Advances in Marine Biology, Volume 26

Advances in MARINE BIOLOGY VOLUME 26 This Page Intentionally Left Blank Advances in MARINE BIOLOGY VOLUME 26 Edit...

Author: Academic Press

29 downloads 1736 Views 16MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Advances in

MARINE BIOLOGY VOLUME 26

This Page Intentionally Left Blank

Advances in

MARINE BIOLOGY VOLUME 26 Edited by

J. H. S. BLAXTER Dunst afnage Marine Research Laboratory, Oban, Scotland and

A. J. SOUTHWARD The Laboratory, Citadel Hill, Plymouth, England

Academic Press Harcourt Brace Jovanovich, Publishers London San Diego New York Boston Sydney Tokyo Toronto

ACADEMIC PRESS LIMITED 24/28 Oval Road London NWI 7DX United States Edition published by ACADEMIC PRESS INC. San Diego, CA 92101

Copyright 0 1990 by ACADEMIC PRESS LIMITED All Rights Reserved

No part of this book may be reproduced in any form by photostat, microfilm, or any other means, without written permission from the publishers British Library Cataloguing in Publication Data Advances in marine biology.-Vol. 26 1. Marine biology-Periodicals QH91.AI 574.92’05 ISBN 0-12-026126-X ISSN 0065-2881

Filmset by Bath Typesetting Limited and printed in Great Britain at T. J. Press (Padstow) Ltd, Cornwall

CONTRIBUTORS TO VOLUME 26 R. 0. BRINKHURST, Ocean Ecology Laboratory, Institute of Ocean Sciences, Sidney, British Columbia, Canada.

B. J. BURD,Galatea Research Inc., Brentwood Bay.

D. H. CUSHING,I98 Yarmouth Road, Lowestoft, Sufsolk, U.K. NR32 4AB. T. HAUG, Department of Marine Biology, Tromse Museum, University of Tromse, Norway. I. HOLMEFJORD, The Agricultural Research Council of Norway, Institute of Aquaculture Research, Sunndalsera, Norway. E. KJBRSVIK, Norwegian College of Fishery Science, University of Tromse, Tromse.

A. MANGOR-JENSEN, Institute of Marine Research, Austevoll Aquaculture Research Station. Storebe. J. D. NEILSON, Marine Fish Division, Canada Department of Fisheries and Oceans, Biological Station, St. Andrews, New Brunswick, Canada.

A. NEMEC, International Statistics and Research Corp., Brentwood Bay. R. I. PERRY,Marine Fish Division, Canada Department of Fisheries and Oceans, Biological Station, St. Andrews. New Brunswick, Canada.

V

This Page Intentionally Left Blank

CONTENTS CONTRIBUTORS TO VOLUME 26

..

..

..

..

..

V

Biology of the Atlantic Halibut, Hippoglossus hippoglossus (L., 1758) TOREHAUG

I. 11.

111.

IV. V.

VI.

VII.

v

.. .. Introduction .. .. .. Identity . . .. .. .. A. Taxonomic status .. .. B. Subspecies/races ,. .. C . Morphology .. .. .. Distribution .. .. .. A. Total area .. B. Differential distribution .. Reproduction . . .. .. A. Sexual maturity .. .. B. Spawning .. .. .. Pelagic Phase .. .. .. A. Embryonic . , .. .. .B. Larvae .. .. .. Immature Phase .. .. A. Occurrence and migrations . . B. Age/size composition and sex ratio C . Growth .. .. .. D. Feeding .. .. .. Mature Phase . . .. .. A. Occurrence and migrations . . B. Sex ratio and age/size composition C . Longevity and growth .. D. Feeding .. .. .. E. Proximate body composition F. Energy economics . . .. Parasites, Diseases and Pollution A. Parasites .. .. .. B. Diseases .. .. .. C . Pollution .. .. .. vii

..

..

.. ..

..

..

.. .. .. ..

..

.. ..

.. .. .. .. ..

.. .. ..

..

.. ..

..

..

..

..

..

.. ..

..

.. ..

.. .. .. .. .. ..

.. .. ..

.. .. .. .. ..

.. .. .. .. .. ..

.. .. .. .. .. .. .

1

.*

.. ..

..

..

..

.. ..

.. ..

..

..

..

.. .. .. ..

.. ..

.. .. .. ..

..

..

.. .. ..

.. .. ..

.. .. .. .. ..

.. .. ..

.. .. .. .. .. .. .. .. .. .. .. .. .. .. ..

..

2 3 3 3 4 5 5 6 7 7 9 11 11 13 22 22 23 25 26 27 27 29 31 33 35 37 38 38 41 42

viii

CONTENTS

.. .. IX. Exploitation .. *. A. Development in total catches * . B. Catch by area .. .. ,. .. .. .. C. Catch by nations D.Over-exploitation . . .. .. * . E. Recruitment and mortality . . F. Management .. .. .. X. Rearing Experiments and Aquaculture A. The first experiments .. *. .. B. Brood stocks of adult spawners C. Egg incubation and hatching .. D.Rearing of yolk-sac larvae . . .. E. Start-feeding of larvae .. .. F. Growth experiments with juvenile halibut XI. Acknowledgements .. .. .. .. ,. .. .. XII. References

..

..

.. ..

..

.. .. .. .. .. .. .. .. .. .. .. ..

..

.. .. ..

..

., .. .. .. .. ..

.. .. .. ..

.. .. .. .. .. .. .. .. .. .. . .. .. .. .. .. *

42 42 45 47 49 50

51 53 53 54 55 57 60 62 63 63

Egg Quality in Fishes E. KIBRSVIK, A. MANGOR-JENSEN AND I. HOLMEFJORD

..

..

11. Egg Quality Characteristics .. .. A. Fertilization processes .. .. B. Physical and physiological properties .. .. C. Morphology . . .. .. .. D. Egg size .. .. E. Chemical content .. .. .. F. Chromosomal aberrations . . .. .. .. 111. Egg Quality of Wild Fish IV. Factors of Importance for Egg Quality . . .. A. Overripening and storage of eggs .. .. B. Broodstock management V. Conclusions and Recommendations .. .. .. .. Acknowledgements .. .. References .. ..

,.

I. Introduction

..

..

..

*.

.. ..

.. .. .. .. .. ,.

.. .. .. .

I

..

..

.. ..

..

.. .. ..

.. .. ..

.. ..

..

..

.. ..

.. ..

.. .. .. ..

.. ..

..

.. .. .. ..

71 72 72 76 76 78 79 85 87 91 91 96 103 105 105

CONTENTS

1x

Die1 Vertical Migrations of Marine Fishes: an Obligate or Facultative Process? J. D. NEILSON AND R. I. PERRY I. Introduction .. .. .. .. .. .. .. 11. The Evidence for Endogenous Rhythms .. .. .. A. What cyclic events can be identified as zeitgebers controlling vertical migrations? . . .. .. .. .. .. B. Do variations in local environment affect vertical position? C. Do ontogenetic variations in vertical migratory behaviour occur, and are they consistent with chronobiological theory? 111. Discussion .. .. .. .. .. .. .. IV. Acknowledgements .. .. .. .. .. .. V. References .. .. .. .. .. .. ..

115 122 122 145 148

150 155 156

The Development and Application of Analytical Methods in Benthic Marine lnfaunal Studies BRENDAJ . BURD.AMANDA NEMEC AND RALPH0. BRINKHURST

.. .. .. I. Introduction .. .. 11. Collection of Data .. .. .. .. A. Sampling devices .. .. .. .. B. Sieving of samples .. .. .. .. C. Sampling effort .. .. .. .. D. Temporal sampling design . . .. .. 111. Analysing the Data Matrix .. .. .. A. The data matrix .. .. .. ,. B. The subjective approach: Community concepts C . Descriptive univariate community indices . . D. Statistical inference . . .. .. .. E. Multivariate data analyses . . .. .. F. Time-series analysis . . .. .. .. IV. Summary .. .. *. .. .. V. References .. .. .. .. ..

.. .. .. ..

..

.. ..

..

..

.. *.

..

*. .. ..

..

.. .. .. ..

..

.. ..

..

.. .. .. .. .. ..

170 172

173 174 175 177 178 178 186

192 212 214

231 232 234

CONTENTS

X

Plankton Production and Year-class Strength in Fish Populations: an Update of the Match/Mismatch Hypothesis D. H. CUSHING

I.

Introduction

..

..

..

11. The Original Hypothesis .. 111. The Member/Vagrant Hypothesis

..

.. ..

..

.. ..

..

.. ..

.. ..

..

IV. Extension to Waters Equatorward of 40" Latitude . . .. .. .. .. A. Background .. .. .. .. .. .. .. B. Events in upwelling areas . . .. .. .. .. .. C. Lasker events .. .. *. .. .. D. Spawning and food . . .. .. .. .. .. .. V. Testing the Hypothesis . . .. .. .. .. .. VI. Some Examples .. .. .. .. A. The original approach .. .. B. The spawning distributions of seven species in the Southern .. .. .. .. Bight of the North Sea .. .. C. The link between recruitment and larval abundance .. .. .. .. D. The Baltic . . .. .. .. .. .. E. The gadoid outburst .. .. .. .. F. The great salinity anomaly of the 1970s . . G. Catches of spiny lobsters and the time of onset of the spring outburst off Tasmania .. .. .. .. H. Larvae of the Arcto-Norwegian cod .. .. .. .. *. .. .. .. VII. Discussion .. .. .. .. .. .. VIII. Acknowledgements .. .. ,. .. .. .. .. .. IX. References .. I

Taxonomic Index

..

..

..

..

Subject Index

..

..

..

..

..

..

..

..

..

Cumulative Index of Titles Cumulative Index of Authors

..

,

250 25 I 256 259 259 259 26 1 262 264 265 265 270 271 273 274 276 278 279 281 285 285

..

295

..

299

. . 309 ..

..

313

Biology of the Atlantic Halibut, Hipp oglossus hipp oglossus (L., 1758) Tore Haug Department of Marine Biology, Tromss Museum, University of Tromso, Tromss, Norway

I. 11.

111.

IV.

V. VI.

VII.

.. Introduction .. .. .. Identity . . .. .. .. A. Taxonomic status . . .. B. Subspeciesiraces .. .. C. Morphology .. Distribution .. .. .. A. Total area . . .. .. .. B. Differential distribution .. Reproduction ... .. .. A. Sexual maturity .. B. Spawning . . .. .. .. Pelagic Phase .. .. .. A. Embryonic . . .. .. B. Larvae .. .. .. Immature Phase . . .. A. Occurrence and migrations . . B. Age/size composition and sex ratio C. Growth .. .. .. D. Feeding .. .. .. Mature Phase . . .. .. A. Occurrence and migrations . . B. Sex ratio and age/size composition C. Longevity and growth .. D. Feeding .. .. ..

.. .. .. .. ..

..

..

..

..

..

.. ..

.. ..

..

.. .. .. ..

..

I

.

..

..

..

..

.. .. ..

.. .. ..

..

..

..

..

..

..

.. ..

.. .. ..

.. .. ..

..

..

.. ..

.. .. ..

..

..

..

.. .. .. .. .. ..

.. .. .. ..

.. ..

..

..

..

..

..

.. ..

..

.. .. .. ..

.. .. ..

.. .. ..

.. ..

..

~.

.. ..

..

.. .. .. ..

..

.. ..

..

.. ..

.. .. ..

.. ..

..

..

..

..

.. ..

..

.. ..

..

.. ..

.. ..

..

..

..

.. .. .. .. .. ..

.. ..

.. ..

.. .. ..

2 3 3 3 4 5 5 6 7 1

9 11 11 13 22 22 23 25 26 21 21 29 31 33

Copyright 0 1990 6’, Acudemic Press Limirrd 411 rights of reproduction in any form reserved

ADVANCES IN MARINE BIOLOGY VOLUME 26 ISBN 0-12-026126-X

1

2

T. HAUG

E. Proximate body composition .. F. Energy economics . . .. .. VIII. Parasites, Diseases and Pollution .. A. Parasites .. .. .. .. B. Diseases .. .. .. .. C. Pollution .. .. .. .. IX. Exploitation .. .. .. .. A. Development in total catches .. B. Catch by area .. .. .. C. Catch by nations . . .. .. D. Over-exploitation . . .. .. E. Recruitment and mortality . . .. F. Management .. .. .. X. Rearing Experiments and Aquaculture . . A. The first experiments .. .. B. Brood stocks of adult spawners .. C. Egg incubation and hatching .. D. Rearing of yolk-sac larvae . . .. E. Start-feeding of larvae .. .. F. Growth experiments with juvenile halibut XI. Acknowledgements .. .. .. XII. References .. .. .. ..

.. .. .. ..

.. ..

..

..

.. ..

.. .. .. .. ..

.. .. .. ..

.. ..

..

.. ..

.. .. .. .. ..

.. .. .. .. ..

..

..

.. ..

..

.. .. .. .. ..

.. ..

.. .. ..

..

.. ..

..

..

..

..

.. .. .. ..

..

.. .. ..

..

..

.. .. .. ..

.. .. .. ..

..

.. ..

..

..

.. ..

.. ..

.. ..

..

..

..

.. .. ..

..

.. ..

..

.. .. .. ..

.. .. .. .. ..

.. ..

35 37 38 38 41 42 42 42 45 47 49 50 51 53 53 54 55 57 60 62 63 63

I. Introduction The Atlantic and Pacific halibut (Hippoglossus hippoglossus and H . stenolepis, respectively)are two morphologically similar flatfish forms inhabiting the boreal and subarctic waters in their respective oceans. Until recently, there were disagreements as to the taxonomic status of these two forms, which had carried the same scientific name until Schmidt (1930) described halibut taken off Sakhalin Island as the separate species H . stenolepis. He compared scale morphology, fin ray numbers and body shape between the Sakhalin form and that of the Pacific part of North America, and concluded that, although identical to each other, they were different from the Atlantic form which he considered a separate species. This was, however, not confirmed by Vernidub (1936) who, based on a larger sample, only gave subspecies status to the Pacific halibut which she named H . hippoglossus stenolepis. In his encyclopaedic documentary on the Pacific halibut, Bell (1981) went even further and concluded, from a critical review of existing published and his own unpublished morphological data, that there was very little basis for designating the Pacific halibut as H . stenolepis, the “narrow-scaled” halibut (steno = narrow, lepis = scale), a separate species from the Atlantic halibut. More recent studies, however, using genetically determined electrophoretically detectable protein variants to test for genetic differences, have

BIOLOGY OF THE ATLANTIC HALIBUT

3

revealed a genetic difference between Atlantic and Pacific halibut of a magnitude that confirms the treatment of the two taxa as separate species (Grant et al., 1984). Pleuronectiform flatfishes are thought to have evolved in the Pacific Ocean in the early Tertiary period because of their greater present-day diversity in the Pacific Ocean (Nikolskii, 1954). Based on calculations of Nei’s genetic distance and the molecular clock hypothesis, Grant et al. (1984) suggested that the ancestors of Atlantic halibut invaded the North Atlantic from the Pacific through the Bering Strait in the Pliocene between 4.5 and 1.7 million years ago, and that the two forms have since been reproductively isolated from one another. In this review I regard the Pacific halibut as a species separate from the Atlantic halibut, and the following is an account of the biology, fishery and potential for aquaculture of the Atlantic species.

II. Identity A.

Taxonomic Status

According to a recent classification given of the Pleuronectiformes (flatfish) by Ahlstrerm et al. (1984) and Hensley and Ahlstrerm (1984), the genus Hippoglossus belongs to the subfamily Pleuronectinae within the family Pleuronectidae, the latter being one of the five families of the suborder Pleuronectoidei. Though there has been unanimity that the genus Hippoglossus belongs to the family Pleuronectidae, there are several versions of the generic and specific names for the halibut. While the present valid name of the Atlantic halibut is Hippoglossus hippoglossus (L., I758), Andriyashev ( 1 954) summarizes the following objective synonyms used: Pleuronectes hippoglossus L., 1758 Hippoglossus vulgaris Flemming, 1828 Hippoglossus hippoglossus Jordan and Everman, 1898 Hippoglossus americanus Gill, I864

B. Subspeciesf Races Tagging experiments in Norwegian waters have revealed that mature halibut show a remarkable homing response, returning to the same area for repeated spawning (Devold, 1938, 1943; God0 and Haug, 1988a). Biochemical genetic studies give no convincing evidence, however, of genetically distinct breeding units, but rather a general impression that the North Atlantic population of

4

T. HAUG

halibut is very homogeneous in terms of genetic variation, even when geographically distinct areas (Norwegian coast, Faroes, Greenland; see Fig. 1) are considered (Mork and Haug, 1983; Haug and Fevolden, 1986; Fevolden and Haug, 1988). It is suggested by these workers that migrations, especially of young fish, and egg/larval dispersal may cause gene flow that contributes to the maintenance of low-level genetic differentiation between stocks.

GREENLAND SQRBYSUND MALANGEN

FIG.1. Cluster dendrogram based on Nei’s unbiased genetic identity coefficients for comparison of four polymorphic gene loci in Atlantic halibut from sampling sites on the coast of Norway (Sereysund, Malangen and Folla), and in Greenland and Faroese waters. The genetic identity values indicate a very homogeneous population. Reproduced with permission from Fevolden and Haug (1988).

The genetic homogeneity observed among halibut from different areas is also supported from studies of morphological characters. Meristic as well as morphometric parameters do not discriminate between fish from different areas (McCracken, 1958; Haug and Fevolden, 1986). Apparently, there is no evidence that subspecies or races of halibut exist in the North Atlantic.

C. Morphology Andriyashev (1954) gives the following description of the genus Hippoglossus:

BIOLOGY OF THE ATLANTIC HALIBUT

5

Body elongate, covered with small cycloid scales and smaller supplementary scales. Lateral line with a steep bend above pectoral fin. Both eyes on the right side of head. Jaws large, symmetrical with large pointed teeth directed posteriorly; teeth on upper jaw in two rows and on lower jaw in one row. Vomer toothless. Infrapharyngeal teeth sharp, in two rows; the inner series with enlarged teeth. Anal spine present (in adult specimens overgrown by skin). Pectoral fin better developed on eye side. Caudal fin weakly emarginate. Vertebrae 49-53.

Furthermore, Andriyashev (1954) states that in H. hippoglossus the main body scales are surrounded by a ring of numerous small scales, while the colour of the eyed side is dark brown, darker in adult specimens and lighter in young ones. The blind side is usually white, although on very rare occasions ambicoloration of various amounts, in extreme cases with incomplete rotation of eye and hooked dorsal fin, have been observed (Gudger and Firth, 1937). In some very rare cases, the general rule of dextral pigmentation/eye localization is broken and reversed sinistral specimens may occur. After 6-8 years of apparently uniform growth between the sexes, female halibut start to grow much more rapidly than males (Jespersen, 1917; Devold, 1938; Joensen, 1954; McCracken, 1958; Mathisen and Olsen, 1968; Jakupsstovu and Haug, 1988). Although some statistically significant differences have been confirmed by univariate and multivariate analysis of morphometric variables, no visual morphological differences between the sexes is evident (Haug and Fevolden, 1986). Haug and Fevolden give the following relative sizes (percentages of total length) for various halibut body proportions: head length, 23.1-23.6%; dorsal fin base, 73.3-74.5%; anal fin base, 50.3-51.1 %; right (eye side) pectoral fin, 11.2-12.5%; left (white side) pectoral fin, 8.6-9.6%; pelvic fins, 4.7-5.3%; upper jaw, 7.0-7.1 YO;lower jaw, 6.5-6.6%. The number of fin rays, which show no significant variation between the sexes, are 91-109 in the dorsal fin, 64-80 in the anal fin, 11-17 in the pectoral fins and 5-6 in the pelvic fins (Haug and Fevolden, 1986).

111. Distribution A.

Total Area

The halibut is distributed (Fig. 2) in parts of the Arctic Ocean and in the northern part of the Atlantic Ocean, occasionally as far south as the Bay o f Biscay and New York on the eastern and western side of the Atlantic, respectively (Andriyashev, 1954). According to Andriyashev, it is commonly found in the southwestern parts of the Barents Sea as far north as Bear Island, and occasionally on the west coast of Spitzbergen. Furthermore, it is

T. HAUG

6

particularly numerous along the Norwegian coasts, off the Faroes and Iceland, and off southern Greenland, and it may also be encountered in the North Sea and western part of the Baltic Sea. Along the east coast of North America, the halibut is distributed from Hudson Strait southward to the southern Grand Bank and St. Pierre Bank, particularly on the continental slope and in the channels running between the fishing banks (Bowering, 1986). 8WN

7WN

5WN

NORTH ATLANTIC OCEAN 4WN

30.N

709N

BWW

soow

4OoW

3CW

2ww

ICW

FIG. 2. Map showing the distribution of halibut (hatched area) in the North Atlantic.

B. Direrential Distribution Immature and mature halibut occupy different habitats. It appears that certain coastal areas (at a depth of 20-60m) may serve as nursery areas where the young halibut are stationary during their first 4-6 years (VedelTining, 1938, 1947; Sigurdsson and Fridriksson, 1952; Sigurdsson, 1956;

BIOLOGY OF THE ATLANTIC HALIBUT

7

Haug and Sundby, 1987; God0 and Haug, 1988a; Stobo et al., 1988). When leaving the nursery areas, a period of intense migration-often to very distant areas, and to shallow as well as deep waters - seems to occur, and there is likely to be considerable mixing of stocks at this life stage (Bowering, 1986; God0 and Haug, 1988a; Stobo et al., 1988). The halibut congregate for spawning in winter on well-defined deepwater spawning grounds, where almost exclusively sexually mature fish occur. Outside the spawning season the halibut leave the spawning grounds and may then be found in deep as well as shallow waters and in inshore as well as offshore areas (Devold, 1938; Mathisen and Olsen, 1968; Olsen, 1969; Haug and Tjemsland, 1986; Jakupsstovu and Haug, 1988).

IV. Reproduction A. Sexual Maturity Male Atlantic halibut reach sexual maturity at a younger age and smaller size than the females. In very substantial material collected in Faroese waters between 1983 and 1986, it was observed that the average (50% level) ages, lengths and total weights (Fig. 3) at which the two sexes matured were 4.5 years, 55 cm and 1.7 kg in males and 7 years, 110-1 15 cm and about 18 kg in females (Jakupsstovu and Haug, 1988). Comparing results from different areas and periods, it appears that substantial variations occur in age at first sexual maturity. Jespersen (1917) concluded that halibut in Icelandic waters do not mature before the age of 9 or 10 years at the earliest. McCracken (1958) came to the same conclusion after studies of halibut caught in North American waters in the 1940s. Investigations in this area in 1972-84 have revealed that males and females matured as early as 4 and 6 years, respectively, and the 50% levels of maturity were 8 and 12 years in the two sexes (Bowering, 1986). In halibut caught in 1936-38 and in 1955-60 in northern Norway, the range and 50% levels of sexual maturity were 7-17 and 12 years for males and 8-18 and 13 years for females (Devold, 1938; Haug and Tjemsland, 1986). Resampling in northern Norway in 1981-85, however, revealed a substantial reduction in the age at first maturity, in that the 50% level was then attained at 7 years in males and 8 years in females (Haug and Tjemsland, 1986). Contrasting variations in age and in average length at maturity, has shown only very small variations throughout the areas and periods investigated (Bowering, 1986; Jakupsstovu and Haug, 1988). In northern Norway, where the most conspicuous changes in age at maturity seem to have occurred, it was suggested that the reduction in age at maturity was due to an observed

T. HAUG

8

increased growth rate following a decline in halibut density due to exploitation (Haug and Tjemsland, 1986). Therefore, it seems that sexual maturity in the halibut is more a function of growth rate and size than of age. According to Roff (1982), this is a common feature in fish species that mature at relatively old ages.

;I,/, 100

‘ 6 0 , , , , , ,

+ z

Y

0

40

20 0

A----?100 120 140

80

TOTAL LENGTH (5cm groups) ....~... ~

20 0

l,,,, A ,

5

7

‘

7

7

,

I

.

.

.

15 20 d5 3b TOTAL WEIGHT (kg)

10

,,.,,....,.... 35

40

FIG.3. Sexual maturity in Atlantic halibut from Faroese waters. Percentage of sexually mature male (solid lines) and female (broken lines) fish with respect to age (a), total length (b), and total (“round”) weight (c). In (c), males are given in 0.5-kg weight groups and females in 3.0-kg weight groups. Reproduced with permission from Jakupsstovu and Haug (1988).

BIOLOGY OF THE ATLANTIC HALIBUT

9

B. Spawning

I . Gonadal development The halibut has a huge potential for egg production, the largest females being able to produce several million eggs (Fulton, 1891; Haug and Gulliksen, 1988a). Spawning females usually have hyaline as well as non-hyaline oocytes in their ovaries (Kjrarsvik et al., 1987; Jakupsstovu and Haug, 1988). The percentage of hyaline oocytes may vary since spawning is intermittent during the breeding season, with only a portion of the maturing oocytes taking in fluid and being discharged at any one time (Haug and Gulliksen, 1988a). The latter has been confirmed from experiments with adult halibut in captivity, which usually release various sized batches of eggs at more or less synchronized intervals (Rabben, 1987). The regression of gonad weight on eviscerated body weight varies with the stage of maturity in both male and female halibut, i.e. fish have gonads of different relative weight depending on their state of maturity. This is particularly conspicuous in females where especially large specimens in the last stage of vitellogenesis develop larger ovaries proportional to eviscerated body weight than smaller females (Haug and Gulliksen, 1988b).

2. Spawning time From studies of the seasonal variations in gonad maturity of adult fish and of the relative abundance of planktonic eggs on the spawning grounds (Fig.4), Kjrarsvik et al. (1987) concluded that halibut spawning in the northernmost parts of Norway peaked at the end of January/beginning of February, although spawning occurred from December to March, and minor temporal variations in peak occurrence were observed. These findings are also supported by studies of the development of the gonadosomatic index by Haug and Gulliksen (1988b). There are clearly inter-area differences in halibut spawning times throughout the North Atlantic. In Faroese waters, spawning increases in intensity (based on the maturity stages of adult fish) in January and February, with the peak possibly extending into early spring (Jakupsstovu and Haug, 1988). These results are consistent with more fragmentary observations on halibut spawning made previously in the Iceland/Faroes/North Sea area (Jespersen, 1917; Vedel - T h i n g , 1936; Rae, 1959), and also with results obtained on the Nova Scotia/Gulf of St Lawrence/Newfoundland Banks area (McCracken, 1958; Kohler, 1967). The biological significance of earlier spawning off the north of Norway compared with more southerly sites is as yet unknown.

T. HAUG

10 40

1984185

35

m0-3d

30

@4-lOd all-18d

25

20 15 10

5 Lo

0

JANUARY

FEBRUARY

MARCH

1985186

JANUARY

FEBRUARY

MARCH

FIG.4. Relative abundance of Atlantic halibut eggs on the spawning ground in Malangen, northern Norway, during December to March in 1984-86. No eggs were found in December. The figure shows the frequencies of the various developmental stages: blastulae (0-3 days), gastrulae until closure of blastopore (4-10 days), and embryo with closed blastopore ( I 1-18 days). Reproduced with permission from Kjarsvik er al. (1987).

3. Spawning areas and hydrographic conditions In Norwegian coastal waters, halibut spawn over a soft clay, or mud bottom in deepwater locations (300-700 m : Devold, 1938; Kjnrrsvik et al., 1987). During spawning, the bottom temperatures and salinities vary between 5 and 7°C and 34.5 and 34.9%0,respectively (Kjnrrsvik et al., 1987). From the occurrence of planktonic eggs and larvae (Jespersen, 1917; Cox, 1924; Jensen, 1926; Rollefsen, 1934; Vedel-Thing, 1936; McIntyre, 1958) and the occasional presence of adult fish with running or spent gonads in commercial catches (Hjort, 1905; Jespersen, 1917; Devold, 1938), it seems likely that spawning also occurs in deepwater slope areas along the continental shelf in various parts of the North Atlantic. Aggregations of adult halibut on certain continental slope spawning grounds were not documented until

BIOLOGY OF THE ATLANTIC HALIBUT

11

1982-83, when a Faroese trawler made very good catches of large halibut in a restricted area at a depth of 700-1000 m along the southwestern slope of the Faroe Bank (Jakupsstovu, 1986). During spawning in this area, fairly homogeneous conditions with respect to hydrography were observed from the surface to the bottom and the spawning appeared to occur at temperatures of 7.5-8.1"C and salinities of 35.0-35.2%0 (Jakupsstovu and Haug, 1989).

V. Pelagic Phase A. Embryonic 1 , Natural occurrence

After spawning, the fertilized eggs gradually move upwards in the water column. In northern Norwegian waters, where the vertical distribution of halibut eggs has been quite thoroughly studied (Haug et al., 1984, 1986), the eggs were primarily found floating at intermediate depths, at temperatures of 4.5-7.O"C and salinities of 33.8-35.0%0. These findings are consistent with previous observations in northern Norwegian coastal waters (Devold, 1938, 1939, 1943), as well as in certain North Atlantic deepwater areas along the continental slope (Vedel-Thing, 1936; McIntyre, 1958), although recent investigations in Faroese continental slope waters indicate that the eggs in this area develop in somewhat higher temperatures (7.5-8.1 "C; see Jakupsstovu and Haug, 1989). According to Haug et al. (1986), the vertical distribution of the eggs is influenced by biological factors (the eggs achieve a somewhat higher specific gravity in their latest developmental phase), as well as the prevalent hydrography on the spawning ground.

2. Size The halibut egg is among the largest of planktonic fish eggs (Russell, 1976). In northern Norway, their diameter varies between 3.06 and 3.49 mm (Haug et al., 1984), which is in agreement with observations of pelagic halibut eggs from other areas (Vedel-Thing, 1936; Jakupsstovu and Haug, 1989). Kjarrsvik et al. (1987) observed that the mean diameter of the eggs tended to decrease during the spawning season. 3. Development Laboratory experiments have revealed that unfertilized halibut eggs possess a rather soft chorion which hardens considerably during the first hours after

12

T. HAUG

fertilization (Lsnning et al., 1982). In light and electron microscope studies of halibut eggs, Lsnning and co-workers have observed that the chorion of unwetted eggs looks somewhat wrinkled, and the cortical alveoli are near the surface of the yolk, After fertilization, the wrinkles and cortical alveoli disappear, and a very small perivitelline space is formed between the chorion and the yolk. Small and regularly distributed pores are visible on the egg surface, while sections of the eggs show a rather thin, homogeneous chorion with 18 concentric lamellae. The mean thickness of the chorion is 9.1 pm. Sections of the yolk show strong aggregations of some yolky material. The incubation period is strongly dependent on temperature, which has an influence even within the first cell divisions (Pittman et al., in press a). In experiments with artificially fertilized eggs, Rollefsen (1934) observed an incubation period of 16 days at 6°C (Fig. 5), Lsnning et al. (1982) observed hatching after 18 days at 5 ° C while Blaxter et al. (1983) reported incubation times to 50% hatching of 20, 18 and 13 days at temperatures of 4.7, 5 and 7"C, respectively. In water temperatures of 6"C, Rollefsen (1934) observed that the first division of the germinal disc took place after 6 h, and that subsequent divisions occurred at intervals of about 3 h. At 5"C, Lsnning et al. (1982) found that halibut eggs reached the 16-cell stage after 23 h, and that gastrulation started after 4 days; the developing blastopore was seen to be characteristically oblong. The closure of the blastopore occurs around day 10 at 5°C (Haug et al., 1986). Blaxter et al. (1983) and Lsnning et al. (1982) found the developing embryo to be characteristically bent during the latter part of organogenesis. The embryo has no pigmentation (Rollefsen, 1934; Blaxter et al., 1983).

7.(ld,Qh.) 8.(4d.Qh.) 9.(6d.Qh.) 10.(7dS13h.)1 1 (10d.) 12. (16d.) FIG. 5. Sketches of egg development of the Atlantic halibut. The age of the eggs is given in hours (h) and days (d) in parentheses. Reproduced with permission from Vedel-Thing (1936).

Hatching appears to be more complicated than in many other fish species, in that the larvae emerge from the egg after dividing the chorion into two well-defined parts (Fig. 6). Helvig (1988) showed that the hatching

BIOLOGY OF THE ATLANTIC HALIBUT

13

glands are located in a narrow ring on the anterior part of the yolk sac, so that the secretion of a hatching enzyme creates a ring of weak material along which the chorion fractures.

FIG.6 . Atlantic halibut larva emerging from the chorion. Enzymes from a ring of hatching glands have created a weakened zone which fractures as the larva struggles to escape. Reproduced with permission from Helvig (1988).

B. Larvae 1. Natural occurrence

In situ records of halibut larvae are scarce. Schmidt (1904), Jespersen (1917), Cox (1924), Jensen (1926), Rollefsen (1934), Vedel-Thing (1936) and Haug and Sundby (1987) together report only 57 Atlantic halibut larvae, between 1 1.9 and 34.0 mm in size, from the plankton. Most of these larvae were taken in open Atlantic waters, southwest of the Greenland/Iceland/Faroes/Scotland ridge, although specimens have also been observed off Nova Scotia (Canada), to the west of Greenland, and in fjords on the coast of Norway. In a major multinational ichthyoplankton survey in the northwest Atlantic between April and July 1963 (Smidt, 1968), only three larvae (c. 12 mm) were recorded in May. A limited number, between 26 and 35 mm in length, were also found in July. These records all came from the Davis Strait, southwest of Greenland and between Greenland and Iceland. Most previous records of Atlantic halibut post-larvae have been made in the upper 5-50m of the water column, generally with the smaller specimens occurring at greater

14

T. HAUG

depths than the older ones (Vedel-Thing, 1936). On a spawning ground (Saraysund) in northern Norway, the hydrographical conditions in the middle of March, which is probably the time when the yolk sac has been absorbed, were characterized by temperatures between 2.4 and 2.7"C and salinities between 34.0 and 34.3%0(Haug and Sundby, 1987; Haug et al., 1989a). The reason for the scarcity of records of pelagic stages is undoubtedly that they are scattered at low density and with uniform distribution over a very large expanse of water. Due to these very scarce field data, all our knowledge on yolk-sac larvae stems from observations made during rearing experiments with artificially fertilized eggs, although some excellent descriptions of post-larval morphology based on individuals from nature exist (see Schmidt, 1904).

2. Yolk-sac larvae Vesicles are usually seen within the yolk sac both pre- and post-hatching. Their biological significance, however, remains unknown (Rollefsen, 1934; Lanning et al., 1982; Pittman et al., 1987). At hatching, the larvae have a total length of about 6-7 mm, and appear very premature. The yolk sac is very large and pigment, functional eyes and mouth are lacking. Hatching occurs long before the trunk of the embryo surrounds the egg (Rollefsen, 1934; Lanning et al., 1982; Blaxter et al., 1983; Pittman et al., 1987). Pittman et al. (1987, in press b) observed the morphological development of halibut larvae reared from hatching to beyond metamorphosis in large (1 1.5-m3) plastic bags and in 15-m3 silos at temperatures of 4, 6 and 9°C. They give detailed information about the increase in length and describe important features of the morphological and behavioural development of the larvae (Figs 7 and 8). Based on this information, Pittman et al. (in press a) have proposed four stages in the yolk-sac phase of Atlantic halibut: Stage I . Yolk present; notochord straight; no stomodaeum; no chromatophores; eyes symmetrical and pressed against the yolk sac; a pair of external branchial pits posterior to head; intestine straight; no pectoral fins; simple heart; body generally lies passive and almost vertical in the water column with head downwards; infrequent swimming, mean speed 0.5 mm/s. Duration 5-7 days. Stage 2. Yolk present; yolk sac changing from an elliptical form to a pearshaped one; notochord straight; mouth present as a small opening; eyes symmetrical, circular and lightly pigmented to pigmented with an unpigmented dorsal gap; branchial pits expand to short opercular flaps; intestine

15

BIOLOGY OF THE ATLANTIC HALIBUT

thickening and beginning to loop; pectoral fins develop from buds to paddlelike fins; heart differentiating into four chambers; liver and gall bladder develop; gall bladder may be coloured light green for a short period; body nearly horizontal in water; mean swimming speed increases from 0.5 to 3.0 mmjs when active. Duration 17-20 days.

intestine rotating

I H L

rj+

yolk

pylorus

15-x

mouth open

J w

0

teeth observed

startfeeding

a a 10n z

Pelvic fins

iunciionai

65 110)

140

180

240

(20)

(30)

140)

310 150)

400

(60)

480 170)

550 180)

630 190)

; 3 1100)

FIG.7. Growth of halibut larvae in mm standard length vs day-degrees and days (in parentheses) during rearing in 1 1.5-m3 plastic bags with temperature stable around 6°C. The appearance of important developmental features is indicated by arrows. Reproduced with permission from Pittman el al. (1987).

Stage 3. Yolk present; yolk sac reducing to tubular form on right side of intestine, connected to medial side of liver; notochord straight; mouth becomes functional; eyes symmetrical and fully pigmented; gill cage developing with some filaments on the arches; intestine fully looped; rectal sphincter develops; ventricular cardiac muscles thicken; mean length of active periods increases from about 5 s to over 1 min; larvae can assume “C” and “S” forms in the horizontal plane in this stage for a few seconds at a time. Duration 15-20 days. Stage 4. Yolk present; little or no yolk left; notochord straight; hypural thickening; hypural spines appear; mouth functional, buccal valve expands

A

C

B

D

FIG. 8. Atlantic halibut yolk-sac larvae in developmental stages I1 ( A X ) and 111 (D). The larvae were reared at a temperature of 6°C.(A) 14 days old: note vesicles behind in the yolk sac. (B) 19 days, 11.5 mm total length. A faint outline of the prospective liver is seen just behind the pectoral fin,and yolk-sac vesicles are clearly separated from the yolk. (C) 25 days, 12 mm in total length. The jaw has become more pointed. (D) 46 days, 14.5 mm in total length. The striped area behind the rotated intestine is the yolk remains. Reproduced with permission from Pittman et al. (1987).

BIOLOGY OF T H E ATLANTIC HALIBUT

17

outward when buccal cavity contracts; eyes symmetrical; gill filaments on arches; some black melanophores along margin of ventral finfold and notochord; ratio of standard length to myotome height is still greater than 15; larva displays searching activity and attacks prey items but mean duration of active periods drops. Duration 4-7 days if not coinciding with active feeding. The yolk is absorbed at somewhat less than 50 days post-hatching, a stage at which the larvae are 11.5-13.0 mm long (Blaxter et al., 1983; Pittman et a[., 1987). Possibly, the larvae may need exogenous food before the yolk is absorbed. Following observations on the development and behaviour of live larvae, Pittman et al. (in press b) discuss whether a functional digestive system is established even earlier than day 30. Likewise, Pittman et al. (in press a) suggested, from their analyses of relative protein synthesis in halibut larvae, that exogenous food uptake is necessary between 25 and 30 days after hatching. Elin Kjrarsvik (pers. comm.) supports this view, and concludes from feeding experiments and morphological studies, that halibut larvae (at 5-7°C) may be capable of digesting food particles from between days 23 and 26 after hatching. The many neuromasts shown in Fig. 8C as circles along the notochord, are in fact very prominent conical humps, increasing from 2 to 4 at hatching to 11-12 at the end of the yolk-sac stage. Scanning electron microscopy shows the presence of kinocilia and sterocilia, indicating that the humps are free neuromast organs that probably endow the larvae with some sensory capability before the development of vision (Blaxter et a[., 1983). Tytler and Blaxter (1988) observed that halibut yolk-sac larvae can drink from the earliest post-hatching stages, thus suggesting that osmoregulation and buoyancy mechanisms similar to the adult are present from the early stages of ontogeny. These authors show that the gut of the yolk-sac larvae is patent soon after hatching and that a stomodaeum (primitive mouth) and proctodaeum (primitive anus) may have functional sphincters before these can be detected by gross anatomical studies. It is apparent that from the time the mouth becomes functional (at stage 3), to the time of starting feeding, the lateral areas of the mouth are enclosed by an oral membrane (Fig. 9A). Pittman et al. (1987) suggest that the oral membrane is necessary for the development of proper articulation of the jaws (Fig. 9b), and that early degradation of the membrane, by bacterial or other means, may compromise both the mouth function and larval viability that is often observed in rearing experiments. In their rearing experiments, Blaxter et al. (1983) observed that chromatophores were laid down 40-50 days post-hatching, mainly as ventral, lateral and dorsal lines along the larval body, and were also found in the gut region.

18

T. HAUG

A

B

FIG.9. Mouth development in Atlantic halibut yolk-sac larvae reared at 6°C.(A) Apparent bone structure of the mouth at about 30 days. Arrow indicates opening in the oral membrane. (B) Apparent bone structure in a functional mouth, around 50 days, larva 13mm long. Melanophores are drawn over the brain and around the lower jaw. Further explanations are given in the text. Reproduced with permission from Pittman er al. (1987).

3. Post-larvae

Following the usage of Russell (1976), feeding larvae are here denoted as post-larvae. Pittman et al. (1987) first observed branchial respiratory movements at 30 days, and the gills appear to be fully developed around 60 days (Fig. 7). The pelvic fins at first appear as dark patches at about 80 days, and very quickly become functional fins approximately 1.5 mm in length. The left eye starts migrating at about 80 days. This occurs at a length (Fig. 7)

BIOLOGY OF THE ATLANTIC HALIBUT

19

which corresponds well with the field observations of Schmidt (1904), where the left eye had begun to migrate on a specimen 16.25 mm long. Pittman et al. (1987) observed that larvae at this stage still search for prey in an upright position. This position gradually becomes horizontal, although the larvae continue to attack the prey from an upright position. In rearing experiments, the eye has migrated halfway around the head by 130 days. This stage of development was observed by Schmidt (1904) on a 29.5-mm larva in the field. During start-feeding in rearing experiments (Pittman et al., 1987), the gut continued to grow and rotate one and a half times, filling the gut cavity (Fig. 10). Between 100 and 105 days, the pylorus began to differentiate as an elongated appendage on the interior circumference of the intestinal loop, from which arose three pyloric caecae. As the jaws became more complex, the bones thickened, and four small teeth could be seen on either side of the jaws on one large larva by 85 days.

FIG.10. Atlantic halibut larva at the start of feeding at about 65 days (reared at 6“C), 16.8 mm standard length and 2.4 mm myotome height. Note the increasing complexity of the myomeres, the prospective pylorus (mass in the centre of the rotated intestine), the pseudobranch posterior to the eye, and the haemal and neural fin rays beginning in the peripheral muscle layer. Reproduced with permission from Pittman et al. (1987).

In experiments with halibut larvae reared in mesocosms, Berg and Oiestad (1986) observed first feeding to appear around day 40 post-hatching, at a time when the larvae were about 12mm long. Schmidt (1904) gave a comprehensive description of pelagic halibut larvae, based on 19 captured specimens between 13.5 and 34mm long. He characterized the larvae by their large heads with a “pug nose” outline to the snout, a straight outlined lower jaw, and very sparse pigmentation. The development of the morphology and pigmentation of halibut post-larvae of various sizes are

20

T. HAUG

described in Fig. 11. Schmidt (1904) observed that at a length of 34 mm, the marbled pattern of the adult pigmentation was already appearing. The true rays began to develop in the dorsal and anal fins in larvae 18-20 mm long. In specimens 24.5-25 mm long, Schmidt counted 103 dorsal rays and 80 anal rays. The pelvic fins, which began to appear in specimens 22-23 mm long, had a little black pigment on them at lengths of 24.5-25.0 mm. The pectoral and caudal fins remained unpigmented. In a specimen 16.25mm long, flexion had taken place and 50 vertebrae were present.

FIG. 11. Pelagic post-larvae of the Atlantic halibut, varying in size from 13.5 to 35mm. Redrawn after Schmidt (1904)

In addition to black pigmentation, the presence of minute red chromatophores has also been described (Schmidt, 1904; Vedel-Thing; 1936). These small orange-red chromatophores are scattered over the head and abdominal region with a somewhat higher density on the snout and tip of the lower jaw. On the body they are distributed along the myotomes, with some also on the myomeres mediolaterally. They are most abundant on the anterior part of the body, especially ventrally. They occur along the dorsal and anal

BIOLOGY OF THE ATLANTIC HALIBUT

21

fin rays out to their margins, and on the interspinal rays, mostly on the anal fin. There are none on the caudal fin except for one ventrally at its base. According to Vedel-Tining ( 1936), the largest post-larva described was 34 mm long, while the smallest bottom stage known was 47 mm long (Fig. 12); the transition from pelagic to bottom life therefore probably takes place at a length of 34-40mm. In mesocosm-rearing experiments, Berg and 0iestad (1986) observed that on day 75 post-hatching, most larvae were in an early stage of metamorphosis, and that on day 90 most of them were able to rest on the bottom although metamorphosis was not yet completed. Apparently, bottom settling occurs before the final completion of pigmentation, eye movement and formation of fins, and the transformation to “real” flat-fish seems to take place on the bottom during growth from 34 to 47 mm (Vedel-Thing, 1936).

FIG.12. The smallest bottom stage of the O-group ever recorded in nature (Faxa Bay, west of Iceland) at 47.1 mm long, showing the eyed side above and the blind side below. Reproduced with permission from Vedel-Thing (1936).

22

T. HAUG

VI. Immature Phase A. Occurrence and Migrations Young halibut are quite localized in apparently well-defined nursery grounds, coastal areas 20-60 m deep with a sandy bottom. Bottom temperatures in such areas off Ssrrya, northern Norway are 7-8°C in July and .u2.5-3SoC in March and May (Haug and Sundby, 1987). In nursery areas between the Faroe Islands (Vedel-Thing, 1938, 1947), off western Iceland (Sigurdsson and Fridriksson, 1952; Sigurdsson, 1956), off Nova Scotia in eastern Canada (Stobo et al., 1988) and in certain Norwegian coastal areas (Haug and Sundby, 1987; God0 and Haug, 1988a), juvenile halibut are frequently encountered as a by-catch in other fisheries using otter trawls and Danish seines. Bottom stages of the 0-group have been demonstrated only near Iceland and in the Faxa Bay nursery area (Vedel-Thing, 1936; Sigurdsson, 1956, 1976). Here, and also in the Faroe fjords and banks, the I-group starts to appear in the catches in late summer (Joensen, 1954; Sigurdsson, 1956), while in halibut nursery areas in northern Norway this age group has not been recorded in the catches (Tjemsland, 1960; Haug and Sundby, 1987). It has been suggested that the very rare occurrence of the smallest bottom stages in catches made on nursery grounds could be due to their distribution on hard and rough bottoms, which makes trawling impossible (VedelThing, 1936), or to the low catching efficiency of the gear for these small fish, e.g. due to large mesh sizes or escape (Haug and Sundby, 1987). In general, emigration from nursery areas seems to start when the fish are 3-4 years old (Joensen, 1954; Sigurdsson, 1956, 1976; Haug and Sundby, 1987; God0 and Haug, 1988a). It seems to occur randomly and to waters of different depths. Although most migrants tagged on nursery areas have been recaptured at various distances from the tagging sites within the same main region where they were tagged (see, e.g. Fig. 13), some very long-distance migrations, even crossing deepwater areas in the North Atlantic, seems to be quite common. Such long-distance migration has been documented from nursery areas in northern Norway to the White Sea, Iceland and Greenland (Fig. 13), from the Faroes to the North Sea and Iceland (Vedel-TBning, 1938, 1947), and from Iceland to the Faroes, Greenland and Newfoundland (Jonsson, 1978; Bowering, 1986). The highly migratory capability, apparently still quite random, of “half-grown” halibut, i.e. fish that have left the nursery areas but are still immature, is further confirmed from tagging experiments in the Barents Sea (Devold, 1943), on the coasts of Norway (God0 and Haug, 1988a) and Greenland (God0 and Haug, 1988b), and on the east coast of North America (Martin and McCracken, 1950; McCracken

BIOLOGY OF THE ATLANTIC HALIBUT

23

and Martin, 1955; McCracken, 1958; Jensen and Wise, 1961; Kohler, 1964; Stobo et al., 1988).

FIG. 13. Migration patterns of small, immature halibut leaving their nursery areas. From a tagging experiment carried out in Lofoten, northern Norway, where 300 fish were tagged in 1961 and 1962. Of these, 180 were recaptured: 140 individuals in the tagging area, while the remaining 40 had emigrated in various directions as shown (after God0 and Haug, 1988a).

B. AgelSize Composition and Sex Ratio

Due to the absence of halibut of the earliest bottom stages (see above) and the emigration patterns of “half-grown’’ specimens from the nursery areas, the most numerous age groups encountered in trawl and Danish seine catches on these grounds are the 11- and 111-groups (see Fig. 14; Jespersen, I9 17, 1926; Joensen, 1954; Sigurdsson, 1956; Tjemsland, 1960; Haug and Sundby, 1987). Measurements have shown that most specimens are shorter than 55 cm in length (see Fig. 15). A slightly greater percentage of males over females (59:51%) has been observed in the catches (Joensen, 1954;

.

% 70

-Intra-territorial. _ _ _ _ _Extra-territorial. -

60

50 40

30 20

10 YearGroup :

I

n

m

H

v m m m

FIG. 14. Age composition of Atlantic halibut in nursery areas. Relative quantity of the different age groups of fish captured in trawls in Faxa Bay, Iceland, inside and outside the territorial limits (i.e. the inner and outer parts of the Bay) in 1924. Reproduced with permission from Jespersen (1926).

5 Coastal banks A

FIG.IS. Size composition of Atlantic halibut in nursery areas. The percentage and length distribution of halibut from trawl catches made in fjords, bays and on coastal banks in Faroese waters. Reproduced with permission from Joensen (1954).

25

BIOLOGY OF THE ATLANTIC HALIBUT

Sigurdsson, 1956; Tjemsland, 1960). Joensen (1954) suggested that this might be due to faster growing females leaving the nursery areas before the males.

C. Growth The size range of the few bottom-stage 0-group halibut specimens observed (as measured in June-July) was 44-70 mm (Vedel-Tining, 1936; Sigurdsson, 1956). In both young male and female halibut, there are considerable variations in size, giving overlapping size distributions between age groups (Fig. 16). The growth rate of young Atlantic halibut has been shown to vary from area to area in the North Atlantic, both on a small scale (e.g. among different fjords) and on a larger scale (e.g. between the Faroes and Iceland: Jespersen, 1917, 1926; Joensen, 1954; Sigurdsson, 1956). 1

401

AGE IV (N=149)

20

0

AGE Ill (N- 1469)

AGE I (N=36) 40

is

30

45

LENGTH (3cm groups)

FIG. 16. Length distribution of different year classes of I - to 4-year-old halibut caught in trawl in Faxa Bay, Iceland, in June-July in 1908 and 1909 (based on data from Jespersen, 1917).

26

T. HAUG

Furthermore, the rate of growth also seems to exhibit variations with time: in the Faxa Bay, west of Iceland, the rate of growth of 1- to 3-year-old halibut clearly increased between 1908-1909 and 1925 (Jespersen, 1926) and even more so between 1925 and 1948-50 (Sigurdsson, 1956). Similar trends have been observed in northern Norway during the period from 1938-39 to 1985 (Haug and Sundby, 1987). Factors such as changes in temperature, stock depletion, depletion of food competitors, and fluctuations of food supply may have contributed to these differences. Young halibut do not grow at the same rate throughout the year (Fig. 17). The average increase in length of the I-, 11-, and 111-groups, as observed in the Faxa Bay during 1936-50, was clearly most rapid in summer and autumn and more or less stationary during winter and spring (Sigurdsson, 1956). CM 56 54 -

52

I

I

I

I

I

I

-

5048 -

4644 42 40 38 36 34 32 30 -

-

-

28 26

-

:'If

I

I

,

12

I

I

I l l

I

I

I

I

I

I

I

1.

uu ~

M J J 4 S O N D J f M A M J J A S O N D J F M A M J J A S O N D

FIG. 17. Growth of juvenile Atlantic halibut. Average length of age groups I, I1 and I11 as measured in various parts of the year in Faxa Bay, Iceland. Data are taken from various months between 1936 and 1950 and allocated to the three age groups. Reproduced with permission from Sigurdsson (1956).

