "predator switch-over" hypothesis, 329, 436, 635, 637-640, 681, 682 predators, 304, 306, 313-316, 318-321, 329, 437, 440-443, 445-447, 451-456, 461, 471-473, 481, 506, 518-520, 525, 527, 540, 546-549, 557-560, 562-565, 576, 587-589, 592, 594-596, 606, 607, 610-616, 619-622, 626-629, 630, 637-640, 652, 656, 670-679, 680, 681, 683-686, 689, 690, 693-696, 702, 703, 712, 718-720, 730, 731 productivity, 305-310, 328, 329, 578, 579, 602-615, 618, 629, 635, 645-656, 657-666, 673, 679-686 (see also breeding success, brood, chicks, nest, reproductive success) promiscuity, 314 (see also polygamy/polygyny, pair bond) renesting, 505, 506, 511, 512, 527, 530-533, 576, 595, 597-599, 659 reproduction, 310-318, 439, 440, 473-555, 575-577, 622, 702 (see also breeding behavior, brood-rearing behavior, nesting behavior) reproductive success, 305-310, 328, 329, 425-427, 432, 434, 437, 460, 495, 503, 578, 579, 602-615, 618, 627-629, 635, 636, 645-666, 673, 679-686, 730, 731 (see also breeding success, brood, chicks, nest, productivity) roosting, 311, 318-321, 545, 564-567, 576, 621 safe site, 475, 489-494, 534, 575, 712 self-regulation, 329, 423-438, 635, 640-645 sex, 302-306, 308, 321-323, 427, 439, 440, 460, 461, 552, 556, 576, 611, 624-629, 642, 666, 667, 674, 676, 704, 720-722
sexual selection, 310-318, 329, 425-428, 434-436, 437, 439, 440, 451-460, 494, 495, 498, 533-537, 666-672, 677, 679 "sexy-son" hypothesis, 451 snow, 314, 320, 324, 326, 329, 445, 446 space, 686, 726-731 space-cover tradeoff, 674, 694, 696, 726-731 spacing behavior, 300, 310-318, 323, 329 423, 428, 434, 441, 443-446, 471, 472, 482, 499, 505, 517-524, 542, 567-577, 607, 614, 635, 636, 640-645, 659, 663-672, 674, 679, 683-688, 702, 726-731 (see also territorial behavior) stabilizing density, 691-696, 731 surplus birds, 316-318, 329, 428, 436, 578-583, 635, 640, 641, 656 (see also nonbreeding) survival, 303, 306-311, 314, 329, 425, 427, 428, 451-453, 460, 473, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-635, 642, 647, 655, 656, 660, 662, 665-667, 669, 672, 678-680, 683-686, 691, 693-703, 720, 721 (see also mortality) ten-year cycle, 423-438, 656-679 territorial behavior, 310-318, 329, 428, 440-446, 452-458, 460, 471, 472, 476, 479, 497-504, 517-524, 567, 579, 635, 640-645, 655, 659, 666, 668, 669, 683-691 territory, 312-314, 316-318, 329, 427, 429, 430, 432, 434, 440-446, 451-457, 497, 498, 567, 640, 642, 644, 655, 666, 668, 669, 686-691 "threshold-of-security" hypothesis, 634-636 "waiting flock," 427, 428, 436 weather, 302, 306, 308, 310, 311, 320, 328, 329, 430, 432, 433, 473, 481, 507, 516, 517, 527, 576, 592, 599, 602-605, 610, 612, 614, 635-637, 659, 665, 674, 676, 677-679, 681, 682, 691, 712 weights, 457, 543, 545, 582-584, 659, 708-710 "winter bottleneck" hypothesis, 634-637 winter cover, 550-555, 576, 718 Ruffed Grouse (Chaps. 3, 4, 13, 14, 15, 16) abundance, 78, 83-87, 89, 94, 96-98, 108-121, 125-128, 152, 156, 157,
INDEX Ruffed Grouse (cont'd.) 578-731 (see also density, numbers, population size) advertisement, 79-81, 83-85, 117, 118, 120, 461, 463, 464, 474-504, 567, 575, 576, 659 (see also display) age, 81, 85, 87-92, 95-99, 103, 105, 110, 111, 113, 127, 140, 155, 464, 465, 477-479, 493-503, 505, 528, 529, 534, 537-539, 546, 547, 555, 575, 578-581, 588, 590, 599-601, 604-609, 614, 617, 626, 627, 629, 645-648, 656, 659-665, 667, 669, 681, 683-685, 704 aggressive behavior, 82, 98-106, 108, 116-121, 127, 498, 523, 630, 640, 641, 665-671, 677, 679, 689, 690 alleles, 665 (see also genetics/genotypes) ambush, 524-527, 576, 602 antipredator behavior, 117, 120, 137, 154, 156, 157, 440, 463, 464, 472, 474, 481, 482, 489-494, 499, 505-512, 517-533, 540, 542-546, 550-552, 555-567, 575-577, 581, 584-592, 686, 687, 705, 709, 717, 728, 729 approachability, 82, 98, 101, 104, 105, 107-109, 113 aspen, 79, 107, 117, 122-125, 133, 142-157, 558-560, 691, 703-705 assortment, 574, 666-672, 685 body condition, 581-584, 640, 641, 681 (see also weights) breeding behavior, 439, 440, 473-504, 533-537, 575-583, 595, 640, 641, 644-656, 702, 711-718 (see also advertisement, aggressive behavior, display, "drumming," female choice, mate choice, mating system, sexual selection, spacing behavior, territorial behavior) breeding success, 81, 87, 90, 92-98, 110, 111, 120, 121, 578, 579, 602-615, 618, 626-629, 635, 645-673, 677, 679-686, 693, 694, 715 brood, 81, 87, 90, 92, 95-98, 110, 114, 116-118, 516, 517, 537-550, 578-581, 584-586, 602-615, 618, 619, 627-629, 645-648, 656, 659-670, 673, 678, 679, 681-685, 730, 731 (see also chicks) brood-rearing behavior 104, 105, 117, 474, 505, 508, 509, 537-548, 550, 575, 576, 584-586, 622, 712-716
791
brood break-up, 548, 550 budding (see foraging) buffer species, 111, 112, 121, 506, 601, 630, 635-640, 674, 681, 682, 693, 694 burrowing, 128, 129, 153, 564-567 (see also roosting) calls, 474, 475 carrying capacity, 636 chicks, 81, 90, 92-98, 110, 114, 116-118, 120, 121, 516, 517, 537-548, 575-578, 584-586, 602-615, 618, 619, 626-629, 640, 645-653, 656, 659-670, 672-673, 677-679, 681-685, 714, 730, 731 (see also brood) Chitty hypothesis, 78, 98, 105, 641, 657, 658, 670, 678 clutch, 98, 516, 517,575,578,581-592,605, 618, 619, 622-624, 629, 660, 662, 673, 681, 684, 701, 702, 730 (see also eggs) color phase 76, 78, 79, 81, 83-90, 94, 95, 99-109, 113-117, 119, 121, 572, 670, 671 (see also plumage) communication, 474, 475 (see also advertisement, display) competition, 119-121, 133, 140, 144,475, 477-482, 499, 555, 567, 579, 659, 665, 670, 674, 677-679, 683-685, 693 conifer, 107, 556-558, 565, 711, 718 courtship, 534 (see also advertisement, displays) cover, 107, 117, 463, 505, 550-561, 564-567, 576, 587, 588, 599, 602-605, 619-622, 636, 656, 669, 672-677, 680-686, 693, 694, 710-718, 727-729, 730, 731 (see also habitat) cover-seeking behavior, 562-567, 576, 595 cover-space tradeoff, 726-729 "crouch-or-fly," 562, 563 cycles, 78, 111, 112, 575, 626, 632, 637-639, 656-679, 683-685 decision-making behavior, 497, 503, 520, 530 demography, 78, 83-98, 108, 110-121, 127, 128, 578-685 density, 83-87, 94, 96-98, 108, 110-121, 125, 127, 128, 152, 156, 157, 461, 531, 532, 566-577, 605-609, 614, 615, 629-696, 730, 731 density-dependent selection, 657-679, 681-685
792
INDEX
Ruffed Grouse (cont'd.) density intolerant/density tolerant, 119, 500, 502, 567-577, 664-672, 679, 685, 729 (see also phenotypes) desertion, 117, 505, 506, 527, 530-533, 576, 592 diet (see food) dimorphism, 494, 495, 543 disease, 615, 676, 681, 682, 686, 704 dispersal, 105, 475, 500, 502, 548, 550, 571, 579, 634-636, 640-642, 659 dispersion, 84, 462-464, 482, 517-524, 544, 570-577, 640, 658, 670, 687-689, 704 (see also spacing behavior) display, 117, 118, 474, 475, 477-482, 490-492, 503 (see also advertisement, distraction behavior, "drumming") distraction behavior, 83, 101, 104, 105, 117, 517-520, 542, 546-549, 576, 657 distribution, 84, 462-464, 544, 640, 658, 687-689 (see also dispersion, latitude, spacing behavior) docile behavior, 98, 117, 119, 120, 665-670, 679 dominance, 665 "drumming," 79, 80, 83-85, 475, 490-493, 671 drumming log, 84, 476, 487-492, 704, 726 eggs, 515-517, 524, 581-592, 629, 640, 679 (see also clutch) experience, 505 female choice, 472, 533-537, 666-672, 677, 679, 683, 685, 691 (see also mate choice) fidelity, 476, 477, 505, 517, 526, 527, 553, 555, 575 fire effects, 569, 570, 592, 678, 728, 730 fitness, 118, 119, 439, 440, 473, 489, 494, 495, 505, 533, 567-577, 584, 679, 680 flocking, 142, 155, 156, 545, 561-564, 576, 711 food, 79, 107, 110, 116, 117, 122-125, 127, 133, 139, 142-157, 473, 510, 514-517, 537, 539-541, 550-552, 555-561, 610, 612, 614, 615, 634-637, 640, 644, 676, 681, 683-686, 689-691, 703-705, 708-711, 730, 731 foraging, 122-157, 537, 539-541, 550-552, 555-561, 576, 705, 708 genetics/genotypes, 98, 104, 119, 120, 133,
155, 567-571, 574, 641, 657-659, 663, 665-672, 674, 677-679, 681-685 growth rates, 494, 495, 539, 612, 614, 677 habitat, 79, 107, 117, 118, 123, 463, 490-492, 507, 511, 512, 525, 526, 537, 542-545, 550-561, 565, 567-577, 587, 588, 595, 599, 602-605, 614, 619-622, 657, 658, 669, 672-677, 680-686, 693, 694, 711, 718, 727-731 (see also cover) hatching, 92-96, 110, 513-515, 539, 578, 592-609, 629, 677, 719 heterozygous/homozygous, 574, 659, 665, 670, 679 home range, 440-443, 461-463, 472, 505-509, 520-523, 531, 542, 545, 546, 687-691, 731 hunting, 81, 85, 87, 88, 90, 93, 96, 97, 105, 106, 110, 111, 113, 114, 119, 140, 642, 658, 659, 663, 671, 676, 678, 686, 696-704, 731 incubation, 518, 519, 524, 576 inversity, 605-609, 693, 694 latitude, 443, 496, 582-584, 592, 594, 628, 657, 658, 672-676, 680 laying, 515-519, 524, 576, 588, 589, 640, 677 limiting factors, 679-686, 696-731 "least-costly male" hypothesis, 472, 627, 628 longevity, 532, 574, 575, 586, 592 management, 644, 686-731 mate, 666-672, 677, 679, 683-685 mate choice, 472, 474, 533-537, 576, 666-672, 677, 679, 683-685, 691 mating system, 439, 440, 461-465, 472, 642, 665-672 microtines, 681, 682 mortality, 81, 88-91, 95, 98 110-115, 117, 119-121, 463, 464, 532, 537-548, 550-560, 564-566, 574-579, 581, 586-592, 609-636, 642, 645, 647, 656, 660, 662, 665-669, 672-676, 678-680, 683-686, 691, 693-705, 720, 721, 730, 731 (see also survival) movements, 462, 500, 502, 528, 529, 537-540, 545, 546, 550-556, 614, 641-643, 670, 704 (see also dispersal) natural selection, 473, 495 nest, 441-443, 463, 464, 503-533, 544, 588-609, 617, 618, 627-629, 647, 659,
INDEX Ruffed Grouse (cont'd.) 661, 662, 672-696, 719, 722, 730, 731 nesting behavior, 92, 94, 116, 117 441-443, 463, 464, 472, 474, 483-488, 503-533, 555, 567, 575, 576, 578-583, 614, 622-624, 666, 687, 688, 704, 711, 718, 726-729 nonbreeding, 475, 534, 578-583, 635, 640-642, 644 (see also surplus birds) numbers, 78, 83-87, 89, 94, 96-98, 108-121, 125-128, 152, 156, 157,461, 566 (see also abundance, density, population size) nutrition, 128, 133, 514-517, 560, 581-584, 614, 640, 641, 644, 677, 681, 682 (see also food) parental investment, 116, 439, 440, 474, 533, 595, 624, 667 permanent sites, 83, 84, 487-489 phenotypes, 500, 502, 567-577, 659, 664-672, 677, 683-685 philopatry, 505, 526, 527, 553, 555, 575 (see also fidelity) plumage, 464, 465, 548 (see also color phase) polygamy/polygyny, 439, 440, 461-465, 642, 670 (see also mating system) polymorphism, 78, 105, 567-577, 643, 659, 664, 679 population dynamics, 83-98, 108-121, 127, 128, 156, 157, 578-685 population size, 78, 83-98, 108-121, 125-128, 152, 156, 157, 578, 696, 730, 731 (see also density, numbers) predation, 111, 112, 116, 120, 126, 137, 440-443, 463, 464, 472, 490, 492, 530-533, 540, 546-549, 557-560, 576, 581, 584-615, 626-630, 634-640, 642, 666, 672-686, 693-695, 704, 705, 711, 720-722, 728, 729, 731 predator control, 718-726, 730, 731 predator-cover complex, 587, 588, 619-624, 656, 680-686, 693, 711-726, 731 "predator switch-over" hypothesis, 111, 112, 120, 121, 630, 633, 635, 637-640, 681, 682 predators, 79, 111, 112, 116, 121, 137, 156, 157, 441-443, 463, 464, 472, 473, 481, 490, 492, 506, 518-520, 525, 527, 540, 546-549, 555-560, 562-565, 576,
793
587-589, 592, 594-596, 606, 607, 610-616, 619-622, 626-629, 630, 637-640, 656, 670, 680, 681, 683-686, 689, 690, 693-696, 702, 703, 712, 718-726, 730, 731 productivity, 81, 87, 90, 92-98, 110, 111, 120, 121, 495, 503, 578, 579, 602-615, 618, 626-629, 635, 645-673, 677, 679-686, 693, 694, 715 (see also brood, chicks, nest) promiscuity, 461-465 (see also mating system) renesting, 92, 505, 506, 511, 512, 527, 530-533, 576, 595, 597-599, 659 reproduction, 81, 87, 90, 310-318, 439, 440, 473-555, 575-577, 622, 702, (see also breeding behavior, brood-rearing behavior, nesting behavior) reproductive success, 81, 87, 90, 92-98, 110, 111, 120, 121, 495, 503, 578, 579, 602-615, 618, 627-629, 635, 636, 645-673, 679-686, 693, 694, 715, 730, 731 (see also brood, chicks, nest, productivity) roosting, 545, 564-567, 576, 621 (see also burrowing) safe site, 464, 475, 489-494, 534, 537, 712 self-regulation, 98, 635, 640-645 sex, 87, 89, 91, 92, 113, 439, 440, 552, 556, 576, 611, 624-629, 642, 704, 720-722 sexual selection, 439, 440, 487-489, 494, 495, 498, 533-537, 666-672, 677, 679 silent males, 119, 120 496, 497, 534, 630, 647, 648 snow, 113, 122, 125-130, 138, 139, 153, 156, 157, 492, 493, 512-515, 550, 557-560, 564-567, 576, 584, 621, 705, 711 space, 686, 726-731 space-cover tradeoff, 674, 694, 696, 726-731 spacing behavior, 116, 441, 463, 464, 482, 499-502, 505, 517-524, 542, 567, 577, 607, 614, 635, 636, 640-645, 683-688, 702, 726-731 (see also territorial behavior) stabilizing density, 691-696, 731 succession, 569, 570, 681, 682 surplus birds, 578-583, 635, 640, 641, 652, 656 (see also nonbreeding)
794
INDEX
Ruffed Grouse (cont'd.) survival, 81, 88-98, 110-117, 119-121, 463, 464, 473, 488, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-635, 642, 647, 656, 660, 662, 665-667, 669, 672, 679, 680, 683-685, 691, 693-705, 720, 721 (see also mortality) ten-year cycle, 78, 111, 112, 656- 679, (see also cycles) territorial behavior, 462, 476-479, 497-504, 517-524, 567, 579, 635, 640-645, 659, 666, 668, 669, 683-691 territory, 462, 497, 498, 567, 640, 642, 644, 666, 668, 669, 686-691 "threshold-of-security" hypothesis, 634, 636 weather, 79, 81, 88, 91-96, 110-115, 121, 122, 133, 139, 473, 481, 507, 516, 517, 527, 564-566, 576, 592, 599, 602-605, 610-612, 614, 635-637, 659, 665, 674, 676-679, 681, 682, 691, 712 weights, 88, 103, 116, 516, 543, 545, 564-566, 582-584, 659, 708-710 "winter bottleneck" hypothesis, 634, 637 winter cover, 550-555, 576, 718 Sage Grouse (Chaps. 7, 13, 14, 15, 16) abundance, 193, 236-239, 578-696, 730, 731 (see also density, numbers) advertisement, 242, 253-269, 465, 474-504, 567, 575, 576 (see also display) age, 240, 243, 244, 248, 250, 267, 268, 477-479, 493-505, 528, 529, 534, 537-539, 546, 547, 555, 575, 578-581, 588, 590, 599-601, 604-609, 614, 617, 626-629, 645-648, 656, 681, 683-685, 704 aggressive behavior, 242, 253-260, 265-269, 498, 523, 640, 641, 689, 690 ambush, 524-527, 576, 602, 718, 718 antipredator behavior, 440, 465-472, 474, 481, 482, 489-494, 499, 505-512, 517-533, 540, 542-546, 550-552, 555-567, 575-577, 581, 584-592, 686, 687, 709, 717, 727-729 assortment, 574, 685 bimaturation hypothesis, 494, 495 body condition, 581-584, 640, 641, 681 breeding behavior, 243-269, 439, 440, 465-504, 533-537, 575-583, 595, 702,
711-718 (see also advertisement, aggressive behavior, display, lek, female choice, fighting, mate choice, mating system, sexual selection, spacing behavior, territorial behavior) breeding success, 242-253, 259-264, 268, 269, 578, 579, 602-615, 618, 626-629, 635, 645-656 (see also reproductive success) brood, 516, 517, 537-550, 578, 579, 581, 584-586, 602-615, 618, 619, 627-629, 645-648, 656, 681-685, 730, 731 brood-rearing behavior, 474, 505, 508, 509, 537-548, 575, 576, 584-586, 622, 712-716 brood break-up, 548-550 buffer species, 506, 601, 635-640, 681, 682, 693, 694 burrowing, 564, 567 (see also roosting) calls, 474, 475 carrying capacity, 636 chicks, 516, 517, 537-548, 575, 576, 578, 584-586, 602-615, 618, 619, 626-629, 640, 645-648, 656, 681-685, 714, 730, 731 (see also brood) clutch, 516, 517, 575, 578, 581-592, 605, 618, 619, 622-624, 629, 681, 684, 701, 702, 730 (see also eggs) communication, 474, 475 (see also advertisement, display) competition, 243, 253-269, 475, 477-482, 499, 555, 567, 579, 683-685, 693 copying, 266, 269 courtship, 242, 253-260, 265-269, 534 (see also advertisement, display) cover, 242, 466, 505, 550-561, 564, 567, 576, 587, 588, 599, 602-605, 619-622, 636, 656, 680-686, 693, 694, 710-718, 727-731 (see also habitat) cover-seeking behavior, 562-567, 576, 595 cover-space tradeoff, 726-729 "crouch-or-fly," 562, 563 cues, 240, 241, 251-269 (see also copying) cycles, 575, 626, 632, 637-639, 656-679, 683-685 decision-making behavior, 497, 503, 520, 530 demography, 578-685 density, 531, 532, 567-577, 605-609, 614, 615, 629-656, 680-696, 730, 731