D . Feeding Apparently, there is a change in the feeding habits of halibut with age (Scott, 1910; Jespersen, 1926; McIntyre, 1953; Rae, 1958; Kohler, 1967). The smallest specimens ( < 30 cm) seem to have a diet which is almost exclusively

BIOLOGY OF THE ATLANTIC HALIBUT

27

composed of crustaceans, especially hermit crabs, prawns, small crabs and mysids. As the halibut grow, the diet includes more and more fish. The food found in the stomachs of medium-sized, immature fish (30-60cm long) is usually a mixture of fish and crustaceans (Fig. 18). In his very thorough studies in Faroese waters, McIntyre (1953) suggested that medium-sized halibut fed more intensively in summer than in spring, and he also concluded that the summer diet included more crustaceans (particularly Eupagurus hernhardus) than the spring diet in which fish (gadoids and sand eels, Ammodytes spp.) were more prevalent (Fig. 18). The importance of small gadoids and, to an even larger extent, of sand eels in the diet of mediumsized halibut is also confirmed by an investigation on fish caught in the North Sea and North Atlantic and landed in the Aberdeen fish market (Scott, 1910), fish caught in Icelandic waters (Jespersen, 1926; McIntyre, 1953) and fish caught by Scottish research vessels in the North Atlantic (Rae, 1958). SUMMER

SPRING G ESMARKll (2) 6.20/.

OTHER GADOIDS (4) 6.8%

CRUSTACEA (4)

CEPHALOPODA (2)5.3% PANDALUS (12) 6.1%

MUNIDA (4)4,2% HYAS (3) 3 4 %

FIG. 18. Feeding of juvenile Atlantic halibut. Seasonal changes in the stomach contents of young Faroese halibut, ranging in size between 31 and 60cm. The volumes of the various constituents of the diet are expressed as percentages of the total food volume in spring (left) and summer (right). The number of stomachs in which each group occurred is shown in parentheses (after McIntyre, 1953).

VII. Mature Phase A. Occurrence and Migrations On the halibut spawning grounds during the spawning season (see Sections IV.B.2 and 3), sexually mature fish are almost exclusively present (Devold,

28

T. HAUG

1938; Mathisen and Olsen, 1968; Haug and Tjemsland, 1986; Jakupsstovu and Haug, 1989). After spawning, mature halibut leave the spawning areas and migrate in all directions (Fig. 19), apparently in search of food, and are dispersed in both shallow and deep waters, and in inshore as well as offshore areas (Devold, 1938, 1943; God0 and Haug, 1988a). Thus, in Norwegian waters, good catches are often obtained in shallow and more intermediate depths in summer, probably due to an increased number of large, mature halibut moving from the deep spawning grounds to the shallower fishing grounds to feed (Devold, 1938). Similar observations have been made in North American waters (McCracken, 1958; Bowering, 1986).

I

OBEAR ISLAND 74’1

1 -

BARENTS SEA

FIG. 19. Migration patterns of sexually mature halibut outside the spawning season. From a tagging experiment carried out in 195660 on a spawning site in Finnmark, northern Norway, where 456 fish were tagged. Of these, 46 were recaptured: 21 individuals in the tagging area during spawning 1-3 years after tagging, while the remaining 25 were recaptured outside the spawning season, in many cases at a considerable distance from the spawning site (after God0 and Haug, 1988a).

Results from Norwegian experiments show that adult halibut, which were tagged and released on spawning grounds in the 1950s, were recaptured 1-3 years later, either within the area of tagging during the spawning season or outside the area of tagging outside the spawning season (Fig. 19). The data also suggest a remarkable “homing” response, i.e. the fish return to the same

BIOLOGY OF THE ATLANTIC HALIBUT

29

area for repeated spawning in several successive years (Devold, 1938, 1943; God0 and Haug, 1988a). A limited dispersal and the annual return migration to spawn has also been suggested for adult halibut in Canadian waters based on tagging results (Stobo et af., 1988). None of the tagging experiments conducted so far has been able to demonstrate that the halibut progeny return to the spawning ground of their parents, i.e. if the homing is natal (God0 and Haug, 1988a). The mechanisms underlying the migrations of halibut are still unknown, although it has been suggested that along the continental slope spawning area southwest of the Faroe Bank, an extremely narrow front between the cold and warm water currents might help the adults to identify the spawning area (Jakupsstovu and Haug, 1989).

B.

Sex Ratio and AgelSize Composition

It has been observed, both in gill net and trawl catches made in deepwater spawning areas, that the number of male halibut caught has exceeded the number of females (Mathisen and Olsen, 1968; Haug and Tjemsland, 1986; Jakupsstovu and Haug, 1988). Mathisen and Olsen (1968) found that the gill net fishing mortality of males was considerably higher than for females on the spawning grounds, and suggested sexual differences in behaviour that might lead to differences in vulnerability in this type of fishery. In addition, the younger age at which the males reach sexual maturity (see Section 1V.A) could affect the relative abundance of males and females and lead to an excess of males on the spawning grounds. Sex ratios different from 1 : I , with an excess of either males or females, have also been recorded for commercial catches of halibut outside the spawning grounds. Seasonal/spatial segregation of the sexes seems plausible, although more needs to be known about halibut behaviour and ecology before a basis for the sex ratio differences in the catches can be elucidated (Joensen, 1954; McCracken, 1958; Rae, 1959; Haug and Tjemsland, 1986). From the apparent random dispersal of adult halibut after spawning (Fig. 19), it appears that the spawning seasons and grounds are the only times and areas where large aggregations of adult specimens occur. The composition by size and age of halibut caught on various spawning grounds in northern Norway clearly show that these fish are almost exclusively mature specimens (Fig. 20). Recruitment to the spawning areas in the 1980s occurred mainly at an age of 5-7 years, and at a length of 60-75 cm in males and 124-140 cm in females, which coincides with the attainment of sexual maturity of the species (Haug and Tjemsland, 1986). Similar observations were made in continental slope spawning areas southwest of the Faroe Bank (Jakupsstovu and Haug, 1989). As seen from Fig. 20, the decrease in age at sexual

30

T. HAUG

maturity of halibut from northern Norway in the past 25 years (see Section 1V.A) has led to earlier recruitment to the adult stock and an increase in the presence of young fish on the spawning grounds in the 1980s compared with the late 1950s.

-

9 bp

15

10

F 5

Malangen/~sterBlen/Andfiord

1981-85

n

10

100

150 ' 200

250

TOTAL LENGTH (5cm groups)

10

20

30

'

40

AGE (years)

FIG.20. Size (left) and age (right) composition (% frequency) of male (solid lines) and female (broken lines) halibut caught in various northern Norwegian spawning areas in the late 1950s and early 1980s. Reproduced with permission from Haug and Tjemsland (1986).

Age and length distributions from catches made outside the spawning season generally include larger numbers of young and small fish, but only rarely the youngest age groups (I-111), which are usually encountered in the nursery areas. Thus, in Faroese waters, the most abundant age groups in longline fisheries before the 1950s were 5- to 8-year-old fish (Joensen, 1954), while on the east coast of North America, 6- to 9-year-old fish dominated the catches in the 1970s (Bowering, 1986). The intermixing of mature and larger immature fish thus seems to occur only during feeding (e.g. Joensen, 1954; McCracken, 1958; Bowering, 1986).

31

BIOLOGY OF THE ATLANTIC HALIBUT 260240220200180. -160-

E -

v

I 140, ! - -

9 120-

Y 100-80. 604020-

1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

4

6

8

10

12

14

16

18

20

22

24

26

28

AGE (years)

FIG.21. Observed growth in length of male (solid line) and female (broken line) halibut are given for caught in a spawning area southwest of the Faroe Bank in 1983-86. Means & S.D. each age group. Reproduced with permission from Jakupsstovu and Haug (1988).

C. Longevity and Growth In the heavily exploited Norwegian halibut stock, Haug and Tjemsland (1986) observed an obvious deficiency in the abundance of old fish. Bowering (1986) reported that from North American waters during 1974-82, old fish also occurred less frequently compared with the 1940s (McCracken, 1958). While fish older than 20 years used to be quite numerous (see Fig. 20), exploitation has meant that fewer halibut now reach this age in many areas. In the spawning stock in the continental slope area southwest of the Faroes, fish older than 20 years of age are also quite rare now, although fish as old as 50 years have occasionally been observed (Jakupsstovu and Haug, 1989). The growth curves for this population (Figs 21 and 22) reveal a clear sexual difference in growth rate. Especially after the age of 6-7 years, the length and weight of the female far exceed those of the male. Such intersexual differences in halibut growth are not only typical for the Faroese population (Jakupsstovu and Haug, 1988), but have also been observed in all other areas of the North Atlantic where halibut growth has been investigated (Jespersen, 1917; Devold, 1938; Joensen, 1954; McCracken, 1958; Mathisen

32

T. HAUG

and Olsen, 1968; Haug and Tjemsland, 1986; Bowering, 1986). In fact, the total weight of male halibut seldom exceeds 50 kg, whereas females as large as 333 kg total weight have been recorded (Mathisen and Olsen, 1968). In female halibut, it appears that the initial growth rate is not only maintained, but even accelerates when they attain maturity. A comparison of total and gutted weights has revealed that most of the yearly weight increment is due to an increase in muscle tissues, i.e. the “edible portion” of the fish (Jakupsstovu and Haug, 1988). 180-

160.

140J

FIG.22. Observed growth in total weight of male (solid line) and female (broken line) halibut caught in a spawning area southwest of the Faroe Bank in 1983-86. Means f S.D.are given for each age group. Reproduced with permission from Jakupsstovu and Haug (1988).

According to Joensen (1954) and McCracken (1958), considerable variation occurs in the growth rate of halibut from various parts of the North Atlantic - the Faroese population having the highest growth rate. This was confirmed by Jakupsstovu and Haug (1988), who also concluded that the

BIOLOGY OF THE ATLANTIC HALIBUT

33

growth rate measured in the 1980s in the Faroese population was even more rapid than the growth observed in the northern Norwegian population in the 1980s, which seemed to be affected by heavy exploitation. The observed changes in growth in northern Norway during a period when the population has most probably declined due to heavy exploitation might suggest that this population parameter is subjected to some degree of density-dependence (Haug and Tjemsland, 1986).

D. Feeding While stomach contents obtained from young halibut reveal a more or less mixed diet, the diet of adults is almost entirely composed of fish (Scott, 1910; McIntyre, 1953; Rae, 1958; Kohler, 1967). It is known that young halibut move out from the coastal areas as they grow (see God0 and Haug, 1988a). According to McIntyre (1 953), this movement should be considered together with the change in size of prey. He suggested that the small food organisms in coastal waters are insufficient for the fast growing halibut, which therefore begin to emigrate. In general, the number of organisms which serve as food for large halibut is limited. Off Iceland, McIntyre (1953) observed that adult halibut fed on only 11 species of organisms, of which the redfish (Sebastes marinus) constituted over 75% of the food by volume. Based on comparisons of halibut stomach contents and the general abundance of prey as determined from trawl catches, McIntyre concluded that the adult halibut most probably selects redfish in favour of other prey items, and he suggested that redfish may play a role in the life-history of the halibut comparable to that of capelin (Mallotus villosus) and herring (Clupea harengus) in the life-history of cod (Gadus morhua), and of sand eels in the life-history of many smaller fish. With the wide distribution and the varied physical conditions experienced by halibut in the North Atlantic, it is hardly surprising that stomach analyses have revealed regional variations in diet. In eastern Greenland waters, the diet includes arctic and arctic-boreal forms such as Gymnellis and Lycodes (Iversen, 1936), as well as redfish (Rae, 1958). In Faroese waters and on the Rockall grounds and adjacent waters, the redfish seem to play a less prominent role as food for adult halibut. In the former area, gadoids, clupeoids and flatfishes occur quite frequently (Scott, 1910), whereas in the latter area the diet includes several species with a southerly distribution, such as sea bream (Pagellus centrodontus), hake (Merluccius merluccius) and blue whiting, Gadus poutassou (Rae, 1958). It can be seen from Fig. 23 that in Icelandic waters redfish form the major part of the halibut’s food throughout the whole year, whereas tusk (Brosme

T. HAUG

34

hrosme) are consumed in some quantity in spring and autumn, and Nephrops norvegicus in winter. McIntyre's (1953) data from Iceland suggested more intensive feeding of adult halibut in summer than in spring. From the very high frequency of empty stomachs occurring in fish caught in gill nets during spawning, Devold (1938) concluded that halibut .feed very little during the spawning season. Haug and Gulliksen (1988b), however, observed very little loss of body weight during gonad development and spawning, and suggested that feeding probably does not cease completely at this time. It may be that locomotor activity, particularly in females, is reduced to such an extent that there is a minimum energy requirement during spawning. In view of the diet preferred by the halibut, McIntyre (1953) concluded that the halibut must leave the bottom in pursuit of prey.

WINTER FISH ( 10)

\

SPRING NEPHROPS NORVEQICUS (2) 13.6%

SEBASTES MARINUS (44) 64.6%

SEBASTES MARINUS (16) 71.9%

FISH (36)

(9)2.3% HADDOCK (1) 4.4%

SUMMf"

A'

'TUMN

\

FIG.23. Seasonal changes in stomach content of adult Atlantic halibut from Icelandic waters. The volumes of the various constituents of the diet are expressed as percentages of the total food volume in each season. The number of stomachs in which each food group occurred is shown in parentheses (after McIntyre, 1953).

BIOLOGY OF THE ATLANTIC HALIBUT

35

E. Proximate Body Composition 1. Water, lipids and proteins

The chemical composition of different tissues varies considerably, especially in lipid and water content (Morawa, 1957; Mannan et al., 1961; Haug et al., 1988). Haug and co-workers, who analysed males and females separately, found that in general the lipid level of spawning females exceeded that of spawning males. In the males, the mean white myotomal muscle lipid content was 6-8.5% of dry weight (tending to decrease in an anteriorposterior direction, Fig. 24), whereas in females 13.6% was observed. The lipid content of red myotomal muscle, fin base tissue and liver was much higher than that of white muscle in both sexes: 39-55% of dry weight. The water content in white myotomal muscle varied from 72 to 77%, while red myotomal muscle, anal fin base tissue and liver generally contained 56-64% water.

FIG.24. Water and total lipids content of various tissues from various parts of the body of mature Atlantic halibut males caught during spawning. N is the number of fish analysed. Tissue types are: WM, white muscle; RM, red muscle; AN, anal fin base “notch”; L, liver. Reproduced with permission from Haug el al. (1988).

The total lipid content observed in halibut fillets (white myotomal muscle) places it in an intermediate position in relation to other species. It is fatter than typically lean fishes such as the cod, but considerably leaner than typically fat fishes such as capelin and herring (see Lambertsen, 1978). The lack of large fat reserves is consistent with McIntyre’s (1953) suggestion that the halibut is an opportunistic feeder.

36

T. HAUG

As observed by Mannan et al. (1961), the variation in content of crude protein was considerably less among tissues than lipid and water. Body wall and red myotomal muscle had a protein content of 16.3-19% (of wet weight), while the corresponding values from white myotomal muscle were 17.5-20.3 YO. 2. Lipids composition In mature halibut, neutral lipids (mainly triacylglycerols) constituted about 75-80% and polar lipids 20-25% of the total lipid pool (Haug et al., 1988). Polar lipids are generally the main constituent of fish eggs, and in ovulated halibut eggs the total lipid pool consisted of ~ 7 0 % of polar lipids (mainly phosphatidyl choline) and -N 30% neutral lipids (Falk-Petersen et al., 1986). A comparison of the fatty acid composition of neutral lipids of halibut with those of redfish reveals considerable similarities, and thus does not contradict the view that redfish may serve as food for halibut (Haug et al., 1988). Haug and co-workers examined the red and white myotomal muscle, anal fin base tissue and liver of halibut, and found no variation in fatty acid composition of the polar lipid fraction between these tissues. The saturated fatty acid content of polar lipids was dominated by 16:O (16.5%), while amounts of 14:0, 15:O and 18:O were low. The major mono-unsaturated fatty acid was 18:1 (1: 13%). The content of w-3 polyunsaturated fatty acid (PUFA) was about 46% with 20:5 and 22:6 dominating. The major saturated fatty acid in the neutral lipid fraction was 16:O (12-15%), 18:l was the major mono-unsaturated fatty acid (15-20%) and 20:5 0-3 and 22:6 w-3 were the dominating PUFAs (15-25%). The composition of the neutral lipid fraction of liver differed from the other tissues with respect to its content of 20: 1 and 22: 1 (lower in liver) and 22:6 0-3 (higher in liver). The relative fatty acid composition of neutral and polar lipids in ovulated halibut eggs (Falk-Petersen et al., 1986) is comparable with the tissues (except the liver) of mature fish. During development, the relative composition of the various lipid classes changes considerably in the egg, and FalkPetersen et al. (1989) showed that PUFAs were used in preference to saturates during the early stages of embryogenesis, whereas the opposite was the case during the later embryonic stages. The fatty acid profile in yolk-sac larvae prior to first feeding includes the same major components as ripe eggs (Bolla, 1988).

3. Vitamins During an investigation of vitamin A potency of the liver oils obtained from various species of fish, Lovern et al. (1933) found a remarkably high content

BIOLOGY OF THE ATLANTIC HALIBUT

37

of vitamin A in halibut. Large variations occurred from fish to fish (range 0.17-10%). Well-marked seasonal fluctuations occurred with an increase during summer and autumn. They also demonstrated both regional and fishsize variations. The latter was confirmed by Lovern and Sharp (1933), who found that the livers of older fish provided oil with a higher vitamin A content than that of younger fish. They were not able to find any relationship between the high vitamin A potency of halibut liver and the diet of the fish. The biological significance of the high vitamin A content of halibut liver is as yet unresolved. In an attempt to find a correlation between the occurrence of vitamins in the liver and skeletal muscles and the activity of the fishes, Brzkkan (1959) studied the relative distribution of the B vitamins niacin, riboflavin, pantothenic acid, vitamin B,, and thiamine in a number of fishes including halibut. With the exception of niacin, which showed highest concentrations in white muscle, all other B vitamins were present in much higher concentrations in the fatty liver and red muscles than in the relatively lean white muscle. Brzkkan (1959) proposed that the main function of red muscle was to carry out some metabolic activities normally taking place in the liver rather than performing normal musculature activity. The range (between organs) of the various B vitamins as measured in pg/g of halibut tissues given by Brzkkan (1959) were 30-44 for niacin, 3.6-10.8 for pantothenic acid, 0.6-7.3 for riboflavin, 0.009-1.0 for vitamin Biz, and 0.4-2.5 for thiamine. F. Energy Economics The lipid content of the halibut liver is high, and it is particularly rich in neutral lipids, the energy “fuel” for living organisms (Haug et af., 1988). During maturation and preparation for spawning, the relative weight of the liver is significantly reduced (Fig. 25), and Haug and Gulliksen (1988b) suggested that the liver must be an important energy source for halibut at this time. The view is supported by studies on the fatty acid composition of livers from spawning halibut, in which it was demonstrated that energy reserves usually deposited in the liver were depleted (Haug et af., 1988). In flatfish, it is generally accepted that both the liver and the carcass itself are the main storage sites for energy reserves (Love, 1970). In mature halibut, however, the carcass seems to be very little affected by the energy expenditure involved in the seasonal gonadal development (Fig. 25). Certainly, males show a small decrease in relative body weight during maturation and spawning, but no such reduction was observed in females (Haug and Gulliksen, 1988b). The observed differences between males and females

38

T. HAUG

in mobilizing muscle tissue into reproductive tissue seems to support suggestions of sexual differences in the physiology and growth/energy strategies in halibut. This is also manifested in female growth, which far exceeds that of the male, particularly after the attainment of sexual maturity (Mathisen and Olsen, 1968; Jakupsstovu and Haug, 1988), and in generally higher lipid levels in several tissues of the female compared with the male during spawning (Haug et al., 1988).

VIII. Parasites, Diseases and Pollution A. Parasites 1. Overall parasitic fauna In a study involving complete parasitological dissection, Zubchenko (1980) found a total of 23 species of parasites in 10 specimens of halibut caught along the continental slope of the Canadian Labrador coast. In the Barents Sea, 13 of these 23 parasites were found on 11 hosts (Polyansky, 1955). Polyansky found 18 species of parasites, internal as well as external, to be peculiar to the halibut. The following groups were represented: myxosporidians, cestodes, trematodes, nematodes, acanthocephalans and crustaceans. The parasites with the highest prevalence were the same on both sides of the Atlantic Ocean. Thus, the incidence of infection of parasitic species with a direct development was particularly high for the myxosporidian Ceratomyxa drepanopsettae and the crustacean Hatschekia hippoglossi. Among the remaining parasites with a complex developmental cycle, the incidence of infection by the cestode Scolex pleuronectis, the trematode Derogenes varicus, the nematode Contracaecum aduncum and the acanthocephalan Echinorhynchus gadi were reported to be high. Quite heavy infestation by S. pleuronectis in western Atlantic halibut was also reported by Ronald (1958a), while Polyansky (1955) found the larvae of Anisakis spp. to occur frequently in halibut from the Barents Sea. Because the halibut is a commercially important species, two aspects of halibut parasitology have attracted detailed attention. First, is the occurrence of the plerocercoid larvae of the cestode Grillotia erinaceus, which in certain areas may represent a commercial problem since it affects the flesh of the host (Rae, 1958). The second is the occurrence and prevalence of the ectoparasites Lepeophtheirus hippoglossi (Crustacea) and Entobdella hippoglossi (Trematoda). The latter (Fig. 26) has proved to be a potential pest in halibut aquaculture (Schram and Haug, 1988).

39

-

5 0.5

Y

X

w P

z

a w 2 -I

42

I

306

1-7

+-?

0

--

38

l3

Y

K

0.5

E0

I

308

a U 2

0

5z

0 0

0

imm.

5

6

i

0

FIG. 25. Variation in relative liver weight (Li) and condition factor ( K ) with the maturity stages 5 and 6 (maturing), 7 (running) and 8 (spent) in male (solid lines) and female (broken lines) halibut. Standard deviations (bars) and the number of fish examined are given. Imm = immature fish. Reproduced with permission from Haug and Gulliksen (1988b).

40

T. HAUG

FIG. 26. Entobdella hippoglossi. Ventral view of an adult specimen. Scale bar = I mm. Reproduced with permission from Schram and Haug (1988).

2. Grillotia erinaceus A comprehensive investigation by Rae (1958) revealed that halibut on certain northeast Atlantic grounds were parasitized by encysted plerocercoid larvae of the cestode G . erinaceus. The most seriously affected areas were found to the west of the British Isles, while halibut from more northerly grounds stretching from the North Sea, northwards along the coast of Norway and westwards through the Faroes and Iceland to Greenland, are free from infestation. Zubchenko (1980) did not report G . erinaceus as being present along the Labrador coast either, although no information exists as to the size of the halibut he examined. According to Rae (1958), immature halibut are not affected, and infestation is also insignificant in fish less than 90 cm in length. The larval parasites may occur in any part of the infected fish, although the heaviest concentrations of cysts are usually found on the outer wall of the stomach, on the mesenteries of the body cavity and on the surface of the liver. The life-cycle of G . erinaceus is by no means resolved, but it is known that skates and rays act as hosts to the sexually mature worm, while the encysted larval stage is abundant in a number of common food fishes (Johnstone, 1912). The early larval life of the parasite is most probably spent in copepods (Ruskowski, 1934). According to Rae (1958), the presence of the plerocercoid larvae in large fish such as the halibut can only be explained by the assumption that the copepods carrying the initial larval phase are first eaten by small fish. These are in turn eaten by the halibut before the larvae have had time to progress beyond the procercoid stage. Although the presence of the plerocercoid larvae in large halibut may represent a blind end to their life-cycle, since skates and rays are unlikely to eat these big fish, a

BIOLOGY OF THE ATLANTIC HALIBUT

41

possible role of larger sharks as final hosts (e.g. Greenland shark, Somniosus microcephalus) cannot be excluded (Rae, 1958). The regional differences observed in the food of halibut seem to provide a possible explanation as to why G . erinaceus is found in some areas and not in others. It is also possible that halibut are grouped in regional stocks as adults. The latter suggestion is also confirmed by tagging experiments (God0 and Haug, 1988a; Stobo et al., 1988). 3. Common ectoparasites

The skin of adult halibut is very often infected by the two easily visible ectoparasites L. hippogfossi, a motile parasitic caligid copepod, and E. hippoglossi, a semi-sessile monogenean trematode (Fig. 26). Schram and Haug (1988) reported that 66-78% of adult halibut in northern Norway were infected by the copepod, while the infection rate of the trematode was 50-60%. The large individuals tended to carry most parasites, and immature fish were free of them. This led Schram and Haug (1988) to suggest that halibut recruits are not infected until the adult population congregates for spawning. Ronald (1958b) also reports the frequent occurrence of this parasitic copepod on the skin of adult halibut in the Gulf of St Lawrence, Canada. The monogenean trematode E. hippoglossi has a single-host life-cycle, and Kearn (1974) found that the eggs hatch at night, that the ciliated oncomiradicium larvae are able to swim actively for at least 24 h at 4"C, and that some of them may be able to live without a host for over 3 days. Schram and Haug (1988) observed greater numbers of this trematode on female halibut. They suggested that this might be due to the more active spawning behaviour of males, making settlement by the ciliated oncomiradicium larvae more difficult.

B . Diseases There is no information available on the natural occurrence of diseases in halibut. Rae (1958) observed that halibut heavily infested by G . erinaceus could be in both poor and in good condition. He was unable to draw firm conclusions as to whether infestation by this parasite was lethal or whether it at least affected the general condition of the fish. In captive fish, it has been observed that infestation by the ectoparasite E. hippoglossi may be heavy enough to cause starvation and, ultimately, the death of the host (Ivar Holmefjord, Sunndals~ra,Norway, pers. comm.). Natural infestations, as observed by Schram and Haug (1988), are by no means as dramatic as

42

T. HAUG

those of captive halibut, and it seems rather doubtful that this parasitic trematode is lethal under natural conditions. Vibriosis attacks have been recorded quite frequently in captive halibut (e.g. Naas et al., 1987), showing that halibut may be susceptible to this disease. Whether vibriosis occurs in conditions of reduced stress in nature is, however. unknown.

C . Pollution To my knowledge the only analyses of pollution effects on halibut are those of the German Ministry of Public Health performed between 1974 and 1977. These were aimed at controlling a number of food fishes, particularly with respect to mercury (Kriiger et al., 1975; Kriiger and Nieper, 1978; Priebe, 1978). It was clearly shown that mercury, and also DDT, tended to accumulate in the flesh of the long-living halibut. In smaller halibut (less than 20 kg total weight), average mercury values of 0.32 p.p.m. were observed, while in halibut larger than 20 kg the average value was 1.O p.p.m., with a considerable number of large specimens exceeding the permissible German tolerance limit of 1.0p.p.m. (Kriiger et al., 1975; Kruger and Nieper, 1978). In the case of DDT, however, the German permissible maximum value (2.0 p.p.m.) was not exceeded, even by the larger specimens (Priebe, 1978).

IX. Exploitation A. Development in Total Catches In the northeast Atlantic, the halibut has always been an attractive species for European fishermen, and extensive directed fisheries - in some cases allied to a severe depletion of stocks - took place in many of the areas shown in Fig. 27 in the first half of the twentieth century (Devold, 1938; Bell, 1981). Natural fluctuations in stock sizes may have occurred (cf. McIntyre, 1952), but these were insignificant compared to the effect of the fisheries. The major fluctuations in stock abundance and changes in the fisheries may, thus, largely be attributed to the removal of man (Bell and Pruter, 1958). According to Bell and Pruter, this is particularly well illustrated by the response of the halibut stocks in the northeast Atlantic to the two major wartime closures or Kriegsschonzeiten. On each of these occasions, the abrupt reduction in total removals was followed by a sharp and temporary increase in stock abundance and catch, and long-liners emerged as the

43

BIOLOGY OF THE ATLANTIC HALIBUT 80"

70"

60"N

50"

40"

30" 70"W

60'W

50-W

40'W

30" w

20"W

1 O'W

FIG.27. Atlantic halibut fisheries areas in the northeast (I-VII) and northwest (VIII-X) Atlantic. The divisions used here are based on ICES (northeastern) and ICNAF/NAFO (northwestern) subareas and are defined as follows. I, Barents Sea/Spitzbergen (ICES subareas I and IIb); 11, Norwegian Sea (ICES subarea Ha); 111, east of UK (ICES subareas 111, IV and VIId,e,f); IV, west of UK (ICES subareas VI and VIIa,b,c,g,h,j,k); V, Faroes (ICES subarea Vb); VI, Iceland (ICES subarea Va); VII, east Greenland (ICES subarea XIV); VIII, west Greenland/Labrador (ICNAF/NAFO subareas 0, 1 and 2); IX, Newfoundland (ICNAF/ NAFO subarea 3); X, Gulf of St Lawrence/Nova Scotian banks (ICNAF/NAFO subareas 4,5 and 6).

dominant element in directed fisheries. After the two world wars, both stock abundance and total catches soon fell, and trawling supplanted lining as the major gear. After the immediate post-war increase in catches (Fig. 28), a decrease occurred in 1951-1955, followed by a sharp increase up to 1960. The latter is mainly due to the entry of the Soviet fleet to the halibut fisheries during this period. During the 1960s, a tremendous expansion occurred in Greenland halibut (Reinhardtius hippoglossoides) fisheries in the northeast Atlantic (see God0 and Haug, 1989). Unfortunately, this led to some statistical confusion, because Greenland halibut and Atlantic halibut were

44

T. HAUG

pooled in the fisheries statistics by some nations in certain years. In the 1960s, the ICES Bulletin Statistique gives the amount of halibut taken in 1962, 1963, 1965 and 1966, but Bell (1981) emphasized that these records most probably also included certain amounts of Greenland halibut. Thus, the true amount of Atlantic halibut landed from the northeast Atlantic during the 1960s is rather uncertain. Since the 1960s, however, stocks have continued to decline. Before 1970, almost all total yearly catches from the area were in excess of 10,000 tonnes, whereas the level after 1970 decreased from 7000 to about 3000 tonnes in 1981 and 1982. Recently, a slight increase has taken place, with a little over 4000 tonnes being landed in 1985.

.-. .----a

25-

t

AlIanIIc NW Allanllc

20-

c

-0

-85

NE

15-

/

g //Pi-.v 5

10-

.-

FIG.28. Total catches of Atlantic halibut from the northeast and northwest Atlantic after the Second World War. The lack of some data from the northeast Atlantic in the 1960s is due to pooling of catches of Greenland and Atlantic halibuts in the fisheries statistics by some nations. Data are from the ICES Bulletin Statistique (northeast Atlantic) and the ZCNAFINAFO Statistical Bulletin (northwest Atlantic).

The halibut was an important species for North American fishermen even at the beginning of this century, American boats having been involved in salt-halibut fisheries off western Greenland in the 1860s (Bell and Pruter, 1958). According to Nickerson (1978), as many as 4500 tonnes of halibut were harvested annually from the northwest Atlantic Ocean in the middle 1880s, whereas the yield in the 1920s and 1930s seems to have fluctuated between c. 3000 and 6000-9000 tonnes (Jespersen, 1938; McCracken, 1958). Following low-intensity fishing during the Second World War, there was a sharp increase in catches to a little over 7000 tonnes in 1950 (Fig. 28). Then after 2 years of decreased yields, an almost continuous increase in catches

BIOLOGY OF THE ATLANTIC HALIBUT

45

prevailed up to 1960 when almost 7000 tonnes were again taken. After 1960, catches declined to a level slightly above 2000 tonnes, which was maintained during the 1970s. In the 1980s, increases in catch have again occurred, with the most recent years yielding in excess of 4000 tonnes.

B. Catch by Area The most important halibut fishing areas in the northeast Atlantic after the Second World War have been the eastern Norwegian and Barents Seas, and the Icelandic and Faroese grounds (Fig. 29). After the First World War, the fisheries in the eastern Norwegian and Barents Seas (areas I and I1 in Fig. 27) expanded, in particular due to the extra effort put into the fisheries in the northern coastal and bank areas (Devold, 1938). Thus, landings in Norwegian ports increased from = 1000 tonnes per year in 1918 to well over 5000 tonnes in 1933 (Haug, 1984). Due to the increased stocks of fish during the Second World War (Mathisen and Olsen, 1968), 6000 tonnes of halibut were soon being landed again in Norwegian ports (Haug, 1984). After a peak in 1950, the yearly catches in the Norwegian Sea fluctuated around 3000 tonnes until the mid- 1960s. According to the ICES Bulletin Statistique, halibut catches increased dramatically in 1965. However, the inclusion of Greenland halibut in the statistics in the late 1960s calls into question their validity. Since 1970, Norwegian Sea catches have fallen, with present landings being less than 500 tonnes per year. In the early post-war years, the annual yield from the Barents Sea/ Spitzbergen was less than from the Norwegian Sea (750-2250 tonnes, see Fig. 29). An increase in catches seems to have occurred from 1959. Again, the pooling of Greenland and Atlantic halibut make the statktics for the 1960s unreliable. Since 1970, the catches taken in this area have been negligible. In the areas to the east and west of the U.K. (areas I11 and IV in Fig. 27), the eastern area (mainly the North Sea) has traditionally been the most important. Prior to the 1920s, yearly halibut catches of 2500-4000 tonnes were taken from this area, whereas the landings from 1922 to 1939 fell almost continuously to a level of z 1500 tonnes (Holden, 1978). As can be seen from Fig. 29, the general post-war tendency has been for a further decline in the catches, and the present catch level for the North Sea area is no more than 250 tonnes. The catch from the coastal and bank areas to the west of the U.K. has seldom exceeded 500 tonnes/year. The Icelandic grounds (area VI in Fig. 27) have always been important for European fishermen. Before the First World War, the total yield from these grounds decreased from 8000 to 3000 tonnes, then between 1918 and 1927 it

T. HAUG

46

remained between 5000 and 7000 tonnes, after which it decreased and stabilized at N 3000 tonnes (Jespersen, 1938). After the Second World War, the yield increased, with a peak of over 7500 tonnes in 1951 (Fig. 29). Subsequently, the catch has fluctuated considerably, with most catches after 1965 being less than 3000 tonnes/year and with a present day annual yield of N 2000 tonnes.

._____. I SpitzbJBarents Sea

la-,

- II Norw. Sea

8.

8.

I , , , 4

....

..

.....

.-.

,

.._.....

Vlll W Greenl./Labrador

FIG.29. Catches of Atlantic halibut in various parts of the northeast and northwest Atlantic. The subarea designations I-X refer to the map in Fig. 27. Some data are lacking from subareas I and I1 for the 1960s due to the pooling of Greenland and Atlantic halibut in the fishery statistics. Data are from the ICES Bulletin Siatisiique (northeast Atlantic) and the ICNAF/ NAFO Statistical Bulleiin (northwest Atlantic).

Prior to the First World War, the total annual yield from the halibut fisheries around the Faroes (area V in Fig. 27) fell quite evenly from just

BIOLOGY OF THE ATLANTIC HALIBUT

47

above 2000 tonnes in 1907 to less than 500 tonnes in 1915 (Jespersen, 1938). During the inter-war years, the yield fluctuated around 1000 tonnes/year. However, there was a fairly steady increase in the total annual yield after the Second World War, with a peak of more than 2700 tonnes in 1960 (Fig. 29). There has since been a steady decrease in the annual yield, with present-day catches around 700 tonneslyear. The catches from eastern Greenland (area VII in Fig. 27) were not given separately in the fishery statistics until 1958. As seen from Fig. 29, the annual yield of halibut from this area is quite low. In the northwest Atlantic, the Gulf of St LawrenceINova Scotian banks areas (area X in Fig. 27) have given the best halibut yields, closely followed by the Newfoundland area (area IX in Fig. 27) after the Second World War (Fig. 29). These two areas were also important halibut grounds prior to the war, and yielded together relatively stable landings of = 2500 tonnes/year, the majority being taken on hooked long-lines in directed fisheries (McCracken, 1958). For the Gulf of St Lawrence/Nova Scotia area, 1950 was the peak year with more than 4000 tonnes being taken, after which catches fluctuated around 2500 tonnes until 1965 (Fig. 29). In 1965-77, catches declined to = 1000 tonnes/year, whereas recent years have seen yields above 2000 tonnes per year. In the Newfoundland area 1951 and 1960 were the peak years, with more than 2500 tonnes being landed in each year. 0 in 1968-77, After 1960, catches declined and remained low ( ~ 7 5 tonnes below 500 tonnes in 1978-81) until 1981, after which there was a sharp increase to over 2000 tonnes. Except in a few years with particular low-level yields in the other northwest Atlantic areas, halibut landings from the Labradorlwest Greenland areas (area VIII in Fig. 27) have never attained the same magnitudes as those from the areas further south after 1945 (Fig. 29). Yearly output was above 1000 tonnes in 1954-61, with a subsequent decrease to less than 100 tonnes in the early 1970s. Catches then increased to levels between 500 and 1000 tonnes in the early 1980s, whereas the most recent yields have again dropped to below 100 tonneslyear. According to the ICNAFINAFO Statistical Bulletins, halibut fisheries in the Labrador area prior to the Second World War were of very little importance, whereas the western Greenland areas had been important since the latter half of the nineteenth century (see Jespersen, 1938; Bell and Pruter, 1958). C. Catch by Nations After the Second World War, Norway and the U.K. exploited most halibut from the northeast Atlantic (Fig. 30). At the beginning of this century,

T. HAUG

48

Iceland and Germany were also major exploiters (Devold, 1938; Jespersen, 1938). The main grounds for the U.K. fishermen have always been Icelandic waters, with substantial quantities also being taken in the North Sea and Faroese waters. The main fishing areas for the Norwegian fishermen have always been Norwegian and adjacent waters (see also Devold, 1938).

-sovlet

.-----Norway , .UK

,-.-;--

6.

‘

-

p f-

,A

;,

,/I’

-.-I*

, $l ,’

($‘>.

‘

\-.-.

<

I

II.\<

‘,,.\

;. . ,

\,

:

ts

-. .

. \

\---...,__-._..\

2:

is.

../

8 -

~

. . ’,/.. . _. ..__. .......-/-. . .. \-.-,*+-\. . . . , . .-..__. . ’\\\

I

.-.

.------. GDR ,

Iceland

,

-

~

,

,

..Faroes

: P>\:/\>” .--_._ _____

1 2

.

-

’ .

. . ’

,-.‘,-’./

__*_-C

6.

.

__r.

~

..

-“ . . . . . .d. ...,.._._._ .. . . ~

’

/ ” \ , , /

~..*.~,.~-~.-~-,~~~.~.~

.-

Canada

FIG. 30. Catches of Atlantic halibut in the northeast (above and middle) and northwest (below) Atlantic by major fishing nations. In the northeast Atlantic, some Soviet and GDR data are lacking for the 1960s due to the pooling of Greenland and Atlantic halibut in the fishery statistics. In the northwest Atlantic, the U.K. catches are only given until 1967, after which they were negligible. Data are from the ICES Bulletin Statistique (northeast Atlantic) and the lCNAFjNAF0 Statistical Bulletin (northwest Atlantic).

Iceland, with its local fishery, and the G.D.R., whose main halibut fisheries take place in Icelandic waters, have maintained their role as two of the more important halibut fishery nations since the Second World War (Fig. 30). During this period, the Faroe Islands and the U.S.S.R. have also played a fairly prominent role in halibut fisheries. The Faroe Islands have operated within their own area, while the U.S.S.R. takes halibut catches

BIOLOGY OF THE ATLANTIC HALIBUT

49

almost exclusively in the eastern Norwegian and Barents Seas. Soviet catches increased markedly from 1956 to 1965 (Fig. 30). However, the previously mentioned inclusion of Greenland halibut in the Atlantic halibut statistics in some years makes it impossible to assess the relative importance of the U.S.S.R. in the 1960s. After 1970, however, Atlantic halibut exploitation by the Soviets has ceased almost completely and is now negligible. With the exception of the Icelanders, who seem to have been able to maintain a more or less stable halibut fishery since the Second World War, all other nations have experienced a notable decrease in total yield from the northeast Atlantic grounds during this period. As seen from Fig. 30, most halibut landed from the northwest Atlantic after the Second World War were taken by the Canadian fleet, whose area of operation has always been the Gulf of St Lawrence and the Newfoundland and Nova Scotian banks. Canadian dominance in the northwestern Atlantic halibut fisheries was also conspicuous from 1928 up to the Second World War. Prior to this, the U.S.A. also played a prominent role in these fisheries, both in Greenland and in American home waters (Bell and Pruter, 1958; McCracken, 1958; Nickerson, 1978). The U.S.A. share of halibut catches had, however, almost ceased by 1940, and in the post-war years the annual yield has seldom exceeded 100 tonnes. According to Nickerson (1978) as many as 14 nations have been involved in the halibut fisheries of the northwest Atlantic. Of the non-American countries, Norway, the F.R.G. and the U.K. made the largest catches after 1947 (Fig. 30). The U.K. also took considerable amounts of halibut before the Second World War, particularly from Greenland waters (see Jespersen, 1938; Bell and Pruter, 1958). After the war, Norwegian fishermen took substantial amounts of halibut between 1954 and 1961 (off western Greenland and Newfoundland), whereas the F.R.G. has dominated the nonAmerican catches in the northwest Atlantic (particularly off western Greenland) after 1975. According to the ICNAFINAFO Statistical Bulletin, Greenlanders themselves also made some good halibut catches in their own home waters in the early 1980s.

D. 0 ver-exploitation

It has been noted that for halibut fisheries, increased effort is usually quickly followed by reduced stocks. This sequence of events is so consistent that one may predict the future of the fishery from its history (Bell and Pruter, 1958). The species may be particularly exposed to high fishing intensities from directed halibut fisheries using long-lines and gill-nets (see, e.g. Devold, 1938; McCracken, 1958; Mathisen and Olsen, 1968; Nickerson, 1978). There

50

T. HAUG

is also a danger of by-catches of smaller, immature halibut by vessels fishing for cod, haddock and other demersal species (see Jespersen, 1917; McCracken, 1958; Sigurdsson, 1976; Haug, 1984). By weight, these incidental landings may be negligible, but the numbers of halibut landed may often exceed the numbers landed from halibut fisheries and thus might reduce recruitment. Apparently, over-exploitation may also contribute to changes in the biological make-up of the halibut stocks. In the eastern Norwegian Sea, the yield from the fisheries declined substantially between the 1950s and the 1980s (Fig. 29A). Associated with this decrease in yield, was a marked increase in the growth rate and a substantial reduction in the age at first maturity in halibut caught in coastal areas off northern Norway (Fig. 31). Such observed changes in population parameters are generally associated with density-dependent factors, and a significant reduction in the halibut stocks may well have contributed to this. Certainly, the observed variations in the biological parameters of halibut should not be explained by one single factor alone, but rather by a set of factors. For example, similar changes in growth and maturation were observed both in cod and Greenland halibut in this area in the same period, and it has been suggested that, in addition to reduced stocks, enhanced temperatures could also be a contributory factor (see God0 and Haug, 1989). The simultaneous low abundance of several major demersal species may also have improved feeding conditions for the survivors, including Atlantic halibut.

E. Recruitment and Mortality There is very limited information available on the life-history of wild halibut until the juveniles appear on the nursery grounds at the age of almost 2 years. This is a serious obstacle that has prevented any estimation of larval recruitment of the species. Furthermore, it appears that no systematic attempts have been made to secure recruitment data for older fish (e.g. by virtual population analysis) in any part of the North Atlantic, although it has been suggested on a very gross scale that year-class strength may vary considerably (Sigurdsson, 1956). The information available on halibut mortality is scarce. From observations made in Greenland waters, it is known that late pelagic stages of the species are eaten by cod (Iversen, 1936; Magnusson and Palsson, 1988). The size attained by halibut makes it plausible that the fish rapidly reach a size that eliminates most enemies, leaving only large predators such as big sharks or whales (Andriyashev, 1954; Rae, 1958). Most probably, therefore, the mortality of large halibut is dominated by fishing.

BIOLOGY OF THE ATLANTIC HALIBUT

51

Based on age composition data from gillnet fisheries in northern Norway in 1955-65, Mathisen and Olsen (1968) calculated a total instantaneous mortality (natural mortality fishing mortality) for the stock as 0.39 for 0 females. The lower mortality in females was linked to males and ~ 0 . 2 for differences in vulnerability to the gill nets caused by sexual differences in behaviour on the spawning grounds where the gillnet fisheries took place.

+

FIG.31. Reduction in age at maturity in (a) male and (b) female halibut from northern Norway during the past 25-30 years. Broken lines indicate fish caught in 1 9 5 M 0 , solid lines indicate fish caught in 1981-85. Redrawn with permission after Haug and Tjemsland (1986).

F. Management Unlike the Pacific halibut fisheries, almost no exploitation of Atlantic halibut has been subjected to regulation except during the two world wars (when the fisheries almost ceased) and during periodic economic conditions which made the fishery unprofitable (Bell, 1981). With the general reduction of all stocks that occurred after the Second World War, many of the fisheries

52

T. HAUG

directed at halibut are no longer economic, particularly in the northeast Atlantic. The catches taken in this area today are mainly by-catch by trawlers and other vessels exploiting other demersal species. The complexity and size of trawl fisheries in Europe calls into question whether any practical measures can ever be taken to restore halibut productivity (see Bell, 1981). Considerable amounts of halibut are probably also taken incidentally in the northwest Atlantic, although the Canadians have been able to maintain a profitable but unregulated directed fishery, particularly in JCNAF/NAFO subareas 3 and 4 (Newfoundland, Gulf of St Lawrence, Nova Scotia, areas IX and X in Fig. 27) (see McCracken, 1958; Bowering, 1986). The only regulations imposed upon Atlantic halibut fisheries at present are aimed at gillnet fisheries for mature specimens on spawning grounds in Norwegian coastal areas. These regulations were introduced after the alarming reduction of spawning stocks after only 2 years of gillnet fishing, which began in 1936, due to the effectiveness of the nets (Devold, 1938). In addition to the general protection of immature fish less than 50 cm in length, the regulations from 1938 included both protection of spawners by a total prohibition of gillnet fisheries during the most intensive spawning period and mesh size regulations. In order to obtain a better basis for these regulations, detailed studies of the gillnet fishery and the spawning stocks of halibut were carried out in a selected area in northern Norway in 1955-65 (Tjemsland, 1960; Olsen and Tjemsland, 1963; Mathisen and Olsen, 1968; Olsen, 1969). These studies, which included the establishment of selection curves for the gill nets, analysis of age composition, growth and mortality rates in this particular gillnet fishery, and analysis of fishing effort and landing data, resulted in only minor changes in the 1938 regulations concerning adult fish. Better protection of immature fish was obtained, however, by prohibiting the marketing of fish less than 65 cm in length (Haug, 1984). These regulations have been maintained practically unchanged in the Norwegian halibut fisheries up to the present date, but the prevailing trend for the catches to decrease in Norwegian waters since 1965 has not changed (Haug, 1984). It may seem, therefore, that fisheries other than those affected by the regulations are responsible for the over-exploitation of the species in this area. In addition, the regulations may not be efficient, e.g. the by-catches of small fish may not be released overboard or else they do not survive the stress of capture. In Canadian waters, there are several management initiatives currently under development, including a minimum retained fish length of 81 cm. Of fish less than this size and obtained in exploratory trawl fishing, Neilson et al. (1989) reported survival rates of between 35 and 77%. They concluded that the ability to survive stress of capture would depend greatly on such factors as the type of gear used, handling time, total catch, fish length, maximum depth fished and trawl duration.

BIOLOGY OF THE ATLANTIC HALIBUT

53

X. Rearing Experiments and Aquaculture A.

The First Experiments

The first known experiments on the artificial fertilization of halibut eggs were carried out by Rollefsen ( I 934) at Trondheim’s Biological Station, Norway. Eggs were obtained from captive fish living in the station aquarium, and Rollefsen was able to incubate eggs through the embryonic period. He also succeeded in keeping some of the larvae alive 10 days after hatching. Norwegian hatching and rearing experiments with Atlantic halibut recommenced in 1974. Until 1982, all eggs used in the experiments were obtained from females captured with gill nets on the spawning grounds. The eggs were immediately fertilized in sea water, and then transported to laboratories where they were incubated in still or circulating water, at various temperatures, with or without the addition of antibiotics, in light or dark, and in various sized incubators (Solemdal et al., 1974; Llanning et al., 1982; Blaxter et al., 1983). A general feature observed in all these early halibut rearing experiments were very high salinities for neutral buoyancy (35.5-37%), in excess of the salinities of 35 %O usually recorded in Norwegian waters. According to RiisVestergaard (1 982), the negative buoyancies observed in many experiments were not due to any osmotic malfunctions, since halibut eggs are well able to cope with the osmotic stress of their natural environment. From the observed negative buoyancies in 35%0 salinity, Blaxter et al. (1983) suggested that halibut eggs might develop near or on the sea bed. Ichthyoplankton trawl surveys on the spawning grounds have, however, revealed that the eggs have a mesopelagic vertical distribution, and are usually found floating in salinities of 34.2-34.6960 (Haug et al., 1984, 1986). General stress, arising from the rough treatment given to mature fish caught in gill nets, the rapid ascent from the great depths on the spawning grounds to the surface, and the mechanical effects of stripping on newly caught fish, may all have contributed to the observed density differences between natural and artificially spawned eggs (Haug er al., 1984). The survival of eggs and larvae in the first rearing work depended to some degree on the conditions of the experiments. None of the experiments were successful in producing any numbers of vigorous offspring capable of feeding, growth and survival. In experiments by Solemdal er al. (1974), the longest living larvae survived for 60 days, but only three specimens showed signs of food (Artemia nauplii) intake. Blaxter et al. (1983), who offered the larvae zooplankton as food, observed that very few larvae fed, and only two specimens were observed to have passed metamorphosis to reach a length of 24 mm after 90 days.

54

T. HAUG

B. Brood Stocks of Adult Spawners With the expansion in sea-ranching and aquaculture in Europe in recent years, there has been a growing interest in the introduction of new species, of which the Atlantic halibut has been of particular interest. This gave rise to the increased use of the species in rearing experiments in the 1980s. However, the quality and quantity of gametes stripped from newly caught fish on the spawning grounds were highly unsatisfactory, and the lack of stable broodstock was a severe obstacle to further development and progress in the cultivation experiments. For this reason, broodstocks of adult halibut were established, and the supply of gametes to the ongoing Norwegian (and other) rearing experiments has been ensured since 1983. The fish used as broodstocks are adult specimens caught on long-lines outside the spawning season and kept in large tanks supplied with a continuous flow of sea water, preferably with salinities and temperatures similar to those recorded near the bottom on the spawning grounds. Adult halibut can be kept alive, fed and stripped through several spawning seasons in captivity (Rabben et al., 1986). The halibut is a multiple spawner, i.e. it gives off various sized batches of ovulated eggs according to its ovulatory rhythm at more or less regular intervals throughout the spawning season. Stripping is the main method for obtaining eggs (Rabben, 1987). During stripping, the female fish is usually led towards a board covered with a very smooth “mattress” of neoprene (to avoid severe mucus loss) and gently lifted above the water surface. The fish is kept unanaesthetized, and eggs, milt and sea water are mixed simultaneously in a wet fertilization method at a ratio of 100:1:100 (Rabben et al., 1986; Rabben, 1987). Rabben emphasizes that stripping of a multiple spawner must be very stressful, and that disturbance of the fish in either of the crucial stages of oogenesis will influence the quality and quantity of eggs. With females kept one or more years in captivity, obviously swollen ovaries have been observed to regress without egg release. Careful methodology and proper knowledge of all aspects of the general reproductive biology of the halibut, and of the ovulatory rhythm of each particular fish, is therefore necessary to obtain acceptable results. Natural spawning in very large basins is now being investigated, and the first spawnings took place in 1987 (Rabben, 1987). This may represent a labour-saving method for the production of large egg volumes with a minimum of human effort. In the domestication process of halibut, however, there is an obvious need for selective breeding. Controllable fertilization methods using stripping may then be essential. Parasites can represent a problem in monocultures of fish, and in captive broodstock halibut very heavy infestations by the ectoparasite Entobdella

BIOLOGY OF THE ATLANTIC HALIBUT

55

hippoglossi (Fig. 25) have been observed (see Schram and Haug, 1988). If repeated a t 3- to 4-month intervals, formaldehyde treatment has proved to be an effective control. Also, treatment with fresh water seems to kill the adult trematodes (Yngvar Svendsen, University of Tromser, Norway, pers. comm.). The presence of sand or gravel as a substratum in the broodstock tanks seems to reduce the vulnerability of fish towards Entobdella attacks (Ivar Holmefjord, Sunndalsma, and Ingvar Huse, Austevoll, Norway, pers. comm.). Other problems experienced with captive halibut include erosion of the fins and tail, wounds of varying severity, blindness and sunburn. Experience from Icelandic experiments with halibut in captivity has shown that the broodstock is very sensitive to direct sunlight, and should be shielded and given only very limited amounts of light in land-based aquaculture (Bjerrnsson, 1986).