INDEX Sage Grouse (cont'd.) density-dependent selection, 681-685 density intolerant/density tolerant, 567-577, 680-685, 729 desertion, 505, 506, 527, 530-533, 576, 592 diet (see food) dimorphism, 494-496, 543 disease, 615, 681, 682, 686, 704 dispersal, 548, 550, 571, 579, 634-636, 640-642 dispersion, 245, 246, 252, 253, 465-472, 482, 517-524, 570-577, 640, 687-689, 704 display, 242, 253-269, 474, 475, 477-482, 496, 503 (see also advertisement, distraction behavior) distraction behavior, 517-520, 542, 546-549, 576 distribution, 245, 246, 252, 253, 465-472, 640, 687-689 (see also dispersion, latitude, spacing behavior) dominance, 259, 263, 264 (see also hierarchy) eggs, 515-517, 524, 581-592, 629, 640, 679 (see also clutch) experience, 505 female choice, 240-269, 465-472, 533-537, 683, 685, 691 (see also mate choice) fidelity, 243-248, 250, 251, 268, 476, 477, 505, 517, 526, 527, 553, 555, 575 fighting, 243, 253, 255-260, 265-269 fire effects, 569-570, 592, 728-730 fitness, 439, 440, 465, 466, 473, 489, 494, 495, 505, 533, 567-577, 584, 680 flocking, 545, 554, 561-564, 576, 711 food, 466, 473, 510, 514-517, 537, 539-541, 550-552, 555-561, 581, 610, 612, 614, 615, 634-637, 640, 644, 681, 683-686, 689, 690, 703, 704, 708-711, 730, 731 foraging 466, 537, 539-541, 550-552, 555-561, 576, 708 genetics/genotypes, 465, 567-571, 641, 681-685 growth rates, 494-496, 539, 612, 614 habitat, 466, 467, 473, 483-486, 493, 505, 507, 511, 512, 525, 526, 537, 542-545, 550-561, 565, 567-577, 587, 588, 595, 599, 602-605, 614, 619-622, 680-686,
795
693, 694, 711-718, 727-731 (see also cover) hatching, 513-515, 539, 578, 592-609, 629, 719 hen clusters, 243, 247, 251-253, 259-265, 268 hierarchy, 240, 241, 259, 263, 264, 266, 268 home range, 440-443, 466, 467, 472, 505-509, 520-523, 531, 542, 545-546, 687-691, 731 hunting, 642, 686, 696-704, 731 incubation, 518, 519, 524, 576 interruptions (copulation disruptions), 257, 258, 261, 266, 536 introductions, 729 inversity, 605-609, 693, 694 latitude, 443, 496, 582-584, 592, 594, 628, 680 laying, 515-519, 524, 576, 588, 589, 640 "least-costly male" hypothesis, 465-472, 627, 628 lek, 240-269, 439, 440, 465-472, 476, 477, 483, 484, 493, 498, 502-504, 535, 574, 647, 671, 687, 688, 727, 728 limiting factors, 679-686, 696-731 longevity, 532, 574, 575, 586-592 (see also mortality) management, 644, 686-731 master cock, 240 mate, 465-472, 683-685 mate choice, 240-269, 465-472, 474, 495, 533-537, 576, 683-685, 691 mating center, 244-246, 250-253, 259-264, 267 mating success, 240, 247-264 mating system, 240, 241, 439, 440, 465-472 microtines, 681, 682 migration, 550-556, 576, 641-643 mortality, 243, 244, 250, 251, 473, 495, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-636, 642, 645, 647, 656, 679, 680, 683-686, 691, 693-703, 720, 721, 730, 731 movements, 528, 529, 537-540, 545, 546, 550-556, 614, 641-643, 704 natural selection, 473, 495 nest, 441-443, 466-472, 503-533, 578, 588-609, 617, 618, 627-629, 647, 679-696, 719, 722, 729-731
796
INDEX
Sage Grouse (cont'd.) nesting behavior, 441-443, 466-472, 474, 483-488, 503-533, 555, 567, 575, 576, 578-583, 614, 622-624, 687, 688, 704, 711-718, 726-729 nonbreeding, 475, 534, 578-583, 635, 640-642, 644, 652, 656 nonrandom mating, 248, 259-264 (see also mating success) numbers, 578-731 (see also abundance, density, population size) nutrition, 514-517, 581-584, 614, 640, 641, 644, 681, 682 (see also food) parental investment, 439, 440, 474, 533, 595, 624 phenotypes, 567-577, 683-685 philopatry 505, 526, 527, 553, 555, 575 (see also fidelity) plumage, 548 polymorphism, 567, 577, 643 population dynamics, 578-685 population size, 193, 236-239, 578-696, 730, 731 (see also density, numbers) predation, 440-443, 465, 472, 495, 511, 527, 530-533, 540, 546-549, 557-560, 576, 581, 584-615, 626-630, 634-640, 642, 680-686, 693-695, 704, 711, 720-722, 728, 729, 731 predator control, 718-726, 730, 731 predator-cover complex, 587, 588, 619-624, 656, 680-686, 711-726, 731 "predator switch-over" hypothesis, 635, 637-640,681,682 predators, 441-443, 465-467, 472, 473, 481, 506, 518-520, 525, 527, 540, 546-549, 557-560, 562-565, 576, 587-589, 592, 594-596, 606, 607, 610-616, 619-622, 626-630, 637-640, 656, 680, 681, 683-686, 689, 690, 693-696, 702, 703, 712, 718-726, 730, 731 productivity, 578, 579, 602-615, 618, 629, 635, 645-656, 679-686 (see also breeding success, brood, chicks, nest, reproductive success) promiscuity, 439, 440, 465-472 (see also mating system) renesting, 505, 506, 511, 512, 527, 530-533, 576, 595, 597-599 reproduction, 439, 440, 465-555, 575-577, 622, 702 (see also breeding behavior,
brood-rearing behavior, nesting behavior) reproductive success, 247-264, 495, 503, 578, 579, 602-615, 618, 627-629, 635, 636, 645-656, 679-686, 730, 731 roosting, 545, 564-567, 576, 621 safe site, 475, 489-494, 534, 575, 712 self-regulation, 635, 640-645 sex, 439, 440, 552, 556, 576, 611, 624-629, 642, 704, 720-722 sexual selection, 240, 247-251, 254-269, 439, 440, 465-472, 482-486, 494, 495, 498, 533-537 "sexy-son" hypothesis, 465, 535 site hypothesis, 251-253, 259, 264, 267, 268, 535 snow, 512-515, 550, 557-560, 564-567, 576, 584, 621, 711 space, 686, 726-731 space-cover tradeoff, 694, 696, 726-731 spacing behavior, 240, 241, 243-269, 441, 465-471, 482, 505, 517-524, 542, 567-577, 607, 614, 635, 636, 640-645, 683-688, 702, 726-731 (see also territorial behavior) stabilizing density, 691-696, 731 "strutting," 242, 253-260, 265-269 succession, 475, 534, 569, 570, 681, 682 surplus birds, 578-583, 635, 640-642, 644, 652, 656 survival, 243, 244, 250, 251, 473, 495, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-636, 642, 645, 647, 656, 679, 680, 683-686, 691, 693-703, 720, 721, 730, 731 territorial behavior, 240, 241, 243-269, 465-472, 476-479, 497-504, 517-524, 567, 579, 635, 640-645, 683-691 territory, 240, 241, 243-253, 259, 264, 267-269, 465-472, 497, 498, 567, 640, 642, 644, 686-691 "threshold-of-security" hypothesis, 634-636 "waiting male" hypothesis, 536, 537 (see also "least-costly male" hypothesis) weather, 473, 481, 507, 516, 517, 527, 572, 574, 576, 592, 599, 602-605, 610-612, 614, 635-637, 681, 682, 691, 712 weights, 478, 543, 545, 582-584, 708-710 "wing-beating," 242, 253-260 "winter bottleneck" hypothesis, 634-637 winter cover, 550-555, 576, 718
INDEX Sharp-tailed Grouse (Chaps. 5, 13, 14, 15, 16) abundance, 159-162, 531, 532, 567-731 advertisement, 164, 187, 188, 465, 474-504, 567, 575, 576, 659 age, 477-479, 493-503, 505, 528, 529, 534, 537-539, 546, 547, 555, 575, 578-581, 588, 590, 599, 600, 601, 604-609, 614, 617, 626, 627, 629, 645-648, 656, 659-665, 667, 669, 681, 683-685, 704 aggressive behavior, 498, 523, 640, 641, 665-671, 677, 679, 689, 690 alleles, 665 (see also genetics/genotypes) ambush, 493, 524-527, 576, 602, 717, 718 antipredator behavior, 187-192, 440, 465-472, 474, 481, 482, 489-494, 499, 505-512, 517-533, 540, 542-546, 550-552, 555-567, 575-577, 581, 584-592, 686, 687, 709, 717, 727-729 assortment, 574, 666-672, 685 body condition, 581-584, 640, 641, 681 (see also weights) breeding behavior, 164, 439, 440, 465-504, 533-537, 575-583, 595, 640, 641, 644-656, 702, 711-718 (see also advertisement, aggressive behavior display, female choice, mate choice, mating system, sexual selection, spacing behavior, territorial behavior) breeding success, 578, 579, 602-615, 618, 626-629, 635, 645-673, 677, 679-686, 693, 694, 715 brood, 172, 175-178, 516, 517, 537-550, 575, 576, 578, 579, 581, 584-586, 602-615, 618, 619, 626-629, 640, 645-648, 656, 659-667, 669, 670, 672, 673, 677-679, 681-685, 714, 730, 731 brood break-up, 175, 178, 191, 192, 548, 550 brood-rearing behavior, 172, 175-178, 191, 192, 474, 505, 508, 509, 537-548, 575, 576, 584-586, 622, 712-716 buffer species, 506, 601, 635-640, 674, 681, 682, 693, 694 burrowing, 180-183, 189-192, 564-567 (see also roosting) calls, 474, 475 carrying capacity, 636 chicks, 172, 175-178, 516, 517, 537-550, 575, 576, 578, 579, 581, 584-586,
797
602-615, 618, 619, 626-629, 640, 645-648, 656, 659-667, 669, 670, 672, 673, 677-679, 681-685, 714, 730, 731 Chitty hypothesis, 641, 657, 658, 670, 678 clutch, 516, 517, 575, 578, 581-592, 605, 618, 619, 622-624, 629, 659, 665, 670, 674, 677-679, 681, 684, 701, 702, 730 (see also eggs) communication, 474, 475 (see also advertisement) competition, 475, 477-482, 499, 555, 567, 579, 659, 665, 670, 674, 677-679, 683-685, 693 courtship, 534 (see also advertisement) cover, 158, 171, 178-181, 184-186, 190-192, 466, 505, 550 -561, 564-567, 576, 587, 588, 599, 602-605, 619-622, 636, 656, 669, 672-677, 680-686, 693, 694, 710-718, 727-731 (see also habitat) cover-seeking behavior, 562-567, 576, 595 cover-space tradeoff, 726-729 "crouch-or-fly," 562, 563 cycles, 575, 626, 632, 637-639, 656-679, 683-685 decision-making behavior, 497, 503, 520, 530 demography, 578-685 density, 159, 160, 531, 532, 567-577, 605-609, 614, 615, 629-696, 730, 731 density-dependent selection, 657-679, 681-685 density intolerant/density tolerant, 567-577, 659, 664-672, 677, 679-685, 729 desertion, 172, 176, 187, 188, 505, 506, 527, 530-533, 576, 592 diet (see food) dimorphism, 494, 495, 543 disease, 615, 676, 681, 682, 686, 704 dispersal, 172, 175-178, 187, 188, 191, 192, 548, 550, 571, 579, 634-636, 640-642, 659 dispersion, 158-162, 164-172,465-472, 482, 517-524, 570-577, 640, 658, 670, 687-689, 704, 707, 708 (see also spacing behavior) display, 164, 187, 188, 474, 475, 477-482, 503 (see also advertisement, distraction behavior) distraction behavior, 517-520, 542, 546-549, 576, 659 distribution, 158-162, 164-172,465-472,
798
INDEX
Sharp-tailed Grouse (cont'd.) 640, 658, 687-689, 707, 708 (see also dispersion, latitude, spacing behavior) docile behavior, 665-670, 679 dominance, 665 eggs, 515-517, 524, 581-592, 629, 640, 679 (see also clutch) experience, 505 female choice, 465-472, 533-537, 666-672, 677, 679, 683, 685, 691 (see also mate choice) fidelity, 476, 477, 505, 517, 526, 527, 553, 555, 575 fire effects, 569, 570, 592, 678, 728, 730 fitness, 187, 439, 440, 465, 466, 473, 489, 494, 495, 505, 533, 567-577, 584, 679, 680 flocking, 174, 181-183, 189-192, 545, 561-564, 576, 711 food, 172, 173, 176, 182, 186, 187, 191, 192, 466, 473, 510, 514-517, 537, 539-541, 550-552, 555-561, 581, 610, 612, 614, 615, 634-637, 640, 644, 676, 681, 683-686, 689, 690, 703-711, 730, 731 foraging, 164, 173-176, 178-182, 184-189, 191, 192, 466, 537, 539-541, 550-552, 555-561, 576, 708 genetics/genotypes, 465, 567-571, 574, 657-659, 663, 665-672, 674, 677-679, 681-685 growth rates, 494, 495, 539, 612, 614, 677 habitat, 158, 159, 170, 178-181, 184-186, 189-192, 466, 467, 473, 483-486, 493, 503, 507, 511, 512, 525, 526, 537, 542-545, 550-561, 565, 567-577, 587, 588, 595, 599, 602-605, 614, 619-622, 657, 658, 669, 672-677, 680-686, 693, 694, 711-718, 727-731 (see also cover) hatching, 513-515, 539, 578, 592-609, 629, 677, 719 heterozygous/homozygous, 574, 659, 665, 670, 679 home range, 163, 164, 166-169, 171-174, 184-186, 191, 192, 440-443, 466, 467, 472, 505-509, 520-523, 531, 542, 545, 546, 687-691, 731 hunting, 642, 658, 659, 663, 676, 678, 686, 696-704, 731
incubation, 518, 519, 524, 576 introductions, 729 inversity, 605-609, 693, 694 latitude, 443, 496, 582-584, 592, 594, 628, 657, 658, 672-676, 680 laying, 184-186, 515-519, 524, 576, 588, 589, 640, 677 "least-costly male" hypothesis, 465-472, 627, 628 lek, 160-162, 164-167, 171, 187, 188, 439, 440, 465-472, 476, 477, 483-485, 493, 498, 502-504, 535, 574, 671, 687, 688, 727, 728 limiting factors, 679-686, 696-731 longevity, 32, 574, 575, 586-592 (see also mortality) management, 644, 686-731 mate, 465-472, 666-672, 677, 679, 683-685 mate choice, 184-186, 465-472, 474, 533-537, 576, 666-672, 677, 679, 683-685, 691 mating system, 186, 439, 440, 465-472, 642, 665-672 microtines, 681, 682 migration, 550-556, 576, 641-643, 670 mortality, 473, 532, 537-548, 550-560, 574-579, 581, 586-592, 609-636, 642, 645, 647, 656, 660, 662, 665-667, 669, 672-676, 678-680, 683-686, 691, 693-703, 720, 721, 730, 731 movements, 158, 164-177, 184-186, 188-192, 528, 529, 537-540, 545, 550-556, 576, 614, 641-643, 670, 704 (see also dispersal) natural selection, 473, 495 nest, 166, 169, 184-187,441-443, 465-472, 503-533, 578, 588-609, 617, 618, 627-629, 647, 659, 661, 662, 672-696, 708, 729, 731 nesting behavior, 165, 166, 169, 184-187, 191, 192, 441-443, 466-472, 474, 483-488, 503-533, 555, 567, 575, 576, 578-583, 614, 622-624, 666, 687, 688, 704, 711-718, 726-729 nonbreeding, 475, 534, 578-583, 635, 640-642, 644, 656 numbers, 159-162, 567-731 (see also population size)
INDEX Sharp-tailed Grouse (cont'd.) nutrition, 184-186, 514-517, 581-584, 614, 640, 641, 644, 677, 681, 682 (see also food) parental investment, 187, 439, 440, 474, 533, 595, 624, 667 phenotypes, 567-577, 659, 664, 665-672, 677, 683-685 philopatry, 505, 526, 527, 553, 555, 575 (see also fidelity) plumage, 548 polygyny, 439, 440, 465-472, 642, 670 polymorphism, 567-577, 643, 659, 664, 679 population dynamics, 578-685 population size, 159-162, 567-731 predation, 440-443, 465, 472, 527, 530-533, 540, 546-549, 557-560, 576, 581, 584-615, 626-630, 634-640, 642, 666, 672-686, 693-695, 704, 711, 720-722, 728, 729, 731 predator control, 718-726, 730, 731 predator-cover complex, 587, 588, 619624, 656, 680-686, 693, 711-726, 731 "predator switch-over" hypothesis, 635, 637-640, 681, 682 predators, 187-192, 441-443, 465-467, 472, 473, 481, 506, 518-520, 525, 527, 540, 546-549, 557-560, 562-565, 576, 587-589, 592, 594-596, 606, 607, 610-616, 619-622, 626-630, 637-640, 656, 670-681, 683-686, 689, 690, 693696, 702, 703,712,718-726, 730,731 productivity, 578, 579, 602-615, 618, 626-629, 635, 645-673, 677, 679-686. 693, 694, 715 (see also brood, chicks, nest, reproductive success) promiscuity, 439, 440, 465-472, 642, 670 renesting, 168, 169, 505, 506, 511, 512, 527, 530-533, 576, 595, 597-599, 659 reproduction, 439-440, 465-555, 575-577, 622, 702 (see also breeding behavior, brood-rearing behavior, nesting behavior) reproductive success, 495, 505, 578, 579, 602-615, 618, 626-629, 635, 636, 645-673, 679-686, 693, 694, 715, 730, 731 (see also breeding success, brood, chicks, nest) roosting, 164, 172, 174-176, 178-181,
799
188-192, 545, 564-567, 576, 621 (see also burrowing) safe site, 475, 489-494, 534, 575, 712 self-regulation, 635, 640-645 sex, 439, 440, 552, 556, 576, 611, 624-629, 642, 666, 667, 674, 676, 704, 720-722 sexual selection, 439, 440, 465-472, 482-487, 494, 495, 498, 533-537 "sexy-son" hypothesis, 465, 535 snow, 171, 174, 180-183, 189-192, 512-515, 550, 557-560, 564-567, 576, 584, 621, 711 space, 686, 726-731 space-cover tradeoff, 674, 694, 696, 726-731 spacing behavoir, 158-172, 184-187, 191, 192, 441, 465-471, 482, 505, 517-524, 542, 567-577, 607, 614, 635, 636, 640-645, 659, 663-672, 674, 679, 683-688, 702, 726-731 stabilizing density, 691-696, 731 succession, 569, 570, 681, 682 surplus birds, 475, 534, 578-583, 635, 640-642, 644, 656 survival, 473, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-636, 642, 645, 647, 656, 660, 662, 665-667, 669, 672-676, 678-680, 683-686, 691, 693-703, 720, 721, 730, 731 ten-year cycle, 575, 626, 632, 637-639, 656-679, 683-685 territorial behavior, 160-162, 187, 188, 465-472, 476-479, 497-504, 517-524, 567, 579, 635, 640-645, 659, 666, 668, 669, 683-691 territory, 465-472, 497, 498, 567, 640, 642, 644, 666, 668, 669, 686-691 "threshold-of-security" hypothesis, 634-636 "waiting-male" hypothesis, 536, 537 weather, 171, 189-192, 473, 481, 507, 516, 517, 527, 572-574, 576, 592, 599, 602-605, 610-612, 614, 635-637, 659, 665, 674, 676-679, 681, 682, 691, 712 weights, 543, 545, 582-584, 659, 708-710 "winter bottleneck" hypothesis, 634-637 winter cover, 171, 179-181, 190-192, 550-555, 576, 718