C . Egg Incubation and Hatching In sea water, the fertilization rate of newly stripped halibut eggs fertilized by the wet method may vary, but usually values above 90% may be obtained (Mangor-Jensen and Jelmert, 1986; Rabben et al., 1986; Jelmert and Rabben, 1987). Mangor-Jensen and Jelmert (1986) demonstrated that fertilization rates decreased with decreasing salinity but that high fertilization rates could be obtained at low salinities if the concentration of calcium was kept near that of sea water. Dry fertilization is possible but less effective than wet fertilization. The eggs also maintain their ability to be fertilized, albeit at a lower rate, for several hours after stripping, provided they are stored in the proper conditions (Kjerrsvik, in press). The salinity for neutral buoyancy of fertilized eggs from broodstock halibut varies to some extent. In general, there is a slight negative buoyancy compared with the water used for incubation. This is most probably attributable to the incubation water being taken from the upper layers (not deeper than 50-55 m; see Jelmert and Rabben, 1987) where the salinity is lower than in the deeper bathypelagic sub-pycnocline water to which the specific gravity of halibut eggs is probably preadjusted (see Haug et al., 1984, 1986). When using still water for incubation, which may in some experimental cases be necessary and/or preferred, the salinity of the water has to be raised by adding salt in order to prevent the eggs from sinking to the bottom of the rearing vessel. The most frequently used egg incubation method in present rearing experiments is, however, based on the use of a so-called upwelling incubator (Rabben et al., 1986; Jelmert and Rabben, 1987; see Fig. 32). Here the fertilized eggs are incubated in a slowly rising flow of sea water. Usually, 3 1

T. HAUG

56

of eggs (1: 120,000) are incubated in each container in which the water flow is set at approximately 3I/min. The water is sand-filtered (average 10-pm pore size), cartridge-filtered (5-pm pore size) and UV-treated before being allowed to enter the incubators (Jelmert and Rabben, 1987). I

k---

750ff~n

+I

I

E E

0 0 t

FIG.32. Diagram of the most frequently used halibut egg incubator vessel (not drawn to scale). A, Water outlet; B, sieve of plankton net, 1000 pm; C, water inlet with 250-pm sieve; D, valve; E, support. The volume of the incubator is approximately 2501. Redrawn with permission after Jelmert and Rabben (1987).

Bacterial activity may cause some egg mortality. Jelmert and Rabben (1 987) have shown good correlations between egg mortality and bacterial

activity in the incubators. They observed, however, that removal of dead eggs caused a marked decrease in bacterial activity, and found it difficult to decide whether inherent mortality or bacterial activity was the more important cause of death. Their data leave no doubt, however, that the husbandry techniques are of great importance, and that any accumulation of dead or dying eggs, which may serve as a substratum for bacterial activity, should be

BIOLOGY OF THE ATLANTIC HALIBUT

57

avoided. Given proper care, which includes observations and removal of dead eggs once a day, it appears that quite low egg mortality during the incubation period can be obtained. Thus, Rabben et al. (1986) observed that less than 13% of the eggs died during incubation in some of their experiments. The fertilized eggs are sensitive to mechanical stress before the closure of the blastopore, whereas after this event they appear considerably more robust (Holmefjord and Bolla, 1988). Hatching percentages have been observed to vary in different environments. Rabben and Jelmert (1986) found that in completely dark incubators, a hatching of nearly 99% was obtained giving positively buoyant larvae; in illuminated incubators, hatching was delayed and gave only a final hatching of 70-80%. In the latter experiment, both the eggs and the larvae developed a slight negative buoyancy. Rabben et al. (1986) also observed a lowered frequency of hatching (93%) in still water as compared with circulating water.

D. Rearing of Yolk-sac Larvae In rearing experiments with halibut larvae, the mortality is high, and many of the larvae end up with deformities such as bent notochords, deformed mouths and swollen yolk-sacs (Sendstad, 1984; Pittman et al., 1987, 1989). Opstad and Raae (1986) found that the larvae are very sensitive to mechanical stress, and that such stress is likely to create malformations. The importance of intact neuromast organs has also been stressed, and it has been suggested that mechanical stress may damage these organs to such an extent that the larvae are incapable of survival (0kland et al., 1986). Furthermore, Rabben and Jelmert (1986) found that exposure to light stressed yolk-sac larvae, making them vulnerable to microbial attacks later in the post-hatching period. This was also the conclusion reached by Rabben et al. (1986), who showed that antibiotic treatment had a beneficial effect on a short-term basis, but a harmful effect on the development of larvae late in the yolk-sac period. Excess light also caused the development of a slightly negative buoyancy, indicating disturbances in the permeability of chorion and/or the vitelline membrane. A higher percentage of normal larvae were observed under conditions of darkness compared with other light regimes by Bolla and Holmefjord (1988). These authors also observed that the percentage of normal larvae was lower at 10°C than at 6 and 2°C. The importance of rearing temperatures was also emphasized by Pittman er al. (1989), who found jaw deformations and oedema to be most common at 9"C, while growth and yolk absorption were better at 4°C than either 9 or 6°C. Pittman ef al. also found that the frequency and duration of active periods increased

58

T. HAUG

as the larvae absorbed the yolk supply, and energy concurrently became a limiting factor. It is evident, therefore, that the rearing of halibut yolk-sac larvae should be performed with the minimum of light and mechanical sources of stress should be avoided. During the latter part of the yolk-sac stage, the larval density should be as low as possible, the temperature cold and the disturbances minimal to give large and healthy larvae. Most of the earlier rearing experiments took place in small laboratory units. In small volumes of water the chances of frequent contact with walls and other larvae are increased, and it was soon acknowledged that larger systems might be advantageous, particularly in order to avoid mechanical stress. The first rearing experiments to produce metamorphosed larvae (2 specimens, 30 mm long) are described by Blaxter et al. (1983). Black plastic bags (2-m3 volume) with a conical bottom were used as rearing systems, and they had an increased overall salinity and a higher salinity near the bottom of the bag; zooplankton was supplied as food. In the more recent rearing experiments, this method has been further elaborated. Encouraged by successes in producing juveniles of other marine fish in mesocosms such as plastic bags, ponds and basins (see Berg et al., 1985), present-day halibut units are now considerably larger than those used by Blaxter et a1 (1983). A strong desire for commercially feasible methods of large-scale production of juvenile halibut has also contributed to this development. At present, most halibut rearing experiments are conducted in large systems, two of which are reviewed here. The most often used system consists of large (10-12 m3) floating plastic bags (Berg et al., 1985, 1987; Berg and Oiestad, 1986; Rabben et al., 1986), while more recently a land-based system using 3.5-m3 silos has been developed (Rabben et al., 1987; Pittman et al., 1989). Both holding of yolk-sac larvae (40-50 days) and feeding experiments have taken place in the floating bags, while only holding experiments have been carried out so far in the silos. The experimental units used in the plastic bag experiments (Fig. 33) are described by Berg and 0iestad (1986) and Rabben et al. (1986). The plastic bags are attached to a floating collar and covered with a black plastic conical roof (to avoid light). Each bag is fitted with a conical bottom. A flexible hose attached to the bottom of the cone allows the bag to be drained into a submerged tank. Filtered sea water is used in preference to unfiltered water. In experiments with still water, the bags were supplied with a layer of high salinity water (38-40%0) in the lower part of the cone, and an upper layer of brackish water (25-30%0) near the surface. The larvae can be transferred from the egg incubators to the rearing systems either just before (Rabben et al., 1986).or just after hatching (Berg and Oiestad, 1986). Throughout the course of the experiments, all dead or living larvae trapped in the high

BIOLOGY OF THE ATLANTIC HALIBUT

59

salinity bottom layer are regularly drained out and fixed. One disadvantage of this method is that the enclosures are subjected to considerable diurnal and other variations in temperature. Provided deep water of a stable temperature is available, the plastic bags can be kept in larger plastic basins with a flow of deep water, so stabilizing the temperature in the bags (Berg et al., 1987).

&

TRIANGLE OPENING

-

1

E - 2 I k-

a W

0 3 4 I

//

FIG.33. Plastic bag system (1 1.5 m3) used in rearing experiments with halibut fry. Redrawn with permission after Berg and Oiestad (1986).

The long yolk-sac period of halibut fry (40-50 days at 5-6°C) gives a long holding period during which very large losses of larvae may occur (Fig. 34). The survival rates vary considerably from experiment to experiment, depending on factors such as water quality and egg quality. Nevertheless, as many as 40-45% of the hatched larvae survive holding throughout the whole yolk-sac period until start-feeding in certain floating bag experiments (Berg and Oiestad, 1986; Naas et al., 1987). During the yolk-sac period, there s e e w to be a period of particularly high mortality between days 20 and 30 (Fig. 34). According to Pittman et al. (1987), this period coincides with the development of the mouth, an event which seems to be very susceptible to bacterial activity that may damage the important oral membrane and cause malformation of the jaws.

_i\,

60

100 100.

,

,(\\\,

T. HAUG

‘.

80

t

70.70t

i5 60--

35 50-: 8 Y

40-

302010-7

l000 1 LARVAL AGE (days)

.io.

FIG.34. Relative survival (YO) of larvae based on number of dead larvae found in the bottom water during a 60-day experiment in floating plastic enclosures. The arrow indicates when food was first found in the guts. Redrawn with permission after Berg and Oiestad (1986).

The use of land-based silos in halibut rearing experiments is an alternative method that combines the volume and shape advantages of the floating plastic bags with possible better environmental control (Rabben et al., 1987; Pittman et al., 1989). Basically, the silos (Fig. 35) are operated in the same manner as the floating enclosures, although some very promising results have emerged from experiments with slow upward flows of water in the silos, where survival rates have proved to be considerably higher (Fig. 36) than in floating enclosures (Rabben et al., 1987). Rabben et al. believe that the improved survival rates are due partly to the “lift” given to the larvae, thus keeping them away from the bottom water, and partly to the improved water quality resulting from the continuous supply of clean water.

E. Start-feeding of Larvae A second distinct period of larval mortality is observed in the floating enclosures from days 50 to 60 (Fig. 34), which is the period of start-feeding (Berg and Oiestad, 1986; Pittman et al., 1987). Even though considerable numbers of reared larvae have been observed to eat and grow as well, the

61

BIOLOGY OF THE ATLANTIC HALIBUT

number of fry produced so far has not been substantial. The mean overall survival percentages from hatching to the post-metamorphosis stage have varied from 1.0 to 3.3% (Berg and Oiestad, 1986; Rabben et al., 1986; Naas et al., 1987). All of these experiments were carried out in floating enclosures at Austevoll Marine Aquaculture Station, Norway, and the number of viable, metamorphosed larvae produced there during 1986 and 1987 was a little over 900. Problems in collecting metamorphosed larvae from the enclosures (Rabbeq et al., 1986) and attacks of vibriosis after metamorphosis (Naas et al., 1987; Pittman et al., 1987) resulted in the death of many otherwise viable fry, so considerably reducing the number of potential survivors.

FIG. 35. Diagram of the silos used in the rearing of halibut fry. I , Header tank containing 40% salinity water; 2, mixing tank containing 40960water; 3, submerged pump; 4,mixer; 5, level regulator; 6, silo; 7, overflow sieve; 8, header tank containing deep water; 9, discharge pipe. Reproduced with permission from Rabben e l al. (1987).

6'ooo1 \

0 2.000

P

*\ 1 .ooo{

*\ '0

L o - .

5

1'0

25 DAYS AFTER HATCHING 1'5

20

rn

30

35

70

FIG.36. Halibut larval survival in three different silo experiments. A, Stagnant water with salt-plug until day 10, later flow; B, continuous flow; C, stagnant water with salt-plug throughout experiment. Redrawn with permission after Rabben el al. (1987).

62

T. HAUG

The food given to the larvae is generally cultivated or natural zooplankton, usually at a density near 20 food organisms/litre. The food includes calanoid copepods in various stages, cladocerans and other crustaceans, bivalves and gastropods, all of which have been eaten by the larvae (Berg and Oiestad, 1986; Naas et al., 1987). The size of the prey (not less than 0.4mm) and halibut predator, in the latter particularly the mouth size, combined with the relative abundance of the prey organisms, determines which prey items are eaten (Sendstad, 1984). Berg and Oiestad (1986) suggested that the older larvae were able to graze down zooplankton to a very low level despite fairly low concentrations of suitable prey items. In start-feeding experiments conducted in smaller systems, Bolla (1988) observed that the proportions of different fatty acids remained unchanged in larvae fed collected plankton, whereas larvae fed cultivated feeds showed marked differences from the yolk-sac stage. Depletion of PUFAs were particularly conspicuous in the latter group, and this led Bolla (1988) to conclude that Atlantic halibut larvae are probably unable to synthesize PUFAs from precursors. Since the growing larvae require PUFAs, the fatty acid composition is an important factor for the estimation of the nutritional value of a feed.

F. Growth Experiments with Juvenile Halibut To test if the relatively slow growth rate observed in halibut in nature can be increased in captivity, some growth experiments have been performed in Norway with wild juvenile halibut (2, 3 and 4 years old) captured in Danish seines in halibut nursery areas (Haug et al., 1989b). These fish were given food ad libitum and were kept in conditions which reduced the possibility of energy expenditure. It was seen that their general condition (length-weight relation) was raised to a higher level in captivity compared with the level at capture. Growth was at its most intensive during summer and autumn (as in nature), but the growth rate in captivity was clearly higher. In fact, Haug et al. (1989b) observed that captive halibut grew faster and added more somatic tissue to their bodies during 1 year in captivity than during 2 years in Nature (Fig. 37). Recent observations of hatchery-reared juveniles also indicate that growth rate can be increased considerably during the first years in captivity compared with natural conditions (Rernnestad and Kirdal, 1989). In more detailed feeding experiments, Davenport et al. (in press) observed that juvenile halibut caught at sea preferred few, but large meals. The mean satiation meal was 11.7% of body weight and total gut clearance time was 120 h. These authors also recorded the oxygen uptake of juvenile halibut as

63

BIOLOGY OF THE ATLANTIC HALIBUT

0.07-0.1 1 ml O,/g/h, depending on nutritional state, and concluded that a flow rate of about 2001 sea water/kg fish weight/h is required in a tank containing halibut.

5.01 5 3.0-

P

12.0B

P 1.01 0 k

21‘8s

4ies sies eies TIME (month/year)

iiies

lie7

FIG.37. Results from growth experiments with immature halibut carried out in northern Norway from October 1985 to January 1987. The graphs show the increase in average total weight for fish from each of the three year classes (1981, 1982, 1983) represented in the experiment (based on data from Haug ef al., 1989b).

XI. Acknowledgements Sincere thanks are due to 0. R. God0 and I. Huse who read critically through the manuscript and gave valuable comments. Financial support was received from the Norwegian Council of Fisheries (NFFR), project nos 1405.003and V405.008.

XII. References Ahlstram, E. H., Amaoka, K., Hensley, D. A,, Moser, H. G. and Sumida, B. Y. (1984). Pleuronectiformes: Development. In “Ontogeny and Systematics of Fishes - Ahlstrom Symposium” (H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr and S. L. Richardson, eds), pp. 640670.Special Publication No. 1, American Society of Ichthyologists and Herpetologists. Andriyashev, A. P. (1954). “Fishes of the Northern Seas of the USSR”. Trudy Zoologicheskogo Institut Akademiya Nauk SSSR, No. 53. (Translated from Russian by the Israel Program for Scientific Translations, IPST Cat. no. 836.) Bell, F. H. (1981). “The Pacific Halibut, the Resource, and the Fishery”. Alaska Northwest Publishing, Anchorage.

64

T. HAUG

Bell, F. H. and Pruter, A. T. (1958). Climatic temperature changes and commercial yields of some marine fisheries. Journal ofthe Fisheries Research Board of Canada 15, 625-683. Berg, L. and Oiestad, V. (1986) Growth and survival studies of halibut (Hippoglossus hippoglossus) from hatching to beyond metamorphosis carried out in mesocosms. Council Meeting of the International Council for the Exploration of the Sea F16, 19 pp. (mimeo). Berg, L., Baarery, V., Danielssen, D. S., Meeren, T. V. d., Naas, K. E., Sendstad, K. and Oiestad, V. (1985). Production of juvenile flatfish species in different sized mesocosms. Council Meeting of the International Councilfor the Exploration of the Sea F65,23 pp. (mimeo). Berg, L., Naas, K . and Pittman, K. (1987). Deepwater flowthrough as a temperature stabilizer in rearing of halibut (Hippoglossus hippoglossus) fry. Council Meeting of the International Council for the Exp!oration of the Sea F16, 11 pp. (mimeo). Bjerrnsson, B. (1986). Kveiteoppdrett pa Island. Nordisk Aquakultur 2(6), 6 9 . Blaxter, J. H. S . , Danielssen, D., Moksness, E. and Oiestad, V. (1983). Description of the early development of the halibut Hippoglossus hippoglossus and attempts to rear the larvae past first feeding. Marine Biology 73, 99-107. Bolla, S . (1988). Fatty acid composition of Atlantic halibut larvae fed on enriched Brachionus, Artemia and collected plankton. International Council for the Exploration of the Sea, Early Life History Symposium, Paper No. 112, 10 pp. (mimeo). Bolla, S. and Holmefjord, I. (1988). Effect of temperature and light on development of Atlantic halibut larvae. Aquaculture 74, 355-358. Bowering, W. R. (1986). The distribution, age and growth and sexual maturity of Atlantic halibut (Hippoglossus hippoglossus) in the Newfoundland and Labrador area of northwest Atlantic. Canadian Technical Report of Fisheries and Aquatic Sciences 1432, 34 pp. Brzekkan, 0. R. (1959). A comparative study of vitamins in the trunk muscle of fishes. Fiskeridirektoratets Skrifter Serie Teknologiske Unders0kelser 3(8), 1 4 2 . Cox, P. (1924). Larvae of the halibut (Hippoglossus hippoglossus L.) on the Atlantic coast of Nova Scotia. Contributions to Canadian Biology, Fisheries 1(21), 41 1-412. Davenport, J., Kjsrsvik, E. and Haug, T. (in press). Appetite, gut transit, oxygen uptake and nitrogen excretion in captive Atlantic halibut Hippoglossus hippoglossus (L.) and lemon sole Microstomus kitt (Walbaum). Aquaculture. Devold, F. (1938). The North Atlantic halibut and net fishing. Fiskeridirektoratets Skrifter, Serie Havunders~kelser5, 1-47. Devold, F. (1939). Kveiteunderserkelsene i 1938. Fiskeridirektoratets Skrifter, Serie Havunders~kelser6,85-96. Devold, F. (1943). Notes on halibut (Hippoglossus vufgaris Fleming). Annales Biologiques, Copenhague 1, 35-40. Falk-Petersen, S., Falk-Petersen, I.-B., Sargent, J. R. and Haug, T. (1986). Lipid class and fatty acid composition of eggs from the Atlantic halibut (Hippoglossus hippoglossus). Aquaculture 52, 207-21 1. Falk-Petersen, S., Sargent, J. R., Fox, C., Falk-Petersen, I.-B., Haug, T. and Kjerrsvik, E. (1989). Lipids in Atlantic halibut (Hippoglossus hippoglossus L.) eggs from planktonic samples in northern Norway. Marine Biology 101, 553-556. Fevolden, S. E. and Haug, T. (1988). Genetic population structure of Atlantic halibut, Hippoglossus hippoglossus (L.). Canadian Journal of Fisheries and Aquatic Sciences 45, 2-7.

BIOLOGY OF THE ATLANTIC HALIBUT

65

Fulton, T. W. (1891). The comparative fecundity of sea fishes. 9th Annual Report of the Fisheries Board, Scotland 3, 243-268. God0, 0. R. and Haug, T. (1988a). Tagging and recapture of Atlantic halibut, Hippoglossus hippoglossus, in Norwegian waters. Journal du Conseil pour I'Exploration de la Mer 44, 169-179. God0, 0. R. and Haug, T. (1988b). Tagging and recaptures of Atlantic halibut (Hippoglossus hippoglossus L.) on the continental shelves off eastern Canada and off western and eastern Greenland. Journal of Northwest Atlantic Fisheries Science 8, 25-3 I . God0, 0. R. and Haug, T. (1989). A review of the natural history, fisheries and management of Greenland halibut, Reinhardtius hippoglossoides, in the eastern Norwegian and Barents Seas. Journal du Consrilpour I'Exploration de la Mer 46, 62-75. Grant, W. S., Teel, D. J., Kobayashi, T. and Schmitt, C. (1984). Biochemical population genetics of Pacific halibut (Hippoglossus stenolepis) and comparison with Atlantic halibut ( H . hippoglossus). Canadian Journal of Fisheries and Aquatic Sciences 41. 1083-1088. Gudger, E. W. and Firth, F. E. (1937). Two reversed partially ambicolorate halibuts Hippoglossus hippoglossus. American Museum Novitates, N. Y. 925, 1-10. Haug, T. (1984). Utvikling og reguleringer i det norske kveitefisket. Fiskets Gang 70, 117-121. Haug, T. and Fevolden, S. E. (1986). Morphology and biochemical genetics of Atlantic halibut, Hippoglossus hippoglossus (L.), from various spawning grounds. Journal of Fish Biology 28, 367-378. Haug, T. and Gulliksen, B. (1988a). Fecundity and egg sizes in ovaries of female Atlantic halibut, Hippoglossus hippoglossus (L.). Sarsia 73, 259-261. Haug, T. and Gulliksen, B. (1988b). Variation in liver- and body condition during gonad development of Atlantic halibut, Hippoglossus hippoglossus (L.). Fiskeridirektoratets Skrifter, Serie Havundersekelser 18, 35 1-363. Haug, T. and Sundby, S . (1987). A preliminary report on the natural occurrence and ecology of Atlantic halibut, Hippoglossus hippoglossus, postlarvae and young immature stages. Council Meeting of the International Council for the Exploration q f t h e Sea F38, 29 pp. (mimeo). Haug, T. and Tjemsland, J. (1986). Changes in size- and age-distributions and age at sexual maturity in Atlantic halibut, Hippoglossus hippoglossus, caught in North Norwegian waters. Fisheries Research 4, 145-1 55. Haug, T., Kjerrsvik, E. and Solemdal, P. (1984). Vertical distribution of Atlantic halibut (Hippoglossus hippoglossus) eggs. Canadian Journal of Fisheries and Aquatic Sciences 41, 798-804. Haug, T., Kjnrsvik, E. and Solemdal, P. (1986). Influence of some physical and biological factors on the density and vertical distribution of Atlantic halibut Hippoglossus hippoglossus eggs. Marine Ecology, Progress Series 33, 207-2 16. Haug, T., Ringer, E. and Pettersen, G. W. (1988). Total lipid and fatty acid composition of polar and neutral lipids in different tissues of Atlantic halibut, Hippoglossus hippoglossus (L.). Sarsia 73, 163-168. Haug, T., Kjerrsvik, E. and Pittman, K. (1989a). Observations o n a wild Atlantic halibut larva (Hippoglossus hippoglossus). Journal of Fish Biology 34,799-801. Haug, T., Huse, I., Kjerrsvik, E. and Rabben, H. (1989b). Observations on the growth of juvenile Atlantic halibut (Hippoglossus hippoglossus L.) in captivity. Aquaculture 80, 79-86.

66

T. HAUG

Helvik, J. V. (1988). “Klekkeprosessen hos kveite (Hippoglossus hippoglossus)”. Cand. Scientific thesis, University of Bergen, 87 pp. Hensley, D. A. and Ahlstrerm, E. H. (1984). Pleuronectiformes: Relationships. In “Ontogeny and Systematics of Fishes - Ahlstrom Symposium” (H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr and S. L. Richardson, eds), pp. 670-687. Special Publication No. 1, American Society of Ichthyologists and Herpetologists. Hjort, J. (1905). “Norges Fiskerier I. Norsk Havfiske 2. del. Fiskeforsog og fangsfelter”. Selskabet for de norske fiskeriers fremme, Bergen. Holden, M. J. (1978). Long-term changes in landings of fish from the North Sea. Rapports et ProcPs- Verbaux des RPunions, Conseil Permanent International pour I’Exploration de la Mer 172, 11-26. Holmefjord, I. and Bolla, S. (1988). Effect of mechanical stress on Atlantic halibut eggs at different times after fertilization. Aquaculture 68,369-371. Iversen, T. (1936). Sydestgrmland, Jan Mayen. Fiskeriundersekelser. Fiskeridirektoratets Skrifter Serie Havundersekelser 5( l), 1-1 7 1. Jakupsstovu, H. (1986). Kalvakanningar vid Magnusi Heinasyni november 1983januar 1984. Fiskirannsoknir 1986(3), 3146. Jakupsstovu, S. H. I. and Haug, T. (1988). Growth, sexual maturation and spawning season of Atlantic halibut, Hippoglossus hippoglossus, in Faroese waters. Fisheries Research 6, 201-215. Jakupsstovu, S. H. I. and Haug, T. (1989). Spawning of Atlantic halibut Hippoglossus hippoglossus in deep waters on the continental slope south west of the Faroe Islands. Frodskaparrit 34/35,79-80. Jelmert, A. and Rabben, H. (1987). Upwelling incubators for eggs of the Atlantic halibut (Hippoglossus hippoglossus L.). Council Meeting of the International Council for the Exploration of the Sea FZO, 8 pp. (mimeo). Jensen, A. C. and Wise, J. (1961). Movements of tagged halibut off new England - 11. Transitions of the American Fisheries Society 90, 489490. Jensen, A. S. (1926). Investigations of the “Dana” in west Greenland waters, 1925. Rapports et ProcPs- Verbaux des RPunions, Conseil Permanent International pour I’Exploration de la Mer 39, 85-102. Jespersen, P. (1917). Contribution to the life history of the North Atlantic halibut (Hippoglossus vulgaris Flem.). Meddelelser f r a Kommisjonen f o r Havunders0gelser, Serie Fiskeri 5(5), 1-32. Jespersen, P. (1926). On the halibut in Icelandic waters. Rapports et Proc2s- Verbaux des RPunions, Conseil Permanent International pour I’Exploration de la Mer 39, 103-1 13. Jespersen, P. (1938). Statistical survey of the halibut fishery in the waters round the Farms, Iceland and Greenland. Meddelelser f r a Kommisjonen f o r Danmarks Fiskeri- og Havunders0gelser, Serie Fiskeri 10(5), 1-37. Joensen, J. S. (1954). On the life history of halibut in Faroese waters. Meddelelser fra Danmarks Fiskeri- og Havundersegelser, Ny Serie 1(5), 1-25. Johnstone, J. (1912). Tetrarhynchus erinaceus van Beneden. I. Structure of Larva and adult worm. Parasitology 4, 364-415. Jonsson, G. (1978). Merkingar d ludu. Sjavarjrettir 6(12), 26-28. Kearn, G. C. (1974). Nocturnal hatching in the monogenean skin parasite Entobdella hippoglossi from the halibut, Hippoglossus hippoglossus. Parasitology 68,161-1 12. Kjersvik, E. (in press). Development of Atlantic halibut (Hippoglossus hippoglossus) eggs during different incubation conditions. Journal of Fish Biology.

BIOLOGY OF THE ATLANTIC HALIBUT

67

Kjerrsvik, E., Haug, T. and Tjemsland, J. (1987). Spawning season of the Atlantic halibut (Hippoglossus hippoglossus) in northern Norway. Journal du Conseil Internationale pour Exploration de la Mer 43, 285-293. Kohler, A. C. (1964). Movements of halibut o n the Nova Scotian and Grand Banks. Journal of the Fisheries Research Board of Canada 21, 837-840. Kohler, A. C. (1967). Size at maturity, spawning season, and food of Atlantic halibut. Journal of the Fisheries Research Board of Canada 24, 53-66. Kruger, K. E. and Nieper, L. (1978). Bestimmung des Quecksilber-Gehaltes der Seefische auf den Fangplatzen der deutschen Hochsee- und Kustenfischerei. 2. Mitteilung und Schluss. Archiv fur Lebensrnittelhygiene 29(5), 165-168. Kruger, K. E., Nieper, L. and Auslitz, H.-J. (1975). Bestimmung des QuecksilberGehaltes der Seefische auf den Fangplatzen der deutschen Hochsee- und Kiistenfischerei. 1. Mitteilung. Archiv f i r Lebensmittelhygiene 26(6), 20 1-207. Lambertsen, G. (1978). Fatty acid composition of fish fats. Comparisons based o n eight fatty acids. Fiskeridirektoratets Skrifter, Serie Ernrering 1, 105-1 16. Lernning, S., Kjerrsvik, E., Haug, T. and Gulliksen, B. (1982). The early development of the halibut, Hippoglossus hippoglossus (L.),compared with other marine teleosts. Sarsia 67, 85-91. Love, R. M . (1970). “The Chemical Biology of Fishes”, Vol. I . Academic Press, London. Lovern, J. A. and Sharp, J. G . (1933). CXCVIII. The diet of the halibut and intensity of feeding, in relation to the vitamin A potency of the liver oil. Biochemical Journal 27, 147CL1472. Lovern, J. A., Edisbury, J. R. and Morton, R. A. (1933). CXCVII. Variations in Vitamin A content of fish-liver oils, with particular reference to seasonal fluctuations in the potency of halibut-liver oil. Biochemical Journal 27, 1461-1469. Magnusson, J. V. and Palsson, 0. K. (1988). The feeding habits of cod and redfish in East Greenland waters with special emphasis on predation on juvenile redfish. International Council for the Exploration of the Sea, Early Life History Symposium Paper No. 72, 26 pp. (mimeo). Mangor-Jensen, A. and Jelmert, A. (1986). The effect of ambient salinity on the buoyancy of eggs from the Atlantic halibut Hippoglossus hippoglossus. Council Meeting of the International Council for the Exploration of the Sea F52, 7pp. (mimeo). Mannan, A., Fraser, D. I. and Dyer, J. (1961). Proximate composition of Canadian Atlantic fish I. Variation in the composition of different sections of the flesh of Atlantic halibut (Hippoglossus hippoglossus). Journal of the Fisheries Research Board of Canada 18,483493. Martin, W. R. and McCracken, F. D. (1950). Movements of halibut on the Canadian Atlantic coast. Fisheries Research Board of Canada, Atlantic Progress Report 50, 3-8. Mathisen, 0. A. and Olsen, S. (1968). Yield isopleths of the halibut, Hippoglossus hippoglossus. in Northern Norway. Fiskeridirektoratets Skrifter, Serie Havundersokelser 14, 129-159. McCracken, F. D. (1958). On the biology and fishery of the Canadian Atlantic halibut, Hippoglossus hippoglossus L. Journal of the Fisheries Research Board of Canada 15, 1269-1311. McCracken, F. D. and Martin, W. R. (1955). Recent recoveries of tagged halibut. Fisheries Research Board of Canada, Atlantic Progress Report 61, 4. McIntyre, A. D. (1952). Statistics of the Scottish halibut fishery 1930-1949. Marine Research, Scotland 1952( I ) , 1-24.

68

T. HAUG

McIntyre, A. D. (1953). The food of halibut from North Atlantic fishing grounds. Marine Research, Scotland 1952(3), 1-20. McIntyre, A. D. (1958). Scottish halibut investigations. Annales biologiques, Copenhague 14,31-33. Morawa, F. W. F. (1957). Die Verteilung des Fettes bei einigen Platt-fischen (Flunder, Kliesche und Heilbutt). Zeitschrift fur die Fischerei 5, 269-273. Mork, J. and Haug, T. (1983). Genetic variation in halibut, Hippoglossus hippoglossus (L.) from Norwegian waters. Hereditas 98, 167-174. Naas, K. E., Berg, L., Klungsrayr, J and Pittman, K. (1987). Natural and cultivated zooplankton as food for halibut (Hippoglossus hippoglossus) larvae. Council Meeting of the International Council for the Exploration of the Sea F17, 23 pp. (mimeo). Neilson, J. D., Waiwood, K. G. and Smith, S. J. (1989). Survival of Atlantic halibut (Hippoglossus hippoglossus) caught by longline and otter trawl gear. Canadian Journal of Fisheries and Aquatic Sciences 46, 887-897. Nickerson, J. T. R. (1978). The Atlantic halibut and its utilization. Marine Fisheries Review 40(7), 21-25. Nikolskii, G. V. (1954). "Special Ichthyology". (Translated from Russian by the Israel Program for Translations, Jerusalem, 1961, IPST Cat. no. 233.) Okland, S., Oiestad, V. and Berg, L. (1986). Development of neuromast cells and their cupula in halibut (Hippoglossus hippoglossus L.) and their destruction in different types of rearing systems. Council Meeting of the International Council for the Exploration of the Sea L7, 10 pp. (mimeo). Olsen, S. (1969). The halibut gillnet fishery in west Finnmark. Analysis of landings data from Oksfjord 1955-1956. Fiskeridirektoratets Skrifter, Serie Havundersokelser 15, 186-197. Olsen, S. and Tjemsland, J. (1963). A method of finding an empirical total selection curve for gill nets, describing all means of attachments. Fiskeridirektoratets Skrifter, Serie Havunders0kelser 13,88-94. Opstad, I. and Raae, A. J. (1986). Physical stress on halibut larvae. Council Meeeting of the International Council for the Exploration of the Sea F18, 19 pp. (mimeo). Pittman, K., Berg, L. and Naas, K. (1987). Morphological development of halibut (Hippoglossus hippoglossus) larvae with special reference to mouth development and metamorphosis. Council Meeting of the International Council for the Exploration of the Sea F18,22 pp. (mimeo). Pittman, K.,Skiftesvik, A. B. and Harboe, T. (1989). Effect of temperature on growth rates and organogenesis in the yolksac larvae of halibut (Hippoglossus hippoglossus L,). Rapports et Procb- Verbaux des RPunions, Conseil Permanent International pour I'ExpIoration de la Mer 191,421430. Pittman, K.,Bergh, 0., Opstad, I., Skiftesvik, A. B., Skjolddal, L. and Strand, H. (in press a). Development of eggs and larvae of halibut (Hippoglossus hippoglossus): Effects of the microbial and physical environments at 3", 6", and 9°C. Journal of Applied Ichthyology. Pittman, K., Skiftesvik, A. B. and Berg, L. (in press b). Morphological and behavioural development of halibut (Hippoglossus hippoglossus) larvae. Journal of Fish Biology. Polyansky, Y. I. (1955). Data on the parasitology of fishes from the northern seas of the USSR. Parasites of the Barents Sea. Trudy Zoologicheskogo Institut Akademiya Nauk SSSR 19, 1-170.

BIOLOGY OF THE ATLANTIC HALIBUT

69

Priebe, K. (1978). Der DDT-Gehalt in der Muskulatur von nordostatlantischen Meeresfischen mit hoher Lebenserwartung unter Berucksichtigung des Quecksilbergehaltes. Die Fleischwirtschaft 58( I), 147-1 50. Rabben, H. (1987). A stripping method for Atlantic halibut (Hippoglossus hippoglossus L.) Council Meeting of the International Council for the Exploration of the Sea F40, 6 pp. (mimeo). Rabben, H. and Jelmert, A. (1986). Hatching of halibut (Hippoglossus hippoglossus L.) eggs under different light conditions. Council Meeting of the International Council for the Exploration of the Sea F17, 10 pp. (mimeo). Rabben, H., Nilsen, T. O., Huse, I. and Jelmert, A. (1986). Production experiment of halibut fry in large enclosed water columns. Council Meeting of the International Council for the Exploration of the Sea F19, 27 pp. (mimeo). Rabben, H., Jelmert, A. and Huse, I. (1987). Production experiment of halibut fry (Hippoglossus hippoglossus) in silos. Council Meeting of the International Council .for the Exploration of the Sea F42, 10 pp. (mimeo). Rae, B. B. (1958). The occurrence of plerocercoid larvae of Grillotia erinaceus (van Beneden) in halibut. Marine Research Scotland 1958(4), 1-31. Rae, B. B. (1959). Halibut -observations on its size at first maturity, sex ratio and length/weight relationships. Marine Research Scotland 1959(4),1-19. Riis-Vestergaard, J. (1982). Water and salt balance of halibut eggs and larvae (Hippoglossus hippoglossus). Marine Biology 70, 135-1 39. Roff, D. E. (1982). Reproductive strategies in flatfish: A first synthesis. Canadian Journal of Fisheries and Aquatic Sciences 39, 1686-1 698. Rollefsen, G. (1934). The eggs and the larvae of the halibut (Hippoglossus vulgaris). Det Kongelige Norske Videnskabers Selskab Forhandlinger 7(7), 20-23. Ronald, K. (1958a). The metazoan parasites of the heterosomata of the Gulf of St. Lawrence. IV. Cestoda. Canadian Journal of Zoology 36, 429-434. Ronald, K. (1958b). The metazoan parasites of the heterosomata of the Gulf of St. Lawrence. 111. Copepoda parasitica. Canadian Journal of Zoology 36, 1-6. Rennestad, I. and Kirdal, A. (1989). Den ferste indikasjon p i vekstkurve for oppdrettskveite. Fiskets Gang 75( l), 15-16. Ruskowski, J. S. (1934). Etudes sur le cycle evolutif et sur la structure des Cestodes de Mer. Pt. 3. Le cycle evolutif de Tetrarhynque Grillotia erinaceus (van Beneden 1958). MPmoires de I’AcadPmie Polonaise des Sciences et des Lettres, Classe des Sciences Mathematiques et Naturelles, Serie B: Sciences Naturelles 6, 1-9. Russell, F. S. (1976). “The Eggs and Planktonic Stages of British Marine Fishes”. Academic Press, London. Schmidt, J. (1904). On pelagic post-larval halibut (Hippoglossus vulgaris Flem. and H. hippoglossoides (Walb.)). Meddelelser fra Kommisjonen f o r Havunders0gelser, Serie Fiskeri 1(3), 1-12. Schmidt, P. J. (1930). On the Pacific halibut. Dokldy Akademii Nauk SSSR, A 1930(8), 203-208. Schram, T. A. and Haug, T. (1988). Ectoparasites on Atlantic halibut, Hippoglossus hippoglossus (L.), from Northern Norway - potential pests in halibut aquaculture. Sarsia 73, 213-227. Scott, T. (1910). On the food of the halibut with notes on the food of Scorpaena, Phycis blennoides, the garpike and Chimaera monstrosa. 28th Report of the Fisheries Board, Scotland 1909(3), 24-37. Sendstad, K. (1984). “Morfologisk og eksperimentell undersekelse av kveitelarver (Hippoglossus hippoglossus L.) med resultater fra et foringsforsak”. Cand. Scientific thesis, University of Bergen. 131 pp.

T. HAUG

70

Sigurdsson, A. (1956). Contribution to the life history of the halibut at the west of Iceland in recent years (1936-1950). Meddelelser fra Danmarks Fiskeri- og Havunders0gelser, Ny Serie 1(16), 1-24. Sigurdsson, A. (1976). Smaludan. Hafrannsoknir 8, 57-64. Sigurdsson, A. and Fridriksson, A. (1952). On the Icelandic fishery in the Denmark Strait in 1950. Annales biologiques, Copenhague 8, 45-46. Smidt, E. (1968). Report on Greenland halibut and halibut eggs and larvae. Environmental Surveys - Norwestatlant 1-3, 1963, Part I. International Commission for the Northwest Atlantic Fisheries, Special Publication 7, 175-179. Solerndal, P., Tilseth, S.and IZliestad, V. (1974). Rearing of halibut. I. Incubation and the early larval stages. Council Meeting of the International Council for the Exploration of the Sea F41, 5 pp. (mimeo). Stobo, W. T., Neilson, J. D. and Simpson, P. G . (1988). Movements of Atlantic halibut (Hippoglossus hippoglossus) in the Canadian North Atlantic. Canadian Journal of Fisheries and Aquatic Sciences 45, 484-491. Tjemsland, J. (1960). “Kveita i Nord Noreg”. Cand. real. thesis. University of Bergen, 30 pp. Tytler, P. and Blaxter, J. H. S. (1988). Drinking in yolk-sac stage larvae of the halibut, Hip oglossus hippoglossus (L.). Journal of Fish Biology 32, 493494. Vedel-Thing, . (1936). On the eggs and young stages of the halibut. Meddelelser fra Kommisjonen for Danmarks Fiskeri- og Havunders0gelser, Serie Fiskeri 10(4), 1-23. Vedel-Tining, A. (1938). Migrations of small halibut marked in Faroese waters. Journal du Conseil Internationale pour Exploration de la Mer 1, 310-375. Vedel-Tining, A. (1947). Marking experiments on 3-5-year-old halibut in Faroese waters. Annales biologigues, Copenhague 2, 24. Vernidub, M. F. (1936). Some data concerning the Pacific form of Hippoglossus hippoglossus. Trudy Leningradskogo Obshchestva Estestvoispytatelei 55(2), 143182. Zubchenko, A. V. (1980). Parasitic fauna of Anarhichadidae and Pleuronectidae families of fish in the Northwest Atlantic. ICNAF Selected Papers 1980(6),41-46.

w

Egg Quality in Fishes E. Kjersvik,' A. Mangor-Jensen2and I. Holmefjord3 Norwegian College of Fishery Science, University of Tromse, Tromse, 'Institute of Marine Research, Austevoll Aquaculture Research Station, Storebe, and The Agricultural Research Council of Norway, Institute of Aquaculture Research, Sunndalsora, Norway

'

1. 11.

Introduction .. .. .. .. Egg Quality Characteristics ,. .. A. Fertilization processes .. .. B. Physical and physiological properties C. Morphology . . .. .. D. Egg size .. .. .. .. E. Chemical content .. .. .. F. Chromosomal aberrations .. .. Egg Quality of Wild Fish . . . . .. Factors of Importance for Egg Quality . . A. Overripening and storage of eggs ., B. Broodstock management .. .. Conclusions and Recommendations .. Acknowledgements .. .. .. References . . . . .. .. .. I

111. IV.

V. VI .

VII.

.

..

..

..

..

..

..

.. ..

.. ..

..

..

.. ..

..

..

..

.. .. .. ..

.. .. ..

..

..

..

..

..

..

.. .. ..

.. .. ..

..

..

..

.. .. .. ..

..

..

..

..

..

..

.. ..

..

..

..

.. ..

..

..

..

..

.. ..

..

..

..

.. ..

..

.. ..

.. .. .. ..

..

71 72 72 76 76 78

79 85

87 91 91 96 103 105 105

1. Introduction Varying egg quality is one of the limiting factors for successful mass production of fish fry. It may also cause some of the variability in recruitment observed in many wild stocks, since poor egg quality may decrease the survival potential of the hatched larvae. Increased knowledge of these problems could contribute to the establishment of rearing regimes that would both result in a better larval survival, and improve choice and handling of broodstock fish. The problems of egg quality, especially in marine fishes, have, however, been little studied. Copyright 0 1990 by Academic Press Limited All rights ofreproduction in any form reserved

ADVANCES IN MARINE BIOLOGY VOLUME 26 ISBN 0-12-02612bX

71

72

E. KJORSVIK ET AL.

In the literature, the term egg quality has been defined and used in various ways by different authors. Due to the limited knowledge in the field, it has been difficult to pinpoint valid quality criteria. The only definition of egg quality that has general validity is probably the egg’s potential to produce viable fry. The egg’s potential to produce viable fry is determined by several physical, genetic and chemical parameters, as well as the initial physiological processes occurring in the egg. If one of the essential factors is lacking, or is incomplete, egg development will fail at some stage. Thus egg quality should be regarded as determined when the egg has left the female fish and the fertilization process is complete. Effects due to treatment during the incubation period are thus excluded. This definition is, however, of very little practical value, since fry viability obviously cannot be determined before the fry is produced. Good practical criteria for the determination of egg quality should be both possible to identify early in development, and be simple to use. The problems of egg quality have received increasing attention, in relation to cultivation as well as to the assessment of reproduction of wild fishes. This chapter discusses possible criteria for the determination of quality of fish eggs. A further discussion of the factors that seem to be important for the regular production of good quality eggs from a broodstock, and the natural level of egg quality for wild fish is also included.

II. Egg Quality Characteristics In most egg quality investigations, fertilization and hatching rate have been used as important criteria. Survival to specific developmental stages and final production of fry have also been used as measures of egg quality. Other characteristics have been linked to these “standard” measurements, but few studies are directly comparable, or deal with comparisons between species. Results concerning egg quality characteristics are summarized in Table 1, which gives the appropriate references.

A. Fertilization Processes Fertilization rate is a useful parameter to detect poor egg quality. This is used in studies of the nutritional requirements of broodstock (see also Table 3). However, fertilization rate does not always correlate with good survival and development in later embryonic stages for many species, such as carp (Cyprinus carpio: Statova et al., 1982), vendace (Coregonus albula:

EGG QUALITY IN FISHES

73

Dabrowski et al., 1987), Japanese flounder (Limanda Yokohama: Hirose er al., 1979), turbot (Scophthalmus maximus: McEvoy, 1984), cod (Gadus morhua: Kjsrsvik et al., 1984a) and Pacific herring (Clupea harengus pallasi: Hay, 1986). During fertilization and activation, the cortical reaction takes place in all teleost eggs. The cortical alveoli break down, release their contents (colloids) from the cortical layer, and thus start the formation of the perivitelline space (Yamamoto, 1961; Ginzburg, 1972). The mechanisms of the process are not completely understood, but this process is correlated with egg quality in red seabream (Pagellus bogaraveo: Sakai et al., 1985), cod (Kjsrsvik and Lsnning, 1983) and turbot (McEvoy, 1984). Cortical alveoli are present after fertilization (i.e. the cortical reaction is incomplete) in poor quality eggs in these species (Fig. I ) . An incomplete activation process may result in the smaller perivitelline space and lack of increase in egg diameter seen after fertilization in poor quality cod eggs (Kjsrsvik et al., 1984a).

FIG.I . Early cleavage in cod eggs (Gadus morhua L.) at 5°C. Egg diameter ‘5 1.3mm. (A and B) 2 h after fertilization. In good quality eggs, the cortical reaction is complete, and the perivitelline space is under formation (A). In poor quality eggs, cortical alveoli are visible in the cytoplasm and the perivitelline space is poorly developed (B). (C and D) 6 h after fertilization. After the first cleavage, good quality eggs (C) are transparent with a clear symmetrical cleavage. In poor quality eggs (D), cleavage is incomplete and cortical alveoli are still visible.

TABLE1. CHARACTERIS~CS USEDTO DFSCRIBE EGG QUALITY Speciff

Rainbow trout, Salmo gairdneri

Fenillzation

0

Atlantic salmon, Salmo salar Carp, Cyprinus carpio AYU, Plecoglossus altivelis Dover sole, Solea solea

Turbot, Scophthalmus maximus

Normal morphology

.

Hatching Normal Egg larvae sue 0

0

0

0

.

Overnpening

.

Storage

ofeggs

Other charactenstics

Springate et al. (1984) No correlation between viability/egg size Egg transparency; lipid droplets distribution; activation tendency Egg transparency; lipid droplets distribution

0

Springate and Bromage (1985)

Nomura et al. (1974)

Escaffre and Billard (1979)

Sakai et al. (1975) Thorpe et al. (1984)

.

0

SOUrCe

Zonova (1973) Hormonal treatment

Hirose et al. (1977)

Early mortality; buoyancy; egg transparency; lipid droplets distribution Egg shape; buoyancy; egg transparency No correlation between viabilitylegg size

Dinis (1982)

McEvoy (1 984) Devauchelle et al. (1988)

Japanese flounder, Limanda Yokohama Red sea bream, Pagellus bogaraveo

.

Jack mackerel, Trachurus symmetricus Cod, Gadus morhua

Herring, Clupea harengus

0

0

0

Hormonal treatment

Hirose et af. (1979)

Wrinkled egg membrane; cortical reaction; buoyancy; “plasma-bulged” eggs Size-dependent survival after starvation Egg shape; buoyancy; cortical reaction; egg hardening; size of perivitelline space Cortical reaction; yolk osmolarity; egg diameter changes; chromosome aberrations

Sakai et a[. (1985)

Theilacker (1981) Kjorsvik and Lenning (1 983)

Kjrarsvik et al. (1984a)

Dushkina (1975)

76

E. KJORSVIK ET AL.

B. Physical and Physiological Properties The fertilization and activation processes seem to be vital for several aspects of embryonic development. The hardening of the egg chorion is due to an enzyme reaction during the activation process (Hagenmaier et al., 1976), and the egg’s ability to sustain mechanical resistance (egg strength or chorion hardness) is better in good than in poor quality eggs of cod (Kjarsvik and Lanning, 1983), lumpsucker (Cyclopterus lumpus: Kjarsvik et al., 1984b) and halibut (Hippoglossus hippoglossus: Kjarsvik, unpublished). Chorion appearance (such as wrinkled egg membranes) and egg shape are also seen to deviate in poor quality eggs of cod (Kjarsvik and Lanning, 1983), turbot (McEvoy, 1984) and red seabream (Sakai et al., 1985). In later embryonic stages, bacterial growth will cause a decrease in egg strength (Kjarsvik et al., 1984a), and poor quality eggs of Dover sole (Solea solea) seem to be more subject to bacterial contamination (Dinis, 1982). A prolonged fertilization process may also affect membrane characteristics, as the vitelline membrane in teleost eggs changes from high to a very low permeability shortly after fertilization (Potts and Rudy, 1969; Loeffler and Lovtrup, 1970; Potts and Eddy, 1973; Mangor-Jensen, 1987). Marine teleost eggs exhibit an internal yolk osmotic value which is much lower than the surrounding medium, and this is maintained by the low permeability of the vitelline membrane. Higher yolk osmolality levels are reported for poor than for good quality eggs of cod (Fig. 2) and lumpsucker (Kjarsvik et al., 1984a,b), probably as the result of a longer and incomplete cortical reaction in these eggs. Alderdice er al., (1979) also considered variations in osmotic properties of Pacific herring eggs to be due to differences in maturity or egg quality. Changes in specific gravity may partly be due to osmoregulatory variations in egg batches, and buoyancy for pelagic eggs is often better for good than for poor quality eggs, as shown in Fig. 3 (Dinis, 1982; Kjarsvik and Lanning, 1983; McEvoy, 1984; Sakai et al., 1985; Kjarsvik, unpublished observations).

C. Morphology The earliest cells (blastomeres) of an embryo are undifferentiated, and they form the basis for the differentiating embryo. A deviation (or defect) in these cells will influence the further development of the embryo more strongly than will defects occurring in single cells later in development. Cell morphology has been used in egg quality or pollution studies, and morphological characters are generally more sensitive parameters than survival.

77

EGG QUALITY IN FISHES 500-

_ _ _ 120% s.w _ _ _ 60% -100% 4501

FIG.2. Yolk osmolality of good and poor egg batches of cod (Gadus morhua). The eggs were fertilized and reared in 60% sea water (20p.p.t.), 100% sea water (34p.p.t.) and 120% sea water (41 p.p.t.) (from Kjerrsvik el a/., 1984a).

y = 9.727+0.566x r = 0.5378

;OoI CIY

c3 E

0.

0

NORMAL EGGS

(Oh)

FIG.3. Buoyancy of red sea bream eggs in relation to normally developed eggs (redrawn from Sakai el a/., 1985).

78

E. KJORSVIK ET AL.

Characteristics studied include symmetrical early cleavage (Figs 1 and 4) and proportions of deformed embryos (see also Fig. 11) and larvae. The symmetry of the early blastomeres seems to be a consistent early indicator of egg viability for several species, such as herring (Clupea harengus: Dushkina, 1975), Dover sole (Dinis, 1982), red seabream (Sakai et al., 1985), cod (Kjersvik and Lenning, 1983; Kjersvik et al., 1984a), turbot (McEvoy, 1984; Devauchelle et al., 1988) and Japanese flounder (Hirose et al., 1979).