800
INDEX
Spruce Grouse (Chaps. 13, 14, 15, 16) abundance, 578-731 (see also density) advertisement, 461, 463, 464, 474-504, 567, 575, 576 age, 464, 465, 477-479, 493-503, 505, 528, 529, 534, 537-539, 546, 547, 555, 575, 578-581, 588, 590, 599-601. 604-609, 614, 617, 626, 627, 629, 645-648, 656, 681, 683-685, 704 aggressive behavior, 498, 523, 630, 640, 641, 689, 690 ambush, 524-527, 576, 602 antipredator behavior, 440, 463, 464, 472, 474, 481, 482, 489-494, 499, 505-512, 517-533, 540, 542-546, 550-552, 555-567, 575-577, 581, 584-592, 686, 687, 709, 717, 728, 729 body condition, 581-584, 640, 641, 681 (see also weights) breeding behavior, 439, 440, 473-504, 533-537, 575-583, 595, 640, 641, 644-656, 702, 711-718 (see also advertisement, aggressive behavior, female choice, mate choice, mating system, sexual selection, spacing behavior, territorial behavior) breeding success, 578, 579, 602-615, 618, 626-629, 635, 645-656, 715 brood, 516, 517, 537-550, 578, 579, 581, 584-586, 602-615, 618, 619, 627-629, 645-648, 656, 681-685, 730, 731 (see also chicks) brood-rearing behavior, 474, 505, 508, 509, 537-548,575,576,584-586,622,712-716 brood break-up, 548, 550 buffer species, 506. 601, 630, 635-640, 681, 682, 693, 694 burrowing, 564-567 (see also roosting) calls, 474, 475 carrying capacity, 636 chicks, 516, 517, 537-548, 575, 576, 578, 584-586, 602-615, 618, 619, 626-629, 640, 645-648, 656, 681-685, 714, 730, 731 (see also brood) Chitty hypothesis, 434, 435, 641 clutch, 516, 517, 575, 578, 581-592, 605, 618, 619, 622-624, 629, 681, 684, 701, 702, 731 (see also eggs) communication, 474, 475 (see also advertisement)
competition, 475, 477-482, 499, 555, 567, 579, 683-685, 693 conifer, 556-558, 565, 711, 718 courtship, 534 (see also advertisement) cover, 463, 505, 550-561, 564-567, 576, 587, 588, 599, 602-605, 619-622, 636, 656, 680-686, 693, 694, 710-718, 727-731 (see also habitat) cover-seeking behavior, 562-567, 576, 595 cover-space tradeoff, 726-729 "crouch-or-fly," 562, 563 cycle, 575, 626, 632, 637-639, 683-685 decision-making behavior, 497, 503, 520, 530 demography, 578-685 density, 531, 532, 567-577, 605-609, 614, 615, 629-656, 680-696, 730, 731 density-dependent selection, 681-685 density intolerant/density tolerant, 567-577, 680-685, 729 desertion, 505, 506, 527, 530-533, 576, 592 diet (see food) dimorphism, 494, 495, 543 disease, 615, 681, 682, 686, 704 dispersal, 548, 550, 571, 579, 634-636, 640-642 dispersion, 462-464, 482, 517-524, 570-577, 640, 687-689, 704 (see also spacing behavoir) display, 474, 475, 477-482, 490, 491, 503 (see also advertisement, distraction behavior) distraction behavior, 517-520, 542, 546-549, 576 distribution, 462-464, 640, 687-689 (see also dispersion, latitude, spacing behavior) docile behavior (see phenotypes) eggs, 515-517, 524, 581-592, 629, 640, 679 (see also clutch) experience, 505 female choice, 472, 533-537, 683, 685, 691 (see also mate choice) fidelity, 476, 477, 505, 517, 526, 527, 533, 555, 575 fire effects, 569, 570, 592, 728, 730 fitness, 439, 440, 473, 489, 494, 495, 505, 533, 567-577, 584, 680 flocking, 545, 561-564, 576, 711
INDEX Spruce Grouse (cont'd.) food, 473, 510, 514-517, 537, 539-541, 550-552, 555-561, 581, 610, 612, 614, 615, 634-637, 640, 644, 681, 683-686, 689, 690, 703, 704, 708-711, 730, 731 foraging, 537, 539-541, 550-552, 555-561, 576, 708 genetics/genotypes, 567-571, 641, 681-685 growth rates, 494, 495, 539, 612, 614 habitat, 463, 473, 483-486, 490, 505, 507, 511, 512, 525,526,537,542-545, 550561,565,567-577, 587, 588, 595,599, 602-605, 614, 619-622, 680-686, 693, 694,711-718,727-731 (see also cover hatching, 513-515, 539, 578, 592-609, 629, 719 home range, 440-443, 461, 463, 472, 505-509, 520-523, 531, 542, 545, 546, 687-691, 731 hunting, 642, 686, 696-704, 731 incubation, 518, 519, 524, 576 inversity, 605-609, 693, 694 latitude, 443, 496, 582-584, 592, 594, 628, 680 laying, 515-519, 524, 576, 588. 589, 640 "least-costly male" hypothesis, 472, 627, 628 limiting factors, 679-686, 696-731 longevity, 532, 574, 575, 586-592 (see also mortality) management, 644, 686-731 mate, 683-685 mate choice, 461-465, 472, 474, 533-537, 576, 683-685, 691 mating system, 439, 440, 461-465, 472, 642 microtines, 681-682 migration, 550-556, 576, 641-643 mortality, 463, 464, 490, 532, 537-548, 550-560, 574-579, 581, 586-592, 609-636, 642, 645, 647, 656, 679, 680, 683-686, 691, 693-703, 720, 721, 730, 731 (see also survival) movements, 462, 522, 528, 529, 537-540, 545, 546, 550-556, 614, 641-643, 704 (see also dispersal, migration) natural selection, 473, 495 nest, 441, 443, 463, 464, 503-533, 578, 588-609, 617, 618, 627-629, 647, 679-696, 719, 722, 730, 731
801
nesting behavior, 441-443, 464, 472, 474, 483-488, 503-533, 555, 567, 575, 576, 578-583, 614, 622-624, 687, 688, 704, 711-718, 726-729 nonbreeding, 475, 534, 578-583, 635, 640-642, 644, 656 numbers, 578-731 (see also density, population size) nutrition, 514-517, 581-584, 614, 640, 641, 644. 681, 682 (see also food) parental investment, 439, 440, 474, 533, 595, 624 phenotypes, 567-577, 683-685 philopatry, 476, 477, 505, 517, 526, 527, 553, 555, 575 plumage, 464, 465, 548 polygamy/polygyny, 439, 440, 461-465, 642 polymorphism, 567-577, 643 population dynamics, 578-685 population size, 578-696, 730, 731 (see also density) predation, 440-443, 463, 464, 472, 527, 530-533, 540, 546-549, 557-560, 576. 581, 584-592, 626-630, 634-640, 642, 680-686, 693-695, 704, 711, 720-722, 728, 729, 731 predator control, 718-726, 730, 731 predator-cover complex, 587, 588, 619624, 656, 680-686, 693, 711-726, 731 "predator switch-over" hypothesis, 635, 637-640, 681, 682 predators 441-443, 463, 464, 472, 473, 481, 506, 518-520, 525. 527, 540, 546-549, 557-560, 562-565, 576, 587-589, 592, 594-596, 606, 607, 610-616, 619-622, 626-630, 637-640, 656, 680, 681, 683-686, 689, 690, 693-696, 702, 703, 712, 718-726, 730, 731 productivity, 578, 579, 602-615, 618, 629, 635, 645-656, 679-686 (see also breeding success, brood, chicks, nest, reproductive success) promiscuity, 439, 440, 461-465, 642 renesting, 505, 506, 511, 512, 527, 530-533, 576, 595, 597-599 reproduction, 439, 440, 473-555, 575-577, 622, 702 (see also breeding behavior, brood-rearing behavior, nesting behavior)
802
INDEX
Spruce Grouse (cont'd.) reproductive success, 495, 503, 578, 579, 602-615, 618, 627-629, 635, 636, 645-656, 679-686, 730, 731 roosting, 545, 564-567, 576, 621 safe site, 464, 475, 489-494, 534, 575, 712 self-regulation, 635, 640-645 sex, 439, 440, 552, 556, 576, 611, 624-629, 642, 704, 720-722 sexual selection, 439, 440, 487-489, 494, 495, 498, 533-537 silent male, 496, 497, 534, 630, 647, 648 snow 512-515, 550, 557-560, 564-567, 576, 584, 621, 711 space, 686, 726-731 space-cover tradeoff, 694, 696, 726-731 spacing behavior, 441, 463, 464, 482, 505, 517-524, 542, 567-577, 607, 614, 635, 636, 640-645, 683-688, 702, 726-731 (see also territorial behavior) stabilizing density, 691-696, 731 succession, 569, 570, 681, 682 surplus birds, 475, 534, 578-583, 635, 640-642, 644, 656 survival, 463, 464, 473, 490, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-635, 642, 647, 656, 679, 680, 683-685, 691, 693-703, 720, 721 (see also mortality) territorial behavior, 462, 476-479, 497-504, 517-524, 567, 579, 635, 640-645, 683-691 territory, 497, 498, 567, 640, 642. 644, 686-691 "threshold-of-security" hypothesis, 634-636 weather, 473, 481, 507, 516, 517, 527, 576, 592, 599, 602-605, 610-612, 614, 635-637, 681, 682, 691, 712 weights, 543, 545, 582-584, 708-710 "wing-clap," 475 "winter bottleneck" hypothesis, 634-637 winter cover, 550-555, 576, 718 Tetrao tetrix (see Black Grouse) T. urogallus (see Capercaillie) Tympanuchus cupiclo (see Prairie Chicken) T. phasianellus (see Sharp-tailed Grouse) White-tailed Ptarmigan (Chaps. 8, 13, 14, 15,16)
abundance, 423, 578-731 (see also density) advertisement, 274-280, 285-289, 294-298, 474-504, 567 (see also display) age, 298, 304, 427, 477-479, 493-503, 505, 528, 529, 534, 537-539, 546, 547, 555, 575, 578-581, 588, 590, 599-601, 604-609, 614, 617, 626, 627, 629, 645-648, 653-656, 681, 683-685, 704 aggressive behavior, 274, 297-299, 423, 425-427, 429-432, 434, 436, 451-454, 498, 523, 640, 641, 689, 690 alleles, 424, 427, 431-434 "alert," 294, 295 ambush, 524-527, 576, 602 antipredator behavior, 278, 285, 288, 290. 293, 294, 440-461, 471, 472, 474, 481, 482, 489-494, 499, 505-512, 517-533, 540, 542-546, 550-552, 555-567, 575-577, 581, 584-592, 686, 687, 709, 717, 728, 729 approachability, 449, 450 assortment, 574, 685 "attack," 294, 295 body condition, 581-584. 640, 681 (see also weights) breeding behavior, 270. 273-289, 439, 440. 473-504, 533-537, 575-583, 595, 640, 641, 644-656, 702, 711-18 (see also advertisement, aggressive behavior, display, female choice, mate choice, mating system, sexual selection, spacing behavior, territorial behavior) breeding success, 425-428, 432, 434, 436. 438, 459, 578, 579, 602-615, 618, 626-629, 635, 645-667, 679-686, 693, 694, 715 brood, 290-295, 298, 427-430, 432, 434, 516, 517, 537-550, 578, 579, 581, 584-586, 602-615, 618, 619, 627-629, 645-648, 653-655, 681-685, 730, 731 (see also chicks) brood-rearing behavior, 290-295, 298, 450, 454, 459, 474, 505, 508, 509, 537-548. 575, 576, 584-586, 622, 712-716 brood break-up, 548, 550 buffer species, 506, 601, 635-640, 681, 682, 693, 694 burrowing, 564-567 (see also roosting) calls, 474, 475 carrying capacity, 636
INDEX White-tailed Ptarmigan (cont'd.) "challenge," 275-278 chicks, 428, 429, 450, 516, 517, 537-548, 575, 576, 578, 584-586, 602-615, 618, 619, 626-629, 640, 645-648. 653-656, 681-685, 714, 730, 731 (see also brood) Chitty hypothesis, 423-438, 641 "churning," 288, 289 clutch, 425-428, 453, 516, 517, 575, 578, 581-592, 605, 618, 619, 622-624, 629, 681, 684, 701, 702, 730 (see also eggs) communication, 474, 475 (see also advertisement, display) competition, 285, 425-428, 436, 438, 475, 477-482, 499, 555, 567, 579, 683-685, 693 courtship, 274-280, 285-289, 534 (see also advertisement, display) cover, 271, 272, 291, 292, 441, 442, 459. 471, 505, 550-561, 564-567, 576, 587, 588, 599, 602-605, 619-622, 636, 656, 680-686, 693, 694, 710-718, 727-731 (see also habitat) cover-seeking behavior, 562-567, 576. 595 cover-space tradeoff, 726-729 "crouch-or-fly," 562, 563 cycles, 423-438, 575, 626, 632, 637-639, 683-685 decision-making behavior, 440, 497, 503, 520, 530 demography, 578-685 density, 423-438, 516, 531, 532, 567-577, 605-609, 614, 615, 629-656, 680-696, 730, 731 density-dependent selection, 681-685 density intolerant/density tolerant, 423, 425, 432-438, 451, 567-577, 680-685, 729 desertion, 505, 506, 527, 530-533, 576, 592 diet (see food) dimorphism, 494, 495, 543 disease, 615, 681, 682, 686, 704 dispersal, 290, 291, 425, 426, 436, 548, 550, 571, 579, 634-636, 640-642 dispersion, 280-285, 482, 517-524, 570-577, 640, 687-689, 704 (see also distribution, spacing behavior) display, 274-280, 285-289, 294-298, 474, 475, 477-482, 503 (see also advertisement, distraction behavior)
803
distraction behavior, 289, 293, 294, 447, 449-451, 454, 459, 471, 517-520, 542, 546-549, 576 distribution, 270, 280-285, 445, 640, 687-689 (see also dispersion, latitude, spacing behavior) docile behavior, 423, 425, 429, 432, 435-437 dominance, 297, 299 eggs, 431, 515-517, 524, 581-592, 629, 640, 679 (see also clutch) experience, 505 female choice, 427, 451-460, 472, 533-537, 683, 685, 691 (see also mate choice) fidelity, 280-291, 298, 444, 476, 477, 505, 517, 526, 527, 533, 553, 575 fire effects, 569, 570, 592, 728, 730 fitness, 288, 424, 425, 439, 440, 444, 451, 473, 489, 494, 495, 505, 533, 567-577, 584, 680 "flight-scream," 275-278, 298 flocking, 273. 289, 296-299, 545, 561-564. 576, 711 food, 282, 285, 437, 447, 457, 458, 473, 510, 514-517, 537, 539-541, 550-552, 555-561, 581, 610, 612, 614, 615, 634-637, 640, 644, 681, 683-686, 689, 690, 703, 704, 708-711, 730, 731 foraging, 288-290, 537, 539-541, 550-552, 555-561, 576, 708 genetics/genotypes, 423-438, 451, 574, 641, 681-685 "ground scream," 275-278 growth rates, 494, 495, 539, 612, 614 habitat, 271, 272, 291, 292, 296, 441, 442, 444, 447, 454, 459, 473, 483-486, 505, 507, 511, 512, 525, 526, 537, 542-545, 550-561, 565, 567-577, 587, 588, 595, 599, 602-605, 614, 619-622, 680-686, 693, 694, 711-718, 727-731 (see also cover) hatching, 290, 298, 428, 429, 440, 513-516, 539, 578, 592-609, 629, 719 heterozygous/homozygous, 429, 432, 435 home range, 280-285, 291-293, 296, 298, 440-444, 471, 472, 505-509, 520-523, 531, 542, 545, 546, 687-691, 731 (see also territory) hormones, 431, 432, 437, 454
804
INDEX
White-tailed Ptarmigan (cont'd.) hunting, 642, 653-655, 686, 696-704, 731 incubation, 288-290, 298, 431, 454, 518, 519, 524, 576 inversity, 605-609, 693, 694 latitude, 443, 496, 582-584, 592, 594, 628, 680 laying, 289, 515-519, 524, 576, 588, 589, 640 limiting factors, 679-686, 696-731 longevity, 532, 574, 575, 586-592 (see also mortality) management, 644, 686-731 mate, 285-287, 427, 451-460, 683-685 mate choice, 285-287, 427, 451-460, 472, 474, 533-537, 576, 683-685, 691 mating system, 285-290, 427, 439-461, 471, 472, 642 microtines, 681, 682 migration, 550-556, 576, 641-643 molt, 445-448, 454, 457 (see also plumage) monogamy, 285-290, 439-461, 471, 472 mortality, 425, 427, 428, 432, 434, 436, 437, 440, 451-453, 460, 532, 537-548, 550-560, 574-579, 642, 645, 647, 653656, 679, 680, 683-686, 691, 693-703, 720, 721, 730, 731 (see also survival) movements, 280-285, 291-293, 296-299, 436, 528, 529, 537-540, 545, 550-556, 614, 641-643, 655, 704 (see also dispersal, migration) natural selection, 473, 495 nest, 273, 285, 287-290, 440-446, 451-456, 471, 472, 503-533, 578, 588-609, 617, 618, 627-629, 647, 679-696, 719, 722, 730, 731 nesting behavior, 285, 287-290, 298, 440-446, 451-456, 458, 472, 474, 483-488, 503-533, 555, 567, 575, 576, 578-583, 614, 622-624, 687, 688, 704, 711-718, 726-729 nonbreeding, 278, 284, 428, 436, 475, 534, 578-583, 635, 640-642, 644, 656 numbers, 423, 578-731 (see also density, population size) nutrition, 514-517, 581-584, 614, 640, 641, 644, 681, 682 (see also food) pair bond, 285-290, 439-461, 471, 642 parental investment, 438-440, 454, 474, 533, 595, 624