FIG.4. Early development of halibut (Hippoglossus hippoglossus) eggs. Egg diameter u 3 mm. The eight-cell stage is reached = 12 h after fertilization (5°C). Good quality eggs (A) undergo a symmetrical cleavage, whereas poor quality eggs (B) may have very irregular cleavages (photograph by Vidar Vassvik, Institute of Aquaculture Research, Sunndalsera).

Other early-stage morphological characters that may be useful indicators of egg quality are egg transparency and distribution of lipid droplets (Dinis, 1982; McEvoy, 1984), size of perivitelline space and changes in egg diameter after fertilization (Kjersvik et al., 1984a; Sakai et al., 1985). Very poor eggs have also been described as “plasma-bulged’’ (Sakai et al., 1985). Good quality eggs are generally described as transparent (except for oil droplets), perfectly spherical with clear, symmetrical early cleavages.

D. Egg Size It is known that egg and larval sizes are correlated. Larger larvae tend to survive longer without food than those hatched from smaller eggs (Blaxter and Hempel, 1963; Theilacker, 1981; Springate et al., 1984; Escaffre and Bergot, 1984; Knutsen and Tilseth, 1985), which may be an advantage for wild larvae. It seems doubtful that egg size gives any permanent or long-term

EGG QUALITY IN FISHES

79

advantages as far as growth and survival of the larvae are concerned (see also Blaxter, 1988). Under favourable conditions (e.g. in hatcheries), it has been shown that egg size has no direct implications for overall egg quality, larval survival and growth in rainbow trout (Salmo gairdneri: Pitman, 1979; Springate and Bromage, 1985), Atlantic salmon (Salmo salar: Thorpe et al., 1984), catfish (Cluria macrocephalus: Reagan and Conley, 1977) and carp (Kirpichnikov, 1966). The ability of smaller fry to grow at the same rate as the initially larger fry is of considerable commercial importance. Potentially of greater importance is the finding that survival of eggs and fry is not affected by the size of the egg (Springate and Bromage, 1985). This has also been confirmed for Atlantic salmon (Glebe et al., 1979) and carp (Tomita et al., 1980; Zonova, 1973). Springate and Bromage (1985) suggest that where size-dependent survival rates have been reported these might reflect a difference in stage of ripeness of the eggs, and thus an overripening effect rather than a size effect. The authors therefore conclude that small eggs are basically not of a lesser quality than larger eggs. In an extensive work on the vendace (Coregonus albula), Kamler et al. (1982) showed that there is a significant correlation between survival to morula stage and variation in egg diameter in egg groups. However, no such relationships were found by Wilkonska and Zuromska (1982) in a similar investigation. Kamler et al. (1982) used the method of Averaged General Quality Indicator (QIC), originally described by Kolman (1973), for quantification of egg quality. This method is based on egg dry weight, content of protein, lipids and energy, as well as egg size, microscopical egg quality, fertilization and survival. In this scoring test, variation in egg diameter seems to be most important in determining egg quality. There do not appear to be similar studies on marine fish eggs, but published data do not provide evidence suggesting that egg diameter is a good criterion for egg quality.

E. Chemical Content The biochemical composition of a healthy egg reflects the embryonic demands both for nutrition and growth. Certain components are known to be “essential” for an organism (i.e. the organism is unable to synthesize the nutrient), and these components have to be present in certain amounts to satisfy biological demands. Biochemical egg quality parameters may therefore exist, and a biochemical evaluation of egg quality could thus be possible even before fertilization. Unfortunately, the essential components differ from organism to organism, although some compounds like amino acids seem to be practically

80

E. KJORSVIK ET AL.

identical as essential compounds for all phyla (Taylor and Medici, 1966). The fact that there are species-specific needs in terms of nutrition may be kept in mind when a certain component is described as a positive egg quality criterion for one species, while the same component in another work is described as an indicator of poor egg quality. The possibility of finding common biochemical quality criteria for fish eggs may therefore be questionable. Nevertheless, species-specific biochemical egg quality criteria are valuable if they are absolute. The most common chemical parameters are discussed below.

I . Pigments The significance of pigments in eggs of salmonids has been a considerable topic of discussion (see Craik, 1985). Scientific verification of pigments as quality indicators for salmonid eggs is, however, lacking. Pigments accumulate in the salmonid oocyte during maturation, and pigment accumulation reflects feed composition (Hubbs and Strawn, 1957; Craik, 1985). The prevailing theory is that pigments serve as biological antioxidants in a similar manner to vitamin E (Burton and Ingold, 1984). These pigments are mainly astaxantin and lutein, both of which are carotenoids. They are found in the free form (unesterified) in the mature eggs (Steven, 1948, 1949; Longinova, 1967; Kitahara, 1983). The role of carotenoids in teleost eggs has also been discussed by Hubbs and Strawn (l957), who showed that offspring from pale eggs of the darter Efheosfoma had lower viability than those from brightly coloured eggs. The content of egg pigments was shown to be highly dependent on broodstock feeding. Correlations between hatching percentage and content of carotenoids in salmonid eggs have been thoroughly investigated by Murayama and Yanase ( 196l), Borovik (1962), Longinova (1966), Galkina ( I 969) and Georgiev (1971). By reviewing these articles, Craik (1985) concluded that the differences in viability could all be attributed to factors other than pigmentation, such as environmental parameters and overripening. Interspecific differences may, however, also exist. Yarzhombek (1970) showed that the carotenoid content of salmonid eggs varies enormously both between species and within species, and Craik (1985) concluded that this component is possibly the most variable in egg composition. Obviously, good quality eggs may differ markedly in pigment content, but there may be a certain critical low limit (Fig. 5). For rainbow trout, however, this critical level is so low that only broodstock fish fed a carotenoid-free diet produced eggs that reached this limit (Murayama and Yanase, 1961). For marine species, the relationships between pigments and egg quality does not seem to have been studied.

81

EGG QUALITY IN FISHES

0

0

n

n

0

0 0 0 0

1

0 0

)

0

0 0

I

I

2

I

I 4

I I I I I 10 6 8 pg CAROTENOIDIg EGGS I

I

I

12

I

1 14

FIG.5. Relationship between carotenoid content and hatching percentage for rainbow trout eggs. The data were collected from several hatcheries (redrawn after Murayama and Yanase, 1961).

2. Vitamin C (ascorbic acid) Vitamin C (ascorbic acid) is found in the gonads of fish in high amounts (Hilton et al., 1979). The significance of this finding is, however, not known, although a link to the process of vitellogenesis has been proposed (Tolbert, 1979). This hypothesis has also been reinforced by the fact that the content of ascorbic acid in growing gonads has been reported to increase during early vitellogenesis (Seymour, 198 1; Sandnes and Brekkan, 1981; Sandnes, 1984). For rainbow trout, a direct correlation between hatching rate and egg content of ascorbic acid was found by Sandnes et al. (1984), as shown in Fig. 6 . They also found a positive correlation between ascorbic acid content in the broodstock feed and in the eggs. It is, however, not clear whether the ascorbic acid plays a direct role in the development of the egg, or whether the higher content in eggs with better hatching success is a result of improved incorporation of yolk proteins during the early vitellogenesis. Nevertheless, content of ascorbic acid deserves further investigation as a quality indicator in fish eggs.

82

E. KJORSVIK ET AL.

0

70 0

0 50r

8 40

-

30

-

20

-

$0

-

O

0

L

I

I

I

I

5

10

i5

20

25

30

95

40

45

50

ASCORBIC ACID (,ug/g WET WT) FIG.6. The distribution of ascorbic acid in the eggs from individual rainbow trout females related to their hatching. Eggs containing less than 20 pg ascorbic acid/g had a wide range of hatching success. (Redrawn Sandnes el al., 1984.)

3. Inorganic components Differences in composition of inorganic components between high and low viability batches of eggs have been inadequately investigated. Craik and Harvey (1984) carried out an extensive investigation to find the chemical cause of egg quality variations. They pointed out that significant variation in gross biochemical composition of eggs produced by individual females of

EGG QUALITY IN FISHES

83

both rainbow trout and salmon occurred, even though the fish were fed identical diets. These variations in inorganic components such as phosphorus, calcium and iron showed a poor correlation with the hatching percentage. There was, however, a significant positive correlation between the percentage of hatched alevins surviving to first feeding and each of the following: egg wet weight, egg dry weight and absolute levels in the egg of bound lipid, precipitable protein and protein phosphorus. These differences were attributed to overripening processes. An increase in free phosphorus may well be explained by proteolytic activity on phospho-proteins. It has, however, not been established whether pelagic eggs from marine spawners conform to this rule. Hirao et al. (1954, 1955) also found a positive correlation between hatching percentage and egg iron content, but these results were interpreted by Craik and Harvey (1984) as being caused by differences in age of the broodstock. 4. Organic composition of eggs and female age

For pike-perch (Lucioperca lucioperca L.)Savel’yeva and Shuvatova (1 972) found a positive, linear correlation between the protein content of eggs, the fat content of the female muscle, and the survival rate of the embryos. In these fish, larvae from medium and large-sized females from the medium age group were most viable. A direct correlation between egg quality (fertilization, survival of larvae, percentage of deformed larvae) and amino acid composition of carp eggs has been established (Vladimirov, 1974). The influence of the age of females on the amino acid composition of eggs is not known. This parameter seems worthy of study, since age-related variability seems important in connection with several other compounds. Whether the fat content of eggs has an influence on egg quality is a matter of discussion. In studies of fat content of roach (Rutilus rutilus: Kuznetsov, 1976) and bream (Abramis brama: Zhukinsky and Kim, 1981) it was concluded that a high fat content in the egg increased the viability of the larvae. Dabrowski et al. (1987) found a positive correlation between hatching rate and egg lipid content in vendace (Coregonus albula), but concluded that the results needed further examination with more homogeneous samples of fish. Mann (1960) considered, on the other hand, that decreased viability of the eggs of older female rainbow trout was associated with an increase in fat and water content. The influence of female age on lipid metabolism of the oocyte has been established in several works. The greatest content of fat was contained in the eggs of middle-age females in carp (Marthysev et al., 1967), gadid sp. (Shatunovskiy, 1973) and flounder (Shatunovskiy, 1963). After a certain age the synthesis of lipids and the generative metabolism

84

E. KJORSVIK ET AL.

decreases in Gadidae, and the presumed qualitative indices of the eggs deteriorate (Shatunovskiy, 1973). Kim (1974) investigated this phenomenon further in carp, and found that the phosphoglyceride fraction was highest in eggs produced by middle-aged females. By contrast, the content of cholesterol showed the highest concentrations in eggs produced by young and old females. This pattern was also seen in flounder (Shatunovskiy, 1963). Kamler et al. (1982) used the Averaged General Quality Indicator (QIC: Kolman, 1973) to determine the quality of vendace eggs from different lakes. This method was also applied to compare the age differences among female spawners. Kamler and Zuromska (1979) described a clear difference in egg quality depending on the age of the female spawner, and the middle-aged spawners produced eggs of best quality. Poorer egg quality is also found in first-time spawners of the female tench (Tinca tinca: Zuromska and Markowska, 1984).

250

I - 160

-200 -

8

-8 $

3

v

>- 150 -

t

-140 u. 0

8 100 -

-120

D

I-

z 3

0

I-

8 m

&0 0

w

3

a

P

50-

-100

a

01

4

I

I

6

I

I

I

I

8 10 AGE (years)

I

I

12

1

I

14

FIG.7. Fecundity and lipid content of bream eggs as a function of female age. Lipid content is described as a positive quality criterion by the author. U, absolute fecundity; 0 ,lipid content of eggs (redrawn after Kuznetsov, 1973).

Total lipid has been used as an indicator of egg quality. Kuznetsov (1 973) established a connection between fecundity and total content of

lipids in the eggs of the bream (Abramis brama: Fig. 7). Surprisingly, he found that young fish with relatively low fecundity had the highest lipid

EGG QUALITY IN FISHES

85

content in the eggs. These eggs also had a smaller diameter than eggs from the older fish. It was concluded that eggs from the low-fecundity young fish had the best quality, and this result was interpreted as an ecological adaptation. Caution should be exercised when using lipids as quality indicators. In vendace, which is a freshwater species, a high content of lipids was regarded as indicating good egg quality (Kamler et al., 1982). On the other hand, Devauchelle et al. (1982) claim for the marine species turbot, sole and seabass (Dicentrarchus labrax), that a higher total lipid content generally corresponds to a very low viability rate of eggs. They also found that eggs from wild fish contained more proteins and lipids than eggs of captive spawners. Polyunsaturated fatty acids (PUFA) were, however, more frequent in the wild spawn.

F. Chromosomal Aberrations Severe chromosomal errors occurring in fish eggs before the gastrula stage are thought to be lethal (Longwell, 1977). The early stages of fish embryos are especially suited for chromosome studies, since they have large chromosomes and undergo frequent cell divisions. Three types of chromosomal aberrations are generally encountered in early fish embryos (Fig. 8; see also Longwell, 1977). Delayed anaphases imply a delayed division of some of the centromeres, but the chromosomes will in most cases reach the poles at late telophase. Fragments are chromosomes or part of chromosomes remaining in or near the equatorial plane. When anaphase bridges occur, some of the chromosomes do not divide, but remain in the equatorial plane; others do not divide properly, but form a bridge between the dividing chromosomes. Fragments and bridges seem especially indicative of severe chromosomal damage (Fig. 9), and will result in the irregular distribution of chromosome material to the daughter cells (Longwell, 1977; Kjrarsvik er al., 1984a). Delayed anaphases seem to be present in most cod eggs, independent of quality (Kjsrsvik er al., 1984a; Stene, 1987), but whether this phenomenon is reparable by the cells is not known. Clear correlations between survival (quality) and cytogenetic status are found for several species. Evidence is also derived from experimental studies in toxicology (Polikarpov, 1966) and from monitoring polluted areas (Longwell and Hughes, 1981; Longwell, 1982; Dethlefsen er al., 1987; Cameron et al., 1989). Chromosome abnormalities at the early embryonic stage seem to be one of the best indicators of sublethal damage to the embryo, as they give a very efficient measure of the “health” status in a batch of eggs.

E. KJORSVIK ET AL.

86

(A)

(B)

(C)

FIG.8. Examples of chromosome aberrations in fish embryos. (A) Normal anaphase, (B) fragments, (C) anaphase bridges and fragments (from Kjersvik el al., 1984a).

FIG. 9. Mitotic divisions in (A,B) whiting (Merlangius merlangus), (C) cod (Gadus morhua) and (D,E) halibut (Hippoglossus hippoglossus). (A,C) Normal anaphase; (B,D) anaphase with fragments between the two parts of the dividing chromosomes; (E) anaphase with fragments and bridges. (A) and (B) are previously unpublished photographs, reproduced with the permission of P. Cameron, Biologische Anstalt Helgoland, Hamburg. (C) is from Kjersvik el al. (1984a) and (D) and (E) are previously unpublished photographs, reproduced with the permission of A. Stene, College of Marine Studies, Alesund.

EGG QUALITY IN FISHES

87

111. Egg Quality of Wild Fish Natural factors that control mortality of fish eggs and larvae are poorly understood, and very few studies have been published on the natural level of egg viability in the sea. Recent studies have pointed out that observations of early morphological (malformations) and cytogenetical characteristics are useful as indicators of viability in embryos (Figs 10 and 11). Investigations of egg quality in the sea are mainly carried out in connection with studies in polluted areas, such as the New York Bight (Longwell and Hughes, 1981; Longwell, 1982), the Baltic (Graumann, 1986; von Westernhagen et al., 1988), and the North Sea (Dethlefsen et al., 1985, 1987; Cameron et al., 1989). These studies all demonstrate a high correlation between abnormal egg development and pollution load in the environment. In the southern North Sea, the proportions of chromosomal aberrations were positively correlated with the contamination of pollutants in gonads and livers in females, and in certain areas up to 50% of whiting embryos were malformed (Dethlefsen et al., 1987; Cameron et al., 1989; see also Fig. 12). Such methods seem very suitable for monitoring differences in sensitivity between species. Year-to-year fluctuations do occur, however, in areas with constant pollution load.

FIG.10. Normal embryonic development of flounder (Plaiichthysjesus)embryos at different stages, shown as Roman numerals. Egg diameter

= 1 mm (from Cameron el at., 1989).

FIG. 1 1. Morphological malformations of fish embryos observed in the North Sea. The figure shows flounder embryos, but no such malformations are species-specific. (a-f) Malformations in the early development: (a) irregular cleavage, poor cell formation with vesicle inclusions; (b) deformation of blastomeres at the 16-cell stage, the cells also contain many vesicles; (c) blastodisc (B) with “loose” cells; (d) early gastrula with abnormal shape and vesicle formation; (e) gastrulation with vesicles in the blastodisc (B) and in the germ ring (KR); (f) midgastrula containing many loose cells and vesicles near the embryonic shield (E) and the germ ring. (g,h) Malformations during early differentiation: (g) gross malformation with vesicular formation (arrows) at the yolk syncytium (D); (h) embryo with heavy vesicle formation which is unable to close the blastopore (BP). (i-m) Malformations of late embryos: (i) loose cell formation of the embryo and yolk, especially in the head region; (k) poor differentiation in the head region (k), twisting of the body and crippled tail; (I) bent notochord; (m) vesicle formation in advanced embryo. Egg diameter ‘1I mm (from Cameron el al.. 1989).

EGG QUALITY IN FISHES

89

FIG. 12 Regional distribution of morphological malformations of early embryonic stages (stage la) of dab (a) and flounder (b), sampled in the southern North Sea during March 1987. The height of the symbols represents the percentage of malformed embryos, octagons represent values above the mean value, and squares represent values below the mean value (from Cameron ei nl., 1989).

In samples of pelagic cod eggs from an unpolluted area (Fig. 13), up to 20% of eggs in early cleavage stages were morphologically normal (Kjrarsvik et al., 1984a). Similar proportions were found in samples from unpolluted areas in the investigations described above. Whether variations in egg

90

E. KJORSVIK ET AL.

quality occur during a breeding season is not known, but any such variation could have consequences for future farming activity. 1000-

-a

800..

m

f

v)

8 a,

P 0 0

-

0

600-

400.-

G

n

$

200.

ci 0-

-

.

March

c-- April

I ,

'

bMay

FIG. 13. Number of eggs ( O ) ,percentage of cod eggs less than 5 days old ( A ) , and percentage of morphologically abnormal cod embryos (A)from planktonic samples in Balsfjorden 1983 (from Kjerrsvik ef al., 1984a).

The early cleavage stages generally exhibit a high incidence of malformations. It is not known, however, to what extent these results are dependent on environmental variables such as temperature, salinity or parental status (gonad contamination, condition factor), and sublethal damage (e.g. chromosomal aberrations) may even be due to factors like natural mutations. Nevertheless, it appears that the occurrence of developmental abnormalities

EGG QUALITY IN FISHES

91

in fish embryos may be a valuable tool for the assessment of natural egg viability in, for example, important spawning areas.

IV. Factors of Importance for Egg Quality A. Overripening and Storage of Eggs As overripening of eggs will eventually occur in broodstock, it is important

to obtain and fertilize eggs at the correct time after ovulation. This may especially be a problem where fish have to be stripped and the eggs artificially fertilized (Fig. 14), and it has been recognized as a special problem in work with salmonids. Several studies on the overripening process

FIG. 14. Stripping of broodstock halibut. It is important to obtain and fertilize the eggs before overripening processes reduce the viability of the eggs. This is especially important for species where eggs have to be stripped from the fish.

92

E. KJORSVIK ET AL.

in such species describe morphological characters of immature and overripe eggs, as well as defining time-scales for these processes (e.g. Nomura et al., 1974; Sakai et al., 1975; Escaffre and Billard, 1979; Bry, 1981; Billard and Gillet, 1981). Results from several such studies are summarized in Table 2, but few have been published on this problem in marine species. Viability after ovulation of rainbow trout eggs is shown in Fig. 15 (from Sakai et al., 1975). If eggs from rainbow trout were retained in the ovary after ovulation, they were fertile for a much longer period in large fish than in small fish spawning for the first or second time (Fig. 16; Escaffre et al., 1977). TABLE2.

PERIOD OF VIABILITY OF

Species

OVULATED EGGSFOR DIFFERENT SPECIES

Viability after Temperature ovulation ("C)

Source

Striped bass, Roccus saxatilis

Stevens (1966)

Ih

Rainbow trout, Salmo gairdneri

10 days 5-7 days

10-12

Nomura et al. (1974) Sakai et al. (1975); Escaffre and Billard

4-6 days

10

Springate et al. (1 984)

<28 h >76h

15 10

Billard and Gillet

7 days

6.5

Krieger and Olson (1989)

10h

26-3 1

(1979); Bry (1981)

Brown trout, Salmo trutta

(1981)

Arctic char, Salvelinus alpinus

Catfish, Clarias macrocephalus Ayu, Plecoglossus altivelis

Mollah and Tan (1983) Hirose et al. (1977)

24 h

Japanese flounder, Limanda Yokohama

48 h

12 _+ 1

Hirose et al. (1979)

10h

12-14

McEvoy (1984)

> 6h

4.0

Holmefjord (unpublished)

9h

5.0

Kjerrsvik and Lernning

Turbot, Scophthalmus maximus

Atlantic halibut, Hippoglossus hippoglossus"

Cod, Gadus morhua"

(1983)

Pacific herring, Clupea harengus pallasi

2 weeks

8-10

Hay (1986)

48 h 10-13 h

4.0 0.8

Blaxter (1955) Dushkina (1975)

Atlantic herring, Clupea harengus"

Dry storage of eggs after stripping from the fish.

EGG QUALITY IN FISHES

93

70 eyed eggs 90 80

1966 d 1967 % hatched 1966 1967

-0-

-+70 60 50

40

30

DAYS AFTER OVULATION

FIG. 15. Changes in percentage of eyed eggs and hatching during overripening of eggs in rainbow trout. The eggs were periodically stripped after ovulation at 5-day intervals in 1966, and at 3-day intervals in 1967. Changes in viability due to overripening generally occur before any morphological changes are visible (redrawn from Sakai el a/.. 1975).

Hirose et al. (1979) concluded that the eggs of Japanese flounder were viable for -48 h after ovulation (Fig. 17). Although fertilization capacity was high for 3 days after ovulation, hatching rate reached a maximum for eggs fertilized 24 h after ovulation, with a steady decrease for eggs fertilized later. Changes in survival potential were apparent earlier than the morphological changes observed in stripped eggs (later than 3 days after ovulation). In the batch-spawning turbot (McEvoy, 1984), good development and hatching were only obtained when fish were stripped within 10h after ovulation. A “one-off-spawner”, like the Pacific herring, on the other hand, seems to resemble the pattern of the salmonids, as eggs were viable for at least 2 weeks after maturation (Hay, 1986). Eggs stored dry (e.g. in a refrigerator) after stripping from the fish will undergo changes similar to overripening. It is possible to store cod eggs for about 9-10 h (Kjarrsvik and Larnning, 1983) herring eggs for 10-48 h (Blaxter, 1955; Dushkina, 1975), and halibut eggs for at least 6 h after ovulation (Holmefjord, unpublished) without loss of viability of the eggs.

94

E. KJORSVIK ET AL.

0

I I I 0 15 21 DAYS AFTER OVULATION

I

30

FIG. 16. Fertilization rate of eggs from rainbow trout, which were stripped at intervals up to 30 days after ovulation. When the fish were incompletely stripped, the retained eggs were fertile for 30 days in large fish which had already spawned several times, whereas in small fish, spawning for the first or second time, eggs lost their fertility after 15 days (redrawn after Escaffre er al., 1977).

These studies on marine eggs also show that fertilization rate is maintained longer than the ability to develop normally, which means that fertilization rate alone is not a reliable criterion of egg quality. The viability of eggs is generally low when morphological changes due to overripening appear. Overripening is often visible as discoloration or nontransparency, fusion of cortical alveoli and a “dimpled” appearance of the cytoplasm (Fig. 18). Statova et al. (1982) described morphologically degenerative changes in common carp eggs. Similar changes occur in other species, and an ageing scale has been performed for turbot (McEvoy, 1984). Ageing rate of ovulated eggs is, however, probably dependent on ambient temperatures (Billard and Gillet, 1981; Hay, 1986). The chemical changes occurring during maturation and overripening are poorly described for most species. The most detailed study of amino acid compositions of eggs differing in degree of maturation was carried out on the Don population of the Russian sturgeon by Fedorova (1976). Total contents

DAYS AFTER OVULATION FIG. 17. Effects of delay between ovulation and stripping on egg quality in the Japanese flounder (Limanda yokohama). The figure shows fertilization and hatching rate, and frequency of deformed larvae, in eggs from fish treated with HCG or SG-G100 to induce ovulation. Fertilization rate from HCG- (0)and SG-H100-treated fish ( A ) , hatching rate from HCGtreated fish ( 0 ) .and percentage of deformed larvae from HCG-treated fish (A).Vertical bars represent standard error of the mean (redrawn after Hirose et al., 1979).

( a 1 Pre-activation

( b 1 Posk-activation

FIG. 18. Schematic figures of the internal changes seen in an overripened rainbow trout egg. In contrast to the normal egg (left half of eggs), many of the uncollapsed cortical alveoli (CA) penetrate into the blastoderm (B) along with oil globules (0)in the overripened egg (right half of eggs). Y,yolk; E. egg membrane; CL, cortical layer; PC, perivitelline space (redrawn after Nomura el al., 1974).

96

E. KJORSVIK ET AL.

of free amino acids, contents of proteins and protein amino acids in ripe and overripe eggs, and also in the blood serum, were compared with data on fertilization and viability of larvae. The conclusion of this work was that some essential amino acids in the blood serum of the female may serve as indicators for the biochemical status and quality of eggs during their maturation. This method was later adopted by Golovanenko (1975) to assess the quality of eggs from the bream. Changes in lipid content during the ripening and overripening process in eggs have been inadequately investigated. Devauchelle et al. (1988) found that overripe eggs of turbot contained more lipids than viable eggs, while overripe eggs contained similar levels as immature gonads. The viable eggs contained lower amounts of all groups of lipids, especially phospholipids. Craik and Harvey (1984) found that the major changes associated with overripening in eggs of rainbow trout were a loss of dry matter, an increase in water content, and a decrease in the precipitable protein. These, and most of the other changes described, can be interpreted in terms of a proteolytic breakdown of the yolk protein, and a loss of small organic molecules such as amino acids and peptides through the egg membranes. The increased water content may be due to an osmotic swelling of the eggs secondary to the increased content of osmolytes from proteolytic breakdown of large organic molecules. Japanese flounders exhibit a weight increase after ovulation, and this is probably related to the drinking of water by mature fish which contain atretic oocytes (Hirose et al., 1979).

B. Broodstock Management Gonadal growth, fecundity and egg viability are known to be very susceptible to environmental influence, such as temperature, fish nutrition and stress factors (Rosenthal and Alderdice, 1976; Billard et al., 1981; Watanabe, 1985). Maturation, ovulation and spawning of salmonid fish are primarily determined by changes in daylength (de Vlaming, 1972; Billard et al., 1978; Whitehead et af., 1978; Bromage et al., 1982; Bye, 1984), and one effect of photoperiodic alterations on spawning time for rainbow trout is a change in egg size (Bromage et al., 1984). Bromage and Cumaranatunga (1988) have also made an extensive review on how the maturation cycle in broodstock rainbow trout might be manipulated to improve the quality and numbers of eggs produced. The temperature during gametogenesis is important both for successful spawning and egg viability. Statova et al. (1982) concluded that it was important, for good egg quality in carp, that fish be maintained at a stable

EGG QUALITY IN FISHES

97

temperature close to the lower limit for spawning prior to hormonal treatment. Eggs from broodstock turbot have the best viability if spawning takes place within a certain temperature range and, if the spawning period is shifted, egg quality seems to be very susceptible to high temperatures occurring during the end of gametogenesis (Devauchelle et al., 1988). Temperature and egg size were also negatively correlated in turbot (Bromley et al., 1986). 1. Nutrition

It is generally believed that feed quality and quantity, and feeding regime are important for egg viability. Food restriction generally reduces total fecundity and may delay maturation and decrease the proportion of maturing fish, as found for species such as brown trout (Bagenal, 1969a,b), rainbow trout (Springate et al., 1985), roach (Rutilus rufilus: Kuznetsov and Khalitov, 1978), haddock (Hislop et al., 1978), winter flounder (Pseudopleuronectes americanus: Tyler and Dunn, 1976), cod (Kjesbu, 1988) and plaice (Horwood et al., 1989). Such studies indicate a large potential for improved reproduction in captive fish. Changes in egg size, egg weight and egg composition due to various feed levels seem more variable. Kuznetsov and Khalitov (1978) found that female roach reacted to different feeding conditions by a change in absolute fecundity as well as the fat content of the eggs, whereas egg weight and egg diameter remained relatively constant. Springate et al. (1985) found that rainbow trout fed on full rations had larger eggs than fish fed on half rations. There were no differences in fat and protein levels, and survival and growth up to fingerlings were similar for the two groups. In a study on cod, Kjesbu (1988) concluded that the fecundity and condition factor of captive fish were 2.5 and 1.5 times those of wild cod of the same size. Stressed fish spawned irregularly and had a very low fertilization rate. Excessive feeding led to an increase in the total number of eggs, but not to an increase in egg size. However, comparisons of feed level studies are complicated by factors such as duration of the experiments, effects of feeding levels on fish size, effects of fish size and age on the egg size, and effects from variations in egg size in different spawns for multiple (batch) spawners. The effect of feed on yolk composition may be expected to be of particular importance for egg quality, but limited information is available on the effects of broodstock nutrition on egg viability in fishes (for an extensive review on broodstock nutrition, see Watanabe, 1985). Apart from experiments covered by Watanabe and co-workers, there seem yet to be few published results on correlations between broodstock nutrition and egg quality. Requirements for different feed components are poorly documented

TABLE3. EFFECTS FROM BROODSTOCK DIETON EGGQUALITY ~

Feed component

Species

Eggs Buoyancy

Fatty acids

Hatching

Survival

Krill oil Raw krill

Red sea bream Red sea bream

+ +

+ +

Corn oil Casein/corn oil (EFA-deficient)

Red sea bream Red sea bream

-

-

Red sea bream

+

+

Red sea bream

0

0

Red sea bream Red sea bream

-

-

Red sea bream

-

-

Red sea bream AYU Rainbow trout Atlantic salmon Tilapia

+ + + + +

+ + + +

Protein

Cuttlefish meal White fish meal Very low protein Low protein/ high calorie Minerals Low phosphorus