phenotypes, 423, 451, 567-577, 683-685 philopatry, 280-291, 298, 444, 476, 477, 505, 517, 526, 527, 553, 555, 575 plumage, 440, 445-448, 454, 457, 548, 549 polygamy, 459, 460, 642 polymorphism, 423, 567-577, 643 population dynamics, 423-438, 578-685 population size, 423-438, 578-731 (see also density) predation, 293, 440-443, 445-447, 451-456, 461, 472, 527, 530-533, 540, 546-549, 557-560, 576, 581, 584-615, 626-630, 634, 642, 680-686, 693-695, 704, 711, 720-722, 728, 729, 731 predator control, 718-726, 730, 731 predator-cover complex, 587, 588, 619-624, 656, 680-686, 693,711-726,731 "predator switch-over" hypothesis, 436, 635, 637-640, 681, 682 predators, 290, 437, 440-443, 445-447, 451-456, 461, 471-473, 481, 506, 518-520, 525, 527, 540, 546-549, 557-560, 562-565, 576, 587-589, 592, 594-596, 606, 607, 610-616, 619-622, 626-630, 637-640, 656, 680, 681, 683-686, 689, 690, 693-696, 702, 703, 712, 718-726, 730, 731 productivity, 578, 579, 602-615, 618, 629, 635, 645-656, 679-686 (see also breeding success, brood, chicks, nest, reproductive success) renesting, 505, 506, 511, 512, 527, 530-533, 576, 595, 597-599 reproduction, 273-298, 439, 440, 473-555, 575-577, 622, 702 (see also breeding behavior, brood-rearing behavior, nest behavior) reproductive success, 425-427, 432, 434, 437, 460, 495, 503, 578, 579, 602-615, 618, 627-629, 635, 636, 645-656, 679-686, 730, 731 (see also breeding success, brood, chicks, nest, productivity) roosting, 545, 564-567, 576, 621 safe site, 475, 489-494, 534, 575, 712 self-regulation, 423-438, 635, 640-645 sex, 273, 296, 297, 304, 427, 439, 440, 460, 461, 552, 556, 611, 624-629, 642, 704, 720-722 sexual selection, 425-428,434,436,437, 439,440,451-460,494,495,498,533-537
INDEX White-tailed Ptarmigan (cont'd.) "sexy-son" hypothesis, 451 snow, 281-284, 290, 445, 446, 512-515, 550, 557-560, 564-567, 576, 584, 621, 711 space, 686, 726-731 space-cover tradeoff, 694, 696, 726-731 spacing behavior, 273-290, 297-299, 423, 428, 434, 441, 443-446, 471, 472, 482, 505, 517-524, 542. 567-577, 607, 614, 635, 636, 640-645, 683-688, 702, 726, 731 (see also territorial behavior) stabilizing density, 691-696, 731 surplus birds, 278, 284, 428, 436, 475, 578-583, 635, 640-642, 644, 656 survival, 425, 427, 428, 451-453, 460, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-635, 642, 647, 655, 656, 679, 680, 683-685, 691, 693-703, 720, 721 (see also mortality) territorial behavior, 273-290. 297-299, 428, 440-446, 452, 453-458, 460, 471, 472, 476-479, 497-504, 517-524, 567, 579, 635, 640-645, 683-691 territory, 273-290, 297-299, 427, 429, 430, 432, 434, 440-446, 451-457, 497, 498, 567, 640, 642, 644, 655, 686-691 "threshold-of-security" hypothesis, 634-636 "threat-posture," 275-277 weather, 280, 289, 298, 430, 432, 433, 473, 481, 507, 516, 517, 527, 576, 592, 599, 602-605, 610-612, 614, 635-637. 681, 682, 691, 712 weights, 457, 543, 545, 582-584, 708-710 "winter bottleneck" hypothesis, 634-637 winter cover, 550-555, 576, 718 Willow Ptarmigan (Chaps. 10, 11, 12, 13, 14, 15, 16) abundance, 330, 333-337, 360, 370-371, 374-379, 381-419, 423, 578-731 (see also density, population size) advertisement, 357, 358, 474, 504, 567, 575, 576, 659 (see also display) age, 336, 355, 356, 360-362, 364, 374, 377, 378, 382-392, 395, 397, 403, 404, 406-419, 427, 477-479, 493-503, 505, 528, 529, 534, 537-539, 546, 547, 555, 575, 578-581, 588, 590, 599-601, 604-609, 614, 617, 626, 627, 629,
805
645-649, 653-656, 659-665, 667, 669, 681, 683-685, 704 aggressive behavior, 360-372, 377, 378, 415, 423, 425-427, 429-432, 434, 436, 451-454, 494, 498, 523, 640, 641, 665-671, 677, 679, 689, 690 alleles, 424, 427, 431-434, 665 (see also genetics/genotypes) ambush, 524-527, 576, 602 antipredator behavior, 340-356, 371, 373, 377, 378, 440-461, 471, 472, 474, 481, 482, 489-494, 499, 505-512, 517-533, 540, 542-546, 550-552, 555-567, 575-577, 581, 584-592, 686, 687, 709, 717, 728, 729 approachability, 383, 384, 404-408, 449, 450 assortment, 574, 666-672, 685 body condition, 333, 334, 340-342, 368, 369, 371-378, 394-397, 581-584, 640, 641, 664, 681 (see also weights) breeding behavior, 330, 355-372, 377, 378, 439, 440, 473-504, 533-537, 575-583, 595, 640, 641, 644-656, 679-686, 693, 694, 702, 711-718 (see also advertisement, aggressive behavior, display, female choice, mate choice, mating system, sexual selection, spacing behavior, territorial behavior) breeding success, 331, 335-337, 372, 377, 378, 392-414, 417, 418, 425-427, 432, 434, 436, 438, 459, 578, 579, 602-615, 618, 626-629, 635, 645-673, 677, 715 brood, 377, 382, 401, 402, 406-409, 411, 412, 417, 418, 427-430, 432, 434, 516, 517, 537-550, 578, 579, 581, 584-586, 602-615, 618, 619, 627-629, 645-648, 653-656, 659-667, 669, 670, 672, 673, 730, 731 (see also chicks) brood break-up, 548, 550 brood-rearing behavior, 383, 384, 406, 407, 450, 454, 459, 474, 505, 508, 509, 537-548, 577, 576, 584-586, 622, 674, 712-716 buffer species, 506, 601, 635-640, 674, 681, 682, 693, 694 burrowing, 333, 340, 345-347, 349-350, 355, 356, 373, 377, 378, 564-567 (see also roosting) calls, 357, 358, 474, 475, 659, 664
806
INDEX
Willow Ptarmigan (confd.) carrying capacity, 636 chicks, 335-337, 377, 378, 382, 383, 395-403, 406-409, 411, 412, 414, 417, 418 428, 429, 450, 516, 517, 537-548, 575, 576, 578, 584-586, 602-615, 618, 619, 626-629, 640, 645-648, 653-656, 659-667, 669, 670, 672, 673, 681-685, 714, 730, 731 (see also brood) Chitty hypothesis, 374, 413, 423-438, 641, 657, 658, 670, 678 clutch, 382, 393-398, 403, 404, 406-411, 413, 417, 418, 426-428, 453, 516, 517, 575, 578, 581-592, 605, 618, 619, 622-624, 629, 660, 662, 664, 673, 681, 684, 701, 702, 730 (see also eggs) communication, 474, 475 (see also advertisement, display) competition, 355-373, 377, 378, 425-428, 436, 438, 475, 477-482, 499, 555, 567, 579, 659, 665, 670, 674, 677-679, 683-685, 693 courtship, 534 (see also advertisement, display) cover, 331-333, 338, 348, 351-354, 373, 377, 378, 383, 394, 397, 400, 404, 406, 410, 411, 441, 442, 459, 471, 505, 550-561, 564-567, 576, 587, 588, 599, 602-605, 619-622, 636, 656, 669, 672-677, 680-686, 693, 694, 710-718, 727-731 (see also habitat) cover-seeking behavior, 562-567, 576, 595 cover-space tradeoff, 726-729 "crouch-or-fly," 562, 563 cycles, 330, 333-337, 372, 374-378, 379, 408, 416, 423-438, 575, 626, 632, 637-639, 656-679 decision-making behavior, 440, 497, 503, 520, 530 demography, 330, 333-337, 355-372, 374-379, 381-419, 578-685 density, 330, 333-337, 345, 360, 361, 370-372, 374-379, 381, 384, 385, 388-391, 393, 397, 398, 414-419, 423-438, 516, 531, 532, 567-577, 605-609, 614, 615, 629-696, 730, 731 density-dependent selection, 657-679, 681-685 density intolerant/density tolerant, 423, 425, 432-438, 567, 577, 664-672, 679-685, 729, (see also phenotypes)
desertion, 505, 506, 527, 530-533, 576, 592 diet (see food) dimorphism, 494, 495, 543 disease, 382, 385, 615, 676, 681, 682, 686, 704 dispersal, 425, 426, 436, 548, 550, 571, 579, 634-636, 640, 641, 642, 659 dispersion, 312, 322, 323, 328, 363, 367, 482, 517-524, 570-577, 640, 670, 687-689, 704 (see also distribution, spacing behavior) display, 357, 358, 474, 475, 477-482, 503, 664 (see also advertisement, distraction behavior) distraction behavior, 447, 449-451, 454, 459, 471, 517-520, 542, 546-549, 576, 659, 664 distribution, 363, 367, 445, 640, 658, 687-689 (see also dispersion, latitude, spacing behavior) docile behavior, 423, 425, 429, 432, 435-437, 665-670, 679 dominance, 665 eggs, 383, 393-401, 403, 404, 406-411, 416-419, 431, 515-517, 524, 581-592, 629, 640, 679 (see also clutch) experience, 505 fecundity, 425-427, 432, 434, 437 female choice, 427, 451-460, 472, 533-537, 666-672, 677, 679, 683, 685, 691 (see also mate choice) fidelity, 444, 476, 477, 505, 517, 526, 527, 553, 555, 575 fire effects, 569, 570, 592, 678, 728, 730 fitness, 424, 425, 439, 440, 444, 451, 473, 489, 494, 495, 505, 533, 567-577, 584, 679, 680 flocking, 345-347, 354-356, 361-370, 373, 377, 378, 545, 561-564, 576, 711 food, 331, 333, 337-340, 344, 352, 355, 356, 372, 373, 375, 377, 378, 383, 402, 404, 406, 408, 409, 411, 412, 414, 417, 418, 437, 444, 457, 458, 473, 510, 514-517, 537, 539-541, 550-552, 555-561, 581, 610, 612, 614, 615, 634-637, 640, 644, 676, 681, 683-686, 689, 690, 703, 704, 708-711, 730, 731 foraging, 331, 333, 337-340, 348, 349, 355, 356, 373, 377, 378, 411, 412, 466, 537, 539-541, 550-552, 555-561, 576, 708
INDEX Willow Ptarmigan (cont'd.) genetics/genotypes, 412-415, 423-438, 451, 567-571, 574, 641, 657-659, 663, 665-672, 674, 677-679, 681685 growth rates, 411, 494, 495, 539, 612, 614, 677 habitat, 331-333, 348, 352, 353, 366-368, 373, 377, 378, 381-383, 394, 397, 400, 401, 404, 406, 410-412, 417, 418, 441, 442, 444, 447, 454, 459, 494, 511, 512, 525, 526, 537, 542-545, 550-561, 565, 567-577, 587, 588, 595, 599, 602-605, 614, 619-622, 657, 658, 669, 672-677, 680-686, 693, 694, 711-718, 727-731 (see also cover) hatching, 381-383, 395-401, 403, 404, 406-408, 411, 417, 418, 428, 429, 440, 513-516, 539, 578, 592-609, 629, 664, 677, 719 heterozygous/homozygous, 429, 432, 435, 574, 659, 665, 670, 679 home range, 440-444, 471, 472, 505-509, 520-523, 531, 542, 545, 546, 687-691, 731 (see also territory) hormones, 431, 432, 437, 454 hunting, 381, 642, 653-655, 658, 659, 663, 676, 678, 686, 696-704, 731 incubation, 382-384, 406, 407, 431, 454, 518, 519, 524, 576 inversity, 605-609, 693, 694 latitude, 443, 496, 582-584, 592, 594, 628, 657, 658, 672-676, 680 laying, 364, 381, 382, 393, 394, 397, 408-410, 413, 417, 418, 515-519, 524, 576, 588, 589, 640, 677 limiting factors, 679-686, 696-731 longevity, 532, 574, 575, 586-592 (see also mortality) management, 644, 686-731 mate, 386, 427, 451-460, 666-672, 677, 679, 683-685 mate choice, 371, 372, 427, 451-460, 472, 474, 533-537, 576, 666-672, 677, 679, 683-685, 691 mating system, 363, 366, 371, 372, 377, 378, 386, 427, 439-461, 471, 472, 642, 665-672 microtines, 379, 381, 385-389, 393-396, 398, 399, 401, 404-411, 413, 414, 416-419, 436, 681, 682
807
migration, 339, 348-356, 373, 393, 550-556, 576, 641-643, 670 molt, 445-448, 454, 457 (see also plumage) monogamy, 363, 366, 386, 439-461, 471, 472, 666 mortality, 330, 331, 335-337, 345-356, 368-370, 376-378, 382-390, 393-404, 406-411, 415, 417, 418, 440, 425, 427, 432, 434, 436, 437, 450-453, 460, 532, 537-548, 550-560, 574-579, 581, 586-592, 609-636, 642, 645, 647, 653-656, 678-680, 683-686, 691, 693-703, 720, 721, 730, 731 (see also survival) movements, 331, 337-339, 348-354, 356, 358, 359, 373, 377, 378, 393, 411, 436, 528, 529, 537-540, 545, 546, 550-556, 614, 641-643, 655, 670, 704 (see also dispersal, migration) natural selection, 473, 495 nest, 382, 393-401, 403, 406-411, 417, 418, 440-446, 451-456, 471, 472, 503-533, 578, 588-609, 617, 618, 627-629, 647, 659, 661, 662, 672-696, 719, 722, 730, 731 nesting behavior, 364, 377, 378, 393, 394, 397, 400, 440-446, 451-456, 458, 472, 474, 483-488, 503-533, 555, 567, 575, 576, 578-583, 614, 622-624, 666, 687, 688, 704, 711-718, 726-729 nonbreeding, 428, 436, 475, 534, 578-583, 635, 640-642, 644 (see also surplus birds) numbers, 330, 333-337, 360, 368-372, 374-379, 381-419, 423, 578-731 (see also density, population size) nutrition, 337, 369, 377, 378, 408-410, 412, 514-517, 581-584, 614, 640, 641, 641, 644, 677, 681, 682 (see also food) pair bond, 363, 366, 386, 439-461, 471, 472, 666 (see also mating system) parental investment, 383, 384, 404-408, 438-440, 454, 477, 533, 595, 624, 667 phenotypes, 412, 413, 415, 423, 451, 567-577, 659, 664-672, 677, 683-685 philopatry, 444, 476, 477, 505, 517, 526, 527, 553, 555, 575 plumage, 440, 445-448, 454, 457, 548, 549 polygamy/polygyny, 363, 366, 386, 459, 460, 642, 670 (see also mating system)
808
INDEX
Willow Ptarmigan (cont'd.) polymorphism, 423, 567-577, 643, 659, 664, 679 population dynamics, 330, 333-337, 355-379, 381-419, 423-438, 578-685 population size, 330, 333-337, 355-379, 381-419, 423-438, 578-696, 730, 731 (see also density) predation, 331, 337, 344, 356, 368-370, 373-375, 377, 378, 382, 386, 389, 393, 398-400, 409-411, 414, 415, 417, 418, 440-443, 445-447, 461, 472, 527, 530-533, 540, 546-549, 557-560, 576, 581, 584-615, 626-630, 634-640, 642, 666, 672-686, 693-695, 704, 711, 720-722, 728, 729, 731 predator control, 718-726, 730, 731 predator-cover complex, 587, 588, 619-624, 656, 680-686, 693, 711-726, 731 "predator switch-over" hypothesis, 375, 436, 635, 637-640, 681, 682 predators, 331, 333, 340, 342-344, 347, 348, 356, 368-370, 373-375, 377, 378, 386, 398-400, 410, 411, 417, 418, 437, 440-443, 445-447, 451-456, 461, 471-473, 481, 506, 518-520, 525, 527, 540, 546-549, 557-560, 562-565, 576, 592, 594-596, 606, 607, 610-616, 619-622, 626-630, 637-640, 656, 670-681, 702, 703, 712, 718-726, 730, 731 productivity, 335-337, 372, 377, 378, 384-419, 578, 579, 602-615, 618, 629, 635, 645-666, 673, 679-686 (see also breeding success, brood, chicks, nest, reproductive success) renesting, 505, 506, 511, 512, 527, 530-533, 576, 595, 597-599, 659 reproduction, 353-372, 377, 378, 439, 440, 473-555, 575-577, 622, 702 (see also breeding behavior, brood-rearing behavior, nesting behavior) reproductive success, 335-337, 372, 377, 378, 384-419, 425-427, 432, 434, 437, 460, 495, 503, 578, 579, 602-615, 618, 627-629, 635, 636, 645-666, 673, 679-686, 730, 731 (see also breeding success, brood, chicks, nest) roosting, 333, 340, 345-347, 349, 350,
356, 373, 377, 378, 545, 564-567, 576, 621 (see also burrowing) safe site, 475, 489-494, 534, 575, 712 self-regulation, 338, 355, 370, 371, 374, 377, 378, 423-438, 635, 640-645 sex, 354-356, 362, 364, 377, 378, 386, 427, 439, 440, 460, 461, 552, 556, 576, 611, 624-629, 642, 664, 666, 667, 674, 676, 704, 720-722 sexual selection, 355-372, 377, 378, 425-428, 434, 436, 437, 439, 440, 451-460, 494, 495, 498, 533-537, 666-672, 677, 679 "sexy-son" hypothesis, 451 snow, 333, 337, 339, 340, 351, 377, 378, 393, 394, 397, 400, 401, 408, 410, 417, 418, 445, 446, 512-515, 550, 557-560, 564-567, 576, 584, 621, 711 space, 686, 726-731 space-cover tradeoff, 372, 674, 694, 696, 726-731 spacing behavior, 338, 355-372, 374, 377, 378, 415, 417, 418, 423, 428, 434, 441, 443-446, 471, 472, 482, 505, 517-524, 542, 567-577, 607, 614, 635, 636, 640-645, 659, 663-672, 674, 679, 683-688, 702, 726-731 (see also territorial behavior) stabilizing density, 691-696, 731 surplus birds, 369, 374, 377, 378, 415, 428, 436, 578-583, 635, 640, 641, 656 (see also nonbreeding) survival, 330, 331, 335-337, 345-356, 368-370, 372, 373, 376-378, 382-390, 393-404, 406, 409, 411, 415, 417, 418, 425, 427, 428, 451-453, 460, 532, 537-548, 550-567, 574-579, 581, 586-592, 609-635, 642, 647, 655, 656, 660, 662, 664-667, 669, 672, 678-680, 683-685, 691, 693-703, 720, 721 (see also mortality) ten-year cycle, 330, 372, 374-378, 423-438, 656-679 territorial behavior, 331, 338, 355-374, 377, 378, 381, 415, 428, 440-446, 452-458, 460, 471, 472, 517-524, 567, 579, 635, 640-645, 655, 664, 659, 666, 668, 669, 683-691 territory, 333, 363, 367, 381, 415, 427, 429, 430, 432, 434, 440-446, 451-457,
INDEX Willow Ptarmigan (cont'd.) 497, 498, 567, 640, 642, 644, 655, 666, 668, 669, 686-691 "threshold-of-security" hypothesis, 634-636 "waiting flock," 361-370, 374, 377, 378, 427, 428, 436 weather, 331, 334, 337, 338, 340, 344, 346, 347, 348, 356, 360, 377, 378, 393, 394, 397, 400, 401, 409, 411, 414, 417, 418, 430, 432, 433, 473, 478, 479, 481,
809
507, 516, 517, 527, 576, 592, 599, 602-605, 610-612, 614, 635-637, 659, 665, 674, 676-679, 681, 682, 691, 712 weights, 333, 340-342, 368, 369, 372-378, 393-397, 402, 417, 418, 457, 543, 545, 582-584, 659, 664, 708-710 "winter bottleneck" hypothesis, 634-637 winter cover, 331-333, 338, 348, 351-354, 377, 550-555, 576, 718
Arthur T. Bergerud is an adjunct full-professor in the department of biology at the University of Victoria, British Columbia, where he has taught since 1967. Previously, he served as director of management for the Newfoundland Government's Wildlife division. Bergerud received his master's degree from the University of British Columbia. Some of the jounals to which he contributes are Journal of Wildlife Management, Canadian Journal of Zoology, Arctic, Auk, Oikos, and Oecologia. Michael W. Gratson is currently working on his Ph.D. in zoology at the University of Victoria. He received his bachelor's and master's degrees in wildlife and biology from the University of Wisconsin, Stevens Point, in 1983.