Vitamins Vitamin E Vitamin C

~~~~~~

Larvae

Source

Normality Watanabe (1985) Watanabe et al. (1984b, 1985b)

-

0

Watanabe (1985)

Takeuchi et al. (1 98 1b)

Watanabe (1985) Watanabe et al. (l984a) Watanabe et 01. (1985b) Takeuchi et al. (1 98 la) Sandnes et al. (1984) Eskeiinen (1989) Soliman et al. (1 986)

Pigments Carotenoids

Trace nietals

+. positive effect; -.

Red sea bream Atlantic salmon Rainbow trout

+

Rainbow trout negative effect: 0, no effect observed.

0 0

+

+

0

Watanabe et al. (1984b) Torrisen (1984) Harris (1 984)

+

Takeuchi et al. (198 1b)

0

I00

E. KJORSVIK ET AL.

for growing, immature fish of many species, and information about the more specific needs of maturing fish is sparse. Experiments aimed at providing information of this kind are expensive and time-consuming, but they are surely needed. The major groups of feed components studied are essential fatty acids (EFA), protein sources and a few vitamins, minerals and pigments. Some important results are summarized in Table 3. The importance of different levels of essential fatty acids (EFA) in the feed (Watanabe er al., 1984a, 1985b) is clearly shown. When red sea bream broodstock were fed a diet containing a high content of corn oil (EFAdeficient), the percentage of buoyant eggs, hatching rate and “final productivity of fish seed” were all significantly reduced compared with controls. The clearest difference occurred at the “final productivity of fish seed”. Offspring from groups of broodstock fed corn oil had a survival rate of about I%, whereas the offspring of groups given frozen raw krill had good survival rates (Watanabe et al., 1985b). In another experiment, red sea bream broodstock were first fed diets containing corn oil, and shortly before spawning the diet was changed to raw krill. This resulted in an elevated “rate of normal larvae” from very low to very good survival (Watanabe et al., 1985b). The corresponding analyses of fatty acid distributions showed that the eggs and larvae from broodstock groups receiving raw krill had almost double the proportions of n-3 HUFA than did offspring from broodstock groups receiving corn oil for the whole period. This was also the case even if the raw krill was only given for a short period before and during the spawning season. This indicates that this species has the ability to change the fatty acid distribution in the eggs rather rapidly. Effects of some feed deficiencies may thus be corrected by paying special attention to the quality of feed just prior to spawning. The general feeding of broodstock should, however, be the best possible, because other essential components may need a longer time to be incorporated into the ovaries at concentrations high enough to ensure the production of good quality eggs. Clear correlations between egg content and broodstock feed composition are found for some components, while incorporation of body stores might complicate the pattern for other components. Correlations between essential fatty acids in broodstock feed and the eggs are clearly shown in several experiments (Watanabe er al., 1984c, 1985a,b; Watanabe, 1985), and content of the fat-soluble vitamin E in the feed was reflected in the composition of the eggs (Watanabe et al., 1985a). However, for proximate and mineral composition, there are no marked differences in eggs from broodstock fed different diets (Watanabe et al., 1984c, 1985a). In the review by Watanabe (1985), several other factors, in addition to the essential fatty acids, were evaluated. The effects of protein source (white fish

101

EGG QUALITY IN FISHES

meal, cuttlefish meal) and dietary protein content were compared. “Final productivity of fish seed” was low for white fish meal, very high for cuttlefish meal, and close to zero for the low-protein diet and EFA-deficient diet (Fig. 19). A diet with a low content of phosphorus was also tested in this study; survival was again close to zero. This shows that mineral deficiency in the feed can affect egg viability. Similarly, in a study of rainbow trout, a broodstock diet lacking trace metal supplement gave almost zero hatching of eggs (Takeuchi et al., 1981b).

‘00 Final production

80

60

40

20

0

1

2

3

4

5

DIET FIG. 19. Effects on the egg quality of red sea bream (Pagrus major) of broodstock diets of different composition. Diet I , control (whitefish);diet 2, low protein; diet 3, low phosphorus; diet 4, EFA-deficient; diet 5, cuttlefish meal (drawn from data in Watanabe, 1985)

A few studies have been carried out on the effects of vitamin content in broodstock diets, and positive effects of vitamin E were shown in studies on red sea bream and ayu (Takeuchi et al., 1981b; Watanabe et al., 1985b). A diet containing very low levels of vitamin C resulted in negative effects for rainbow trout and tilapia (Sandnes et al., 1984; Soliman et al., 1986). Also Eskelinen (1989) found the highest survival of Atlantic salmon fry when broodstock diet was supplemented with ascorbic acid, but the variation between the groups was too high to draw any firm conclusion. In the same

I02

E. KJORSVIK E T A L .

study, high contents of alpha-tocopherol did not increase the survival of eggs and fry, but sufficient amounts of alpha-tocopherol in raw materials of foodstuffs and in tissue stores might explain this result. As discussed earlier in this review, pigments are possibly among the most variable components in eggs (Craik, 1985). Regarding the importance of pigments (carotenoids) in broodstock feed, there were no clear effects on egg viability of Atlantic salmon (Torrissen, 1984) and rainbow trout (Harris, 1984). Pigmented broodstock diets appeared to increase the proportion of normal larvae in a study on red sea bream, but the hatching rate of eggs was not increased (Watanabe et al., 1984b). Only severe deficiencies in essential fatty acids, protein, minerals or vitamins have, hitherto, resulted in clear effects on egg quality. Egg production has high biological priority in maturing fish, and the mobilization of body stores of important components can probably cover minor deficiencies in feed composition. When little is known of the specific broodstock nutritional requirements, the practical composition of a broodstock diet could be based on existing knowledge about the general requirements of each species. Extra additions of unstable components such as ascorbic acid should further enhance the nutritional quality of broodstock diets. 2. Induced spawning The control of the reproductive process in commercially important fish species is a primary goal for fish culturists. Many species will not spawn naturally in captivity. This has led to much experimental work on hormonal therapy in fish, and has resulted in the development of techniques to induce efficient and cost-effective spawning (see the review by Donaldson and Hunter, 1983). Hormonal treatment has been used to investigate ovulation and overripening processes in fish eggs, but not many studies report on the direct effect of induced spawning on egg quality. Important questions relating to the hormonal treatment of fish concern types of hormones, minimum dosage requirements, number of treatments, and minimum size of oocytes (ovarian maturity) before treatment (see Bromage, 1988). The correct dosage requirements are not known for many species, and there appear to be marked interspecific differences. There are also conflicting reports regarding the fertility and viability of eggs obtained from individuals undergoing similar treatments, e.g. LHRH-a therapy (Crim et al., 1983; Crim and Glebe, 1984; Barnabe and Barnabe-Quet, 1985). The type of injection (emulsion vs aqueous solution) seems to affect maturation in plaice (Lirnanda yokoharna) and goby (Acanthogobiusfzavimanus: Aida et al., 1978). Several reports indicate the lowest doses that are

EGG QUALITY IN FISHES

103

required give the best egg quality in Dover sole (Ramos, 1986), grey mullet (Mugil cephulus: Lee et al., 1987) and white sea bream (Mylio berdu: Mok, 1985). Several low-dose injections seem to be more stressful for ayu, and give poorer egg quality, than the same total hormone dose given as a single injection (Hirose, 1980). In a study with Japanese flounder, the eggs remained in a good condition for 2-3 days, depending on the type of hormone used (HCG or SG-Gl00) to induce ovulation (Fig. 17; Hirose et ul., 1979). Piscine gonadotropin was apparently more effective than mammalian hormones in inducing ovulation. Similar results are also reported for ayu (Plecoglossus altivelis: Ishida et ul., 1972), medaka (Oryzius Zutipes: Hirose and Donaldson, 1972; Hirose, 1976) and catfish (Heteropneustesfossilis: Sundararaj and Anand, 1972). According to Hirose et al. (1979), these results may be due to the species-specificity of the piscine and mammalian hormones, and may be related to the evolution of hormone receptors and activities.

V. Conclusions and Recommendations This review has focused on the possible indicators of egg quality. Variations in a range of physical, chemical and biological characteristics have been observed in fish eggs, although the association with subsequent viability has been only equivocally demonstrated for many parameters. Regarding egg quality criteria, it seems important to emphasize that biological processes occurring in fish eggs seem to be basically the same for all teleost species, although expressions and timing may differ. The biochemical compositions, on the other hand, seem more variable and also more species-specific. Because the biochemical content of the eggs may be a reflection of broodstock feed, it seems logical that there may be room for a certain variability in composition without this having negative effects on viability. Parameters like egg survival and hatching rate are very crude measures, and morphological malformations and rate of “normal” (viable) larvae are more reliable indicators of quality. The most useful practical early assessment of the quality of a batch of eggs should be made before incubation, and this implies that a good knowledge of normal egg development is necessary. Early morphological patterns (cortical reaction and early cleavage) may be used in the assessment of the egg quality of several fish species. Chromosomal aberrations have been shown to be useful measures of egg quality, and may have potential as a monitoring tool for egg viability in the sea. To date, no universal biochemical parameter for egg quality has been found. There may be species-specific indicators, but none is well documented. Essential components may provide the key, with both free amino

I 04

E. KJORSVIK ET AL.

acids and fatty acids meriting further study. The use of free amino acids as a quality criterion seems interesting, since Vladimirov (1974) found a connection between the content of certain free amino acids in the eggs of sturgeon and survival. The total fat content of eggs may indicate egg quality. Whether this holds true for most species is yet to be proved. Good quality eggs spawned in nature could provide a reference for optimum values of certain chemical components, such as lipids. Large size variations within an egg batch may indicate poor egg quality, but this may be explained by overripening, where overripe eggs have swollen to larger size than ripe eggs. Egg batches containing both overripe and ripe eggs are probably most common in marine batch spawners that have short ovulation intervals. The morphological and physical characteristics of low viability eggs are often similar to those seen in the egg during the overripening process. It seems possible, therefore, that many of the problems with poor quality eggs are due to overripening changes in the eggs after ovulation. Overripening has been studied in the eggs of salmonids and some marine species. Species vary in the time at which the overripening process begins, but overripening seems to occur much faster in batch-spawners than in fish spawning once per season. Changes in biochemical composition take place during the overripening process (probably due to proteolytic breakdown). The changes include a loss of dry matter and an increase in water content (Craik and Harvey, 1984). The most dramatic changes are morphological, but egg viability will have decreased markedly before these morphological changes are visible. The overripening of eggs is a major problem when spawning fish are kept in captivity. Further research into maturation and overripening processes in fish gonads is therefore urgently needed. The age of the females seems to affect the viability period after ovulation, but it also influences the quality of the eggs in most species that have been investigated. Middle-aged females produce the best eggs, but this rule is not without exceptions. Some fish only spawn once in their lifetime and, in the case of bream, the youngest spawners seem to produce the best eggs (Kuznetsov, 1973b). Surprisingly little is known of how exogenous factors like poor feeding regimes and unfavourable conditions during maturation and spawning influence egg survival. Some species seem to adapt more easily to an "unnatural" environment than do others. Devauchelle et al. (1988) concluded that flatfishes were more difficult to keep in captivity than roundfishes. Malnutrition will certainly affect egg quality, although lower limits for most essential components are not known for most species. There is a lack of studies aimed at determining interactions between broodstock feed composition, the biochemical content of eggs and egg quality. More atten-

EGG QUALITY IN FISHES

105

tion should be paid to the effect of temperature on gametogenesis and spawning. This is a major factor in the determination of egg quality for turbot and carp. Other environmental conditions (e.g. light, salinity, and factors causing stress) will certainly affect egg quality. How egg quality affects the functionality of surviving larvae is also largely unknown, Investigations of egg quality characteristics should involve comparative tests of several parameters for many species, e.g. fertilization rate and activation processes in the eggs, buoyancy of pelagic eggs, morphological characteristics (cleavage pattern, distribution of oil globules, etc.), chorion strength, rate of normal embryos and larvae, and survival and hatching rates. Cytogenetic studies on chromosomal aberrations will certainly increase the sensitivity in such studies, as has been shown in several recent publications. Studies on egg quality should also include chemical analyses of osmolality and composition of total fat, fatty acids, free amino acids, minerals and vitamins. Several other poorly understood aspects should also be further studied. These studies should be aimed at investigating correlations between environmental factors, spawning fish condition and feeding, and egg quality. A better knowledge of overripening processes after ovulation is also required. More attention should be directed towards the problem of producing good quality eggs, rather than large quantities of eggs.

VI. Acknowledgements Thanks are due to the Norwegian Fisheries Research Council for financial support. This chapter was written as part of the Norwegian research program “New Species for Aquaculture”. We are grateful to P. Cameron, A. Stene and V. Vassik for allowing us to refer to their unpublished results. Thanks are also due to Hilkka Falkseth for technical assistance, to Malcolm Jobling for correcting the English text, and to the many people who answered our enquiries.

VII. References Aida, K., Izumo, R. S., Satoh, H. and Hibiya, T. (1978). Induction of ovulation in plaice and goby with synthetic LH-releasing hormone. Bulletin of the Japanese Society of Scientijic Fisheries 44, 445450. Alderdice, D. F., Rao, T. R. and Rosenthal, H. (1979). Osmotic responses of eggs and larvae of the Pacific herring to salinity and cadmium. Helgolunder wissenschaftliche Meeresuntersuchungen 32, 508-538. Bagenal, T. B. (1969a). The relationship between food supply and fecundity in brown trout Salmo trutta L. Journal of Fish Biology 1, 167-182. Bagenal, T. B. (1969b). Relationship between egg size and fry survival in brown trout Salmo trutta L. Journal of Fish Biology 1, 349-353.

106

E. KJORSVIK ET AL.

Barnabe, G. and Barnabe-Quet, R. (1985). Advancement and improvement of induced spawning in the sea-bass Dicentrarchus labrax (L.) using an LHRH analogue injection. Aquaculture 49, 125-132 (in French). Billard, R. and Gillet, C. (1981). Vieillisement des ovules et potentialisation par la temperature des effets des micropolluants du milieu aqueux sur les gametes chez la truite. Cahiers: Laboratoire d’Hydrobiologie de Monterau 12, 35-41. Billard, R., Breton, B., Fostier, A., Jalabert, B. and Weil, C. (1978). Endocrine control of the teleost reproductive cycle and its relation to external factors. In “Comparative Endocrinology” (P. Gaillard and L. Boer, eds), pp. 3748. Elsevier/North Holland, Amsterdam. Billard, R., Bry, C. and Gillet, C. (1981). Stress, environment and reproduction in teleost fish. In “Stress and Fish” (A. D. Pickering, ed.), pp. 185-208. Academic Press, London. Blaxter, J. H. S. (1955). Herring rearing-I. The storage of herring gametes. Marine Research 3, 1-12. Blaxter, J. H. S. (1988). Pattern and variety in development. In “Fish Physiology” (W. S. Hoar and D. J. Randall, eds), Vol. XI, pp. 1-58. Academic Press, London. Blaxter, J. H. S. and Hempel, G. (1963). The influence of egg size on herring larvae (Clupea harengus). Journal du Conseil: Conseil International pour I’Exploration de la Mer 28, 21 1-240. Borovik, E. A. (1962). The problem of the role of carotenoids in embryonic development of rainbow trout. 2nd Byelorussian Zoological Conference, Abstracts of Proceedings, pp. 32-34. Bromage, N. R. (1988). Propagation and stock improvement. In “Intensive Fish Farming” (J. Shepherd and N. Bromage, eds), pp. 103-153. BSP Professional Books/Blackwell Scientific, Oxford. Bromage, N. R. and Cumaranatunga, R. (1988). Egg production in the rainbow trout. In “Recent Advances in Aquaculture” (J. F. Muir and R. J. Roberts, eds), pp. 63-1 38. Croom Helm/Timber Press, London and Sydney/Portland, Oregon. Bromage, N. R., Whitehead, C., Elliott, J., Breton, B. and Matty, A. (1982). Investigations into the importance of daylength on the photoperiodic control of reproduction in the female rainbow trout. In “Reproductive Physiology of Fish” (C. Richter and H. Th. Goos, eds), pp. 233-236. Pudoc Press, Wageningen. Bromage, N. R., Elliott, J. A., Springate, J. R. C. and Whitehead, C. (1984). The effects of constant photoperiods on the timing of spawning in the rainbow trout. Aquaculture 43, 2 13-223. Bromley, P. J., Sykes, P. A. and Howell, B. R. (1986). Egg production of turbot (Scophthalmus maximus L.) spawning in tank conditions. Aquaculture 53, 287293. Bry, C.(1981). Temporal aspects of macroscopic changes in rainbow trout (Salmo gairdneri) oocytes before ovulation and of fertility during the post-ovulatory period; effect of 17a-hydroxy-20~-dihydroxyprogesterone. Aquaculture 24, 153160. Burton, G. W. and Ingold, K. U. (1984). P-carotene-an unusual type of lipid antioxidant. Science 224, 569-573. Bye, V. (1984). The role of environmental factors in the timing of reproductive cycles. In “Fish Reproduction: Strategies and Tactics” (G. W. Potts and R. J. Wootton, eds), pp. 187-205. Academic Press, London.

EGG QUALITY IN FISHES

107

Cameron, P., Berg, J., von Westernhagen, H. and Dethlefsen, V. (1989). Missbildungen bei Fischembryonen der siidlichen Nordsee. In “Warnsignale aus der Nordsee” (J. L. Lozen, W. Lenz, E. Rachor, H. von Westernhagen and B. T. Waterman, eds), pp. 281-294. Verlag Paul Parey, Berlin. Craik, J . C. A. (1985). Egg quality and egg pigment content in salmonid fishes. Aquaculture 41, 6 1-88. Craik, J. C. A. and Harvey, S. M. (1984). Egg quality in rainbow trout: The relation between egg viability, selected aspects of egg composition, and time of stripping. Aquaculture 40,1 15-1 34. Crim, L. W., Sutterlin, A. M., Evans, D. M. and Weil, C. (1983). Accelerated ovulation by pelleted LHRH analogue treatment of spring-spawning rainbow trout (Salmo gairdneri). Aquaculture 35, 299-307. Crim, L. W. and Glebe, B. D. (1984). Advancement and synchrony of ovulation in Atlantic salmon with pelleted LHRH-analog. Aquaculture 43, 47-56. Dabrowski, K . , Luczynski, M., Czeczuga, B. and Falkowski, S. (1987). Relationships among coregonid fish reproductive effort, carotenoid content in egg and survival of embryos. Archiv fur Hydrobiologie 19 (suppl.), 29-18. Dethlefsen, V., Cameron, P. and von Westernhagen, H. (1985). Untersuchungen iiber die Haufigkeit von Missbildungen in Fischembryonen der siidlichen Nordsee. Information die Fischwirtschaft 32, 22-27. Dethlefsen, V., Cameron, P., von Westernhagen H. and Janssen, D. (1987). Morphologische und chromosomale Untersuchungen an Fischembryonen der siidlichen Nordsee in Zusammenhang mit der Organochlorkontamination der Elterntiere. Veroffentlichungen des Instituts fur Kusten- und Binnenjscherei Hamburg 91, 57 PP. Devauchelle, N., Brichon, G., Lamour, F. and Stephan, G. (1982). Biochemical composition of ovules and fecund eggs of sea bass (Dicentrarchus labrax), sole (Solea vulgaris) and turbot (Scophthalmus maximus). In “Reproduction Physiology of Fishes”. Proceedings of the International Symposium on Reproductive Physiology of Fish, pp. 155-157. Devauchelle, N.. Alexandre, J. C., Corre, N. L. and Letty, Y. (1988). Spawning of turbot (Scophthalmus maximus) in captivity. Aquaculture 69, 159-184. Dinis, M. T. (1982). Methods of incubating Dover sole (Solea solea L.) eggs. Relatorios de Actividades do Aquario Vasco da Gama 12, 8 pp. Donaldson, E. M. and Hunter, G. A. (1983). Induced final maturation, ovulation and spermiation in cultured fish. In “Fish Physiology” (W. S. Hoar and D. J. Randall, eds), Vol, 9B, pp. 351403. Academic Press, London. Dushkina, L. A. (1975). Viability of herring (Clupea) eggs and fertilizing capacity of herring sperm stored under various conditions. Journal of Ichthyology 15, 423429. Escaffre, A.-M. and Bergot, P. (1984). Utilization of the yolk in rainbow trout alevins (Salmo gairdneri Richardson): Effects of egg size. Reproduction Nutrition Development 24, 449460. Escaffre, A.-M. and Billard, R. (1979). Evolution de la fecondades ovules de truite arc-en-ciel (Salmo gairdneri) laisses dans la cavite abdominale au cours de la periode post-ovulatoire. Bulletin Francais de Pisciculture 272, 455-164. Escaffre, A.-M., Petit, J. and Billard, R. (1977). Evolution de la quantite d’ovules recoltes et conservation de leur aptitude a ttre fecondes au cour de la periode postovulatoire chez la truite arc-en-ciel. Bulletin Francais de Pisciculture 265, 134-142.

108

E. KJORSVIK ET AL.

Eskelinen, P. (1989). Effects of different diets on egg production and egg quality of Atlantic salmon (Salmo salar L.). Aquaculture 79,275-28 1. Fedorova, L. S. (1976). Physiological and biochemical characters of the reproductive products and larvae of the sturgeons during artificial rearing. Journal of Ichthyology 16, 427436. Galkina, Z. I. (1969). The effect of size and color intensity of eggs on embryonic development and growth of young rainbow trout. Izvestiya Gosudarstveunogo Nauchno-lssledovatel’skogo Instituta Ozernogo i Rechnogo Rybnogo Khozvaistva 68, 173-186. Georgiev, G . S. (1971). Carotenoids and vitamin A content in Salmo irideus eggs and their significance in the initial periods of embryogenesis. Folia Balcania 2, 1-1 I . Ginzburg, A. S. 1972. “Fertilization in Fishes and the Problem of Polyspermy”. Keeter Press, Jerusalem, 366 pp. Glebe, B. D., Appy, T. D. and Saunders, R. L. (1979). Variation in Atlantic salmon (Salmo salar). International Council for the Exploration of the Sea C . M . 19791 M:23. 11 pp. (mimeographed). Golovanenko, L. F. (1975). Qualitative evaluation of the sexual products of the bream Abramis brama from biochemical and histochemical indicators. Journal of Ichthyology 15, 499-502. Graumann, G. B. (1986). Morphological anomalies in the Baltic sea fishes at early stages of ontogenesis. In “Symp. Ecol. Invest. Baltic Sea Environ.” (Ekologija Baltijskogo morja), pp. 282-29 I . Gidrometoizdat, Leningrad. Hagenmaier, H. E., Schmitz, J. and Fohles, J. (1976). Zum Vorkommen von Isopeptidbindungen in der Eihulle der Regenbogenforelle (Salmo gairdneri Rich.). Hoppe-Seyler ’s Zeitschrift fur physiologischen Chemie 357, 1435-1438. Harris, L. E. (1984). Effects of a broodfish diet fortified with canthaxanthin on female fecundity and egg color. Aquaculture 43, 179-183. Hay, D. E. (1986). Effects of delayed spawning on viability of eggs and larvae of Pacific herring. Transactions of the American Fisheries Society 115, 155-161. Hilton, J. W., Cho, C. Y., Brown, R. G . and Slinger, S. S. (1979). The synthesis, halflife and distribution of ascorbic acid in rainbow trout. Comparative Biochemistry and Physiology A 63, 447-453. Hirao, S., Yamada, J. and Kikuchi, R. (1954). Violaxanthin in rainbow trout egg. Bulletin of the Japanese Society of ScientiJic Fisheries 20, 431-434. Hirao, S., Yamada, J. and Kikuchi, R. (1955). Relations between chemical constituents of rainbow trout eggs and hatching rate. Bulletin ofthe Japanese Society of Scientific Fisheries 21, 240-243. Hirose, K. (1976). Endocrine control of ovulation in medaka (Oryzias latipes) and ayu (Plecoglossus altivelis). Journal of the Fisheries Research Board of Canada 33, 989-994. Hirose, K. (1980). Effects of repeated injections of human chorionic gonadotropin (HCG) on ovulation and egg qualities in the ayu, Plecoglossus altivelis. Bulletin of the Japanese Society of ScientiJic Fisheries 46, 813-8 18. Hirose, K. and Donaldson, E. M. (1972). Biological study of ovulation in vitro of fish - 111. The induction of in vitro ovulation of Oryzias latipes oocytes using salmon pituitary gonadotropin. Bulletin of the Japanese Society of ScientiJic Fisheries 38, 97-100. Hirose, K., Ishida, R. and Sakai, K. (1977). Induced ovulation of ayu using HCG, with special reference to changes in several characteristics of eggs retained in the body cavity after ovulation. Bulletin of the Japanese Society of ScientiJic Fisheries 43,409416.

EGG QUALITY IN FISHES

109

Hirose, K., Machida, Y. and Donaldson, E. M. (1979). Induced ovulation of Japanese flounder (Limanda Yokohama) with HCG and salmon gonadotropin, with special references to changes in the quality of eggs retained in the ovarian cavity after ovulation. Bulletin of the Japanese Society of Scientific Fisheries 45, 3 1-36, Hislop, J. R. G., Robb, A. P. and Gauld, J. A. (1978). Observations on effects of feeding level on growth and reproduction in haddock, Melanogrammus aeglefinus (L.) in captivity. Journal of Fish Biology 13, 85-98. Horwood, J. W., Greer Walker, M. and Witthames, P. (1989). The effect of feeding levels on the fecundity of plaice (Pleuronectes platessa). Journal of the Marine Biological Association of the United Kingdom 69, 81-92. Hubbs, C. and Strawn, K. (1957). Survival of F1 hybrids between fishes of the subfamily Etheostominae. Journal of Experimental Zoology 134, 33-62. Ishida, R., Hirose, K. and Donaldson, E. M. (1972). Induction of ovulation of ayu, Plecoglossus altivelis with salmon pituitary gonadotropin. Bulletin of the Japanese Society of Scientific Fisheries 38, 1007-1012. Kamler, E. and Zuromska, H. (1979). Quality of eggs and production of Coregonus albula (L.) in three Masurian lakes. Polskie Archiwum Hydrobiologie 26, 595-623. Kamler, E., Zuromska, H. and Nissinen, T. (1982). Bioenergetical evaluation of environmental and physiological factors determining egg quality and growth in Coregonus albula (L.). Polskie Archiwum Hydrobiologie 29, 71-121. Kim, Ye. D. (1974). Age and annual dynamics of amino acid content in mature sexual products of the carp. In “Qualitative Differences in the Early Ontogeny of Fishes” (trans.). Naukova Dumka Press, Kiev. Kirpichnikov, V. S. (1966). Goals and methods in carp selection. In “Selective Breeding of Carp and Intensification of Fish Breeding in Ponds” (V. S. Kirpichnikov, ed.). Bull. St. Sci. Res. Inst. Lake River Fish. 61, 4 1 4 2 . Kitahara, T. (1983). Behaviour of carotenoids in chum salmon (0. keta) during anadromous migration. Comparative Biochemistry and Physiology B. 76, 97-101. Kjesbu, 0.S. (1988). The spawning activity of cod (Gadus morhua L.). Journal of Fish Biology 34, 195-206. Kjarsvik, E. and Lanning, S. (1983). Effects of egg quality on normal fertilization and early development of the cod, Gadus morhua L. Journal of Fish Biology 23, 1-12. Kjarsvik, E., Stene, A. and Lanning, S. (1984a). Morphological, physiological and genetical studies of egg quality in cod (Gadus morhua L.). In “The Propagation of Cod Gadus morhua L.” (E. Dahl, D. S. Danielssen, E. Moksness and P. Solemdal, eds), Flodevigen rapportserie 1, 67-86. Kjarsvik, E., Davenport, J. and Lanning, S. (1984b). Osmotic changes during the development of eggs and larvae of the lumpsucker, Cyclopterus lumpus L. Journal of Fish Biology 24, 3 1 1-32 I . Knutsen, G . M. and Tilseth, S. (1985). Growth, development and feeding success of Atlantic cod larvae Gadus morhua related to egg size. Transactions of the American Fisheries Society 114, 507-51 1. Kolman, R. (1973). “Quantitative Evaluation of Quality” (trans.). Panstw. Wyd. Ekonomiczne, Warsaw. Krieger, K. G . and Olson, R. N. (1989). Some factors involved in gamete quality in Arctic char (Salvelinus alpinus) broodstock of the Nauyuk Lake strain. Proceedings of the Aquaculture International Congress and Exposition, Vancouver, British Columbia, Canada, 6 9 September 1988, p. 74. Kuznetsov, V. A. (1973). Fecundity of the bream Abramis brama L. and the quality of its eggs. Journal of Ichthyology 13, 669479.

110

E. KJORSVIK ET AL.

Kuznetsov, V. A. (1976). The importance of egg quality in the reproduction of carp and perch populations with regulated river discharge. In “The Ecological Physiology of Fishes” (trans.), Part 2. Naukova Dumka Press, Kiev. Kuznetsov, V. A. and Khalitov, N. K. (1978). Alterations in the fecundity and egg quality of the roach, Rutilus rutilus in connection with different feeding conditions. Journal of Ichthyology 18, 63-70. Lee, C. S., Tamaru, C. S., Miyamoto, G. T. and Kelley, C. D. (1987). Induced spawning of grey mullet (Mugil cephalus) by LHRH-a. Aquaculture 62, 327-336. Loeffler, C. A. and Lovtrup, S.(1970). Water balance in the salmon egg. Journal of Experimental Biology 52, 29 1-298. Longinova, T. A. (1966). Carotenoids of rainbow trout. In “All-Union Conference on the Ecological Physiology, Abstracts of Proceedings”, pp. 48-50. Academy of Science of the U.S.S.R. and the U.S.S.R. Ministry of Fisheries, Moscow. Longinova, T. A. (1967). Carotenoids of rainbow trout in the developments of gonads and of the eggs. In “The Metabolism and Biochemistry of Fishes”, pp. 336340. Vysshaya Shkova Press, Moscow. Longwell, A. C. (1977). A genetic look at fish eggs and oil. Oceanus 20, 45-58. Longwell, A. C. (1982). Cytological-cytogenetical perspectives on petroleum effects in the marine environment - chromosome aberrations. Background Paper NAS Review: Petroleum in the Marine Environment 1981, 66 pp. Longwell, A. C. and Hughes, J. B. (1981). Cytological, cytogenetical and developmental state of Atlantic mackerel eggs from the sea surface of the New York Bight, and prospects for biological effects monitoring with ichthyoplankton. Rapports et Procis- Verbaux des Reunions: Conseil International pour I’Exploration de la Mer 179, 275-29 1. Mangor-Jensen, A. M. (1987). Water-balance in developing eggs of the cod Gadus morhua L. Fish Physiology and Biochemistry 3, 17-24. Mann, G . (1960). Gonadengewicht und Eizahlen bei Regenbogen-forellen. Informationen fur die Fischwirtschaft, B7, p. 2. Martyshev, F. G., Anisimova, M. M. and Privezentsev, Y. A. (1967). “Age Selection in Carp Rearing”. Kolos Press, Moscow. McEvoy, L.-A. (1984). Ovulatory rhythms and over-ripening of eggs in cultivated turbot, Scophthalmus maximus L. Journal of Fish Biology 24, 437-448. Mok, T. K. (1985). Induced spawning and larval rearing of the white seabream, Mylio berda. Aquaculture 44, 4 1 4 9 . Mollah, M. F. A. and Tan, E. S. P. (1983). Viability of catfish (Clarias macrocephalus, Gunther) eggs fertilized at varying post-ovulation times. Journal of Fish Biology 22, 563-566. Murayama, S. and Yanase, M. (1961). The amounts of chemical constituents of eyed eggs of rainbow trout from various sources. Bulletin of the Tokai Regional Fisheries Research Laboratory 31, 31 1-316. Nomura, M., Sakai, K. and Takashima, F. (1974). The overripening phenomenon of rainbow trout, I. Temporal morphological changes of eggs retained in the body cavity after ovulation. Bulletin of the Japanese Society of Scientijc Fisheries 40, 977-984. Pitman, R. W. (1979). Effects of female age and size on growth and mortality in rainbow trout. Progressive Fish Culturist 41, 202-204. Polikarpov, G . G. (1966). “Radioecology of Aquatic Organisms: The Accumulation and Biological Effect of Radioactive Substances”. Reinhold, New York. 314 pp. Potts, W. T. W. and Eddy, F. B. (1973). The permeability to water of the eggs of certain marine teleosts. Journal of Comparative Physiology 82, 305-3 15.

EGG QUALITY IN FISHES

111

Potts, W. T. W. and Rudy, P. P. ( I 969). Water balance in the eggs of the Atlantic salmon, Salmo salar. Journal of Experimental Biology 50, 223-237. Ramos, J. (1986). Induction of spawning in common sole (Solea solea L.) with human chorionic gonadotropin (HCG). Aquaculture 56, 239-242. Reagan, R. E. and Conley, C. M. (1977). Effect of egg diameter o n growth of channel catfish. Progressive Fish Culturist 39, 133-1 34. Rosenthal, H . and Alderdice, D. (1976). Sublethal effects of environmental stressors, natural and pollutional, on marine eggs and larvae. Journal of the Fisheries Research Board of Canada 33, 2047-2065. Sakai, K.. Nomura, M., Takashima, F. and Oto, H . (1975). The over-ripening phenomenon of rainbow trout, 11. Changes in the percentages of eyed eggs, hatching rate and incidence of abnormal alevins during the process of overripening. Bulletin of the Japanese Society of Scientific Fisheries 41, 855-860. Sakai, K., Nomura, M. and Takashima, F. (1985). Characteristics of naturally spawned eggs of red sea bream. Bulletin of the Japanese Society of Scientific Fisheries 51, 1395-1399. Sandnes, K. (1984). Some aspects of ascorbic acid and reproduction in fish. In “Ascorbic Acid in Domestic Animals” (J. Wgger, F. J. Tagwerker and J. Moustgaard, eds), pp. 206-2 12. Royal Danish Agricultural Society, Copenhagen. Sandnes, K . and Brekkan, 0. R. (1981). Ascorbic acid and the reproductive cycle of ovaries in cod (Gadus morhua). Comparative Biochemistry and Physiology A , 70, 545-546. Sandnes, K., Ulgenes, Y., Brzkkan, 0. R. and Utne, F. (1984). The effect of ascorbic acid supplementation in broodstock feed on reproduction of rainbow trout (Salmo gairdneri). Aquaculture 43: 167-1 71. Savel’yeva, E. A. and Shuvatova, T. F. (1972). Some characteristics of the parentprogeny relationship in the Don pike-perch [Lucioperca lucioperca (L.)]. Journal of Ichthyology 12, 241-248. Seymour, E. A. (1981). Gonadal ascorbic acid changes in level with ovarian development in the crucian carp, Carassius carassius L. Comparative Biochemistry and Physiology A 70, 451453. Shatunovskiy, M. I . (1963). Dynamics of fat and water content of gonad flesh in the Baltic Sea Pleuronectesjesus Linne and its relation to the peculiarities in the maturation of gonads. Vopr. Ikhtiol. 3, 652-667 (in Russian). Shatunovskiy. M. I. (1973). The role of studies of the metabolism in solving certain questions of the abundance dynamics of fishes. In “The Ecological Physiology of Fishes” (trans.). Moscow. Soliman, A. K . , Jauncey, K. and Roberts, R. J. (1986). The effect of dietary ascorbic acid supplementation on hatchability. survival rate and fry performance in Oreochromis mossambicus (Peters). Aquaculture 59, 197-208. Springate, J. C. R. and Bromage, N. R. (1985). Effects of egg size on early growth and survival in rainbow trout (Salmo gairdneri Richardson). Aquaculture 47, 163172. Springate, J. R. C., Bromage, N. R., Elliott, J. A. K. and Hudson, D. L. (1984). The timing of ovulation and stripping and the effects on the rates of fertilization and survival to eying, hatch and swim-up in the rainbow trout (Salmo gairdneri L.). Aquaculture 43, 3 13-322. Springate, J. R. C., Bromage, N. R. and Cumaranatunga, P. R. T. (1985). The effects of different rations on fecundity and egg quality in the rainbow trout (Salmo gairdneri). In “Nutrition and Feeding in Fish” (C. B. Cowey, A. M. Mackie and J . G. Bell, eds), pp. 371-393. Academic Press, London.

I 12

E. KJORSVIK ET

AL.

Statova, M. P., Talikina, M. G. and Kalinich, R. A. (1982). Physiologicalkhemical characteristics of the eggs of common carp, Cyprinus carpio (Cyprinidae), under conditions of fish farming. Journal of Ichthyology 22, 117-128. Stene, A. (1987). Light microscopical studies of chromosomes in embryos of cod, Gadus morhua L. Journal of Fish Biology 31, 445-450. Steven, D. M. (1948). Studies on animal carotenoids. I. Carotenoids of the brown trout Salmo trutta L. Journal of Experimental Biology 25, 369-387. Steven, D. M. (1949). Studies on animal carotenoids. 11. Carotenoids in the reproductive cycle of the brown trout. Journal of Experimental Biology 25, 295303. Stevens, R. E. (1966). Hormone-induced spawning of striped bass for reservoir stocking. Progressive Fish Culturist 28, 19-28. Sundararaj, B. I. and Anand, T. C. (1972). Effects of piscine and mammalian gonadotropins on gametogenesis in the catfish, Heteropneustes fossilis (Bloch). General and Comparative Endocrinology 3 (suppl.), 688-702. Takeuchi, M., Ishii, S. and Ogiso, T. (1981a). Effect of dietary vitamin E on growth, vitamin E distribution, and mortalities of the fertilized eggs and fry in ayu, Plecoglossus altivelis. Bulletin of Tokai Regional Fisheries Research Laboratory 104, 11 1-122. Takeuchi, T, Watanabe, T., Ogino, C., Saito, M., Nishimura, K. and Nose, T. (1981b). Effects of protein-high calorie diets and deletion of trace elements from a fish meal diet on reproduction of rainbow trout. Bulletin of the Japanese Society of Scientific Fisheries 47, 645-654. Taylor, M. W. and Medici, J. C. C. (1966). Amino acids requirements of green beetles. Journal of Nutrition 88, 176-180. Theilacker, G. H. (1981). Effect of feeding history and egg size on the morphology of Jack mackerel, Trachurus symmetricus, larvae. Rapports et Procis- Verbaux des Rhunions: Conseil International pour I‘Exploration de la Mer 1 7 8 , 4 3 2 4 0 . Thorpe, J. E., Miles, M. S. and Keay, D. S.(1984). Developmental rate, fecundity and egg size in Atlantic salmon, Salmo salar L. Aquaculture 43, 289-305. Tolbert, B. M. (1979). Ascorbic acid metabolism and physiological function. In “Vitamin C: Recent Advances and Aspects in Virus Diseases, Cancer and Lipid Metabolisms” (A. Hanck and G. Ritzel, eds). International Journal f o r Vitamin Research. 19 (suppl.), 127-142. Tomita, M., Iwahashi, M. and Suzuki, R. (1980). Number of spawned eggs and ovarian eggs and egg diameter and per cent eyed eggs with reference to the size of female carp. Bulletin of the Japanese Society of Scientific Fisheries 46, 1077-108 I . Torrissen, 0.J. (1984). Pigmentation of salmonids - effect of carotenoids in eggs and start-feeding diet on survival and growth rate. Aquaculture 43, 185-193. Tyler, A. V. and Dunn, R. S. (1976). Ration, growth, and measures of somatic and organ condition in relation to meal frequency in winter flounder, Pseudopleuronectes americanus, with hypotheses regarding population homeostasis. Journal of the Fisheries Research Board of Canada 33, 63-75. Vladimirov, V. I. (1974). Dependence of embryonic and larval quality in common carp on the age of females, amino acid content in the eggs and their addition into the water during early development. In “Qualitative Differences in the Early Ontogenesis of Fishes” (in Russian), pp. 94-1 13. Vlaming, V. L. de (1972). Environmental control of teleost reproductive cycles. A brief review. Journal of Fish Biology 4, 131-140.

EGG QUALITY IN FISHES

113

Watanabe, T. (1985). Importance of the study of broodstock nutrition for further development of aquaculture. In “Nutrition and Feeding in Fish” (C. B. Cowey, A. M. Mackie and S. G . Bell, eds), pp. 395414. Academic Press, London. Watanabe, T., Arakawa, T., Kitajima, C. and Fujita, S. (1984a). Effect of nutritional quality of broodstock diets on reproduction of red sea bream. Bulletin of the Japanese Society of ScientiJic Fisheries 50, 495-501. Watanabe, T., Itoh, A., Murakami, A., Tsukashima, Y., Kitajima, C . and Fujita, S. (1984b). Effect of nutritional quality of diets given to broodstock on the verge of spawning on reproduction of red sea bream. Bulletin of the Japanese Society of Scientific Fisheries 50, 1023-1028. Watanabe, T., Ohhashi, S., Itoh, A., Kitajima, C. and Fujita, S. (1984~).Effect of nutritional composition of diets on chemical components of red sea bream broodstock and eggs produced. Bulletin of the Japanese Society of ScientiJic Fisheries 50, 503-51 5. Watanabe, T., Itoh, A., Satoh, S., Kitajima, C. and Fujita, S. (1985a). Effect of dietary protein levels and feeding period before spawning on chemical components of eggs produced by red sea bream broodstock. Bulletin of the Japanese Society of Scientific Fisheries 51, 1501-1 509. Watanabe, T., Koizumi, T., Suzuki, H., Satoh, S., Takeuchi, T., Yoshida, N., Kitada, T. and Tsukashima, Y. (1985b). Improvement of quality of red sea bream eggs by feeding broodstock on a diet containing cuttlefish meal or on raw krill shortly before spawning. Bulletin of the Japanese Society of ScientiJic Fisheries 51, 1511-152 1. Westernhagen, H. von, Dethlefsen, V., Cameron, P., Berg, J. and Furstenberg, G . (1988). Developmental defects in pelagic fish embryos from the western Baltic. Helgolander wissenschaftliche Meeresuntersuchungen 42, 13-36. Whitehead, C., Bromage, N., Forster, J. and Matty, A. (1978). The effects of alterations in photoperiod on ovarian development and spawning time in the rainbow trout. Annales de Biologie Animale, Biochemie et Biophysique 18, 10351053. Wilkonska, H. and Zuromska, H. (1982). Effects of enviromental factors and egg quality on egg mortality in Coregonus albula (L.) and Coregonus lavaretus (L.). Polskie Archiwum Hydrobiologie 29, 123-1 57. Yamamoto, T. (196 1). Physiology of fertilization in fish eggs. International Review of Cytology 12, 361405. Yarzhombek, A. A. (1970). Carotenoids in Salmonidae and their relation to reproduction of these fishes. Trudy vses. nauchnoissled. Inst. morsk. ryb. Khoz. Okeanogr. 69, 234-267 (in Russian). Zhukinsky, V. N. and Kim, Yo. D. (1981). Characteristics of age-related variability in the composition of amino acids and lipids in mature and overripe eggs of the Azov roach Rutilus rutilus and the bream Abramis brama. Journal of Ichthyology 20, 121-132. Zonova, A. S. (1973). The connection between egg size and some of the characters of female carp (Cyprinus carpio L.). Journal of Ichthyology 13, 679-689. Zuromska, H. and Markowska, J. (1984). The effect of sexual products quality on offspring survival and quality in tench (Tinca tinca L.). Polskie Archiwum Hydrobiologie 31, 287-313.

This Page Intentionally Left Blank

Die1 Vertical Migrations of Marine Fishes: an Obligate or Facultative Process? J. D. Neilson and R. I. Perry Marine Fish Division, Canada Department of Fisheries and Oceans, Biological Station, St. Andrews, New Brunswick. Canada

Introduction .. .. .. .. .. .. .. .. The Evidence for Endogenous Rhythms . . . . .. .. .. A. What cyclic events can be identified as zeilgebers controlling vertical migrations? . . .. .. .. .. .. .. .. B. D o variations in local environment affect vertical position? .. C. D o ontogenetic variations in vertical migratory behaviour occur, and are they consistent with chronobiological theory? , . .. .. 111. Discussion . . . . .. .. .. .. .. .. .. IV. Acknowledgements .. .. .. .. .. .. .. V. References . . . . .. .. .. .. .. .. ..

I. 11.

.. ..

.. .. .. ..

..

..

115

122

I22 I45 148

150 I55 156

1. Introduction Die1 vertical migrations are cyclic changes in the position of aquatic organisms in the water column that occur with 24-h periodicity. Such movements occur at various stages of the development of teleosts, although they are often more evident during the first year of life (see Table 1). Despite frequent study, fundamental questions concerning the nature of die1 vertical migrations remain. For example, most life-history traits exhibit phenotypic plasticity in response to environmental factors during development (Stearns, Copyright 0 1990 by Acudemic Press Limited A / / rights of reproduction in any /orm reserved

ADVANCES IN MARINE BIOLOGY VOLUME 26 ISBN 0-12-026126-X

1 I5

116

J. D. NEILSON AND R. I. PERRY

1982; Bervan, 1982). Whether diel vertical migrations of marine fishes vary in such a fashion is not well understood at present. The extent to which patterns of diel vertical migration change with the development of fish is also poorly documented. These problems are of interest from both the biological and fisheries standpoints. From the biological perspective, studies of diel periodicity in vertical migration advance our understanding of the interaction among fish, their predators and prey, and the abiotic environment. Vertical migrations also represent mechanisms through which energy can be transferred among the various depths of the ocean (see, for example, Blaber and Bulman, 1987). Indeed, below IOOOm, diel vertical migrations of zooplankton and micronekton are rare or non-existent, hence vertical movements of deepwater fish such as Notoscopelus elongatus (Angel et al., 1982) may be the only means through which such energy transfers can occur, apart from detrital input and physical processes such as upwelling. From the fisheries viewpoint, an understanding of diel vertical migration is critical for surveys of abundance at all life-history stages. Such surveys of exploited populations of adult fish have long been employed during the stock assessment process and are becoming increasingly significant, particularly in light of the difficulties associated with obtaining reliable commercial fishery statistics (Anon., 1981). Surveys of juvenile (0-group, post-metamorphosis) fish abundance are widely used to assess year-class strength prior to the recruiting cohort entering the fishery (Koeller et al., 1986). Larval fish surveys are also used in surveys of abundance (Sullivan, 1980), sometimes to hindcast spawning stock biomass (Lasker and Sherman, 1981). Should the availability to the sampling gear of any of those life-history stages vary in either a complex or unpredictable fashion as a result of diel vertical migrations, the utility of those surveys would be diminished (Kendall and Naplin, 1981). An example is the study of Walsh (1988), who showed that changes in availability due to vertical migration of juvenile and adult yellowtail flounder (Limandaferruginea) on the Grand Banks were a serious source of bias in the estimation of stock abundance and indices of recruitment. Arimoto er al. (1983) recently reviewed evidence for diel variation in catch rates in set line fisheries off the coast of Japan. Apart from studies of established stocks, Zusser (1958) also noted that a failure to appreciate the nature and extent of diel movements may lead to erroneous conclusions concerning the possibilities of developing commercial fisheries from virgin stocks. Many authors have commented on the apparent constancy in phase and period of vertical migrations. Attempts to explain the regularity in changes in depths occupied by fish have often involved the notion of circadian or endogenous rhythmicity (see, for example, Woodhead, 1966; Gibson, 1973). The term “circadian”, as commonly used, refers to a self-sustained rhythm,

117

DlEL VERTICAL MIGRATIONS

either synchronized or free-running, of 24 h-periodicity. Endogenous rhythms of 24-h periodicity are thought to occur in virtually all organisms, with the possible exception of bacteria and some algae (Schwassmann, 1971; Takahashi and Zatz, 1982). Most endogenous rhythms are synchronized by a natural cyclic phenomenon, or zeifgeber (see Fig. 1; Aschoff, 1965; Takahashi and Zatz, 1982), such as light (Blaxter, 1975). Other candidates for a zeitgeber in the marine environment include temperature, feeding periodicity and tides, among others. The process by which the period of freerunning rhythm is entrained to that of the zeifgeber is known as “phaseshifting” (Takahashi and Zatz, 1982; Fig. 1). In teleosts, the pineal organ is thought to be central to the organization of circadian systems (Kavaliers, 1979a).

c

E

0

s

I

Surface-

Free-running Rhythm

,

Phaseshift

I

-

-m .-0

L L

P)

> c 0

a

m

Bottom

I

E

1

m

a

Entralned Rhythm

0000

I I

1200 T i m e (h)

2400

FIG.I . Diagrammatic representation of how the period of a free-running rhythm becomes entrained through phase-shifting to that of a zeirgeber.

118

J. D. NEILSON AND R. 1. PERRY

However, the extent to which die1 vertical migrations of teleosts is controlled by endogenous rhythmicity is controversial (Nelson and Johnson, 1970; Siegmund and Wolff, 1973; Casimir, 1971; Kavaliers, 1978; Eriksson and van Veen, 1980). This is in contrast to the invertebrate literature, which is replete with examples of die1 vertical migrations apparently driven by endogenous rhythms (see, for example, Cushing, 1951; Bainbridge, 1961; Herman, 1963; Harris, 1963; Ringelberg, 1964; Verwey, 1966; Ringelberg et al., 1967). Even in those investigations, however, the evidence that endogenous circadian rhythms are responsible for diel vertical migrations is often circumstantial and not definitive. The uncertainty surrounding endogenous control of die1 vertical migrations in marine fish is of critical importance. If such movements were under the comparatively precise control of an endogenous rhythm, the phenomenon could be modelled and predictions made with the simplifying assumption that the process was an obligate one. Thus, studies of fish biology and stock abundance predicated on a knowledge of diel vertical migrations would be greatly facilitated. What tests can one employ to detect the occurrence of such rhythmicity? Diagnostic features of the presence of an endogenous circadian system are that rhythms persist (so-called “freerun”) when environmental periodicities are excluded, and that the natural period (z) is close to that of a solar day. However, it is virtually impossible to exclude experimentally all potential zeitgebers, especially those of a subtle geophysical nature that have been purported to exist (Brown et al., 1970), including periodicities in the earth’s movement or magnetic field. Moreover, such tests are certainly not practical under most field situations. A more realistic test might include an examination of whether the postulated endogenous rhythms persist when zeitgebers behave in an aberrant fashion (for example, an interruption of the light-dark cycle during a solar eclipse, or variations in local hydrographic regimes). A further characteristic of endogenous rhythms is that of temperature-independence (Saunders, 1977). By temperature-independent, it is not meant that endogenous rhythms are unaffected by cyclic temperature changes. Indeed, when such cycles have an appropriate period, they may act as powerful zeitgebers for ectotherms (Saunders, 1977; Aschoff, 198I). Temperature-independence implies that patterns of vertical migration do not vary with changes in temperature regimes. If that criterion is not met, it may be possible to conclude that patterns of vertical migrations are not driven by endogenous rhythms. Finally, Woodhead (1966) noted that endogenous rhythms in behaviour tended to “anticipate” the start of the daylight or night period, but were environmentally timed and reinforced by changes in the intensity of the Zeitgeber. In this article we present a review of studies of the die1 vertical migration

DlEL VERTICAL MIGRATIONS

119

of marine fish species. In a limited number of cases, we extend our review to studies of freshwater fish when no comparable information is available for marine forms. The extent to which fish exhibit plasticity in their patterns of diel migrations is examined, both with respect to varying environmental conditions and ontogeny. Most importantly, we also assess whether patterns of die1 vertical migration fit the criteria for endogenous circadian rhythms. Cyclic phenomena in the marine environment that are candidates for zeitgebers are identified. We then examine if the migration pattern linked to the zeitgeber fits the criteria of Saunders (1977) for an endogenous rhythm, i.e. the presumed endogenous rhythm should show a spontaneous periodicity of about 24 h, relative temperature independence, and persistence of the rhythm even when the zeitgeber behaves in an aberrant fashion. Finally, we attempt to address whether the observed die1 periodicity in depth occupied is a result of endogenous effects whereby the periodicity is derived from the fishes’ biochemistry or biophysics, rather than more direct effects of environmental stimuli such as changes in light and temperature providing a cue for certain behaviour. In light of our findings, we then offer some general considerations for the design of field studies of marine fishes, both for purposes of surveys of abundance and for more general studies of their biology. 0

;; 20 Y

60 0 5 0 0 1000 1 5 0 0 2 0 0 0

Time ( h )

FIG.2. Diagrammatic representation of the two primary types of diel vertical migration found in fish.

To facilitate discussion of the various types of diel migrations, it is useful to characterize migrations into two general categories (Fig. 2): fish move up in the water column at the onset of night, and down with the onset of day (type I), and vice versa, fish move down with the onset of night and up with the onset of day (type 11). In Table 1, the occurrence of the two types of diel migrations is documented for many species and life-history stages.

120

J. D. NEILSON AND R. I. PERRY

TABLE1. EXAMPLES OF STUDIES DESCRIBING DIELVERTICAL MIGRATION I N TELEOSTS, BY LIFE-HISTORY STAGE' Type I Common name

Scientific name

Life-history stage

Volga sturgeon Acipenser gueldenstaedti juvenile Bleak Alburnus alburnus juvenile Roach Rutilus rutilus juvenile Rudd Scardinius erythrophthalmus juvenile Bream Abramis brama juvenile Redfish Sebastes sp. juvenile Atlantic cod juvenile Gadus morhua Atlantic cod G. morhua juvenile Atlantic cod G. morhua adult Sand lance Ammodytes sp. larvae larvae Herring Clupea harengus larvae C. harengus larvae C. harengus larvae C. harengus larvae C. harengus larvae C. harengus larvae C. harengus larvae C. harengus larvae C. harengus larvae Capelin Mallotus sp. larvae Horse mackerel Trachurus mediterraneus juvenile lsaza Chaenogobius isaza juvenile Benthosema glaciale juvenile (?) Stomias boa ferox juvenile (?) Xenodermichthys copei juvenile (?) Leuroglossus stilibius larvae Diogenichthys laternatus larvae Merluccius gayi larvae Various spp. larvae Various spp. juvenile Alewife AIosa pseudoharengus larvae Dorosoma cepedianum larvae Yellow perch Perca fr'avescens larvae Smelt Osmerus mordax larvae Morone chrvsoos larvae Haddock Melanogrammus aeglejnus larvae Haddock M . aeglejnus larvae Haddock M . aeglejnus larvae Haddock M . aeglejnus larvae Haddock M . aeglejnus larvae Silver hake Merluccius bilinearis larvae

. '

Source Levin (1982) Bohl (1 980) Bohl (1980) Bohl (1980) Bohl (1980) Beltestad et al. (1975) Perry and Neilson (1988) Koeller et al. (1986) Beamish (1966) Johansen (1925) Potter and Lough (1987) Graham and Sampson (1981) Johannessen (1986) Johansen (1925) Johansen (1925) Lough and Cohen (1982) Potter and Lough (1982) Wales (1975) Wales (1984) Woodhead (1966) Fortier and Leggett (1983) Zusser (1966) Takahashi (1981) Roe and Badcock (1984) Roe and Badcock (1984) Roe and Badcock ( I 984) Sameoto (1982) Sameoto (1982) Sameoto (1982) Leis (1 986) Southward and Barrett (1983) Cole and Macmillan (1984) Cole and Macmillan (1984) Cole and Macmillan (1984) Cole and Macmillan (1984) Cole and Macmillan (1984) Miller et al. (1963) Colton (1965) Perry and Neilson (1988) Bailey (1975) Koeller et al. (1986) Koeller et al. (1 986)

121

DIEL VERTICAL MIGRATIONS

TABLE1. Continued Type I

-

Common name

Scientific name

Life-history stage

Norway pout

Trisopterus esmarkii Pleuronectes platessa P . fresus Pagrus major Agonus cataphractus Liparis sp. Gobius microps G. scorpioides Lampenus Iampetrvormis Centronotus gunnellus Anguilla rostrata

juvenile larvae larvae larvae larvae larvae larvae larvae larvae larvae juvenile

Bailey (1975) Johansen (1 925) Johansen (1925) Kitajima et al. (1 985) Johansen (1 925) Johansen (1925) Johansen (1925) Johansen (1 925) Johansen (1 925) Johansen (1 925) Helfman (1986)

Brevoortia tyrannus Leiostomus xanthurus L . xanthurus Micropogonias undulatus Paralichthys sp.

larvae larvae juvenile juvenile juvenile

Lewis and Wilkens (1971) Lewis and Wilkens (1971) Weinstein et al. (1980) Weinstein et al. (1980) Weinstein et al. (1980)

Red Sea bream

Eel Atlantic menhaden spot spot Atlantic croaker Various flounders

Source

Tvue I1 Common name

Rudd Roach Herring Herring Herring Capelin Gizzard shad Various spp. Sandeels

Scientific name

Life-history stage

Scardinius erythrophthalmus juvenile Rutilus rutilus juvenile Clupea harengus larvae C. harengus larvae C. harengus' larvae Mallotus villosus larvae

Leuroglossus stilbius Lepomis sp. Dorosoma cepedianum Ammodytes personatus A. marinus Pleuronectes platessa

Plaice Vendace Coregonus albula Pinfish Lagodon rhomboides Atlantic cod Gadus morhua Gulf menhaden Brevoortia patronus Gadus macrocephalus

larvae larvae larvae larvae larvae larvae larvae larvae larvae adult larvae larvae

Source

Girsa (1973) Girsa (1973) Wood (1971) Potter and Lough (1982) Graham and Sampson (1981) Beltestad et al. (1975) Ahlstrerm (1 959) Bergmann (1 98 1) Bergmann (1 98 1) Leis (1 986) Yamashita et al. (1985) Ryland (1964) Ryland (1964) Hamrin (1986) Lewis and Wilkens (1971) Brunei (1972) Sogard et al. (1987) Boehlert et al. (1985)

For an explanation of type-I and -11 migrations. see the text. Studies in bold print signify instances where both types of migration have been documented.

"

122

J. D. NEILSON AND R. I. PERRY

While there are many variants around these basic strategies, such as a pre-dawn rise in herring (Clupea harengus) which typically follow a type-I migration (Brawn, 1960), the proposed division provides a useful frame of reference for the following discussion.

II. The Evidence for Endogenous Rhythms A.

What Cyclic Events can be Identijied as Zeitgebers Controlling Vertical Migrations? 1. Light

Of all possible zeitgebers, light has attracted the greatest attention, although many field observations concern the adult stage. One of the earliest observations of die1 differences in depth of teleost larvae was by Hardy in 1922, although it was not reported until much later (Hardy, 1956). He noted variations in the catch of herring larvae between night and day at a single station that were greater than the variations between all other stations on the survey. Without a correction for the day-night differences, he realized that the survey results were valueless. Other early studies of vertical migrations of teleost larvae were by Johansen (1925) and Russell (1926, 1927, 1928). Johansen (1925) suggested larvae did migrate in response to changes in light intensity, but that a more important cause was movement of their principal prey, which in turn responded to changes in light intensity. Russell (1927) suggested light was of prime importance in initiating vertical migrations, and that animals occupied depths that had an optimum light intensity. He hypothesized that as larvae moved away from the optimum, they became stimulated by changes in physiological reactions controlled by photochemical mechanisms to return to the zone of optimum light intensity. He further expected that an endogenous rhythm could occur, synchronized to light. Such rhythms would persist for a time once the zeitgeber had been removed. Woodhead (1966), in reviewing the extensive literature on adult Atlantic herring, concluded that most observations tended to support the hypothesis that schools followed a zone of optimum light intensity in their type-I vertical migrations, although much of the evidence was circumstantial. He further noted, however, that although the basic pattern of vertical migration might be related to changes in surface illumination, there were frequent variations from the norm which suggested that particular “optimum” light intensities might be of little significance to the fish. Blaxter (1975) also reviewed the role of light in controlling the vertical distribution of fish, mainly herring. He too concluded that there was abundant evidence, albeit

DIEL VERTICAL MIGRATIONS

123

of a circumstantial nature, implicating light as a significant regulator of vertical migration. Such evidence included the timing of vertical migration in relation to dusk and dawn, the lack of vertical migration at high latitudes, experiments showing the following of light preferenda and the interference of vertical migration patterns with the use of artificial lights or during eclipses. However, he also found a high degree of variability in the response of fish, both on an interspecific and intraspecific level. For example, studies of the behaviour of herring during eclipses have yielded conflicting results. Skud (1968) showed that juvenile herring caged at the surface off the coast of Maine responded to a partial eclipse as though it was sunset. The response, however, was somewhat variable among individuals. In contrast, Stickney (1 972) found no such response by juvenile herring to a partial eclipse in a similar experiment conducted in 1972, near the site of Skud's work. The work of Blaxter (1975) led him to believe that fish did not have a preferred level of illumination, but initiated migration in response to changes in brightness. Wales (1984) showed that at the larval stage of development, the vertical migratory behaviour of herring did not give indications of an endogenous rhythm. Vertical migration did not reliably persist in constant conditions, nor was there consistent evidence of a circadian rhythm of sensitivity or responsiveness when tested throughout a 24-h cycle. Wales also found that behaviour could be entrained to cycle lengths considerably shorter than 24 h. Among other marine species, examples of the type of variable response to light seen in herring include the work of Beamish (1966), who found that meteorological events modified patterns of diel vertical distribution of adult groundfish off Canada's Atlantic coast. For example, on cloudy days, adult redfish (Sebastes sp.) remained in mid-water, rather than exhibiting their more common type-I vertical migration. Zusser (1958) also described a similar phenomenon for many north-east Atlantic species. Soviet workers apparently continue to support the view that fish have preferenda of illumination (Mel'nikov et al., I98 1). Indeed, these workers are among the few that have proposed a mathematical model which describes diel vertical migrations of fish. In their case, the formulation is: H=

where:

a{cos[l5(T - 12)]*coscp*cos6,

+ sin cp*sin 6 , + sin h , } -log,,E a,

H a

= depth at which a given fish is found (m); = coefficient which averages 9.0 on a clear day and 8.0 on an

h,

=

overcast one; the angle of the sun where the illumination of the water surface is conditionally accepted as 0 (set at 15');

J. D. NEILSON AND R. 1. PERRY

124 cp

6, E a,

= = = =

the geographic latitude of the area (degrees); the sun’s declination (degrees); intensity of illumination (lux); attenuation coefficient.

This equation describes a type-I vertical migration. An expression is also given for the range of depths over which fish might occur:

where H and H ’are the upper and lower limits, respectively, of the depth layer which fish occupy during vertical migrations. Similarly, E’ and E“ are the upper and lower boundaries of the light range in which the fish are found (measured in lux). X , is the nominal transparency of the water as measured with a Secchi disc (m). The maximum depth achieved during the downward migration is also postulated to be related to the transparency of the water, since transparency strongly determines light intensity at depth. Differences in water transparency are suggested to affect diel vertical migrations to a greater extent than surface illumination. The model of Mel’nikov et al. (1981) further predicts that, at high latitudes, the extent of diel migrations are reduced relative to populations occurring at lower latitudes, when other factors are equal. While the authors used acoustic data to develop the model (species and stocks were not given), there appears to have been no validation of the model using data from other sources. While the generality of the model is as yet unknown, the implied rhythmicity with no change in either phase or period is consistent with the presence of endogenous rhythms. Other Soviet workers apparently subscribe to the view that fish behaviour can be adequately modelled using sinusoidal functions such as those described by Mel’nikov et al. (1981), e.g. Yemel’yanov (1976) and Kryuchukov et al. (1984). In a detailed study of a freshwater population of pelagic fish in Lake Turkana, Kenya, Hopson and McLeod (1982) reached a similar conclusion to Mel’nikov ez al. (1981), noting that light intensity and modifications to the light climate resulting from turbidity are the environmental factors chiefly responsible for fishes’ diel vertical migrations. Other studies of the response of teleosts to changing light include that of Carey and Robison (1981), who studied the distribution of adult swordfish (Xiphias gladius) with respect to light in both the Atlantic and Pacific using acoustic telemetry. They concluded that all populations exhibited a type-I vertical migration, and

125

DIEL VERTICAL MIGRATIONS

responded to light. The best example was supplied by a single Atlantic fish which seemed to follow an isolume closely. However, the diel vertical migrations also appeared to reflect the influence of water temperature as well as light (Fig. 3). The conclusion that swordfish responded to light was based on a comparatively small number of fish ( N = 5 and 1 for Pacific and Atlantic, respectively), and the authors seemed somewhat,equivocal in linking the vertical distribution patterns to light solely, mentioning other possibilities such as post-feeding thermotaxis aiding digestion or the need to recover from deep dives. Subsequent work (Carey, in press), where the position of a submersible photometer was co-ordinated with that of a swordfish carrying an acoustic tag, indicated that the vertical movements of the fish were more rapid than the rate of change in isolume depth. It was concluded that the swordfish was probably not changing depth in direct response to light. Roe (1983) examined the vertical distributions of euphausiids and mesopelagic fish in relation to light intensity in the northwestern Atlantic and found that the population of each of the six species of fish occurred at light intensities spanning at least three orders of magnitude.

6

12

18

24

6

12

18

24

8

12

16

T i m e (Hours) 9 Nav

10 Nov

1 1 Nov

FIG.3. Die1 vertical migrations of a single swordfish (Xiphius gludius) off Cape Hatteras, as indicated by an acoustic depth-sensing telemeter. The movements are shown with respect to temperature and time. Note how the type-I migration was apparently modified on 9 November,, as the fish's night-time ascent was limited by a cold stream of water of continental shelf origin. As surface waters warmed with the approach of the fish to the Sargasso Sea, the night-time ascent continued to shallower waters. Reproduced with permission from Carey and Robison (1981).

Morever, there was no evidence that such fish were either restricted to, or followed, an isolume. Roe and Badcock (1984) provided further evidence of the lack of correlation of isolumes with vertical distribution of mesopelagic fishes, with data for a further five species. Examples of such a lack of correlation between light and diel vertical migrations for the juvenile life-

I26

J. D. NEILSON AND R. 1. PERRY

history stage can also be supplied from our own investigations of the diel vertical movements of 0-group gadids in north-west Atlantic waters (Figs 4 and 5). Although the vertical migration of cod appeared to bear some relationship to photoperiod with large catches obtained in surface waters only during darkness, there was no evidence that fish were following a particular isolume. In the case of haddock, catches better reflected the mean tidal current speed than photoperiod (Perry and Neilson, 1988). OZSL E

O W

osoz

0

FIG.4. Distribution of catches of 0-group cod (Gadus morhua) in relation to isolumes through a 48-h period, as determined from stratified sampling at three depths using an International Young Gadid Pelagic Trawl. The light measurements were made using a submersible light meter, with data collected every 4 h, from 0800 to 2000 h only. Dotted lines indicate isolumes extrapolated to local sunrise and sunset. Georges Bank, June 1985.

zzz

00

LLZ ,

SOL

201

-

-

00

ez

Otl

oz (W 2

- o* -

09

FIG.5. Distribution of catches of 0-group haddock (Melanogramrnus aeglefinus) in relation to isolumes through a 48-h period, as determined from stratified sampling at three depths using an International Young Gadid Pelagic Trawl. The light measurements were made using a submersible light meter, with data collected every 4 h, from 0800 to 2000 h only. Extrapolated isolumes are shown as dotted lines. Georges Bank, June 1985.

A particularly interesting contrast of the apparent role of light in mediating diel vertical migrations is provided in two papers dealing with freshwater cyprinids. Bohl (1980) compared the spatial distribution of juvenile roach (Rutilus rutilus), rudd (Scardinius erythrophthalmus), bream (Abramis hrama) and bleak (Alburnus alburnus) through acoustic records obtained

DIEL VERTICAL MIGRATIONS

127

from Bavarian lakes. The depth occupied by the fish was dependent on the water transparency and followed an isolume of about lx. In contrast, Girsa (1973) found that the behaviour and vertical distribution of juvenile rudd and roach was substantially altered by the presence of a predator with diel migrations completely suppressed, an observation also found for some species of tropical reef-dwelling fish (Hobson, 1973). In the case of Girsa’s study, the usual type-I1 migration response and ascent of the juvenile fish as light levels increased was inhibited and did not recur until 6-13 h after the predator had been removed. In the marine environment, a similar response was noted by Nash (1982) in a study of diel variations in the behaviour of small demersal fish on soft sediments on the west coast of Scotland. One such species, Lesueurigobius friesii, tended to be either in its burrow or very close to the entrance during the night. There were also correlations between the time of emergence from the burrow and dawn and disappearance from the mud surface and dusk. However, the presence of large numbers of gadoids caused a reduction in the number of L . friesii caught with a beam trawl, suggesting a modification in cyclic behaviour patterns with the appearance of predators. Similarly, Nesterov and Bazanov (1986) showed that the vertical distribution and behaviour of adult flying fish (Exocoetidae) was modified in the presence of predators. Zusser (1966) provided a further example where the response of fish to light could be modified by a secondary environmental factor, in that case, prey availability. She showed that in adult horse mackerel (Trachurus mediterraneus ponticus), the usual type-I vertical migration thought to be driven by changes in light intensity could be suppressed when prey were plentiful. Batty (1987), in a laboratory study of larval herring behaviour, showed that the presence of prey resulted in a modification of the activity cycle of the larvae, which otherwise was closely associated with light intensity. As noted in the Introduction, if an endogenous rhythm is responsible for the cyclic behaviour, then the rhythm should be intact following the removal or modification of the zeitgeber. Such modifications of the natural photoperiod have been attempted in the laboratory. For example, Kruuk (1963) was able to suppress the diel activity cycle of adult sole (Solea vulgaris) by keeping them in continuous light, but when the fish were maintained in darkness for 2 days, the normal periodicity of activity persisted. Harder and Hempel (1954) and de Groot (1964), working with adult plaice (Pleuronecres platessa) held in constant light, were able to reduce, but not completely suppress, nocturnal activity. Woodhead (1 966) also built convincing cases for the occurrence of endogenous rhythms in marine fish. An anticipatory response was also found to occur in bluegills (Lepomis macrochirus) held in a 24-h light-dark cycle but were briefly exposed to light at a random time in

128

J. D. NEILSON AND R. 1. PERRY

the daily dark cycle (Davis, 1962). Davis found that the duration of the accompanying “light-shock” reaction was longer at times early in the dark period compared with the response later in the dark period. Given Woodhead’s contention (1966) that such anticipatory responses are often indicative of endogenous rhythms, it is possible that light is synchronizing activity cycles in bluegills. Woodhead (1966) has further indicated that laboratory experiments with continuous light or darkness, which may partially or completely inhibit the normal expression of endogenous rhythms, or cause the phase to drift, in no way refutes their presence. This statement is somewhat puzzling, as such tests are often used for the very purpose of showing the presence of an endogenous rhythm (Schwassmann, 1971). Wales (1975) and Kavaliers (1979a) have shown that for a number of teleost species, the control of the vertical migration response to changing brightness is at least partially under extraretinal control. Kavaliers postulated that the pineal organ served to organize the various circadian systems present in fish, and also had a direct photosensory role in the entrainment of circadian rhythms to light-dark cycles. A laboratory investigation was completed by Wales (1984), who monitored the vertical movements of intact and bilaterally enucleated herring larvae during changes in light intensity. Vertical movements to an overhead light source were exhibited with both groups swimming up a column of seawater at surface light intensities below lx. The vertical migration was driven continuously by repeated lightdark cycles. No clear circadian rhythms were evident under constant light or darkness, nor were there diel rhythms of sensitivity or responsiveness to changes in light intensity. The view that the pineal organ serves to integrate circadian systems was supported by the work of McNulty (1982), who stated that in goldfish (Carassius aurafus) the metabolic activities and possible secretory functions of the pineal gland are synchronized to the light-dark cycle. Significantly, McNulty found that die1 changes in the fine structure of the pineal organ including cell, nuclear and nucleolar volume no longer occurred in fish held in continual light for 7 days. Hence, the rhythm did not appear to “free-run” and one of the criteria for an endogenous rhythm mentioned earlier was not met. In contrast to McNulty’s findings, Omura and Ali (1982) found that a diel rhythm of numbers of pineal synaptic ribbons in the killifish (Fundulus heteroclitus) persisted even when fish were kept in continuous darkness. Woodhead (1966) noted the behaviour of fish towards light itself has been shown to change over a 24-h period. Breder (1959) placed fish in a “split” or “choice” tank and noted the percentage of fish choosing an illuminated or darkened chamber. Using a variety of freshwater species, Breder was able to demonstrate a definite diel rhythm in the “light-seeking” responses. Similar experiments were conducted by Kawamoto and Niki (1952) and Kawamoto

DIEL VERTICAL MIGRATIONS

129

and Konishi (1955). A later variant of the experiment was self-selection of photoperiod (Thor, 1972; Colgan, 1975), a class of experiments where the animal is capable of controlling its own light regime. Both Thor and Colgan could find no die1 rhythms in light preferences by the goldfish (Carassius auratus) and the pumpkinseed sunfish (Lepomis gibbosus), and indeed found a great deal of variation among individuals. On the other hand, Kroneld (1976) postulated the occurrence of an endogenous rhythm in his study of the periodic locomotor activity of adult burbot (Lora Iota) at the Arctic Circle. In attempting to interpret his data using existing chronobiological theory, he noted that observed rhythmicities showed considerably more flexibility in relation to the zeirgeber than was expected. We conclude, as did Blaxter (1975) and McFarland (1986), that while it is abundantly clear that light is important in mediating die1 vertical migrations, it cannot with certainty be assigned the role of zeitgeber. It seems equally plausible that light acts upon die1 rhythms of fish in a direct fashion, rather than as a synchronizer of endogenous rhythms. Perhaps the best example of this was a laboratory-based study of Halichoeres poecilopterus by Kabasawa (1982). The movement of the fish, as recorded by an infra-red sensing system, seemed independent of any endogenous rhythmicity, and responded directly to the type of lightaark cycle and to water temperature. Such an observation is inconsistent with the hypothesis of endogenous control, given that chronobiological theory postulates that endogenously driven cycles are temperature-independent. Therefore, the role of light as a zeitgeber, and indeed the occurrence of light-mediated circadian rhythms, can be questioned in such instances, inasmuch as fish often do not appear to be responding in an obligate fashion to changes in illumination.

2. Food Although the studies relating die1 movements to prey distributions are less numerous than those seeking correlations with light, several authors have attempted to relate the vertical distribution of fish to their prey. Takahashi (1981) found that juvenile isaza (Chaenogobius isaza) in Lake Biwa, Japan, followed a type-I vertical migration to the surface at night where they restricted themselves to a thin stratum ranging from the upper layer of the thermocline to the lower layer of the epilimnion. Takahashi conjectured that the reason for the migration of isaza might be connected with an abundance of their zooplankton food in the upper stratum. Bailey (1989) ascribed the die1 vertical movements of 0-group walleye pollock (Theragra chalcogramma) in the eastern Bering Sea to food availability. When prey availability was high, the pollock either did not migrate

130

J. D. NEILSON AND R. 1. PERRY

vertically or their movement from the warm surface layer was delayed several hours. When prey availability was low, even the smaller fish undertook diel vertical migrations, which at prey-rich sites had demonstrated less propensity to undertake such movements. Bailey further noted that his field observations could be reconciled with the laboratory work of Smith et al. (1986), who found that an energetic advantage of diel vertical migration accrued to 0-group walleye pollock when deprived of food. When prey availability was low, growth was enhanced through exposure to low temperatures. However, when prey were abundant, maximum food intake and growth were positively correlated with temperature. Fortier and Leggett (1983) showed that both the vertical distributions and migrations of post-larval capelin (Mallorus villosus) and herring were closely related to the vertical distribution of their prey. Similarly, mesopelagic species, including the lantern fish (Benthosema glaciale) often follow a type-I vertical migration to the upper 100-200 m of water to feed on zooplankton, their principal prey (Gjosaeter and Kawaguchi, 1980; Roe and Badcock, 1984; Sameoto, 1988). A predator of myctophids, Merluccius novaezelandiae, undertakes die1 vertical migrations that closely mirror those of its principal prey, bringing it within 50 m of the surface at night (Bulman and Blaber, 1986). In freshwater, Helfman (1986) described the type-I vertical migration of adult American eels (Anguilla rostrata) in a cave spring, as they fed at night. A similar nocturnal feeding pattern was also described for the zooplanktivorous tidewater silverside (Menidia beryllina) by Wurtsbaugh and Li (1985). They also showed that the patterns of diel migration were not rigid, and described both sex- and size-related differences as well as seasonal changes. Browman and Marcotte (1986) have also described considerable variation in diel feeding behaviour of 30 Atlantic salmon (Salmo salur) alevins followed individually through a 15-h period. Begg (1976) showed a correlation of the type-I movement of the Tanganyika sardine (Limnothrissa miodon) and dominant prey zooplankton. He further observed that the movement of the sardine was closely linked to the photoperiod, including the phase of the moon at night. Hobson (1973) described die1 feeding migrations in tropical fish, and claimed that studies of the daily activities of such animals show the overriding influence of two primary concerns: to eat and to avoid being eaten. In our investigations of 0-group gadids on Georges Bank, we have found that cod and haddock show a greater partitioning of their diet (Fig. 6) and diel vertical migrations (Fig. 7) at a food-poor site associated with thermally stratified waters relative to a comparatively food-rich site located in isothermal waters. This may be a mechanism to reduce competition when food is limited (Perry and Neilson, 1988). Gascon and Leggett (1977) found

131

DlEL VERTICAL MIGRATIONS

a similar response of a freshwater fish community to food limitations in Lake Memphremagog, Quebec. Percentage of Total Numbers in Diet- Mixed Site

(A)

20

30

40

60

YeomyaIa amerleane 0.08 Paguru. l a r v a e

P

Oammarldea

Cad

Calanolda sagltte efegana

~ ~ m a c f nSeP .

(6)

Percentage of Total Numbers in Diet- Stratified Site 0

2

4

r

l

1

6

1

8

10

1

I

W

16

20

I

1

FIG.6. Zooplankton prey composition of gut contents of 0-group Atlantic cod (Gudus morhua) and haddock (Melunogrummus aeglefinus) from a thermally mixed (A) and a stratified (B) site on Georges Bank in 1985. Only zooplankton occurring in 25% of the guts examined are included. Numbers at the end of the histograms represent the value of a chi-squared electivity index for that taxonomic group and fish species. Only those values significantly different from 0 and P > 0.01 are included. From Perry and Neilson (1988).

In apparent agreement with Hobson (1973), Turuk (1973) claimed that the daily vertical movements of adult Atlantic cod are governed by feeding patterns. When feeding on pelagic organisms [capelin, amphipods (Themisto sp.)], cod follow a type-I vertical migration. When feeding on benthic food,

J. D. NEILSON AND R. I. PERRY

132

cod descend to the bottom at night and ascend during the day, a type41 migration. Interestingly, a duality in vertical migration patterns had already been demonstrated for adult Atlantic cod in the Gulf of St. Lawrence by Brunel (1969, who showed that the “classic” type-I migration occurred from mid-July to September. Animals exhibited a type-I1 response somewhat earlier in the year, from May to mid-July. Brunel suggested that pelagic prey swarming in the maximum June daylight acted as visual stimuli, attracting cod to the midwater from the bottom during daytime. Turuk ( 1974) examined the relationship between fatness (defined as the weight of the liver) expressed as a percentage of the total weight of adult Atlantic cod. Cod with minimum fat reserves most often made regular daily vertical migrations. Among other groundfish species, adult beaked redfish (Sebastes sp.) on the Grand Banks of Newfoundland are known to follow a type-I vertical migration in pursuit of their prey (Atkinson, 1989). Similar behaviour has been noted for a Pacific congener, SebastesJlavidus (Lorz et al., 1983). I

I

0

I

1

I

I

I

I

1

400

800

1200

1600

2000

Time ( A D T )

FIG.7. Mean depth occupied (determined from depth-stratified sampling at three depths) by 0-group haddock (Melanogrammus aeglefinus) and cod (Gadus morhua) on Georges Bank in 1985. The stations were occupied for 24 and 48 h for the stratified and mixed sites, respectively. ADT refers to Atlantic Daylight Time. From Perry and Neilson (1988).

133

DIEL VERTICAL MIGRATIONS

When the less common type-I1 vertical migration pattern is found, it can often be attributed to similar movements of prey. For instance, there are several examples of this type of migration behaviour in the genus Ammodytes, occurring over a wide geographic range, e.g. see studies of the larval stages of A . umericanus (Covill, 1959), larval A . murinus (Ryland, 1964) and A . personatus (Yamashita et ul., 1985). Ryland (1964) also found that the type-I1 migration of larval plaice (Pleuronectes plutessu) was related to food availability. Surface

1

Zone o f Maximum Feeding Opportunities a n d ;,;;ation

.-0 c L

m >

(A) Surface

c

0

0,

0 E

m 0

Bottom

(6) Dawn

Dusk

FIG.8. Diagrammatic representation of two types of interactions between visually orienting juvenile fish following a type-I migration and their zooplanktonic prey. In the first instance (A), the prey remain relatively high in the water column. Predators of the juvenile fish also occur in that layer. To optimize energy input against the risk of predation, a crepuscular pattern of feeding occurs - the so-called anti-predation window (Clark and Levy, 1988). The second scenario (B) shows the implications of co-occurring type-I migrations of both juvenile fish and their prey. In this case, a crepuscular activity pattern would also be expected.

From laboratory studies, we know that the periodicity of feeding affects the physiological condition of fish in a direct fashion. For example, Spieler and Noeske (1984) have shown that the time at which an adult goldfish feeds is sufficient to reset its internal clock. Experiments with the formation of

134

J. D. NEILSON AND R. I. PERRY

daily growth increments in early life-history stages of fish, a process otherwise thought to be under the general control of an endogenous rhythm (Campana and Neilson, 1983, have indicated that feeding periodicity affects the number of such rings produced (Neilson and Geen, 1982). Such findings further underscore the potential of feeding periodicity for entrainment of endogenous rhythms. As demonstrated by the studies cited in preceding paragraphs, there is ample evidence that the diel vertical distribution of fish is profoundly influenced by the movement of their zooplanktonic prey. Hence, it would seem that there is sufficient evidence, both from field observations and laboratory studies, to support the view that feeding can be a significant zeitgeber. Our review of the literature suggests that there are at least two possible scenarios describing the interaction of zooplankton prey and visually orienting predators that follow a die1 vertical migration (Fig. 8). In the first scenario, the prey remain relatively shallow in the water column. Predators of the planktivores also occur in that layer. To optimize energy input against the risk of predation, a crepuscular pattern of feeding occurs - the so-called “anti-predation window” (Clark and Levy, 1988). Examples of this type of interaction in the marine environment are given in Gjosaeter and Kawaguchi (1980), Roe and Badcock (1984) and Sameoto (1988). The second scenario shows the implications of co-occurring type-I migrations of both juvenile fish and their prey. Two “feeding windows” are available for the predators: at dawn prior to migration of prey into depths where insufficient illumination for efficient predation exists, or at dusk immediately prior to nightfall. In this case, a crepuscular feeding pattern would also be expected, with two peaks in activity per 24h. Although Fig. 8 shows the sort of interactions that would occur if fish were following a type-I vertical migration, we postulate that two peaks of feeding would also follow from a type-I1 migration. An example of co-occurring type-I migrations is provided by an acoustically telemetered swordfish followed by Carey (in press), shown with respect to the movement of its presumed prey (Fig. 9). Could the two peaks in feeding activity somehow serve as a zeitgeber entraining an endogenous rhythm of vertical migrations? It is well known that if the period of the zeitgeber ( T ) lies outside the so-called primary range of entrainment of the period of the endogenous clock (T), free-running behaviour or aperiodicity results. However, if the period of the zeitgeber is a submultiple of T, say 12 h (as is the case in Fig. 8), then entrainment is possible (Saunders, 1977). Having shown that chronobiological theory allows for a zeitgeber of such a period, it remains to ascertain whether in the wild, periodic availability of prey is sufficiently predictable to constitute a zeitgeber . Esterley (1917) demonstrated that control of diel vertical migration in

DlEL VERTICAL MIGRATIONS

I35

Acurtia could be due to an endogenous rhythm, which would occur regardless of external stimuli (except possibly gravity). Harris (1963) found an endogenous rhythm in locomotor activity of Duphniu mugnu, with a periodicity of 24 h when held under continuous dark, and 28 h when held under conditions of continuous light. This cycle of locomotor activity was believed to induce vertical migration. Enright and Hamner (1967) studied mixed populations of marine zooplankton in large experimental enclosures. Taxonomic groups examined included copepods, amphipods, mysids, decapods, medusae, chaetognaths, and mollusc larvae. They demonstrated that persistent endogenous rhythms were important to the vertical migrations of several species of crustaceans, but the migrations were conditioned by previous exposure to a light-dark cycle. In other organisms, vertical migrations occurred in direct response to light intensity. Enright and Hamner (1967) concluded that although vertical migration of zooplankton may appear to be a single phenomenon, its underlying physiological mechanisms are not uniform among species.

FIG.9. The movement (shown as the white line) of an acoustically telemetered swordfish (Xiphiusgludius)on Georges Bank in 1982, shown with respect to the movement of its presumed prey, as indicated by a %-kHz echogram. Dawn was at 0445 h. Reproduced with permission from Carey (in press).

Most reviews (Cushing, I95 I; Longhurst, 1976; Raymont, 1983) suggest that vertical migration of zooplankton occurs primarily in response to external (i.e. exogenous) stimuli, of which light intensity is seen as the most important: other factors include temperature, salinity, hydrostatic pressure, oxygen concentration and food supply. However, Enright and Hamner (1967) noted that it is difficult to distinguish between the role of a stimulus as

136

J. D. NEILSON AND R. I. PERRY

a direct cause of vertical migration or as a modifier of an internal rhythm. Combinations of these external stimuli may act to modify or override the dominant influence of light intensity. Vertical migrations of zooplankton may be directly affected by factors such as the presence of a thermocline (Southward and Barrett, 1983), to the extent where the same species may have different migratory behaviours on either side of a tidal front (Turner and Dagg, 1983; Sameoto, 1984; Scrope-Howe and Jones, 1986). Raymont (1983) noted that among conspecifics of zooplankton, sexes and stages can have different patterns of vertical migration due to differences in their physiological states, such as gonad condition or their food requirements. Dagg (1985) showed that well-fed Neocalanus plumchrus exhibited diel migration behaviour, whereas food-limited individuals did not. Longhurst (1976) concluded that physiological condition and light adaptation interact with endogenous rhythms in a complex manner to produce the many observed patterns of vertical migration. Given the considerable differences in vertical distribution exhibited even by conspecific zooplankton at a single site (see, for example, Pearre, 1973, 1978), it is difficult to ascribe the role of zeitgeber to such a variable phenomenon. 3. Bioenergetic advantage In instances where fish inhabit thermally stratified waters, it has been suggested that they may gain an energetic advantage through diel migration. How the benefit is obtained is unclear. According to some authors, diel migration to cooler bottom waters may represent an energy conservation strategy (Brett, 1971; Biette and Geen, 1980). However, Wurtsbaugh and Neverman (1988) studied the diel vertical movements of juvenile sculpin (Corms extensus) and concluded that its type-I migration conferred an advantage through elevated digestive rates in warmer near-surface waters, permitting greater feeding and growth. These findings are difficult to reconcile with those of Smith et al. (1986), who found that in the laboratory, the energetic advantage of diel migration for juvenile walleye pollock was only apparent for fish reared under low food availabilities, where growth was favoured by low temperature. Wurtsbaugh and Neverman (op. cit.) manipulated temperature regimes, but did not vary ration, and hence it is difficult to compare the results of the two studies. Alexander (1972) provided a detailed review of the energetics of vertical migration by fishes and found further ambiguities. He noted that while there were instances where the energetic advantage of moving into cooler water were clear, the energy saving did not provide a universal explanation for the vertical migrations of fish. Uotani ( 1 973) noted a diel cycle of swimbladder inflation in larvae of

DIEL VERTICAL MIGRATIONS

137

clupeoid fish, including Engraulis japonicus, Sardinops melanosticta and Etrumeus teres, where fish captured at night near the surface had full swimbladders and those captured during the day had empty ones. Hunter and Sanchez (1976) noted such differences may represent a critical energysaving mechanism for the starvation-prone engraulid larvae, if the benefits exceeded the cost of vertical migration to the surface or swimbladder inflation at night. They were able to describe diel changes in swimbladder inflation in the northern anchovy (Engraulis mordax) similar to those observed by Uotani (1973) for other species. From their laboratory-based observations, they concluded that the diel rhythm of inflation did indeed represent an energy-saving mechanism. Hoss and Phonler (1984) showed that Gulf menhaden larvae (Brevoortia patronus) > 12 mm standard length had deflated swimbladders by day and inflated ones at night, achieved by swallowing air at the surface. However, subsequent studies of the diel migrations of menhaden larvae in the northern Gulf of Mexico (Sogard et al., 1987) showed that the larvae slowly sink at night, despite having gas in their swimbladders, and are distributed at various depths. If thermally stratified waters were a persistent feature, it is conceivable that the periodic change in metabolic rate associated with vertical migration could act as a zeitgeber. However, in many locations, thermal stratification may be a seasonal or even shorter-lived event; in particular, subject to breakdown by wind action (e.g. Kraus, 1977). It is more likely that the energetic advantage, when present, is a consequence of the migration, not a factor controlling it. 4. Tides Endogenous rhythms are, of course, not limited to those of 24-h periodicity. For animals that live in the littoral zone, the regular rise and fall of the tide can potentially provide a powerful zeitgeber (Enright, 1965). Studies of intertidal fishes have found that several species show marked rhythms of locomotor activity at near-tidal (circatidal) periodicity. Several relevant studies are reviewed by Gibson (1978). Such examples may be divided into those species with persistent endogenous rhythms of circatidal periodicity, and those in which the periodicity does not persist for more than 2-3 cycles under constant conditions. Perhaps the clearest examples of persistent circatidal rhythms of locomotor activity in intertidal fishes are provided by the Blenniidae. Gibson (1965, 1967) described an overt circatidal rhythm of Blennius pholis that persisted for several days under constant conditions in the laboratory. Coryphoblennius galerita (Gibson, 1970) is another species which demonstrated continuous circatidal activity rhythms under constant conditions for

138

J. D. NEILSON AND R. I. PERRY

up to 5 days, the duration of the longest experiment. Periodicity of activity was greater than 12 h, in keeping with the phase of the tide, with activity strongest at times corresponding to high tide in the natural environment. Activity patterns of the sand goby (Pomatoschistus minutus) were also found to exhibit a persistent circatidal rhythm when held under constant conditions; however, peak activity occurred at the predicted time of ebb tide in their natural environment (Gibson and Hesthagen, 198I). Such differences in the timing of maximal activity with different phases of the tide may relate to the environmental characteristics in which the species occur (Gibson, 1982). By being most active on the ebb tide, fish on sandy shores may avoid becoming stranded at low tide. In contrast, fish on rocky shores find refuge in the tide pools during low tide, becoming most active at high tide when they are able to move and forage more extensively. These differences in activity for fishes on both rocky and sandy shores will certainly influence their vertical distributions in the water column. Another example of a species with a persistent circatidal rhythm of locomotor activity is the hogchoaker (Trinectes maculatus), in particular from the Patuxent Estuary, Maryland (O’Connor, 1972). Rhythmic activity patterns persisted under conditions of constant light, with maximal activity corresponding to slack tide. This example is particularly interesting as the study was initially undertaken to investigate the reasons this species was available to trawl gear at only certain times of the day. Several species display rhythmic activity with a circatidal period in their natural environment, but under constant laboratory conditions these patterns degrade within a few cycles, generally to one of circadian periodicity (Gibson, 1971, 1978; Fig. 10). The tide pool cottid (O!igocottus rnaculosus) from the Pacific coast of Canada exhibits a short-lived (I-day) circatidal activity pattern (Green, 1971). Gibson (1967) suggested the short persistence of the circatidal rhythms of the littoral fishes Acanthocottus bubalis and Ciliata mustela could be related to their ecologies. Both species often occur sublittorally, where an endogenous circatidal activity pattern would be of little need, for example to prevent stranding. The North Sea plaice (Pleuronectes platessa) provides another example of a species that exhibits circatidal activity patterns when it lives as a juvenile in shallow water, but quickly reverts to a circadian rhythm when held under constant conditions (Gibson, 1973). The intensity of the tide varies between geographic locations, with some regions having minimal tidal fluctuations. These variations enable useful comparisons to be made between the response of the same or related species to the tide, and provide a test of the magnitude of tidal changes required for it to act as a possible zeitgeber. Littoral and shallow water fishes in the Baltic, which has a minimal tidal range, seem to regulate their activities by

139

DIEL VERTICAL MIGRATIONS

light (Gibson, 1982). Young flounder (Platichthysflesus)in the Baltic appear to have a bi- or tri-modal circadian rhythm in the laboratory (Muus, 1967), but on beaches with stronger tidal variations they present (short-lived) circatidal rhythms under constant conditions (Gibson, 1976, 1978). The absence of a circatidal activity pattern in the sea catfish (Arius felis) occurring along the coast of Texas, which has a tidal range of 0.3 m, has also been noted (Steele, 1985). Naylor (1976) suggested that genera naturally occurring in tidal and non-tidal areas appear to have predominant circadian activity rhythms and rapidly lose any endogenous circatidal rhythmicity in non-tidal conditions. However, Gibson and Hesthagen (1981) report that sand goby (Pomatoschistus minutus) collected from the Oslofjord, Norway (an area of small and unpredictable tidal range), showed a weakly persistent circatidal rhythm without any endogenous circadian component, although there was considerable variability between individual fish.

r

4

1200 2400 High Tide

4

1200

4

2400

4

1200

4

2400

41 2 0 0

(h)

FIG.10. The mean hourly activity of two Coryphoblennius guleritn held in complete darkness for 3 days. The arrows indicate the predicted time of high tide. Reproduced with permission from Gibson (1970).

While the influence of the tide on locomotor activities of littoral fishes is quite clear, it is less evident how the tide can affect, or even be detected by, fishes offshore. Yet there is now good evidence that many species of fish, and particularly the early life-history stages, can effectively utilize the two-layer circulation system of estuaries for transport from, or maintenance within, a specific area. Rogers (1940) appears to have been the first (Iles and Sochasky, 1985) to provide an example of the retention of fish larvae (smelt,

140

J. D. NEILSON AND R. I. PERRY

Osmerus mordux) within an estuary (Miramichi River). He suggested this was due to the interaction of the type-I vertical migratory behaviour and the two-layer estuarine circulation (i.e. fresher water flowing out of the estuary in the upper layer, and saltier water flowing into the estuary in the deep layer). Laprise and Dodson (1989) have reanalysed Rogers’ (1 940) results, and found them to be consistent with the hypothesis of retention by tidal vertical migration, which predicts that larvae will occur near the surface during flood tides and closer to the bottom during ebb tides. The result is relatively little net downstream advective transport. Laprise and Dodson (1989) further examined the hypothesis of retention of 0. mordax larvae by tidal vertical migration in the well-mixed upper estuary of the St. Lawrence River, where there is a strong seaward residual current in the surface layer and a slow upstream current in the deep layer. They found that these larvae were near the surface during flood tides, where maximum upstream currents occurred, and in the slower deep currents during ebb tides. They proposed that this type of tidal migratory behaviour may be typical of Canadian Atlantic smelt populations, with the adaptive significance being the control of horizontal larval distributions and ultimately maintenance of population isolation. Fortier and Leggett (1982, 1983) investigated the concept of circatidal vertical migrations of larval capelin and herring in the St. Lawrence estuary. In their 1983 paper, they concluded that the average position of post-larval Atlantic herring was close to the depth of null longitudinal velocity, and that vertical migrations followed a semi-diurnal cycle which brought the postlarvae into the upper layer during flood tide. In contrast, capelin larvae exhibited only shallow vertical migrations (within the upper 20 m), with older larvae occurring closer to the surface. This vertical distribution tended to increase their downstream transport. Several species of fish spawn offshore but have nursery areas within estuaries (Boehlert and Mundy, 1988). In contrast to the above studies, where the problem is retention of larvae within the estuary, larvae of offshore spawning species must first move into the estuary, and then remain there in the nursery area. Weinstein et al. (1980) studied this problem in an inlet in North Carolina for three species of commercially important fishes: the spot (Leiostomus xunthurus), Atlantic croaker (Micropogonias undulutus) and flounders (Puralichthys sp.). They found that post-larval Atlantic croaker tended to remain oriented to the bottom at all times, and so accumulated in large numbers at the head of the estuary. In contrast, postlarval spot and flounders showed vertical migratory behaviour with responses to both photoperiod and tide. By day and during ebb tides, these species remained on or near the bottom, but at night they moved off the bottom, occurring higher in the water column during flood tides. It was

141

DIEL VERTICAL MIGRATIONS

suggested that this shallow night-time and flood-tide distribution enhanced the movement of these species into their marsh nursery areas, where they would settle back to the bottom during daylight or ebb tide periods. The model devised by Weinstein et al. (1980) is presented in Fig. 11. These aLthors further suggest the tidal response may be particularly important in well-mixed estuaries where there is negligible upstream flow in the deep layers. The North Sea plaice (Pleuronectes platessa) is another species which spawns offshore, but with nursery grounds in near-shore and estuarine areas. Larval plaice also show an interaction of photoperiod and tide, such that they move into midwater and near-surface depths in greatest numbers during night-time flood tides, and so are transported preferentially into the estuary (de Veen, 1978; Rijnsdorp et al., 1985). Spot- P 8 r a l l c h t h y r Day Surface

Surface

SP.

................. . . . . . ........... ....... Night

Atlantic Croakor Day

. . . Night

Net Non-tidal Flow (Upper L a w )

Downstream

Upstream

FIG. 11. A model of movements with the tide for three taxa of post-larval fishes in a North Carolina estuary, showing different strategies of tidal-stream transport. Note the interaction of both tidal and photoperiod response. Spot and flounders use shallow tidal areas, whereas Atlantic croakers move to the upper reaches of the estuary, presumably by remaining deep in the water column at all times. Reproduced with permission from Weinstein e? al. (1980).

The clearest examples of vertical behaviour patterns interacting with tidal currents in open waters are provided by selective tidal stream transport (Greer Walker et al., 1978). Its essential feature is that the fish are in the water column and are transported in a certain direction during one stage of the tide, and drop out of the water column and experience relatively little displacement during the alternate stage of the tide. This type of interaction was originally described by Creutzberg (1961) for elvers of the European eel (Anguilla anguilla), and has since been observed for adult sole (Solea solea), Atlantic cod (Gadus morhua) and dogfish (Scyliorhinus caniculus) on the

142

J. D. NEILSON AND R. I. PERRY

European continental shelf (Arnold and Cook, 1984) as well as the “glass eel” life-history stage of the American eel (Anguilfarostruta) (McCleave and Kleckner, 1982). Tidal stream transport appears to be a particularly important behavioural mechanism assisting ripening and spent North Sea plaice in their migrations to and from spawning grounds in the Southern Bight, and assisting juvenile plaice to enter their nursery grounds (Arnold and Cook, 1984). With such diverse examples of the response of locomotor activity and vertical migratory behaviour to tidal fluctuations, from littoral fishes to those in open water on the continental shelf, what characteristics of the tide may have the potential to act as zeitgebers? In littoral fishes, suppression by the tide of activities such as foraging, respiration and movement, may have an important influence on entraining circatidal rhythmicity (Boehlert and Mundy, 1988). Cyclic changes in hydrostatic pressure may also play an important role, as indicated by Gibson (197 1) for Blennius phofis,by Gibson et al. (1978) for post-larval plaice, and by Green (1971) for Ofigocottus maculosus. However, none of these fishes have swimbladders, and therefore the mechanism of depth perception is unclear. Green (1971) and Gibson et al. (1978) both suggest that potential piezoelectric properties of otoliths reported by Morris and Kittleman (1 967) may allow pressure sensitivity. Other properties which undergo cyclical changes induced by the tide include light, turbidity, turbulence, temperature and salinity. However, Green (1971) suggests that temperature, salinity and turbulence would be unreliable as zeitgebers due to the variable influence of sea and weather conditions. He found that Oligocottus maculosus collected from low and high tidepools exhibited the same circatidal rhythm, despite differences of up to 4 h between times of flooding in the natural environment. The same arguments can be made for light and turbidity. Boehlert and Mundy (1988) comment that since tides differ in magnitude and phase over the range of most species, any zeitgeber must be associated with the local conditions. They further note that some species, such as English sole (Purophrys vetulus), may use different cues in different areas, since some estuarine nursery areas for English sole have freshwater input, whereas others have almost none. They suggest that a suite of physical factors associated with the tide may serve as the zeitgeber. Such problems of differences in timing of the tide between low and high areas of the littoral zone do not apply to estuarine and open water areas. Even in these areas the factors associated with cyclical tidal changes that may act as zeitgebers to induce tidal vertical migration, have not been clearly identified. Such factors include olfactory cues of from-land or from-sea currents, detection of flow direction or reversals, variations in turbulence or turbidity (Laprise and Dodson, 1989), and induction of electric fields in the water (McCleave and Kleckner, 1982). The mechanisms synchronizing

DIEL VERTICAL MIGRATIONS

143

vertical migration with the tides in selective tidal stream transport of adults in the North Sea are also not clear (Arnold and Cook, 1984). Fortier and Leggett (1983) concluded that tidal vertical migrations of herring larvae in the St. Lawrence estuary were related to fluctuations in the depth occupied by their invertebrate prey. Vertical movements of many invertebrate species are known to be associated with tidal fluctuations (e.g. Palmer, 1973; Cronin and Forward, 1979). However, in their study of tidal vertical migration by smelt larvae in the well-mixed region of the St. Lawrence estuary, Laprise and Dodson (1989) found no association of the vertical migrations of larvae with either their food or predators. In our study of vertical migrations of juvenile gadids on the tidally energetic Georges Bank (Perry and Neilson, 1988), we found an apparent inverse relationship at a well-mixed site between the number of haddock caught and the mean tidal current speed, with fish leaving the bottom during periods of comparatively slack water. We suggested that migration was in phase with that of the preferred prey Neomysis americana, which also moved up in the water column during periods with slower currents. Blaxter (1978) has raised the question of the value to fish of an endogenous rhythm, in particular when that rhythm is sufficiently regular and predictable as, for example, light or the tide. Gibson and Hesthagen (1981) have suggested that circatidal rhythms, especially when peak activity is phased with the ebb tide, prepare the animals for the most favourable tidal phase to prevent stranding on the beach. Boehlert and Mundy (1988) noted that if a single cue is used as a stimulus for pelagic behaviour, then there is no mechanism for the fish to detect the end of the flood tide, resulting in considerable downstream transport until a change in that cue was detected. Therefore, entrainment of a circatidal rhythm to that cue would allow the fish to settle from the water column without the necessary change of zeitgeber.

5. Predation avoidance

Type-I1 vertical migration responses are frequently linked to predation avoidance by copepods (see, for example, Ohman et al., 1983; Gliwicz, 1986). Such responses can also be found in marine teleost populations. For example, Yamashita et al. (1985) attributed the type-I1 vertical migration response of Japanese sand-eel larvae to a mechanism for reducing predation. These larvae cease swimming at night and sink. As noted by Yamashita et al. (1989, such inactivity may reduce attacks by piscivorous predators such as carnivorous copepods and chaetognaths that detect their prey through vibrations and are known to attack mainly at night (Lillelund and Lasker,

144

J. D. NEILSON AND R. I. PERRY

1971; Bailey and Yen, 1983). The larvae may also lessen the risk of predation by leaving the comparatively shallow depth strata that carnivorous animals frequent at night in the course of their type-I migration. Bailey (1975) also postulated that the presence of the carnivorous spurdog (Squulus ucunthius) at all times in well-lit surface waters may cause 0-group North Sea gadoids to move down at dawn to avoid predation. In a study of northern anchovy larvae (Engruulis mordux), Hunter and Sanchez (1976) described a nightly pattern of swimbladder inflation followed by a slow sinking of the larvae. Similar to the conclusions of Yamashita et ul. (1985), they suggested that the reduced activity produced by slower sinking speeds could reduce detection by predators that detect the movements or turbulence of their prey. Hunter and Sanchez (1976) also noted that the vertical migration of the larvae could result in exposure to different and possibly less hazardous predators at night. The diel vertical migration of juvenile sockeye salmon (Oncorhynchus nerku) in freshwater lakes has been much studied, and is typically type-I. The dawn and dusk distribution is related to feeding as the fish are visual predators, but, during the day, the sockeye occur far below regions of concentrated zooplankton abundance. In developing a model describing the diel vertical migration of the sockeye, Clark and Levy (1988) state that, through maintaining a crepuscular feeding pattern, this species is optimizing the trade-off between food intake and the risk of predation that occurs near the surface. Indeed, an interesting prediction of the model is that in order to avoid predation, the sockeye salmon are adopting vertical migration and feeding patterns that result in less than the maximum possible growth rate, an observation that was borne out by field data for several populations summarized by Brett (1983). On the other hand, some vertical migratory behaviour is counter-intuitive with regard to predator avoidance. Bailey (1989) noted that the vertical migration of juvenile walleye pollock puts them at risk of predation by conspecifics when they follow a type-I vertical migration into deeper strata occupied by adults during daylight hours. In such instances, Bailey (1 989) noted the benefits of migration must then outweigh the risks associated with cannibalism. While it is clear that, under certain circumstances, avoidance of predation may influence or directly cause vertical migration, we conclude that the risk cannot be considered a zeitgeber, because the action of predators - like that of prey availability described in the previous section - is unlikely to be invariant. Through mechanisms such as satiation, density-dependence and seasonal changes in relative abundance, it seems probable that the risk of predation will vary considerably, rendering it an unlikely candidate for a zeitgeber.

DIEL VERTICAL MIGRATIONS

145

6. Effect of commensal species If gelatinous zooplankton undertake regular die1 vertical migrations, obligate commensal species must follow. Extensive examples of commensal relationships between marine fish and medusae were documented by Mansueti (1963), and subsequently for gadids by Bailey (1975), Koeller et al. (1986) and Bailey (1 989). There are reports that at least some gelatinous zooplankton undergo a die1 vertical migration. For example, medusae of Aurelia sp. follow a die1 vertical migration, apparently closely linked to the extent of underwater illumination (Yasuda, 1973). Mansueti (1963) notes that the European whiting (Gadus merlangus) and Trachurus sp. have commensal relationships with Aurelia. Another common gelatinous zooplankter is Pelagia noctiluca, which is sometimes found at the surface in substantial aggregations (J. Purcell, Horn Point Environmental Laboratory, Maryland, pers. comm.). According to Franqueville (197 l), Pelagia noctiluca exhibits a type-I vertical migration. Fish reported by Mansueti (1963) as commensal with P. noctiluca include Zcichthys lockingtoni (medusafish), Trachurus symmetricus (Pacific jack mackerel) and Psenes sp. (driftfish). Hamner and Schneider (1 986) report that Chrysaora melanaster also follow a type-I migration. Congeneric forms of C. melanaster are known to be commensal for many different fish species. Unfortunately, no vertical depth distribution data appear to be available for Cyanea sp., one of the most commonly noted commensals of marine fish (Mansueti, 1963). If gelatinous zooplankton follow an endogenous rhythm during their vertical migrations, such a rhythm might also be falsely attributed to closely associated fish. Mansueti (1963) notes that the association between fishes and gelatinous zooplankton is an ephemeral ecological phenomenon resulting from a series of extrinsic chance factors in which jellyfishes are essentially passive hosts and the fish are active opportunists. Rather than promote a regularity of vertical migration, Colton and Temple (1961) found that the co-occurrence of various gadoids and the near surface-dwelling Cyanea sp., limited the vertical distributions of these fishes to comparatively shallow water. Thus vertical migrations can be affected by the presence of another factor, contrary to expectation, if patterns of vertical distribution of fishes were governed by an endogenous rhythm.

B. Do Varations in Local Environment Afect Vertical Position? Vertical migration patterns of fishes can be modified by conditions in the local environment. Temperature is the most obvious, particularly when

146

J. D. NEILSON A N D R. 1. PERRY

sharp vertical gradients are present. Ofori-Adu (1978) showed that sardines and other small pelagic fishes occurred in the upper layers during periods when the thermocline was weak and close to the surface. Presumably, such vertical distributions were limited during periods with well-defined thermoclines. On Georges Bank, Miller et al. (1963) found that while larval haddock could occur throughout the upper 50m of the water column, over 80% of the larvae occurred within the depth range of the thermocline. Colton (1965) found that juvenile haddock on Georges Bank also remained near the thermocline, with slight day-night changes in mean depth. In an investigation of summer ichthyoplankton in the mid-Atlantic Bight, Kendall and Naplin (1981) found the larvae of some fish species remained above the thermocline (notably bluefish Pomatomus saltatrix and Gulfstream flounder Citharichthys arctifrons), while larvae of other species remained within or below the thermocline. They noted, however, that at least some larvae of the latter group did migrate across the thermocline at night, experiencing at times a change from 18 to 22 “C. The effect of an inverse thermocline (temperature increasing with depth) on limiting the evening ascent of a deep scattering layer was noted by Davies (1976) in a subarctic Pacific region. He found the ascent of the scattering layer was arrested at the thermocline separating the cold surface layer from the warmer deep layer. While the composition of the scattering layer was not positively identified, it was believed to be similar to acoustic “signatures” of fish schools in coastal waters. In assessing the impact of local temperature conditions on modifying vertical migration patterns, the effect of temperature acclimation must be considered. The influence of thermal history on the distributions of fishes with regard to temperature has been well established (e.g. Fry, 1971), including its effect on the distribution of juvenile gadids within a temperature gradient (Tat’yankin, 1974). The ability of a temperature gradient of a certain magnitude to inhibit vertical migration in one location compared with another may depend in part on the temperature to which fish in each area are acclimated. Furthermore, the ability of fishes to tolerate different temperatures can vary seasonally, depending on the seasonal behaviour patterns. This was demonstrated by Olla et al. (1985) for bluefish, which became distributed in progressively colder water from summer to fall, coincident with their southerly fall migration. Similar to seasonal effects, ontogenetic changes can cause differences in the response of fish to local conditions, in particular temperature. Hamrin (1986) examined the die1 vertical distributions of juvenile and adult vendace (Coregonus albula) in thermally stratified lakes. He found that adults occurred in deep cold water during the day, migrating upwards at night but were limited by the warm (18 “C) epilimnion. In contrast, juvenile vendace in

DIEL VERTICAL MIGRATIONS

147

the same lake occurred at warmer mid-depths by day, and were able to undergo a 10°C change at night as they migrated into the epilimnion. The ontogenetic transitions of the red hake (Urophycis chuss: Steiner and Olla, 1985), walleye pollock (Bailey, 1989) and juvenile rockfish (Sebastes diploproa: Boehlert, 1977) from pelagic to demersal habitats also occur when a thermocline is present. This transition may require acclimation to the deeper, colder water. Such acclimation may involve gradual and periodic forays below the thermocline, with fish returning to warmer waters to recover, giving the appearance of periodic vertical migrations of increasing amplitude. Vertical gradients of temperature, however, often represent gradients of other physical and biological properties, which may influence the local pattern of vertical migration by fishes. Dissolved oxygen concentrations can be a limiting factor for some fishes, particularly for fast swimmers such as the tunas (Sund et al., 1981) in areas where oxygen concentrations are low (such as the eastern tropical Pacific). Barkley et al. (1978) have modelled the relatively narrow depth range in the tropical Pacific Ocean suitable for skipjack tuna given an upper temperature limit and a deep oxygen minimum. Suthers and Gee (1986) have also noted the effect of low dissolved oxygen concentrations on juvenile yellow perch (Perca javescens) distributions in a prairie marsh. Rudstam and Magnuson (1985) modelled the restriction of vertical distributions of yellow perch and cisco (Coregonus artedii) due to temperature and oxygen gradients in Wisconsin lakes. They noted that fish distributions were often correlated with gradients of these properties, but that variation in food ration between oligotrophic and eutrophic lakes produced differences between observed vertical distributions and those predicted by the model. Olla et al. (1985) found that food presented at normally debilitating temperatures caused juvenile bluefish to enter these temperatures and feed successfully. Such starvation-induced changes in behaviour towards preferred temperatures were also noted by Javaid and Anderson (1967) for salmonids. The combined influence of temperature structure and food availability at the thermocline was also noted by Southward and Barrett (1983) for juveniles of a flatfish species off Plymouth, and by Perry and Neilson (1988) for juvenile gadids on Georges Bank. The results of the latter study contrast somewhat with those of Colton (1965), discussed earlier. We found that at a thermally stratified site, haddock were distributed predominantly at the thermocline (Fig. 12), coincident with the largest zooplankton biomass. The importance of prey in modifying the vertical distributions of young fish should not be underestimated, considering the comment by Olla ef al. (1985) that food acquisition is particularly critical for rapidly growing juveniles. Shanks (1 983) has suggested that slicks associated with tidally forced

148

J. D. NEILSON A N D R. 1. PERRY

internal waves may concentrate zooplankton and fish, transporting them shoreward and potentially vertically as well. Such vertical accumulation and transport has been noted for zooplankton (Haury et al., 1983). In the littoral zone, Green (1971) found that Oligocortus maculosus adults show weakly persistent circatidal rhythms in the laboratory, with activity periods phased to the time of high tide. In the field, however, activity was also regulated by temperature, turbulence and light intensity, such that under certain combinations of these factors the fish may be totally inactive at high tide. Changes in local conditions induced by the tidal cycle can be extensive (these impacts are discussed in the earlier section on tides).

FIG. 12. Mean depths (shown as a dashed horizontal line, determined from depth-stratified sampling at three depths) of 0-group haddock (Melunogrurnmusueglefinus), shown with respect to the thermocline on Georges Bank in 1985. Data represent successive sampling over 24 h.

We have presented several studies of the ability of local conditions to modify die1 vertical behaviour pattern of fishes, suggesting such patterns are not under strong endogenous control. In contrast, there are relatively few studies that show no influence of local conditions on vertical migrations. Dalpadado and Gjssaeter (1987) indicated that mesopelagic fish caught in the Red Sea showed a type-I pattern of vertical migration similar to that observed in other areas, despite the unusual hydrographic features including extremely high salinities. On the Scotian Shelf, vertical migrations of 0-group silver hake were unaffected by sharp thermal gradients of 10C" (Koeller et al., 1986). C. Do Ontogenetic Variations in Vertical Migratory Behaviour Occur, and are they Consistent with Chronobiological Theory? It has long been known that different life-history phases of teleosts show

DIEL VERTICAL MIGRATIONS

149

varying patterns of endogenous behaviour (see, for example, Johnson and Muller, 1978). Willis and Pearcy (1982) have also reported that among mesopelagic fish caught off Oregon, the extent of die1 migration was often a size-related phenomenon, with the range of migration positively correlated with fish size. Other recent works have shown that even among animals of the same cohort, differences in patterns of vertical distribution can be expected (Fortier and Leggett, 1983). Grave (1981) found that with increasing size, mackerel larvae (Scomber scombrus) and early juveniles carry out more distinct die1 vertical migrations. Yamashita et al. (1985) showed that sand-eel larvae (Arnmodytes personatus) less than 5 mm standard length did not demonstrate die1 vertical migrations, but type-I1 migratory behaviour became evident once larvae reached 5-6mm in length. The authors attributed the change to the transition from endogenous to exogenous feeding. The exogenous feeding occurred mostly during daylight hours and at night the larvae ceased swimming and sank. The authors speculated that such cessation of activity may be an effective anti-predator strategy, as carnivorous copepods and chaetognaths which forage at night rely on the detection of pressure changes associated with the movement of their prey. Perry and Neilson (1988) reported that smaller members of the cohorts of pelagic 0-group cod and haddock were distributed at shallower depths and undertook less extensive migrations than did larger fish. After completion of their pelagic existence, 0-group Atlantic cod adopt a demersal life-style. Once on the bottom, smaller fish undertake less extensive migrations than do larger ones (Pearcy et al., 1979; Brunel, 1965, 1972). In contrast, haddock around Sable Island, Nova Scotia, showed well-defined die1 vertical migration as 0-group fish, but not as I-year-olds (Scott, 1984). In the eastern Bering Sea, smaller members of a cohort of 0-group walleye pollock undertook less extensive vertical migrations than did the larger fish (Bailey, 1989). Brewer and Kleppel( 1986) demonstrated that large (> 12 mm) northern anchovy larvae undertook a type-I migration to the surface, whereas smaller fish did not, a result consistent with those shown by Hunter and Sanchez (1976). 0-Group Anguilla sp. larvae also undergo more distinct vertical migrations as they increase in size (Schoth and Tesch, 1984), as do herring larvae (Johannessen, 1986). Similarly, the time-series analyses of Laprise and Dodson ( 1989) indicated that the cross-correlations of depth occupied by the larvae with tidal current became more precise as smelt larvae grew. The observation supported the view that through vertical migrations of increasing amplitude, the older larvae used tidal currents more efficiently to maintain their horizontal position. Potter and Lough (1987) followed the development of vertical migration behaviour of sand lance (Ammodytes sp.) and noted that whereas recently hatched larvae were found throughout the water column on Nantucket Shoals and Georges Bank, later stages were

150

J. D. NEILSON A N D R. I. PERRY

caught mostly at night and their abundance increased with depth. Kryuchukov et al. (1984) have shown age-related changes in the diel rhythm of locomotor activity of the mirror carp (Cyprinus carpio), and found that periodicity in locomotor activity was absent until after the fish were 2 months old. Such changes probably reflect the progressive recruitment of sensory cells and locomotory capability as teleost larvae develop (Blaxter, 1986). Our analysis of ontogenetic effects on vertical migration indicates that studies demonstrating age-related effects on diel vertical migration are numerous. The typical pattern of change with development includes a greater amplitude of vertical movement as ichthyoplankton gain greater locomotory and sensory capability through the first year of life, as mentioned above. A notable exception was the study of Boehlert et al. (1985), who described the vertical distribution of larvae of 33 species of fish off the Oregon coast in spring and summer. They found no trends regarding changes in depth distribution with increasing size of the larvae. Ontogenetic effects observed in marine fish are consistent with chronobiological theory. Saunders (1977) notes that most vertebrate species appear to be arrhythmic as embryos, with periodicity becoming evident soon after birth or hatching. Davis (1981) further indicated that the amplitude of endogenous rhythms sometimes tends to increase over time, a phenomenon known as “maturation” of the rhythm. An example of such an effect is the work of Gibson et al. (1978), who showed an instance where the effect of zeitgeber disruption decreased with age. Similarly, Campana (1984) showed that the influence of light as a zeitgeber on larval plainfin midshipmen (Porichthys notatus), decreased as fish aged. In a practical sense, ontogenetic effects give further scope for variation in diel vertical migrations, particularly given that inter-annual differences in length and age are to be expected among members of cohorts for many stocks.

111. Discussion Eriksson (1978) noted that the number of studies describing circadian rhythms have increased rapidly since the Cold Spring Harbor Symposium in 1960. That has also been the case for fish. However, the evidence for endogenous rhythms in fish is not as clear as that obtained for other organisms (see, Eriksson, 1978). The apparent plasticity of endogenous behaviour in fish is indicated in Eriksson (1978, table l), where reports are listed of dual phasing (noctural vs diurnal) of diel patterns of activity. In some laboratory-based studies reviewed by Eriksson, opposite patterns were noted for individual fish, depending on the time of the experiment. Whether

DIEL VERTICAL MIGRATIONS

151

such differences would be manifested in varying patterns of die1 migration was not examined, but it demonstrates the apparent latency of the phenomenon. Table 1 of this chapter builds upon the list provided by Eriksson (1978), with those studies shown in bold face type having counterparts where vertical migration behaviour of opposite phase was observed for the same species. Eriksson hypothesized that such biphasic capacity gives fish flexibility in fulfilling their actual ecological needs in a semi-opportunistic way. Muller (1978) also noted a high degree of flexibility of the circadian system of temperate fish to environmental conditions. McFarland (1986) suggests that the extent of flexibility in circadian systems of teleosts is positively correlated with latitude. At higher latitudes, the flexibility which Muller (1978) and Eriksson (1978) spoke of would be advantageous, as it would provide protection from deviant behaviours resulting from free-running endogenous rhythms that became uncoupled to photopheriod as a zeitgeber, e.g. as the seasons progressed. McFarland ( 1986) gives examples of such behaviour, including a diurnally active fish remaining in the water column too long, or a nocturnally active fish beginning its evening ascent into the water column too soon. On the other hand, less flexible rhythmic behaviour might be expected in the tropics, where photoperiod varies to a lesser extent (as do other potential zeitgebers). This hypothesis seems to be of general significance to our understanding of fish behaviour and warrants further evaluation. Further evidence for the latency of teleosts’ response to cyclic cues is provided by an examination of the herring literature. As noted by Seliverstov (1974), Lough (1975) and Graham and Sampson (1981), who made day-night comparisons of catches as evidence of die1 migration, large catches can occur either by day or night. Sjoblom and Parmanne (1978) found that larval herring of the Gulf of Finland occurred in relatively deep water by day and closer to the surface at night during early summer, but that this behaviour was reversed later in the summer. Such changes were not explicable by the amount of light present or by the size of the larvae. Stephenson and Power (1988) examined the vertical distribution of larval herring in the tidally active Bay of Fundy, and found that their movements were also not explicable in terms of reaction to photoperiod but appeared to have a semi-diel component linked to an undetermined factor. In a subsequent study (Stephenson and Power, in press), they returned to the same site 2 years later and found no evidence of a semi-diel component. These authors concluded that inter-annual differences in vertical migration behaviour exist, even at a given location. Bailey (1 975) found that the tendency to undertake vertical migrations appeared labile in 0-group haddock in the North Sea, depending on some factor which varied from area to area. In the freshwater environment of Lake Sibaya, South Africa, Bowen and

152

J. D. NEILSON AND R. I. PERRY

Allanson (1982) found that juvenile Tilapia mossambica moved daily from deep offshore waters to shallow littoral areas where they fed for several hours before returning to deep water. When the littoral zone was free of debris, juvenile T. mossambica occurred in the littoral zone only during daylight hours. Following elevation of the lake level and resulting accumulation of debris providing cover for ambush predators, T. mossambica were abundant in the littoral zone only at night, when the risk of predation was lessened. A further example from freshwater populations is provided by Levy (1987), who found it necessary to invoke a multi-factor hypothesis (light, food intake and predation avoidance) to explain die1 changes in the depth occupied by populations of juvenile sockeye salmon (Oncorhynchus nerka) in their rearing lakes. Such lability may also account for the virtual absence of attempts to model the die1 vertical distribution of fish populations, which, to our knowledge, has resulted in only one published work for marine populations (Mel’nikov et al., 1981), and another for an anadromous species, sockeye salmon during the lake-rearing phase of their lifehistory (Clark and Levy, 1988). We believe that recent advances in chronobiological theory can be invoked to explain the plasticity in fishes’ vertical migration behaviour. Investigations on fish activity rhythms (Eriksson, 1978; Kavaliers, 1978) have given support to the idea that an underlying multi-oscillatory system is responsible for the flexibility of circadian systems under natural conditions as well as the commonly occurring arrhythmic behaviour of fish held in constant environmental conditions. For example, Wales (1975) showed that light-dependent behaviour in larvae of herring and plaice (Pfeuronectes platessa) could be divided into retinally evoked phototaxes and an extraretinally evoked kinesis of unknown origin, although the photosensitive pineal gland has been suggested as having a role in the circadian organization in teleosts (Kavaliers, 1979b). Hence, at least two photosenses are likely to be involved. Douglas (1982) reached a similar conclusion, noting that control of photomechanical changes in the retina of rainbow trout (Safmogairdneri) has both an endogenous component, which maintains a crepuscular rhythm of light adaptation, and a direct effect of light, that maintains light adaptation throughout a normal day. A growing body of literature suggests that multiple oscillators temporally integrate animals with their environments (see, for example, Winfree, 1980; Moore-Ede et af., 1982). Such systems have also been described for teleosts held under laboratory conditions. For example, Spieler and Noeske (1984) have described the interaction of photoperiod and feeding schedule as zeitgebers. They found that feeding schedule, when it provides a consistent cue, overrides the effect of photoperiod to entrain cyclic activity in goldfish. Batty (1987) held herring larvae in the laboratory, and also showed that prey

DIEL VERTICAL MIGRATIONS

153

availability as well as photoperiod could influence diel cycles of activity. In an earlier work, Pearre (1978) apparently recognized the potential of multiple oscillators driving vertical migrations when he proposed that a model based on hunger as a mediator and light as a synchronizer could explain many instances of apparently anomalous die1 migrations, and relate diel to longer-term migratory behaviour. Boehlert and Mundy (1988) have suggested that a suite of physical factors influences endogenous rhythms (in their case, a circatidal rhythm of utilization/avoidance of transport by currents) rather than a single zeitgeber, an argument consistent with the views of Gibson (1973) mentioned previously. Boehlert and Mundy further claim that such a multi-oscillatory system is likely, given the lack of conclusive experimental identification of a single zeitgeber for tidal rhythms (Gibson er al., 1978; Cronin and Forward, 1979, 1983). That view finds further support from the work of Kingsford (1988), who notes that general models of larval fish behaviour are usually unrealistic, given the vast array of physical and biological processes that might modify them. To illustrate the effect of a hypothetical multi-oscillatory system on the die1 vertical distribution of fish, one could visualize an array of zeitgebers such as shown in Fig. 13. Under certain circumstances, perhaps with a predictable supply of food provided at certain points in the tidal cycle, such as shown earlier for 0-group haddock on Georges Bank, a fish would switch from being entrained by a light-dark cycle to entrainment by the tidal cycle, with a resulting period (T) equivalent to the duration of the tidal cycle, not 24 h. As shown earlier and summarized in Fig. 13, modifications to the “parent” rhythm could also occur through the effects of thermoclines, meteorological events and other cyclic stimuli. A further example is the die1 vertical migrations of larval and juvenile soft sculpin (Gilbertidia sigalutes) reported by Marliave (1981). In that instance, larvae entrain to photoperiod (crepuscular, then nocturnal). During the noctural period, larvae entrain secondarily to food density (J. B. Marliave, Vancouver Public Aquarium, Vancouver, B.C., pers. comm.) The multi-oscillatory hypothesis could also account for the findings of Zusser (1966) who showed that in adult horse mackerel, the usual type-I vertical migration, thought to be driven by changes in light intensity, could be suppressed when prey were plentiful. Similarly, the multi-oscillatory hypothesis could account for the complex vertical movements exhibited by estuarine fish (see, for example, Weinstein et al., 1980; Fig. 11) which respond to both light and tide. These field studies are consistent with the laboratory findings of Spieler and Noeske (1984), stated previously. Using the terminology of Pittendrigh (1981), the slave oscillator (tide or food availability) is subordinate to the pacemaker entrained by photoperiod, but can override the effect of the pacemaker under certain circumstances.

154

J. D. NEILSON AND R. I. PERRY

Potential Environmental Modifiers

Resultant Rhythms

Thermocllnea

+ -o " c

0

of Significance 7 .$>>::

NO

.................. .......... >: ....... ........

~ . a

' P a r e n t ' Rhythm Entralned by Photoperlod

e 100%

Full Moon

V,

->

.a. Cloudless Night,

Meteorological E v e n t a

r=z4h

t

Tldea of Slgnlflcance 7

I-

; Others 7

FIG. 13. Diagrammatic figure showing a subset of the available :ei/gebers and other environmental factors in the marine environment, and how they influence die1 vertical migrations of fish.

We conclude that die1 vertical migrations are often a facultative process that can be significantly influenced by local conditions, including variations in hydrography. This fact may mask any easily detectable and consistent expression of endogenous rhythmicity. But what are the implications for fisheries-related investigations? With regard to early life-history stages, Perry and Neilson (1 988) noted that adequate prediction of pelagic-dwelling juvenile haddock and cod vertical distributions for survey purposes required a knowledge of the thermal structure of the water column, distribution of prey and tidal current speed. However, at what point do the extensive data requirements for conducting a useful and reliable survey of abundance of pelagic early life-history stages dictate that investigators reconsider their utility? We suggest that, for many species, the sources of variation affecting patterns of die1 vertical migration are numerous and their effect so profound as to obviate the utility of midwater surveys of abundance in a quantitative sense. As an alternative. investigators could consider studying different life-

DIEL VERTICAL MIGRATIONS

155

history stages. For example, Mahon and Neilson (1987) examined the gut contents of 0-group haddock on Georges Bank and found that die1 migration was not an important phase of the ecology of fish that had adopted a demersal life-style. By concentrating on such stages of the life-history, workers could create more tractable problems, inasmuch as abundance and its variation would be estimated in a two-dimensional rather than a threedimensional field. However, problems could arise with that approach if inter-annual variation in the timing of ontogenetic migration to the bottom were expected. A further option would be to consider sampling designs that include most of the water column, such as oblique or stepped-oblique hauls. Even in those instances, variation in patterns of vertical distribution could cause problems if a significant proportion of the population was, at times, near the sea bed beyond the range of the sampling gear. Changing the focus of a study to a different life-history stage has the added complication of requiring different sampling gear, given that larger stages are often more adept at evading capture (Brander and Thompson, 1989; Potter et al., I 990). For adult populations, the above-noted sources of variation in patterns of vertical migration for early life-history stages would also apply to pelagic teleosts such as herring. However, with the exception of the studies by Brunel (1965) and Beamish (l966), comparatively few studies have addressed variability in patterns of vertical migration of adult groundfish. The lack of information may reflect the assumption that once groundfish adopt a demersal existence, their excursions off bottom are infrequent and not a potential source of error in surveys of abundance. Given the study by Walsh (l988), who showed vertical migrations of juvenile and adult yellowtail flounder caused a serious source of bias in estimation of stock abundance and indices of recruitment, that assumption might warrant closer scrutiny for other species formerly believed to undertake inconsequential vertical migrations.

IV. Acknowledgements We thank Darlene Warren and Marilynn Rudi for assistance in obtaining some of the less readily available literature cited here. W. McMullon and F. Cunningham provided the figures from work originating from us, while F. G. Carey, R. N. Gibson and M. P. Weinstein graciously permitted the reproduction of their figures here. Jennifer E. Purcell kindly provided many of the references on interactions of gelatinous zooplankton and marine fish. We would also like to thank J. H. S. Blaxter, T. D. Iles, R. G. Halliday, A. J. Southward, R. W. Rangely and several anonymous referees for their constructive reviews of an earlier version of this manuscript.

156

J. D. NEILSON AND R. I. PERRY

V. References Ahlstrom, E. H. (1959). Vertical distribution of pelagic fish eggs and larvae off California and Baja California. Fishery Bulletin, United States 60, 107-146. Alexander, R. M. (1972). The energetics of vertical migration by fishes. Symposia of the Society for Experimental Biology 26, 273-294. Angel, M. V., Hargreaves, P., Kirkpatrick, P. and Domanski, P. (1982). Low variability in planktonic and micronektonic populations at 1,000 m depth in the vicinity of 42”N, 17”W: Evidence against diel migratory behaviour in the majority of species. Biological Oceanography 1, 287-3 19. Anon. (1981). Workshop report. In “Bottom Trawl Surveys” (W. G. Doubleday and D. Rivard, eds), Vol. 58, pp. 7-8. Canadian Special Publication of Fisheries and Aquatic Sciences. Department of Fisheries and Oceans, Ottawa. Arimoto, T., Inoue, Y. and Ogura, M. (1983). Die1 variation of catch in coastal setline. Bulletin of the Japanese Society of ScientiJic Fisheries 49, 1175-1 181. Arnold, G. P. and Cook, P. H. (1984). Fish migration by selective tidal stream transport: First results with a computer simulation model for the European continental shelf. In “Mechanisms of Migration in Fishes” (J. S. McCleave, G. P. Arnold, J. J. Dodson and W. H. Neill, eds), pp. 117-261. North Atlantic Treaty Organization Conference Series No. 4, Marine Science. Plenum Press, New York. Aschoff, J. (1965). The phase-angle difference in circadian periodicity. In “Circadian Clocks” (J. Aschoff, ed.), pp. 262-278. North Holland Press, Amsterdam. Aschoff, J. (1981). Freerunning and entrained circadian rhythms. In “Handbook of Behavioural Neurobiology” (J. Aschoff, ed.), pp. 81-93. Plenum Press, New York. Atkinson, D. B. (1989). Die1 movements of beaked redfish and the implications of these for stratified random bottom trawl estimates of biomass and abundance. North American Journal of Fisheries Management 9, 163-1 70. Bailey, K. M. (1989). Interaction between the vertical distribution of juvenile walleye pollock Theragra rhalcogramma in the eastern Bering Sea, and cannibalism. Marine Ecology Progress Series 53, 205-213. Bailey, K. M. and Yen, J. (1983). Predation by a carnivorous marine copepod, Euchaeta elongata Esterly, on eggs and larvae of the Pacific hake, Merluccius productus. Journal of Plankton Research 5, 71-82. Bailey, R. S. (1985). Observations on diel behaviour patterns of North Sea gadoids in the pelagic phase. Journal of the Marine Biological Association of the United Kingdom 55, 133-142. Bainbridge, R. (1961). Migrations. In “Physiology of the Crustacea” (T. H. Waterman, ed.,), Vol. 2, pp. 431-463. Academic Press, London. Barkley, R. A., Neill, W. H. and Gooding, R. M. (1978). Skipjack tuna, Karsuwonus pelamis, habitat based on temperature and oxygen requirements. Fishery Bulletin, United States 76, 653-662. Batty, R. S. (1987). Effect of light intensity on activity and food-searching of larval herring, CIupea harengus: A laboratory study. Marine Biology 94, 323-327. Beamish, F. W. H. (1966). Vertical migration by demersal fish in the Northwest Atlantic. Journal of the Fisheries Research Board of Canada 23, 109-139. Begg, G. W. (1976). The relationship between the diurnal movements of some of the zooplankton and the sardine Limnothrissa miodon in Lake Kariba, Rhodesia. Limnology and Oceanography 21, 529-539.

DIEL VERTICAL MIGRATIONS

157

Beltestad, A. K., Nakken, 0. and Smedstad, 0. M. (1975). Investigations on die1 vertical migration of 0-group fish in the Barents Sea. Fiskeridirektoratets Skrifter Serie Havundersmkelser 16, 229-244. Bergmann, H. (1981). Distribution and abundance of larval fishes in Lake Sangchris (1976). Illinois Natural History Survey Bulletin 32, 61 5-630. Bervan, K. A. (1982). The genetic basis of latitudinal variation in the wood frog Rana sylvatica. I. An experimental analysis of life history traits. Evolution 36, 962-983. Biette, R. M. and Geen, G. H. (1980). Growth of underyearling sockeye salmon (Oncorhynchus nerka) under constant and cyclic temperatures in relation to live zooplankton ration size. Canadian Journal of Fisheries and Aquatic Sciences 37, 203-2 10. Blaber, S. J. M . and Bulman, C. M. (1987). Diets of fishes of the upper continental slope of eastern Tasmania: Content, calorific values, dietary overlap and trophic relationships. Marine Biology 95, 345-356. Blaxter, J. H. S. ( I 975). The role of light in the vertical migration of fish - a review. In “Light as an Ecological Factor 11” (G. C. Evans, R. Bainbridge and 0. Rackham, eds), pp. 189-210. Blackwell Scientific, Oxford. Blaxter, J. H. S. (1978). Summary of symposium on “Rhythmic Activity of Fishes”. In “Rhythmic Activity of Fishes” (J. E. Thorpe, ed.), pp. 285-288. Academic Press, London. Blaxter, J. H. S. (1986). Development of sense organs and behaviour of teleost larvae with special reference to feeding and predator avoidance. Transactions of the American Fisheries Society 115, 98-1 14. Boehlert, G . W. (1977). Timing of the surface-to-benthic migration in juvenile rockfish, Sebastes diploproa, off Southern California. Fishery Bulletin. United States 75, 887-890. Boehlert, G . W. and Mundy, B. C. (1988). Roles of behavioral and physical factors in larval and juvenile fish recruitment to estuarine nursery areas. American Fisheries Society Symposium 3, 51-67. Boehlert, G. W., Gadomaski, D. M. and Mundy, B. C. (1985). Vertical distribution of ichthyoplankton off the Oregon coast in spring and summer months. Fishery Bulletin, United States 83, 61 1-622. Bohl, E. (1980). Die1 pattern of pelagic distribution and feeding in planktivorous fish. Oecologia 44, 368-375. Bowen, S. H. and Allanson, B. R. (1982). Behavioral and trophic plasticity of juvenile Tilupia mossambica in utilization of the unstable littoral habitat. Environmental Biology of Fishes 7, 357-362. Brander, K . and Thompson, A. B. (1989). . , Die1 differences in avoidance of three vertical profile sampling gear by herring larvae. Journal qf Plankton Research 11, 775-784. Brawn, V. M. (1960). Seasonal and diurnal vertical distribution of herring (Clupea harengus L.) in Passamaquoddy Bay, N.B. Journal of the Fisheries Research Board of Canada 17, 699-7 1 1. Breder, C. M. (1959). Studies on social groupings in fish. Bulletin of the American Museum of Natural History 117, 397481. Brett, J. R. (1971). Energetic responses of salmon to temperature. A study of some thermal relations in the physiology and freshwater ecology of sockeye salmon (Oncorhynchus nerka). American Zoologist 11, 99-1 13. Brett, J. R. (1983). Life energetics of sockeye salmon, Oncorhynchus nerka. In “Behavioral Energetics: The Cost of Survival in Vertebrates” (W. P. Aspey and S. I. Lustick, eds), pp. 29-63. Ohio State University Press, Columbus.

158

J. D. NEILSON AND R. 1. PERRY

Brewer, G . D. and Kleppel, G . S. (1986). Die1 vertical distribution of fish larvae and their prey in nearshore waters of southern California. Marine Ecology Progress Series 27, 2 1 7-226. Browman, H. I. and Marcotte, B. M. (1986). Diurnal feeding and prey size selection in Atlantic salmon, Salmo salar, alevins. In “Contemporary Studies on Fish Feeding” (C. A. Simenstad and G . M. Cailliet, eds), pp. 269-284. Junk, Dordrecht. Brown, F. A., Jr, Hastings, J. W. and Palmer, J. D. (1970). “The Biological Clock: Two Views”. Academic Press, London, 93 pp. Brunel, P. (1965). Food as a factor or indicator of vertical migrations of cod in the western Gulf of St. Lawrence. International Commission of the Northwest Atlantic Fisheries Special Publication 6, 4 3 9 4 8 . Brunel, P. (1972). The Gas@ cod ecosystem in the Gulf of St. Lawrence. 111. The daily and seasonal vertical migrations of cod (Gadus morhua) in 196G62. Le Naturaliste Canadien 99, 287-357. Bulman, C. M. and Blaber, S. J. M. (1986). Feeding ecology of Macruronus novaezelandiae (Hector) (Teleostei: Merlucciidae) in south-east Australia. Australian Journal of Marine and Freshwater Research 37, 621-639. Campana, S. E. (1984). Interactive effects of age and environmental modifiers on the production of daily growth increments in otoliths of plainfin midshipman, Porichthys notatus. Fishery Bulletin, United States 82, 165-177. Campana, S. E. and Neilson, J. D. (1985) Microstructure of fish otoliths. Canadian Journal of Fisheries and Aquatic Sciences 42, 1014-1032. Carey, F. G . (in press). Further acoustic telemetry observations of swordfish. In “Planning the Future of Billfishes” (R. H. Stroud, ed.) Part 2. National Coalition for Marine Conservation, Savannah, Georgia. Carey, F. G . and Robison, B. H. (1981). Daily patterns in the activities of swordfish, Xiphias gladius, observed by acoustic telemetry. Fishery Bulletin, United States 79, 277-292. Casimir, M. J. (1971). Zur Morphologie, Histochemie, Tagesperiodik und Biologie der Operculardruse bei Labriden und Scariden (Pisces). Marine Biology 8, 1 2 6 146. Clark, C. W. and Levy, D. A. (1988). Die1 vertical migrations by juvenile sockeye salmon and the antipredation window. American Naturalist 131, 272-290. Cole, R. A. and MacMillan, J. R. (1984). Sampling larval fish in the littoral zone of western Lake Erie, USA. Journal of Great Lakes Research 10, 15-27. Colgan, P. (1975). Self-selection of photoperiod as a technique for studying endogenous rhythms in fish. Journal of Interdisciplinary Cycle Research 6, 203-21 1. Colton, J. B., Jr (1965). The distribution and behaviour of pelagic and early demersal stages of haddock in relation to sampling techniques. International Commission of the Northwest Atlantic Fisheries Special Publication 6, 3 17-333. Colton, J. B., Jr and Temple, R. F. (1961). The enigma of Georges Bank spawning. Limnology and Oceanography 6, 280-291. Covill, R. W. (1959). Food and feeding habits of larvae and postlarvae of Ammodytes americanus. Bulletin of the Bingham Oceanographic Collection, Yale University 17, 125-146. Creutzberg, F. (1961). On the orientation of migrating elvers (Anguilla vulgaris Turt.) in a tidal area. Netherlands Journal of Sea Research 1, 257-338. Cronin, T. W. and Forward, R. B. (1979). Tidal vertical migration: An endogenous rhythm in estuarine crab larvae. Science 205, 102G1021.

DIEL VERTICAL MIGRATIONS

159

Cronin, T. W. and Forward, R. B. (1983). Vertical migration rhythms of newly hatched larvae of the estuarine crab, Rhithropanopeus harrisii. Biological Bulletin (Woods Hole) 165, 139-153. Cushing, D. H. (1951). The vertical migration of planktonic Crustacea. Biological Reviews of the Cambridge Philosophical Society 26, 158-192. Dagg, M. J. (1985). Effects of food limitation on die1 migratory behaviour in marine zooplankton. Archiv fur Hydrobiologie 21, 247-255. Dalpadado, P. and Gjrasaeter, J. (1987). Observations on mesopelagic fish from the Red Sea. Marine Biology 96, 173-183. Davies, I. E. (1976). Arrest of deep scattering layer migration by a positive thermal gradient. Nature, London 260, 599-600. Davis, F. C. (1981). Ontogeny of circadian rhythms. In “Handbook of Behavioral Neurobiology” (J. Aschoff, ed.), pp. 257-274. Plenum Press, New York. Davis, R. E. (1962). Daily rhythm in the reaction of fish to light. Science 137, 430432. de Groot, S. J. (1964). Diurnal activity and feeding habits of plaice. Rapports et Proces- Verbaux des Rtunions, Conseil International pour I’Exploration de la Mer 155, 48-51. de Veen, J. F. (1978). O n selective tidal transport in the migration of North Sea plaice (Pleuronectes platessa) and other flatfish species. Netherlands Journal of Sea Research 12, 115-147. Douglas, R. H. (1982). An endogenous crepuscular rhythm of rainbow trout (Salmo gairdneri) photomechanical movements. Journal of Experimental Biology 96,377388. Enright, J. T. (1965). Entrainment of tidal rhythm. Science f Washington, D . C . ) 147, 864-867. Enright, J. T. and Hamner, W. M. (1967). Vertical diurnal migration and endogenous rhythmicity. Science (Washington, D . C . ) 157, 937-941. Eriksson, L.-0. (1978). Nocturnalism versus diurnalism - dualism within fish individuals. In “Rhythmic Activity of Fishes” (J. E. Thorpe, ed.), pp. 69-89. Academic Press, London. Eriksson, L.-0. and van Veen, T. (1980). Circadian rhythms in the brown bullhead, Ictalurus nebulosus (Teleostei): Evidence for an endogenous rhythm in feeding, locomotor, and reaction time behaviour. Canadian Journal of Zoology 58, 18991907. Esterley, C. 0. (1917). The occurrence of a rhythm in the geotropism of two species of plankton copepods when certain recurring external conditions are absent. University of California Publications in Zoology 16, 393-400. Fortier, L. and Leggett, W. C. (1982). Fickian transport and the dispersal of fish larvae in estuaries. Canadian Journal of Fisheries and Aquatic Sciences 39, 1 1501163. Fortier, L. and Leggett, W. C. (1983). Vertical migrations and transport of larval fish in a partially mixed estuary. Canadian Journal of Fisheries and Aquatic Sciences 40, 1543-1555. Franqueville, C. (1971). Macroplancton profond de la Mediterrante. Tithys 3, 1 I56.

Fry, F. E. J. (1971). The effect of environmental factors on the physiology of fish. In Fish Physiology. Vol. VI: Environmental Relations and Behaviour” (W. S. Hoar and D. J. Randall, eds), pp. 1-98. Academic Press, London.

160

J. D. NEILSON AND R. 1. PERRY

Gascon, D. and Leggett, W. C. (1977). Distribution, abundance, and resource utilization of littoral zone fishes in response to a nutrient/production gradient in Lake Memphremagog. Journal of the Fisheries Research Board of Canada 34, 1105-1 117. Gibson, R. N. (1965). Rhythmic activity in littoral fish. Nature 207, 544545. Gibson, R. N. (1967). Experiments on the tidal rhythm of Blenniuspholis. Journal of the Marine Biological Association of the United Kingdom 47, 97-1 11. Gibson, R. N. (1970). The tidal rhythm activity of Coryphoblennius galerita (L.) (Teleostei, Blenniidae). Animal Behaviour 18, 539-543. Gibson, R. N. (1971). Factors affecting the rhythmic activity of Blennius pholis L. (Teleostei). Animal Behaviour 19, 336-343. Gibson, R. N. (1973). Tidal and circadian activity rhythms in juvenile plaice, Pleuronectes platessa. Marine Biology 22, 379-386. Gibson, R. N. (1976). Comparative studies on the rhythms of juvenile flatfish. In “Biological Rhythms in the Marine Environment” (P. J. DeCoursey, ed.), pp. 199-213. University of South Carolina Press, Columbia. Gibson, R. N. (1978). Lunar and tidal rhythms in fish. In “Rhythmic Activity of Fishes” (J. E. Thorpe, ed.), pp. 201-213. Academic Press, London. Gibson, R. N. (1982). Recent studies on the biology of intertidal fishes. Oceanography and Marine Biology Annual Review 20, 363-414. Gibson, R. N. and Hesthagen, I. H. (1981). A comparison of the activity patterns of the sand goby Pomatoschistus minutus (Pallas) from areas of different tidal range. Journal of Fish Biology 18, 669-684. Gibson, R. N., Blaxter, J. H. S. and de Groot, S. J. (1978). Developmental changes in the activity rhythms of the plaice (Pleuronectes platessa, L.). In “Rhythmic Activity of Fishes” (J. E. Thorpe, ed.), pp. 169-186. Academic Press, London. Girsa, I. I. (1973). Alteration in the behavior and vertical distribution of certain juvenile cyprinids in relation to illumination intensity and the presence of a predator. Journal of Ichthyology 13, 449-454. Gjosaeter, J. and Kawaguchi, K. (1980). A review of the world resources of mesopelagic fish. Food and Agriculture Organization Fisheries Technical Paper 193.

Gliwicz, M. Z. (1986). Predation and the evolution of vertical migration in zooplankton. Nature, London 320, 746-748. Graham, J. J. and Sampson, D. B. (1981). Factors affecting the depth distribution of larval herring (Clupea harengus L.) in coastal Maine waters. North Atlantic Fisheries Organization Scientific Council Report Document 8 I/IX/ 138, 9 pp. Grave, H. (1981). Food and feeding of mackerel larvae and early juveniles in the North Sea, Rapports et ProcPs- Verbaux des Reunions, Conseil International pour I’Exploration de la Mer 178, 454459. Green, J. M. (1971). Field and laboratory activity patterns of the tidepool cottid Oligocottus maculosus Girard. Canadian Journal of Zoology 49, 255-264. Greer Walker, M., Harden Jones, F. R. and Arnold, G . P. (1978). The movements of plaice (Pleuronectes platessa) tracked in the open sea. Journal du Conseil, Conseil International pour I’Exploration de la Mer 38, 58-86. Hamner, W. M . and Schneider, K. (1986). Regularly spaced rows of medusae in the Bering Sea: Role of Langmuir circulation. Limnology and Oceanography 31, 171177. Hamrin, S. F. (1986). Vertical distribution and habitat partitioning between the different size classes of vendace, Coregonus albula, in thermally stratified lakes. Canadian Journal of Fisheries and Aquatic Sciences 43, I61 7-1625.

DIEL VERTICAL MIGRATIONS

161

Harder, W. and Hempel, G. (1954). Studien zur Tagesperiodik der Aktivitat von Fischen I . Versuche an Plattfischen. Kurze Mitteilungen aus dem Znstitut fur Fischereibiologie der Universitat Hamburg 5 , 22-3 I . Hardy, A. C. (1956). “The Open Sea. Its Natural History: The World of Plankton”. Collins, London. Harris, J. E. (1963). The role of endogenous rhythms in vertical migrations. Journal of the Marine Biological Association of the United Kingdom 43, 153-166. Haury, L. R., Wiebe, P. H., Orr, M. H. and Briscoe, M. G. (1983). Tidally generated high-frequency internal wave packets and their effects on plankton in Massachusetts Bay. Journal of Marine Research 41, 65-1 12. Helfman, G . S. (1986). Die1 distribution and activity of American eels (Anguilla rostrata) in a cave-spring. Canadian Journal of Fisheries and Aquatic Sciences 43, 1595-1605. Herman, S. S. (1963). Vertical migration of the opossum shrimp, Neomysis americana. Smith. Limnology and Oceanography 8, 228-238. Hobson, E. S. (1973). Die1 feeding migrations in tropical reef fishes. Helgolundrr wissenschafrliche Meeresuntersuchungen 24, 36 1-370. Hopson, A. J. and McLeod, A. A. (1982). The photo-environment and patterns of migration and distribution in openwater fish. In “Lake Turkana: A Report on the Findings of the Lake Turkana Project, 1972-1975” (A. J. Hopson, ed.), Vol. 2, pp. 637-641. Goverment of Kenya and the Ministry of Overseas Development, London. Hoss, D. E. and Phonlor, G. (1984). Field and laboratory observations o n diurnal swim bladder inflationdeflation in larvae of gulf menhaden, Brevoortia patronus. Fishery Bulletin, United States 82, 5 13-5 17. Hunter, J. R. and Sanchez, C. (1976). Die1 changes in swim bladder inflation of the larvae of the northern anchovy, Engraulis mordax. Fishery Bulletin, United States 74, 847-855. Iles, T. D. and Sochasky, J. B. (1985). Drift and current selection, transport and retention, disperal and homing: A review of concepts and principles relevant to the temporal and spatial organization of marine systems. International Council f o r the Exploration of the Sea, Council Meeting L:14, 18 pp. Javaid, M. Y. and Anderson, J. M. (1967). Influence of starvation on selected temperature of some salmonids. Journal of the Fisheries Research Board of Canada 24, 1515-1519. Johannessen, A. (1986). Recruitment studies of herring (Clupea harengus, L.) in Lindaaspollene, Western Norway, 1-3. Fiskeridirektoratets Skrifter Serie Havundersokelser 18, 139-240. Johansen, A. C. (1925). On the diurnal vertical movements of young of some fishes in Danish waters. Meddelelser Fra Kommissionen .for Havundersogelser Serie Fiskeri 8(2), 27 PP. Johnson, T. and Muller, K. (1978). Different phase position of activity in juvenile and adult perch. Naturwissenschaften 65, 392. Kabasawa, H. (1982). On activity patterns of Halichoeres poecilopterus under the conditions of artificially changed environments. Biennial Report of the Keikyu Aburatsubo Marine Park Aquarium. 11, 4 6 5 1 . (Abstract only.) Kavaliers, M. (1978). Seasonal changes in the circadian period of the lake chub, Couesius plumbeus. Canadian Journal of Zoology 56, 259 1-2596. Kavaliers, M . (1979a). The pineal organ and circadian organization of teleost fish. Revue Canadienne de Biologie 38, 281-292.

162

J. D. NEILSON A N D R. I. PERRY

Kavaliers, M. (1979b). The pineal organ and circadian rhythms of fishes. In “Environmental Physiology of Fish” (M. A. Ali, ed.), Vol. 35, pp. 631-645. Plenum Press, New York. Kawamoto, N. Y. and Konishi, J. (1955). Diurnal rhythm in phototaxis of fish. Report of Faculty of Fisheries, Prefectural University of Mie 2, 7-17.

Kawamoto, N. Y. and Niki, T. (1952). An experimental study on the effect of leading fish by fish attraction lamps. Faculty of Fisheries, Prefectural University of Mie 1, 175-196.

Kendall, A. W. and Naplin, N. A. (1981). Diel-depth distribution of summer ichthyoplankton in the middle Atlantic Bight. Fishery Bulletin, United States 79, 705-726.

Kingsford, M. J. (1988). The early life history of fish in coastal waters of northern New Zealand: A review. New Zealand Journal of Marine and Freshwater Research 22, 463479.

Kitajima, C., Tsukashima, Y. and Tanaka, M. (1985). The voluminal changes of swim bladder of larval red sea bream Pagrus major. Bulletin of the Japanese Society of Scientijc Fisheries 51, 159-764. Koeller, P. A., Hurley, P. C. F., Perley, P. and Neilson, J. D. (1986). Juvenile fish surveys on the Scotian Shelf: Implications for year-class size assessments. Journal du Conseil, Conseil International pour I’Exploration de la Mer 43, 59-76.

Kraus, E. B. (1977). “Modelling and Prediction of the Upper Layers of the Oceans”. Pergamon Press, Oxford. Kroneld, R. (1976). Phase shift of swimming activity in the burbot Lota Iota L. (Pisces, Gadidae) at the Arctic Circle. Physiological Zoology 49, 49-55. Kruuk, H. (1963). Diurnal periodicity in the activity of the common sole, Solea vulgaris Quensel. Netherlands Journal of Sea Research 2, 1-28. Kryuchukov, V. I., Gireyev, N. M. and Tsonif, B. F. (1984). Age-related changes in the diurnal rhythm of locomotor activity of the mirror carp, Cyprinus carpio (Cyprinidae). Journal of Ichthyology 24, 64-71. Laprise, R. and Dodson, J. J. (1989). Ontogeny and importance of tidal vertical migrations in the retention of larval smelt Osmerus mordax in a well-mixed estuary. Marine Ecology Progress Series 55, 101-1 11. Lasker, R. and Sherman, K. (1981). The estimation of spawning stock size from fish egg and larval surveys. Rapports et ProcPs- Verbaux des RPunions, Conseil International pour I’Exploration de la Mer 178, 268-278.

Leis, J. M. (1986). Vertical and horizontal distribution of fish larvae near coral reefs at Lizard Island, Great Barrier Reef, Australia. Marine Biology 90,505-516. Levin, A. V. (1982). Ground selection, diurnal rhythm of vertical distribution and swimming velocities of young Volga sturgeon, Acipenser gueldenstaedti Brandt (Acipenseridae) under experimental conditions. Journal of Zchthyology 22(4), 130136.

Levy, D. A. (1987). Review of the ecological significance of diel vertical migrations by juvenile sockeye salmon (Oncorhynchus nerka.). In “Sockeye Salmon (Oncorhynchus nerka) Population Biology and Future Management” (H. D. Smith, L. Margolis and C. C. Wood, eds), Vol. 96, pp. 44-52. Canadian Special Publication of Fisheries and Aquatic Sciences. Department of Fisheries and Oceans, Ottawa. Lewis, R. M. and Wilkens, E. P. H. (1971). Abundance of Atlantic menhaden larvae and associated species during a diel collection at Beaufort, North Carolina. Chesapeake Science 12, 185-187.

DIEL VERTICAL MIGRATIONS

163

Lillelund, K. and Lasker, R. (1971). Laboratory studies of predation by marine copepods on fish larvae. Fishery Bulletin, United States 69, 655-667. Longhurst, A. R. (1976). Vertical migration. In “The Ecology of the Seas”. (D. H. Cushing and J. J. Walsh eds), pp. 116-137. W. B. Saunders, Philadelphia. Lorz, H. V., Pearcy, W. G . and Fraidenburg, M. (1983). Notes on the feeding habits of the yellowtail rockfish, Sebastes jlavidus, off Washington and in Queen Charlotte Sound. California Fish and Game 69, 33-38 Lough, R. G. (1975). A preliminary report of the vertical distribution of herring larvae on Georges Bank. International Comtnission for the Northwest Atlantic Fisheries Research Document 75/50, 9 pp. Lough, R. G. and Cohen, R. E. (1982). Vertical distribution of recently hatched herring larvae and associated zooplankton on Jeffreys Ledge and Georges Bank, October, 1974. Northeast Fisheries Center, National Marine Fisheries Service. Woods Hole Laboratory Reference 82-10, 96 pp. Mahon, R. and Neilson, J. D. (1987). Diet changes in Scotian Shelf haddock during the pelagic and demersal phases of the first year of life. Marine Ecology Progress Series 37, 123-130. Mansueti, R. (1963). Symbiotic behaviour between small fishes and jellyfishes, with new data on that between the stromateid, Peprilus alepidotus, and the scyphomedusa, Chrysaora yuinquecirrha. Copeia 1963 ( I ) , 40-80. Marliave, J. B. (1981). Vertical migrations and larval settlement in Gilbertidia sigalutes, F. Cottidae. Rapports et Procks- Verbaux des RPunions, Conseil International pour I’Exploration de la Mer 178, 349-351. McCleave, J. D. and Kleckner, R. C. (1982). Selective tidal stream transport in the estuarine migration of glass eels of the American eel (Anguilla rostrata). Journal du Conseil, Conseil International pour I’Exploration de la Mer 40, 262-271. McFarland, W. N. (1986). Light in the sea -correlations with behaviors of fishes and invertebrates. American Zoologist 26, 3 8 9 4 0 1. McNulty, J. A. (1982). The effects of constant light and constant darkness on daily changes in the morphology of the pineal organ in the goldfish, Carassius auratus. Journal of Neural Transmission 53, 277-292. Mel’nikov, V. N., Grudov, K. P. and Butumba, A. (1981). Analysis of diurnal vertical migration of fishes. Biologischeskie Nauki (Moscow) 4, 108-1 12. (Unedited translation available as Canadian Translation of Fisheries and Aquatic Sciences no. 5421, Canada Institute for Scientific and Technical Information, National Research Council, Ottawa, Ontario, K 1A OS2.) Miller, D., Colton, J. B., Jr and Marak, R. R. (1963). A study of the vertical distribution of larval haddock. Journal du Conseil, Conseil International pour 1’E.xploration de la Mer 28, 3 7 4 9 . Moore-Ede, M. C., Sulzman, F. M. and Fuller, C. A. (1982). “The Clocks that Time Us”. Harvard University Press, Cambridge, Mass. Morris, R. W. and Kittleman, L. R. (1967). Piezoelectric property of otoliths. Science 158, 368-370. Muller. K. (1978). The flexibility of the circadian system of fish at different latitudes. In “Rhythmic Activity of Fishes” (J. E. Thorpe, ed.), pp. 91-104. Academic Press, London. Muus, B. J. (1967). The fauna of Danish estuaries and lagoons: Distribution and ecology of dominating species in the shallow reaches of the mesohaline zone. Meddelelser .fra Danmarks Fiskeri-og Havundersogelser 5, 1-3 16.

164

J. D. NEILSON AND R. I. PERRY

Nash, R. D. M. (1982). The diel behavior of small demersal fish on soft sediments on the West Coast of Scotland using a variety of techniques, with special reference to Leseuerigobius friesii Pisces Gobiidae. Marine Ecology (Pubblicazione della Stazione Zoologica di Napoli) 3, 161-1 78. Naylor, E. (1976). Rhythmic behaviour and reproduction in marine animals. In “Adaptation to Environment: Essays on the Physiology of Marine Animals” (R. C. Newell, ed.), pp. 393429. Butterworths, London. Neilson, J. D. and Geen, G. H. (1982). Otoliths of chinook salmon (Oncorhynchus tshawytscha): Daily growth increments and factors influencing their production. Canadian Journal of Fisheries and Aquatic Sciences 39, 134C1347. Nelson, D. R. and Johnson, R. H. (1970). Die1 activity rhythms in the nocturnal bottom dwelling sharks Heterodontus francisci and Cephaloscyllium vantriosum. Copeia 1970, 732-739. Nesterov, A. A. and Bazarlov, S. I. (1986). Vertical distribution and behavior of flying fish (Exocoetidae). Journal of Ichthyology 26(5), 159-162. (Abstract only.) O’Connor, J. M. 119721. Tidal activity rhythm in the hogchoaker, Trinectes maculatus (Bloch and Schneider). Journal of Experimental Marine Biology and Ecology 7. 173-177. Ofori-Adu, D. W. (1978). The occurrence of thermoclines on the shelf of Ghana and the significance of thermocline depth variations to the surface schooling fishes. Fishery Research Unit, Ghana. Marine Fisheries Research Report 6, 46 pp. Ohman, M. D., Frost, B. W. and Cohen, E. B. (1983). Reverse diel vertical migration: An escape from invertebrate predators. Science 220, 1404-1407. Olla, B. L., Studholme, A. L. and Bejda, A. J. (1985). Behaviour of juvenile bluefish Pornatomus saltatrix in vertical thermal gradients: Influence of season, temperature acclimation and food. Marine Ecology Progress Series 23, 165-177. Onura, Y. and Ali, M. A. (1982). Effect of hypophysectomy on the synaptic ribbons in the pineal organ of the killifish Fundulus heteroclitus. Bulletin of the Japanese Society of Scientijc Fisheries 48, 1679-1 684. Palmer, J. D. (1973). Tidal rhythms: The clock control of the rhythmic physiology of marine organisms. Biological Reviews of the Cambridge Philosophical Society 48, 377418. Pearcy, W. G., Hopkins, C. C. E., Grr~nvik,S. and Evans, R. A. (1979). Feeding habits of cod, capelin, and herring in Balsfjorden, Northern Norway, JulyAugust 1978: The importance of euphausiids. Sarsia 64, 269-277. Pearre, S., Jr (1973). Vertical migration and feeding in Sagitta elegans Verrill. Ecology 54, 300-3 14. Pearre, S., Jr (1978). Problems of detection and interpretation of vertical migration. Journal of Plankton Research 1, 29-44. Perry, R. I. and Neilson, J. D. (1988). Vertical distributions and trophic interactions of age-0 Atlantic cod and haddock in mixed and stratified waters of Georges Bank. Marine Ecology Progress Series 49, 199-214. Pittendrigh, C. S. (1981). Circadian systems: General perspectives. In “Handbook of Behavioral Neurobiology” (J. Aschoff, ed.), pp. 57-80. Plenum Press, New York. Potter, D. C. and Lough, R. G . (1982). Vertical distribution of herring larvae (Clupea harengus L.) on Nantucket Shoals, November, 1977, collected by MOCNESS aboard Anton Dohrn 77-03. Northeast Fisheries Center, National Marine Fisheries Service. Woods Hole Laboratory Reference 82-08, 59 pp. Potter, D. C. and Lough, R. G. (1987). Vertical distribution and sampling variability of larval and juvenile sand lance (Ammodytes sp.) on Nantucket Shoals and Georges Bank. Journal of Northwest Atlantic Fishery Science 7 , 107-1 16.

DIEL VERTICAL MIGRATIONS

165

Potter, D. C., Lough. R. G., Perry, R. I. and Neilson, J. D. (1990). Comparison of the MOCNESS and IYGPT pelagic samplers for the capture of 0-group cod (Gadus morhua) on Georges Bank. Journal du Conseil, Conseil International pour I’Exploration de la Mer 46, 121-128. Raymont, J. E. G. (1983). “Plankton and Productivity in the Oceans, Vol. 2: Zooplankton”, 2nd edn. Pergamon Press, Oxford, 824 pp. Rijnsdorp, A. D., van Stralen, M. and van der Veer, H . W. (1985). Selective tidal transport of North Sea plaice larvae Pleuronectes platessa in coastal nursery areas. Transactions of the American Fisheries Society 114, 4 6 1 4 7 0 . Ringelberg, J. (1964). The positively phototactic reaction of Daphnia rnagna Straus: A contribution to the understanding of diurnal vertical migration. Netherlands Journal of Sea Research 2, 3 1 9 4 0 6 . Ringelberg, J., Kasteel, J. V. and Servaas, H. (1967). The sensitivity of Daphnia magna Straus to changes in light intensity at various adaptation levels and its implication in diurnal vertical migration. Zeitschrifttfur Vergleichende Physiologie 56, 3 9 7 4 0 7 . Roe, H. S. J. (1983). Vertical distributions of euphausiids and fish in relation to light intensity in the northwestern Atlantic. Marine Biology 77, 287-298. Roe, H. S. J. and Badcock, J. (1984). The die1 migrations and distributions within a mesopelagic community in the North East Atlantic. 5. Vertical migrations and feeding of fish. Progress in Oceanography 13, 3 8 9 4 2 4 . Rogers, H . M. (1940). Occurrence and retention of plankton within the estuary. Journal of the Fisheries Research Board of Canada 5 , 1 6 4 17 1 . Rudstam, L. G. and Magnuson, J. J. (1985). Predicting the vertical distribution of fish populations: Analysis of cisco, Coregonus artedii, and yellow perch, Perca jlavescens. Canadian Journal of Fisheries and Aquatic Sciences 42, 1 178-1 188. Russell, F. S. (1926). The vertical distribution of marine macroplankton. 111. Diurnal observations on the pelagic young of teleostean fishes in the Plymouth area. Journal of the Marine Biological Association of the United Kingdom 14, 387-414. Russell, F. S. (1927). The vertical distribution of plankton in the sea. Biological Reviews of the Cambridge Philosophical Society 2, 21 3-262. Russell, F. S. (1928). The vertical distribution of marine macroplankton. VIII. Further observations on the diurnal behaviour of the pelagic young of teleostean fishes in the Plymouth area. Journal of the Marine Biological Association of the United Kingdom 15, 829-850. Ryland, J. S. (1964). The feeding of plaice and sand-eel larvae in the southern North Sea. Journal of the Marine Biological Association of the United Kingdom 44, 343346. Sameoto, D. D. (1982). Vertical distribution and abundance of the Peruvian anchovy Engraulis ringens and sardine Sardinops sagax larvae during Nov. 1977. Journal of Fish Biology 21, 171-186. Sameoto, D. D. (1984). Environmental factors influencing diurnal distribution of zooplankton and ichthyoplankton. Journal of Plankton Research 6 , 767-792. Sameoto, D.D. (1988). Feeding of lanternfish (Benthosema glaciale) off the Nova Scotia Shelf. Marine Ecology Progress Series 44, 113-129. Saunders, D. S. (1977). “An Introduction to Biological Rhythms”. Blackie, London, 170 pp. Schoth, M. and Tesch, F. W. (1984). The vertical distribution of small 0-group Anguilla larvae in the Sargasso Sea with reference to other Anguilliform leptocephali. Meeresforschung 30, 188-195.

166

J. D. NEILSON AND R. I. PERRY

Schwassmann, H. 0. (1971). Biological rhythms. In “Fish Physiology” (W. S. Hoar and D. J. Randall, eds), Vol. 6, pp. 371-428. Academic Press, London. Scott, J. S. (1984). Short-term changes in distribution, size and availability ofjuvenile haddock around Sable Island off Nova Scotia. Journal of Northwest Atlantic Fishery Science 5, 109-1 12. Scrope-Howe, S. and Jones, D. A. (1986). The vertical distribution of zooplankton in the Western Irish Sea. Estuarine, Coastal and Shelf’Science 22, 785-802. Seliverstov, A. S. (1974). Vertical migration of larvae of the Atlanto-Scandian herring (Clupea harengus L.). In “The Early Life History of Fish” (J. H. S. Blaxter, ed.), pp. 253-262. Springer-Verlag, Heidelberg. Shanks, A. L. (1983). Surface slicks associated with tidally forced internal waves may transport pelagic larvae of benthic invertebrates and fishes shoreward. Marine Ecology Progress Series 13, 3 11-3 15. Siegmund, R. and Wolff, D. L. (1973). Circadian-rhythmik und Gruppenverhalten bei Leucaspius delineatus (Pisces, Cyprinidae). Experientia 29, 54-58. Sjoblom, V. and Parmanne, R. (1978). The vertical distribution of Baltic herring larvae (Clupea harengus L.) in the Gulf of Finland. Finnish Fisheries Research 2, 5-18. Skud, 9. E. (1968). Responses to marine organisms during the solar eclipse of July 1963. Fishery Bulletin, Fish and WildliJe Service, United States 66, 259-27 1. Smith, R. L.. Paul, R. L. and Paul, A. J. (1986). Effect of food intake and temperature on growth and conversion efficiency in juvenile walleye pollock (Theragra chalcogramma) (Pallas), a laboratory study. Journal du Conseil, Conseil International pour I’Exploration de la Mer 42, 241-253. Sogard, S. M., Hoss, D. E. and Govoni, J. J. (1987). Density and depth distribution of larval gulf menhaden, Brevoortia patronus, Atlantic croaker, Micropogonias undulatus, and spot, Leiostomus xanthurus, in the Northern Gulf of Mexico. Fishery Bulletin, United States 85, 60 1-609. Southward, A. J. and Barrett, R. L. (1983). Observations on the vertical distribution of zooplankton, including post-larval teleosts, off Plymouth in the presence of a thermocline and a chlorophyll-dense layer. Journal of Plankton Research 5, 599618. Spieler, R. E. and Noeske, T. A. (1984). Effects of photoperiod and feeding schedule on die1 variations of locomotor activity, cortisol, and thyroxine in goldfish. Transactions of the American Fisheries Society 113, 528-539. Stearns, S. C. (1982). The role of development in the evolution of life-histories. In “Evolution and Development” (J. T. Bonner, ed.), pp. 237-258. Spinger-Verlag, New York. Steele, C. W. (1985). Absence of a tidal component in the die1 pattern of locomotive activity of sea catfish, Arius felis. Environmental Biology of Fishes 12, 69-73. Steiner, W. W. and Olla, 9. L. (1985). Behavioral responses of prejuvenile red hake, Urophycis chuss, to experimental thermoclines. Environmental Biology of Fishes 14, 167-173. Stephenson, R. L. and Power, M. J. (1988). Semidiel vertical movements in Atlantic herring Clupea harengus larvae: A mechanism for larval retention? Marine Ecology Progress Series 50, 3-1 I . Stephenson, R. L. and Power, M. J. (1989). Observations on herring larvae retained in the Bay of Fundy: Variability in vertical movement and position of the patch edge. Rapports et Procis- Verbaux des Rkunions, Conseil International pour I‘Exploration de la Mer.

DIEL VERTICAL MIGRATIONS

167

Stickney, A. P. (1972). The locomotor activity of juvenile herring in response to changes in illumination. Ecology 53, 438-145. Sullivan, L. (1980). Distribution, abundance and mortality estimates of larval butterfish, Peprilus triacanthus, from MARMAP surveys, 1977 and 1978. lnternational Council for the Exploration of the Sea, Council Meeting G18, 12 pp. Sund, P. N., Blackburn, M. and Williams, F. (1981). Tunas and their environment in the Pacific Ocean: A review. Oceanography and Marine Biology Annual Review 19, 443-512. Suthers, I. M. and Gee, J. H. (1986). Role of hypoxia in limiting die1 spring and summer distributions of juvenile yellow perch (Perca javescens) in a prairie marsh. Canadian Journal of Fisheries and Aquatic Sciences 43, 1562-1570. Takahashi, J. S. and Zatz, M . (1982). Regulation of circadian rhythmicity. Science 217, 1104-1 I 1 I . Takahashi, S. (1981). Vertical distribution and die1 migration of isaza Chaenogobius isaza Pisces in Lake Biwa, Japan. Zoological Magazine 90, 145-151. Tat’yankin, Yu. V. (1974). The effect of the preliminary adaptation temperature on the distribution of juvenile cod (Gadus morhua morhua) and pollock (Pollachius virens) within a temperature gradient. Journal of lchthyology 14, 755-760. Thor, D. H. (1972). Schedules of self-lighting behaviour in the yellow canary, goldfish, and Mongolian gerbil. Journal of General Psychology 87, 23-35. Turner, J. T. and Dagg, M. (1983). Vertical distributions of continental shelf zooplankton in stratified and isothermal waters. Biological Oceanography 3, 1-39. Turuk, T. N. (1973). Diurnal periodicity in feeding and the vertical movements of the Atlantic cod (Gadus morhua L.). Journal of Ichthyology 13, 275-288. Turuk, T. N. (1974). Fatness and vertical migrations of the Atlantic cod (Gadus morhua morhua). Journal of lchthyology 14, 977-982. Uotani, I. (1973). Diurnal changes of gas bladder and behavior of postlarval anchovy and other related species. Bulletin of the Japanese Society of ScientiJic Fisheries 39, 867-876. Verwey, J. (1966). The role of some external factors in the vertical migration of marine animals. Netherlands Journal of Sea Research 3, 245-266. Wales, W. (1975). Extraretinal control of vertical migration in fish larvae. Nature 253,4243. Wales, W. (1984). Photic behaviour and vertical migration in herring larvae. Marine Behaviour and Physiology 1 1 , 139-156. Walsh, S. J. (1988). Die1 variability in trawl catches of juvenile and adult yellowtail flounder on the Grand Banks and the effect on resource assessment. North American Journal of Fisheries Management 8, 373-38 1 . Weinstein, M. P., Weiss, S. L., Hodson, R. G. and Gerry, L. R. (1980). Retention of three taxa of postlarval fishes in an intensively flushed tidal estuary, Cape Fear River, North Carolina. Fishery Bulletin, United States 78, 419-434. Willis, J. M. and Pearcy, W. G. (1982). Vertical distribution and migration of fishes of the lower mesopelagic zone off Oregon, USA. Marine Biology 70, 87-98. Winfree, A. T. (1980). “The Geometry of Biological Time”. Biomathematics Vol. 8. Springer-Verlag, New York, 530 pp. Wood, R. J. (1971). Some observations on the vertical distribution of herring larvae. Rapports et Proces- Verbaux des Reunions, Conseil International pour I’Exploration de la Mer 160. 60-64. Woodhead, P. M. J. (1966). The behaviour of fish in relation to light in the sea. Oceanography and Marine Biology Annual Review 4, 337403.

168

J. D. NEILSON AND R. I. PERRY

Wurtsbaugh, W. A. and Li, H. (1985). Diel migrations of a zooplanktivorous fish (Menidia beryllina) in relation to the distribution of its prey in a large eutrophic lake. Limnology and Oceanography 30, 565-576. Wurtsbaugh, W. A. and Neverman, D. (1988). Post-feeding thermotaxis and daily vertical migration in a larval fish. Nature 333, 846-848. Yamashita, Y.,Kitagawa, D. and Aoyama, T. (1985). Diel vertical migration and feeding rhythm of the larvae of the Japanese sand-eel Ammodytes personatus. Bulletin of the Japanese Society of ScientiJic Fisheries 51, 1-5. Yasuda, T. (1973). Ecological studies on the jellyfish, Aurelia aurita (Linne) in Urazoko Bay, Fukui Prefecture - VII, Diel vertical migrations of the medusa in the early fall, 1969. Publications of the Seto Marine Biological Laboratory 20,491499 (Abstract only.) Yemel’yanov, I. P. (1976). Form of fluctuations in biorhythmology. Novosibirsk: Nauka Press. 126. (Abstract only.) Zusser, S. G. (1958). The daily vertical migration of pelagic fish. Trudy VNIRO 36, 1-32. Zusser, S . G. (1966). Reaction of horse mackerel to light according to the amount of food consumed. Trudy VNIRO 60, 325-349.

The Development and Application of Analytical Methods in Benthic Marine Infaunal Studies Brenda J. Burd,' Amanda Nemec2 and Ralph 0. Brinkhurst3 'Galatea Research Inc., Brentwood Bay, 'International Statistics and Research Corp.. Brentwood Bay and 30cean Ecology Laboratory, Institute of Ocean Sciences, Sidney, British Columbia, Canada

Introduction .. .. .. .. .. Collection of Data .. .. .. .. A. Sampling devices ,. .. .. .. B. Sieving of samples ., .. .. .. C. Sampling effort .. .. .. .. D. Temporal sampling design , . .. .. 111. Analysing the Data Matrix ., .. .. A. The data matrix .. .. .. .. B. The subjective approach: Community concepts C. Descriptive univariate community indices . . D. Statistical inference . . .. ,. .. E. Multivariate data analyses , . .. .. F. Time-series analysis , . .. .. .. IV. Summary , . .. .. .. .. .. V. References . . . . .. .. .. .. I. 11.

..

..

..

.. ..

.. ..

.. .. .. .. .. ..

.. ..

.. .. ..

..

..

..

..

.. .. ..

..

.. ..

.. ..

..

..

.. .. ..

..

.. ..

.. ..

..

..

.. .. ..

..

.. ..

.. .. ..

.. .. .. ..

..

.. .. .. ..

170 172 173 174 175 177 178 178 186

192 212 214 231 232 234

Copyright 0 19W by Academic Press Limited AN rights of rrproduclion in m y form reserved

ADVANCES IN MARINE BIOLOGY VOLUME 26 ISBN 0-12-026126-X

169

170

B. J. BURD E T A L .

I. Introduction The purpose of this review is to examine the development of practical analytical approaches that have traditionally been applied to benthic softbottom macrofaunal (particularly marine) studies. We hope to provide some insight into why certain approaches are used, and the assumptions and resulting limitations of these methods. A thorough review of all methods ever used by benthic researchers is impossible and unnecessary, as specific reviews are available on many topics and will be referred to. Theoretical aspects of ecology are not examined in detail, but reference to recent treatments is included for those who have the time and energy to delve into such issues. Some commonly applied methods and topics are covered in more detail than others, especially if they are considered to be pivotal in the development of benthic analytical studies. It is felt this type of review is useful because of the diffuse nature of the large volume of literature and research in benthic ecology. For ecologists recently embarked on benthic studies, sorting through this literature is a difficult task. In order to design a survey program, it is important to consider why certain theories are currently popular. Are they logical developments of trial and error, honest representations of well thought out hypotheses or just trendy? Our bias towards certain methods may be obvious to the reader, but there is no formal attempt to declare a “right” or “wrong” approach. This is a rapidly changing field and the review can only be considered up to date as of December 1988. In this review, benthic studies include what Hurlbert (1984) described as mensurative or survey research, which is non-experimental. Mensurative studies involve random sampling of the organisms (and related factors) from different stations, selected according to some reasonable survey pattern based on the objectives of the study. Our treatment is limited mainly to the literature on the remote sampling (shipboard) of marine macrobenthic infauna inhabiting soft substrata. Examples of studies and theories derived from meiofaunal, intertidal, freshwater and some land-based study areas are included where they have contributed to marine benthic survey theory. There are three basic issues which have been prominent in benthic ecological research in one form or other over the years. These are:

( I ) What is the most efficient and accurate method of identifying and describing the underlying faunal structure of a sample? How does this depend on the concept of a “community” in benthic ecology? (2) How can the faunal structure of samples be distinguished from each other over time and space? (3) How do natural and anthropogenic habitat variables affect the faunal

BENTHIC MARINE INFAUNAL STUDIES

171

structure of samples, and can these two types of effects be distinguished? From time to time, attempts have been made to synthesize or standardize approaches to benthic sampling and analysis (for recent examples, see Boesch, 1977; Verner et al., 1985; Chapman et al., 1987; Becker and Armstrong, 1988), or to discuss statistical problems, sampling practices and new methods (cf. Green and Vascotto, 1978; Tetra Tech Inc., 1986; GEEP workshop, Mar. Ecol. Prog. Ser. 46, 1988). Recently, Lopez (1988) discussed the comparative aspects of limnological and marine benthic macrofaunal studies, a rare effort indeed. Section I1 of this chapter defines sampling terms and basic sampling considerations. Sampling design is very much dependent on the objective of a study. Therefore it is virtually impossible to describe the “right” or “wrong” way to sample data. The important sampling parameters (size and number of samples) can only be confidently decided upon after preliminary reconnaissance sampling and analysis of data from the study site. Therefore, most methods for sampling design are reviewed along with the appropriate discussion of analytical methods in Section 111. Section I11 discusses the organization (1II.A) and analysis (1II.B-1II.E) of data in benthic studies, starting with the simple methods developed early in benthic ecological study, and progressing to the currently popular computerintensive multivariate methods. The development of each stage in analysis has continued in parallel to some extent. Therefore, the discussion does not attempt to present a chronology of methodological development. The types of analysis in order of discussion include: Section IZI.B: The subjective approach: Community concepts. An understanding of the term “community” and such related concepts as “continuum’’ is perhaps the primary requirement for the analysis and interpretation of benthic survey data. Petersen’s pioneering work in the early 1900s marks the first serious attempt to tackle these issues. Although Petersen used subjective methods to describe and compare benthic communities, he recognized the need for an objective and systematic approach. However, it was only with the development of computer technology that the necessary methods were to become readily available. Consequently, more or less subjective methods were employed until recently, and many continue to be used today. For example, in limnology, the Theinemann school of lake typology, which advocates the use of characteristic taxa to identify communities, is in this philosophical paradigm (Brinkhurst, 1974).

Section 1ZI.C: Descriptive univariate community analysis. Since Petersen’s

172

B. J. BURD E T A L .

time, ecologists have sought to describe communities using graphical and mathematical models which reduce all the data from a given sampling station to a single number, index or function. Univariate models do not recognize the multidimensional effects of species interacting with each other. Nevertheless, their simplicity makes these models popular, particularly in pollution contexts. This phase in the evolution of an objective methodology is dominated by the diversity index in applied aquatic studies in North America, and by the pseudoquantitative Saprobian system once popular in Europe (cf. Leppakoski, 1977), particularly in freshwater studies. Section ZZZ.D: Most of the recent advances in inferential hypothesis testing have focused on computer-intensive descriptive analytical methods, which are often too complex to interpret subjectively. Of particular interest is the recent proliferation of non-parametric simulation techniques for significance testing of large and complex data sets. Section ZZZ.E: Multivariate community analysis. Though some of these methods are quite old, they have gained wider acceptance in recent years than the methods discussed in Section 1II.C. These methods incorporate the multidimensionality of species relationships within benthic assemblages. Section ZZZ..F:Most time-series studies use multivariate methods because of the increased dimensional complexity added by temporal considerations. Time-series studies are still relatively rare, but are increasingly becoming prevalent in the literature as computer-intensive multivariate methods develop. The primary aim of all methods will continue to be the simplification of complex patterns occurring in benthic faunal assemblages so that interpretation and comparison is possible. Much of the audience for simplified reporting of complex taxonomically based data sets consists of environmental managers, engineers and political agencies that no longer accept the “tlitist” testimony of experts, particularly when major industrial or other developments clash with environmental concerns.

II. Collection of Data In this section, we review briefly four important aspects of benthic surveys: the choice of a suitable sampling device, the sieve size, the number and spatial distribution of the samples, and the timing of samples. We make no attempt to give a specific set of guidelines. The purpose is simply to highlight the main issues, many of which continue to be debated, and which are reviewed elsewhere.

BENTHIC MARINE INFAUNAL STUDIES

173

We will refer to the data collected from one unit of sample effort, whether by grab, core, quadrat, photograph, trawl or other, as “the sample unit” or “replicate”. The term “sample” has been used in benthic studies in a variety of different ways, but usually corresponds to the data for a “station” (i.e. it includes all the replicates from a designated site, for which all variables are measured). The reader will also find throughout this chapter that “sample” in inverse analyses means the total complement of a given taxon across all sample units. Therefore, where the distinction is important, the term “station” will be used to indicate the total data complement of the “sample units” for one designated site. A.

Sampling Devices

The vast array of grabs, cores, sleds, trawls, suction samplers, etc., which have been used for benthic survey research preclude an effective review in this chapter. Grabs and cores have traditionally been used for quantitative sampling of infaunal animals since the early 19OOs,whereas sleds, dredges and trawls have been used for qualitative sampling of larger and more dispersed epifauna. Thorson (1957)drew attention to the problems caused by the shock wave created in front of many sampling devices as they approach the bottom. Since that time, there have been a number of good descriptive reviews of sampling devices (Holme, 1964;Hopkins, 1964;McIntyre, 1970;Eleftheriou and Holme, 1984). Statistical tests have been published relating to the relative efficacy of different samplers. Grab or core-type samplers can profoundly affect the quantitative results from coarse sediments and at shallow depths (Wigley, 1967; Christie, 1975; Tyler and Shackley, 1978; Hartley, 1982).Gerlach et al. (1985)pointed out that the loss of meiofaunal animals using remote grabs or cores was very high compared to direct sampling (such as SCUBA). Rutledge and Fleeger (1988) describe a laboratory experiment on the efficiency of sampling the meiobenthos with respect to core penetration rate. Hartley (1982)cites an example of an inter-calibration experiment between laboratories from which he concluded that differences in results were related partially to the differences in design of two different Van Veen grabs. Dybern et al. (1976)reviewed and recommended standard procedures for sampling in the Baltic Sea, in order to avoid sampling discrepancies between studies. Many authors have their own justifications and reasons for using specific sampling devices, or make it clear that convenience or cost is of primary importance. For limnological sampling, reviews include Brinkhurst (1 974) and Edmondson and Winberg (1971), while Hartley and Dicks (1987) provide a review of sampling methods in estuaries.

174

B. J. BURD E T A L .

B. Sieving of Samples An important consideration in quantitative sampling of the benthos is choice of screen size. Historically, benthic fauna have been delimited into three groups based on the size of organisms trapped by different sized screens (see Reish, 1959; Thorson, 1966; Schwinghamer, 1981; Warwick, 1984; Gerlach et al., 1985; Platt, 1985). In general, researchers have recognized up to four size groups usually referred to as microbes (bacteria, etc.), meiofauna (including foraminifera and the smallest invertebrate fauna), macrofauna (most of the biomass of benthic animals) and megafauna (often lumped with macrofauna; low in abundance but with high individual biomass). Over the years, there has been disagreement as to the optimum screen sizes for benthic studies, although most researchers have focused on the middle (macrofauna) group. Reish (1959) indicated that a screen as small as 0.27mm is required to sample 95% of the animals, whereas a 1-mm mesh will sample 95% of the biomass, and therefore all of the megafauna and most of the macrofauna. The 1-mm mesh screen has been applied most often in pollution monitoring studies of macrofauna and in those studies in which the primary concern is to sample most of the biomass of the animals present (cf. Pearson, 1975; Poore and Kudenov, 1978a,b), whereas studies concerned with meiofauna commonly employ 0.063- to 0.1-mm mesh screeens. More recently, Holme and McIntyre (1984) have recommended the lower size limit of 0.5 mm for macrofaunal sampling, based on their belief that the smaller macrofauna are an important component of benthic assemblage structure even if they do not make up a significant portion of the biomass. Rees (1984) also notes that many polychaete species fragment into pieces smaller than 1 mm during shipboard processing, and recommends the use of 0.5-mm screens. Becker and Armstrong (1988) recommend an initial sieving with a 1-mm screen, followed by a secondary sieve with a 0.5-mm screen (the material from the latter may or may not be processed, but is available if required). The choice of screen size obviously depends on the objectives of the study. For example, in areas of gross pollution, there may be no macrofauna. Therefore, the only sensible sampling program may be designed to capture meiofauna, as some meiofaunal species tend to be more tolerant than the macrofauna. Studies of energy flow or respiration may require a more comprehensive sample. The smaller the screen, the greater the cost and time required to process samples, particularly if taxonomic expertise for the smaller groups is not readily available. (For more detail on the functional separation and sizes of meiofaunal and macrofaunal groups, see Section III.C.3.)

BENTHIC MARINE INFAUNAL STUDIES

I75

C. Sampling Efort The balance between volume (or area) of sample unit, number of replicates and number of sample stations is necessarily dependent on the required accuracy, the hypotheses being tested and the overall objectives of the study. The importance of designing the sampling program to suit the statistical methods to be used can hardly be overstressed. Therefore, the design of sampling and analyses must be dynamic and concurrent. For example, some inferential methods require a minimum number of sample replicates for reliability (see Nemec and Brinkhurst, 1988a). Since benthic infauna are relatively immobile, much of the theory that has been developed for sampling plant communities is applicable to benthic communities (Greig-Smith, 1964; Kershaw, 1973). Green (1979), Holme and McIntyre ( I 984), Hurlbert (1984) and Baker and Wolff (1987) review most of the important issues in sampling. Cochran (1963) provides the standard reference on sampling techniques from a statistician’s perspective. Ripley ( 1 981) discusses various spatial sampling schemes, including: ( I ) Uniform random sampling where a sample area is defined and sites within that area picked at random in sufficient quantity to produce an approximately uniform spacing of samples. (2) Stratified random sampling, which is appropriate if some information about the sample area is available, and involves selecting sites from non-overlapping areas (strata) that are usually delineated by environmental factors (e.g. depth, substratum type). Within each stratum the sampling conditions should be as homogeneous as possible, so that the different strata themselves can be compared. Within each area the sampling follows a uniform random pattern as in (I). (3) Systematic random sampling which involves sampling at regular intervals, usually along a gradient (e.g. pollution). At each point along the gradient a sample area or quadrat is selected, in which replicates are selected at random (as in I).

These sampling schemes are three of many which may be acceptable. Random samples may be most amenable to statistical community analysis but non-random or systematic samples are often used to examine spatial distribution of a community (cf. Cliff and Ord, 1981). Clearly, the sampling pattern should be designed to cover adequately the area about which inferences are to be made, so that it can be assumed that the sampling bias has been reduced to an acceptable level. Figure 1 illustrates some of the

B. J. BURD ET AL.

176

......... .......... . ......... ......... ......... ...... ......... ......... ......... ......... Regular

Random

Clumped

0 0 0 0 0 0 0

2

0.

.O

0 0 0 0 0 0 0 0 0

(B)

FIG. I . An illustration of different spatial dispersion patterns in benthic assemblages and the effect of sample size and position on data collection. Most benthic assemblages have clumped dispersions, which produce the greatest sampling problems. The smaller rectangles represent theoretical sample sizes and locations. (A) The animals are evenly or regularly dispersed and sample location or bearing has no real effect on conclusions about species abundance and dispersal. (B) The random distribution of animals could affect conclusions depending on the sample location and size. (C) The clumped or aggregated faunal distribution produces the greatest variation in sample results. This final pattern is the most commonly encountered among mixed species groups in benthic habitats.

problems of sample coverage and size that can arise depending on the spatial dispersion of the fauna (see Section 1II.C.1). Hurlbert (1984) emphasizes the importance of suitable sample replication for testing hypotheses, and warns of the problem of pseudoreplication in ecological studies, i.e. using replicate samples which are not suitable for testing the specific hypothesis. Green (1979) recommends the use of stratified random sampling when there is a large-scale environmental pattern (e.g. a salinity gradient along an estuary), and discusses the use of nested random sampling (i.e. random sampling on several spatial scales within a sample area) when sources of variation are hierarchically related or when the environment is known to be spatially patchy but not on a sufficiently large scale to define strata. For example, hypotheses concerning the spatial aggregation of species or assemblages may require a nested random design, with the use of a series of sampling devices of different sizes, in order to examine the dispersion of animals on different spatial scales. Saila et al. (1976) suggested an optimal allocation of survey

BENTHIC MARINE INFAUNAL STUDIES

177

resources based on stratifed sampling in the New York Bight. Cuff and Coleman (1979) discussed the benefits of a random stratified sample design for (adequately) determining the mean number of individuals per taxon, but concluded that a simple uniform random sampling pattern was just as good. Interestingly, they claimed that if the number of stations was increased at the expense of number of grabs per station, the efficiency of the estimate of mean abundance per taxonomic group (i.e. polychaetes, gastropods, etc.) increased, to an optimum of 1 grab per station. This is not necessarily true for inferences about other aspects of faunal structure, or for statistical hypothesis testing. Primary considerations that govern the choice of volume and number of sample units (replicates) include obtaining representative coverage of the number of species and individuals (and biomass if applicable), and accuracy (or power - see Section III.D.2) of the statistical analysis. Various methods which have been developed for examining optimum sampling effort will be discussed in those sections that pertain to the application of the analytical models on which they are based. Such methods depend upon the availability of data from a previous set of samples, for planning of further sampling. Otherwise, the number of sample units or replicates to be obtained at each station must be determined subjectively. Traditionally, researchers have used between two and five replicates per station. Hartley (1982) and Holme and McIntyre (1984) recommend five replicates of 0.1 m2 area (sampler size) for macrofaunal sampling, but point out that faunal density has an overwhelming influence on accuracy (see Section 1TI.C.1).

D.

Temporal Sampling Design

The timing and frequency of sampling is usually decided by economic or practical considerations. Most benthic assemblages exhibit some degree of seasonal variation, and may vary on shorter time-scales (tidal, daily). Govaere et al. (1 980) described the Nyquist criterion for time-series analysis which states that “sampling frequency must be at least twice the highest frequency of the phenomenon studied”. For “patch” studies (related to diversity mechanisms - see Section III.C.2), the implications can be staggering, because life-cycles may be very short in some species. Therefore, design considerations such as aliasing that are discussed in books on the analysis of time-series data are often not applicable to benthic surveys, because data are not collected with sufficient regularity or frequency to test for periodicity or other temporal effects. Barnard et al. (1986) discuss the trade-off (with respect to estimates of the mean abundances of species) between detailed surveys at a single point in

178

B. J. BURD E T A L .

time and less detailed, long-term surveys (see also Smith, 1978). An extreme example is given by Legendre et al. (1985), in which one station was examined many times to study community successional stages. Long-term sampling on specific sites is particularly difficult in the deep sea for logisitic and cost reasons, particularly since the low abundance of fauna requires large-scale (often semi-quantitative) samples to ensure a reasonable distribution of species (see Gage et al., 1980). In many cases, annual surveys of an area are conducted at a time of year when neither major recruitment nor mortality is occurring in the assemblage. Sampling according to ecological cycles, such as during the spring plankton bloom, is probably more sensible than simple calendar dates, but planning for ship time usually prohibits such flexibility.

111. Analysing the Data Matrix Although there has been little progress in design, both the analytical method and theory of sampling strategy have changed dramatically. Benthic ecologists have borrowed heavily from theory developed in earlier ecological studies of terrestrial habitats, as the similarities between sessile terrestrial and marine communitites are indisuputable. The logical starting point in a review of methodology is the description of the data set itself.

A.

The Data Matrix

Since the early 1900s, researchers have found it useful to set out the data collected (abundance, biomass, presence/absence, etc.) as a tabular data matrix, with taxa as one axis, station identifications as the other, and numeric faunal data as the body of the matrix (as in Table 1). It is often useful to set out environmental data in the same manner. The environmental data matrix has a similar format to the faunal data matrix, with columns representing stations and rows representing different environmental variables. Researchers may standardize or transform the data to avoid having different units of measurement in each row (e.g. percentages, ratios, etc.: see Field et al., 1982). In most studies, the faunal data matrix includes abundance data (counts per sampling unit). The data may be abundances standardized to some surface area sampled, or it may be a presence/absence indicator. There is a great deal of variation in the degree of taxonomic effort applied to the compilation of benthic faunal data. Long and Lewis (1987) and Warwick

179

BENTHIC MARINE INFAUNAL STUDIES

( 1988a) found that for macrobenthic samples, identification to family only

was good enough for broad community identifications based on abundance. TABLE1. EXAMPLE OF A SAMPLE ABUNDANCE DATA MATRIXFOR BENTHICMARINE COLUMBIA" FAUNAFROM BRITISH Taxon

Polychaeta Aglaophamus malmgreni Ampharete acutifrons Ampharete jnmarchica Ampharetidae Amphit ritinae Ancistrosyllis groenfandica Anobothrus gracilis Acesta lopezi Acesta neosuecica Aliia yuadrilobata A ricidea Asychis A utolytus Brada villosa Chone ecuudata Cirrophorus branchiatus Clymenura columbiana Cossura longocirrata Cossura modica Deeamastus gracilis Euchone incolor Eucl.vmene Euclymene zonalis Euclyrneninae Eusdlis assimilis Exogone verugera Galathowenia oculata Gattyana cirrosa Glycera eapitata Glycinde armigera Glyphanostomum pallescens Goniada annulata Podarkeopsis brevipalpa Gyptis Harmothoa lunalata Harmothoe Jasmineira paeijica Kefersteinia cirrata "

5A-1 5A-2 5B-1 5B-2 5C-1 5C-2 5D-1 5D-2 6-1 9-1 0 0 0 0 0 0

0 3 0 0 0 0

0 0 0 0 0

1

0

3 0 0 0 0 0 0 0 I 0 4 0 1 0 0 0 0 0 0 0 0 0 0

5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0

0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Note the number of zero records for this group of stations.

0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0

0 0 0 0 0 2 0

0 0 0 0 0 0 0

1

0 0 0 0 0 0 0 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0

0 0 0 0 0 0

0 0 0 0

0 0 0 0 0

0

0

1

0

0

0 0 0

1

0

0 0

0

0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0

0 0

0 0 0 0 0 0

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1

0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 2 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0

I80

B. J. BURD E T A L .

However, Popham and Ellis (1971) concluded that phyletic or class identifications alone did not delimit associations based on abundance, unless a selected number of dominant species was included (an uncomfortably subjective method). Herman and Heip (1988) suggest that meiobenthos may be diagnostic of community structure at the genus or even higher taxonomic levels, though Warwick (1 988a) is less enthusiastic about grouping of meiofauna at higher taxonomic levels. It is generally accepted that the more detailed the taxonomic identification of samples, the more reliable the interpretation of results (despite the fact that this valuable information is largely ignored during the application of simplistic univariate measures such as diversity indices). In most benthic survey studies, there has always been a presumption that the underlying taxonomy is sound. Ellis (1985) reviews the potential scale and ramifications of this problem. This may have been especially problematical in Petersen’s day. Where adapted to a specific habit and habitat, animals from very different ancestors may be taxonomically linked in a phenetic (similarity-based) classification, though separated by a phylogenetic system (if these are as reliable as claimed by their proponents). Ecologists should be aware of the philosophical basis for supposed higher taxonomic groupings, which may differ according to whether the author is a cladist, a pheneticist, a hybrid or is unaware of this very lively debate in systematics. Changes in names of even the most common taxa plague ecologists, even though taxonomists work with a system designed to move towards stability. Different data sets may contain many examples of the same species hidden under different names. Furthermore, higher taxa may be constituted in a variety of different ways, so that “rolled up” taxonomic groups may not be equivalent. The faunal data matrix may also consist of weight measurements. Much controversy exists in the literature as to which type of weight measurement should be used. Possibilities include wet weight (usually blotted or slightly air-dried), dry weight (oven or freeze-dried) and organic weight (ash-free dry weight or labile organic carbon). In many cases, wet weight is the most feasible of these alternatives, and several authors have published approximate conversion values to organic weight for different taxonomic groups (cf. Thorson, 1957; see also Crisp, 1984). Brey et al. (1988) discuss the conversion of dry and ash-free dry weights of macrobenthic invertebrates to energy units. Such approximate conversions inevitably introduce an error factor. The use of wet or dry weights is not entirely reliable in benthic invertebrate studies because of the difficulty in separating out large masses of inorganic material such as the shells of pelecypods and gastropods. A few such shelled specimens may be weighed separately and a rough correction factor applied for a set of samples. A potential error involved in this method

BENTHIC MARINE INFAUNAL STUDIES

181

is that shells can contain a substantial amount of organic matter (Kuenzler,

1961). Even the careful measurement of organic weights is not entirely satisfactory unless a large area can be sampled quantitatively. This is because the presence of the odd large, rare specimen can greatly increase the weight in one sample unit, making replicates extremely variable. On the other hand, the removal of large specimens can produce an unrealistic result in the biomass analysis since the space requirements of the benthic fauna may be an important factor in community structure. A further complication of using weight data is that, in most cases, the specimens are wet-preserved (alcohol or formalin, etc.), which can cause shrinkage and leaching of organic material. Proper preservation methods for marine macrobenthos are reviewed by Holme and McIntyre (1984) and Ellis (1987). Few authors have attempted to use combined measures of abundance and biomass. Recently, Warwick (1986) suggested a pollution monitoring method which uses a comparison of biomass and abundance data. Combination methods will undoubtedly become more popular as scientists continue the struggle to describe accurately assemblages, and to predict changes in them. Analyses proceed by describing relationships and general trends that exist either (a) between the columns or stations of the population or environmental matrix (Q-mode or normal analysis), or (b) between the rows or taxa of the faunal matrix (R-mode or inverse analysis). Most researchers concentrate on the comparison of sites (normal or Q-mode analysis). Methods are also available which attempt to relate Q- and R-mode analyses, or the environmental and faunal data matrices (see Section 1II.E). 1. Data reductions Data matrices obtained from benthic survey studies commonly include hundreds of species (rows) and several replicates each for a large number of stations (columns). The sheer size of the array may be unmanageable. A review of different strategies for data reduction is given by Stephenson and Cook (1980). Data can be reduced in several ways. One way is to reduce the number of samples in the data matrix by averaging or pooling across replicates. This is often done so that existing computer programs can manage the entire data set. Unfortunately, the loss of information about variance around the mean (abundances, biomasses or other) severely limits the use of inferential statistics and the reliability of interpretations based on comparing mean abundances between stations. Therefore, replicates should only be averaged or pooled once it is determined that there is little variability for a given station (see Hurlbert, 1984). A second method of data reduction affects the number of species (rows).

182

B. J. BURD E T A L .

Many of the species sampled may be extremely rare. Researchers frequently reduce the data set either by eliminating species that are deemed “rare” according to some set of criteria, or by rolling up species into taxonomically higher groups such as genera or families. This is done to make the analysis of data less unwieldy and time-consuming, and to reduce the complexity in results of multivariate studies caused by the inclusion of “unimportant” species. Another rationale for data reduction is to produce a set of symmetrical data matrices (all the same dimensions) to accommodate the type of multivariate analysis being used, particularly when several matrices are being compared. Statistical procedures have been used to test the significance of the relationship between each individual species and an environmental matrix or variable. Species showing a non-significant relationship are subsequently eliminated. Such a set of tests suffers from the multiple comparisons problem (progressive increase in family-wise error rate with increasing number of comparisons) familiar to users of univariate tests such as Analysis of Variance (ANOVA). Multivariate methods often have biases with respect to the relative weighting of species in the analysis. The Bray-Curtis similarity coefficient (see Section 1II.E. l), for instance, places the most emphasis on the abundant taxa, with minor consideration of rare species. Euclidean distance, unlike many other metric similarity measures, is unbounded, and can become infinitely large if there are many zero entries in the data matrix. The resulting distortion of results in a multivariate analysis can be alleviated if an appropriate data reduction or roll-up is performed. The reduction deemphasizes the abundance of common species and increases the emphasis on rare species in the analysis. Of the two, roll-up is probably preferable, to avoid loss of rare species data which may actually be vital in delimiting or defining a community (cf. Brinkhurst, 1987; Burd and Brinkhurst, 1987). Numerically rare but large specimens can be important in community structure and should therefore not be eliminated. For example, Gray and Pearson (1982) pointed out that Stephenson et al. (1972) in this way eliminated one of the original community - defining species in their reanalysis of Petersen’s data. This points out that what constitutes an “abundant” or “non-abundant” species is species-dependent and should be viewed with caution. To overcome this type of problem, Smith et al. (1988) recommend the use of “species standardized abundances” in the data matrix (see next section). 2. Data transformations Prior to descriptive or inferential analysis, raw data (usually abundances) are

BENTHIC MARINE INFAUNAL STUDIES

183

often transformed. This section will discuss such a priori or “primary” transformations, and it should be clarified from the beginning that secondary transformations of manipulated data (e.g. rotations used in classification and ordination analyses for the optimal interpretation of results) are not included in the discussion. Primary transformations are performed for several reasons, which are rarely stated clearly in applied studies. It is unfortunate that transformations are often used arbitrarily without any examination of their effect on the data or their utility. For instance, there are studies in the literature in which two different transformations of a given data set are used prior to two different multivariate analyses (e.g. classification and ordination). Data subjected to different transformations are not readily comparable, unless trends are strong enough to be evident regardless of the treatment or method of analysis (in which case the transformation was probably pointless). Transformations are often used in conjunction with similarity measures (see Sections III.E.2 and 3), which in some cases have biases related to abundant vs rare species (for discussion see Clarke and Green, 1988). Data transformations used to correct biases prior to analysis usually reduce the disparity in emphasis on different species evident in the original abundances. For example, many researchers apply geometric transformations, so that instead of a small species (represented by 100 animals) being ten times more important than a large one (represented by only 10 animals), it is only two times more important (log base 10 transformation), or about three times more important (square root transformation). This may or may not be intended by the researcher, who must then interpret the results for the transformed data in terms of the original animal distribution. Transformations are not necessary for descriptive analyses such as classification and ordination, but Clarke and Green (1988) suggest that data reductions combined with transformations are usually needed to correct problems caused by a large number of zero entries. More importantly, many statistical (parametric inferential) analyses assume that the data follow a normal distribution. Each method of analysis is based on a specific set of assumptions, which may differ from one method to the next. It is important to understand those assumptions and, if necessary, take appropriate action to help ensure their validity. Some standard assumptions include: ( I ) The underlying distribution of the measurement of interest (e.g. animal abundance) is normal or near-normal (the importance of this depends on sample size). This assumption is unlikely to be true in aggregated or clumped assemblages where most species are overdispersed (see Section 1II.C. 1). Hughes and Thomas (1971a,b) point out that in ordination, the proportion of the total variance accounted

184

B. J. BURD E T A L .

for by the first few factor axes is generally increased if the data approximate a multivariate normal distribution, thus the incentive for data transformations. Multi-species density data are rarely multivariate normal. Many parametric tests are robust enough to withstand skewness of data, especially if the other assumptions are met, and the populations being compared have similarly shaped distributions. (2) The variance of the variable of interest is independent of its mean. This can be tested by plotting logarithmic values of the mean (Z) vs the variance (S2) for all species at a station (or group of stations combined in some rational manner) and performing a regression analysis (see Downing, 1979) to obtain the equation S 2 = axb (refer to Taylor’s power law discussion in Section 1II.C.1). If the variance is related to the mean in this manner, a variance stabilizing (power) transformation (in which the exponent is equal to 1 - b/2) can be applied to remove the effect (Downing, 1979). Square root and log transformations tend to be at extreme ends of the transformation scale, and therefore should be used only if the relationship between variance and mean warrants it. By extrapolation, samples of the same benthos taken at different times would not necessarily require the same transformation. L. R. Taylor (1980) also points out the fact that greater error in aggregation estimates (b) is introduced by the lumping of species into higher taxa. (For further comments on this topic, see Downing, 1980, 1981, 1986; Chang and Winnel, 1981.) (3) The variance is additive. In aggregated assemblages the variance is commonly multiplicative. Stabilizing the variance (as in 2) may increase the probability of additivity by alleviating skewness in the distribution of the sample means. One problem with data transformations is that if the best transformation is to be chosen, it will probably be different for every station. Yet it is not feasible to use a series of different data transformations when performing analyses using the combined data from all stations. A common transformation must be selected, by estimating the degree of clumping of the entire data set. Therefore, the usefulness of transformations for stabilizing variance is questionable in analyses of large and diverse data sets. The most commonly used data transformations in benthic studies are the square root, root-root (or fourth root: see Field et al.,1982), cube root and log transformations [log or In (x + 1) for data sets with zero entries]. Reviews of this topic are common, and include Hoyle (1973), Tukey (1 977) and Hoaglin et al. (1983). To eliminate the problem of zero entries in the

BENTHIC MARINE INFAUNAL STUDIES

185

data matrix, researchers often add a small value to each entry (usually 0.1-1) before transformation (depending on the biases of the analytical method to be used: see Clifford and Stephenson, 1975). This is necessary for log transformations, since log 0 is undefined. Such an augmented transformation can produce further interpretation problems. Downing (1979) concluded from an examination of many benthic freshwater studies, that the fourth root (root-root) transformation ( b = 1.5, 1 - 6/2 = 0.25) was of general utility for stabilizing variance in benthic studies. Vezina (1988) suggests that b = 1.22 is more appropriate for stabilizing variance in marine invertebrate assemblages. Josefson (198 1) applied Analysis of Variance (ANOVA) to log-transformed and untransformed abundance and biomass data and found no difference in the results, suggesting some resilience with respect to normality and homogeneity of variance requirements for this type of test. Field ef al. ( I 982) suggest that coding abundances (e.g. using a scale of relative abundance from 0 to 5 for absent to dominant) often has the effect of normalizing data. This seems unlikely, and would require normality testing. An extreme example of data transformation is the conversion of abundance data to binary (presence/absence) data, which is appropriate when only the occurrence of a given species is in question. If there is low confidence that the variance and the underlying distribution of the measurement of interest meet the assumptions of normality, the researcher should either decide upon a useful data transformation or should consider the use of non-parametric inferential statistics which do not require prior knowledge of the underlying frequency distribution (though they may still require a symmetric distribution of data). Whatever the rationale for use, the selection of a transformation should have some ecological basis, though most researchers ignore this aspect entirely. For example, Clarke and Green (1988) argue that log transformations have a sensible basis because they transform the variance (of density measures, etc.) to percentage variance of the measure, and population density tends to vary spatially and temporally on a percentage basis. No corroborating evidence or discussion is given on this point, or on the behaviour of multi-species assemblages. Field et al. (1982) suggest that data “standardization” (Bray and Curtis, 1957; Clifford and Stephenson, 1975; Smith et al., 1988) or “relativization” should be done when quantitative R-mode (inverse) analyses are carried out (see also Boesch, 1973; Hailstone, 1976). In other words, fractions of the maximum abundance sampled over all stations for a species, can be used to replace actual abundances for that species. The authors suggest that species which are functionally interdependent (e.g. host and parasite) may otherwise be separated into different groups because their relative abundances are different. Another alteration which Clifford and Stephenson (1975) include

186

B. J. BURD ET AL.

in this category is standardizing the data by dividing by the standard deviation (Z scores). Such data alterations are dependent on the data from the entire matrix in order to perform the transformations.

B. The Subjective Approach: Community Concepts What is a benthic community and how can we characterize it? From 1911 to 1918, Petersen (cf. 1913, 1914, 1915a,b)-and later Thorson (1957)-described benthic population structure in a subjective manner and introduced the concepts which would become the foundation of benthic survey studies. In particular, the definition and practical usage of the term “community” has been a cornerstone of benthic analytical development. It is appropriate, therefore, to preface the discussion of analytical methods (Sections 1II.D1II.F) with a review of some important community concepts. 1. Community vs continuum

The first quantitative ecological studies on marine benthos were carried out in Northern Europe by Petersen (1913, 1915a,b). Using subjective judgement and his expertise as an ecologist, Petersen described a series of benthic communities which he considered to be relatively stable and cosmopolitan. Petersen was fairly conservative in his conclusions about these communities, describing them as “statistical units only”, although no statistical analysis of data was carried out. The units were dominated by recurrent species which gave the community its name, and were related to typical coastal and sediment types. The search for characterizing species is a persistent theme in benthic ecology, particularly in pollution studies (see next section). The limnological equivalent to this phase is the benthic indicator species concept in both lake typology studies (reviewed by Brinkhurst, 1974) and in early pollution work (Hynes, 1960). Petersen (1914, cited in Thorson, 1957) stated that: the animals, which are not seasonal and which comprise an important part of the whole mass of the community, owing to number or weight will presumably be best suited for characterization of the community.

In 1957, Thorson refined Petersen’s work and developed the “parallel communities” hypothesis, based on studies in Northern Europe. Thorson suggested that Petersen’s communities were not cosmopolitan at the species level, but rather at the genus (or family) level. Therefore, researchers studying benthos in areas outside Petersen’s locales (irrespective of latitude) could expect to find persistent communities with dominants of the same

BENTHIC MARINE INFAUNAL STUDIES

187

genus or family as the classic Petersen communities, but not necessarily the same species. These could be considered “parallel” communities or “community-units” by terrestrial botanists (Whittaker, 1970). Illies and Botosaneanu (1963) used a similar approach in designating stream communities (see also Harrison and Hynes, 1988). Like Petersen, Thorson was concerned that the subjective skill used by experienced ecologists to characterize communities be substantiated by quantitative sampling and data analysis, although few statistical methods were in use in benthic ecology in the 1950s. Thorson (1966) reiterated his theories of parallel communities by describing the bottom types associated with specific communities (regardless of latitude). He also revised his earlier theories (1957) by admitting that communities without dominant species (i.e. with many low abundance species) cannot fit into the parallel community structure. This is evident in tropical and many deep-sea benthic communities. The parallel theory also does not take into account peaks in abundance in the meiobenthos, where these species dominate the fauna. Researchers are still citing examples of Thorson’s parallel macrobenthos communities in various parts of the world (e.g. Shelford, 1935; Buchanan, 1963; Horikoshi, 1970; Ellis, 1971; Masse, 1972; Warwick and Davies, 1977; Govaere et al., 1980; Shackley and Collins, 1984). Horikoshi (1970) described a Thorson MafdanelOphiura community as far away as the Sea of Japan. Buchanan and Moore (1986) described long-term stability in one of Petersen’s Amphiura fififormis communities and cited evidence that biotic and abiotic factors affected this stability. The recurrent and persistent nature of such assemblages of animals suggests that the concept of ecologically significant, interactive groupings of animals cannot be dismissed completely as “coincidence”, as suggested by Gleason (1926). Most researchers subscribe to the belief that the range of cited examples lends credence to the Thorson parallel community theory, but that its simplicity and subjectivity makes it useful only as an overview or reference point for more detailed ecological study. Petersen and Thorson, though concerned with the quantitative description of communities, did not define the term “community” satisfactorily from either the statistical or ecological point of view, except to indicate that a community was a discrete and repetitive unit characterized by certain dominant species and specific habitat types. The concept of discrete communities as depicted by Petersen and Thorson has been challenged over the years (Gleason, 1926; Jones, 1950; Burbanck et a f . , 1956; Lie, 1968; Mills, 1969; Gray, 1974). Many botanists are inclined towards the “continuum” viewpoint (for reviews of the early development of the community concept in terrestrial

I88

B. J. BURD ET AL.

systems, see Whittaker, 1967, 1970), which suggests that species composition changes along gradients of habitat factors rather than forming discrete communities (Fig. 2). For example, if samples are taken from two distinct but homogeneous substratum types, two distinct species groups may be collected and these might be called “communities”. If a third sample, taken between the first two substrata, contained a mixed group of species, this assemblage might be called a “transitional” community by some researchers. Alternatively, the entire set of three samples could be considered a “continuum” of species. The discreteness of species groups depends on the form and distribution of the particular environmental gradients, as well as the biotic variables sampled and their scale of measurement. Therefore, the discreteness of a group of species tentatively labelled a “community” may be simply a sampling artefact, or a convenient descriptive unit (Gray, 1974). Mills ( 1969) provided an excellent review of the community/continuum debate, discussing the classical definitions of the terms “community”, “formation” and “association” as they apply to terrestrial ecology, and the use of such terms as “community” and “biocoenosis” in benthic ecology. He redefined a benthic “community” in accordance with the concept of a climax community in botany. A “major” benthic community is one which is self-sustaining without other communities, and is defined as: a group of organisms occurring in a particular environment presumably interacting with each other and with the environment, and separated by means of ecological survey from other groups (Mills, 1969).

Identifying such a “major” community is no small matter, and since Petersen and Thorson’s descriptions of recurring communities, no real attempts have been made in benthic ecology to identify major communities. Determining the functional boundaries of a “community” requires some definition or recognition of the equilibrium point in the succession of the fauna. DeAngelis and Waterhouse (1 987) review the concept of successional equilibrium and stability, and suggest that it can be identified or predicted only on such a large scale as to be virtually unmeasurable or untestable. Most authors avoid the argument of “continuum” vs “community”, but tend to lean in favour of one or the other, or use a compromise approach incorporating both viewpoints. In practice, the distribution of the animals collected will determine the method of analysis, as will the philosophical leaning of the researcher. Lie (1968) discussed the controversy between the theories of bounded communities and continua based on the overlapping and varied niche requirements of all the species in the sample set. Lie concluded that discrete communities are absent in Puget Sound, where there are strong environmental gradients. This absence of discrete communities is also evident in many polluted areas (Anger, 1975; Pearson and Rosenberg, 1978).

189

BENTHIC MARINE INFAUNAL STUDIES

~-

Environmental Gradient

I 0

B

0

C 0

-0

C

Environmental Gradient FIG. 2. A conceptual drawing of the difference in population distributions along environmental gradients. Each curve represents the frequency distribution of a species. (A) The nine species distributions are grouped into three fairly discrete “communities” along the environmental gradient. with considerable overlap in distributions. (B) The six species curves shown are minimally overlapped, showing a “continuum” of species along the environmental gradient (see Section III.Bl

Allee et at. (1949) and Burbanck et al. (1956) described the “ecotone”, which is a transitional zone o r gradient between two different communities. The breadth of the ecotone varies with the rate of environmental gradient change in physically controlled benthic habitats characterized by great biotic and physical instability. Therefore, estuaries represent ecotones between the surrounding freshwater and marine communities. This concept has been supported by the findings of Kay and Knights (1975), Ristich et al. (1977), Burns ( 1 978) and Maurer et ul. ( 1 978), though the term “ecotone” seems not

190

B. J. BURD E T A L .

to have become common in usage except in descriptions of the transitional fauna along pollution gradients (e.g. Pearson and Rosenberg, 1978; Knox and Fenwick, 1981). The validity of both the “community” and “continuum” concepts has led in present-day practice to a compromise concept of “intergrading communities”. Maurer et al. (1978) describe estuarine faunas as a “mosaic of assemblages”, some of which are distinct, others amorphous, and which are associated with salinity and sediment type. This description highlights the important observation that an assemblage (or community) may be divided up into small patches, with other assemblages intermixed spatially or temporally. Chapman and Brinkhurst (198 I) described benthic communities in estuaries which migrate along a changing salinity gradient. These communities seem to respond to the strong annual rhythmicity in undammed rivers with high altitude sources, such as the Fraser River on the west coast of Canada. Certain identifiable species groups are located further downstream during freshwater intrusions (high run-off periods) than during periods of strong marine incursions and low run-off. Similar cycles of communities shifting in space have been detected in data from other rivers, such as the St. John River estuary (Gillis, 1978), River Tees estuary (Gray, 1976) and estuaries of the Georgia coast (Howard and Frey, 1975). 2. Indicators of anthropogenic impacts on communities Methods similar to those described by Petersen (1913-1915) and Thorson (1957) are still commonly used to select species or assemblages indicative of pollution. There is some validity to this approach, since Pearson and Rosenberg (1978) suggest that certain species (or genus) groups occur in organically enriched habitats everywhere. Leppakoski (1979) concurs with this viewpoint, but cautions that it is based almost entirely on organic pollution work from boreal bottom communities. In limnology, the traditional method used for typing rivers was the Saprobien system (Kolkwitz and Marsson, 1908), which has subsequently been expanded by various authors (Washington, 1984). Traditionally, species used as indicators of organic pollution have included a number of opportunists which are primary colonizers in such disturbed areas as naturally anoxic basins (Rosenberg, 1980). Perhaps the most common of these is the Cupitella cupitata complex (cf. Rosenberg, 1973; Pearson and Rosenberg, 1978; Gray and Christie, 1983; Levings et al., 1983), which is in reality not one, but a series of sibling species (Grassle and Grassle, 1976). Generally, the presence of such species coincides with an increase in the number of deposit feeders which move into areas as organic

BENTHIC MARINE INFAUNAL STUDIES

191

input increases (Pearson and Rosenberg, 1978). The use of such species as pollution indicators is becoming less and less popular, due to the fact that they tend to be ubiquitous, and are often found in even higher abundances in non-polluted areas than in polluted areas (see Botton, 1979; Gray and Pearson, 1982). Washington (1984) reviews the historical development of biotic indices (i.e. methods or indices which use indicator organisms for pollution monitoring), advocating the use of “clean” species which disappear under various conditions of pollution, rather than the use of tolerant species which can be ubiquitous. Meiofauna such as certain nematodes and a few oligochaete species can be very abundant in extremely polluted sites which cannot be inhabited by the macrobenthic opportunists (Nichols, 1977; Leppakoski, 1977, 1979;Elmgren, 1978). These taxa decline in abundance with increasing distance from the pollution source and are of limited value as pollution indicators in sites with less extreme conditions. Meiofaunal ratios (such as nematodes/copepods, known as the N/C ratio) have occasionally been used as gross indicators of pollution (Raffaelli and Mason, 1981). The N/C ratio is simple to use, but is reliable only on sandy intertidal beaches, where nematode populations increase and copepod populations decrease under increasing pollution conditions. (For a review and discussion of the relative behaviour of these meiofaunal groups in different environmental conditions, see Raffaelli, 1987.) Soule and Keppel (1988) discuss much of the current philosophy on the use of indicator organisms in marine systems. The indicator species concept has been replaced with indicator community models in limnology, where the issue is older and has been subjected to greater scrutiny (cf. Hellawell, 1986). In marine situations, more objective analytical methods for identifying indicator species or groups of species are being attempted (see Sections III.C.2 and III.E.4). Some researchers use higher-order taxonomic groups as indicators of pollution, to economize on sampling and processing costs and to reduce taxonomic discrepancies (cf. Pontasch and Brusven, 1988; Warwick, 1988a,b). However, the potential reliability of results at different taxonomic levels is not adequately proven. It is not safe to assume that species or genera which share an evolutionary relationship (and are thus taxonomically linked) also share ecological requirements, and can therefore be lumped together for community structure comparisons. Gray and Christie (1983) point out the importance of recognizing the effect of natural cyclical changes in polluted areas and suggest monitoring a selection of indicator species over large areas to elucidate the distinctions between man-made and natural disturbance. Opportunistic species groups are useful for reflecting changes in environmental factors in this type of study. The aforementioned authors suggest that this type of monitoring is

I92

B. J. BURD ET AL.

not easily done in soft sediments, but decline to say why this may be so. Gray (1985) has suggested that the major thrust of research should be the study of effects of specific pollutants on individual species (or assemblage groups: see Widbom and Elmgren, 1988) and the relationship between these effects and the benthic ecosystem as a whole. There is a great deal of this type of research that has been published and is still on-going, but because it falls within the realm of manipulative or non-survey research, it will not be discussed further. The concept of the assessment and detection of pollution by means of specific indicator species or groups is no longer widely used. The usefulness of indicators lies in rapid and economic preliminary field surveys to identify potential sources of identifiable types of problems. The “indicators” may provide markers or warnings, manifested by lowering of diversity, or an increase in the abundance of a few taxa that ingest bacteria and tolerate heavy silt loads or low oxygen conditions. Bio-indicators (sentinel species) are used in a somewhat different way from organic pollution indicator species. The accumulation of inorganic toxins in benthic communities is traced by monitoring changes in tissue levels within certain key species (for a review, see Phillips and Segar, 1986). The selection of sentinel species is still fairly subjective, but one requirement is that they should be sedentary or sessile.

C . Descriptive Univariate Community Indices Many researchers have sought to describe the underlying structure of communities in an objective manner. This has resulted in the development of various univariate community indices. The methods described have been criticized as being too simplistic, as they attempt to reduce complex multivariate associations to a single number. Despite this, many of these indices have become entrenched in the regulations governing permits for waste disposal. Because of their greater descriptive specificity, multivariate methods (described in Section 1II.E) are rapidly replacing univariate indices. Since many univariate methods are still in common usage in benthic studies, and are often useful for an initial characterization of a community or sampling strategy, a selection of these indices will be described. The coverage is far from exhaustive, and the reader is referred to the other reviews cited throughout the section. In most survey studies, abundance and the number of species are measured. In some cases, biomass has been used to describe communities, although the difficulties involved with using such a variable character have made biomass analyses unpopular. It is only recently that new approaches to

BENTHIC MARINE INFAUNAL STUDIES

I93

modelling the dynamic interactions of biomass and trophic flow in benthic communities have begun to emerge in the literature (see Section III.C.3). Inferential methods for hypothesis testing with univariate data models are commonly found in introductory statistics textbooks and will not be discussed here. These include normal-theory methods such as t-tests, ANOVA, regression and correlation. Many non-parametric methods are used as well, including rank-order techniques for monotonic data and the bootstrap and randomization methods (see Diaconis and Efron, 1983; Felsenstein, 1985). The latter two types are increasing in popularity (see Clarke and Green, 1988) but are still rarely found in introductory statistics textbooks. Both the bootstrap and randomization methods are non-parametric in that they do not require specific assumptions about the distribution or variance of the data.

1. Spatial distributions and sampling design

Ecologists continue to theorize about the relationships between species abundances and the spatial distribution of assemblages. Brown (1 984) discusses niche size with respect to the dispersion and abundance of species on different spatial scales. His discussion begins with the simple assumption that a species is most abundant at the centre of its niche (as defined by all the variables affecting that niche hyperspace), and declines towards the edges. Many methods have been used to describe the spatial distribution of communities. One way of doing this is through “frequency distributions” which relate sample counts (such as number of individuals in a sample) to frequency of counts (such as number of samples: cf. Elliott, 1977; Tukey, 1977; Devore, 1987) in a histogram or curvilinear graph. Each point on the graph typically represents a single sample unit. If the abundance data can be fitted to one of the standard frequency distribution models, the model can then be used to make decisions on the sampling effort required to collect a given percentage of the animals in the assemblage. Elliott (1977) described several commonly used frequency distribution models, including the positive and negative binomial and Poisson distributions, and discussed methods for testing the fit of data to these models. Elliott points out that for benthic invertebrate assemblages, the efficacy of the positive binomial is rare, unless the assemblage is regularly or uniformly dispersed (see Fig. I), in which case the variance is expected to be less than the assemblage mean. Holme (1950) cited an example of uniform spacing in a population of intertidal bivalves, which was possibly related to the foraging behaviour of the species. Wilson et al. (1977) described a uniformly spaced brittle star assemblage. Territoriality may result in a

194

B. J. BURD E T A L .

regular distribution of individuals within a species, but it is almost impossible to conceive of this occurring in multi-species assemblages. The Poisson distribution is a limiting case of the positive binomial, which occurs when the variance is approximately equal to the mean. This implies a random (rather than regular) distribution of individuals within a given area (Fig. 1). Taylor el al. (1978) suggest that Poisson distributions are unlikely in benthic assemblages, except at very low assemblage densities (cf. Clark and Milne, 1955). The negative binomial distribution is one of many possible models which can be used when the assemblage is clumped or contagious (i.e. aggregated: see Fig. 1) and the variance is greater than the mean. This distribution is described in some detail by Elliott (1977), who also mentions several other possible “clumped” assemblage models. Most assemblages will show some degree of clumping, since most abiotic factors affecting that assemblage are rarely uniformly or randomly applied. Apart from describing community structure, frequency distributions are useful for selecting a data transformation to produce a normal data distribution, which is a prerequisite for many parametric statistical analyses. Various methods described in most introductory statistics textbooks (e.g. probability plots, chi-squared goodness of fit, Kolmogorov-Smirnov test for distributions with mean and variance unknown: see Lilliefors, 1967), can be used to assess how well a theoretical distribution (normal, Poisson, etc.) fits the faunal data. It is not always feasible to fit a preconceived distribution to clumped assemblage data. Consequently, indices that measure the degree of clumping or aggregation have been used to describe the spatial distribution of these aggregations of species (for reviews, see Morisita, 1959; Taylor, 1961; Elliot, 1977; Taylor et al., 1978). One currently popular index (“b”) is given by the exponent in the power law discussed by L. R. Taylor (1961), by W. D. Taylor (1980) and by Taylor et al. (1978), which is discussed below. Taylor’s power law relates the sample variance (S2)to the mean faunal density ( 3 ) as described in Section III.A.2 (S2= aZb),where a is constant and b has been used as a dispersion index. As b increases, aggregation or clumping increases (although Taylor suggests that a is also an important indicator of dispersion). The value of this type of function is that it does not presuppose a specific frequency distribution model. This is important since, as L. R. Taylor (1980) pointed out: There is, in practice, no good biological reason to expect the same statistical frequency distribution to fit data for the same species at different (population) densities.

Downing (1986) examined the arguments for explaining the dependence of

BENTHIC MARINE INFAUNAL STUDIES

195

the value of h on the species or group being examined. He argues that b is not species-specific, but is environmentally related, supporting a demographic model for its application. Gage and Geekie (1973a,b) used an index similar to Taylor’s h to examine the deviation from the random (Poisson) to the spatial dispersion of fauna in a series of benthic samples in Scottish sealochs. They found that faunas from shallow, current-swept areas of muddy sands were more aggregated than deeper faunas found in the soft mud of quiet waters. Taylor’s h and other currently used indices of dispersion (see Elliott, 1977) are all sensitive to the size of the sample quadrat. Such indices are also sensitive to the number of replicates and sample coverage, and therefore should be used with caution when comparing faunas sampled using different methods (Elliott, 1977; Downing, 1986). This problem is related to the fact that the dispersion pattern may change as a larger and larger chunk of the substratum is examined. For example, in a small sample quadrat the assemblage may seem to be randomly dispersed, whereas a larger sample would reveal a pattern of high density separated by gaps of low density. Based on empirical data, Downing (1979) suggested a “universal” power law for all freshwater benthic taxa. The suggested value for Taylor’s b was 1.5. More recently, Vezina (1988) applied the power law to marine benthic assemblages and concluded that h was consistently close to 1.2 for all size groups and sediment types examined. Thus he concluded that marine benthos are generally less aggregated than freshwater benthos. Interestingly, the variance vs density relationship in marine benthos showed little variation between studies, suggesting that pre-sampling may be unneccesary in many cases. Figure 3 illustrates the comparison between the generalized relationships proposed by Downing and Vezina. The attraction of using Taylor’s power law to obtain some generalized “b” value is that this can be useful for deriving “variance stabilizing transformations” (Downing, 1979; and see Section III.A.2). Dispersion measures can be used to examine how representative the sampling pattern is with respect to the number of species and number of individuals in the assemblage. Measures of aggregation give a good indication of the evenness of the assemblage. This information can be used to predict the size of the sample unit required to collect a representative number of individuals, and the total number of sample units required to obtain acceptable within-site variance (Elliott, 1977; Downing, 1979; Holme and McIntyre, 1984). Downing (1979) utilized the relationship between variance and mean given by Taylor’s power law (Taylor, 1961) to devise an equation for estimating the sampling precision (=standard error) of the mean density of animals. The equation is based on the sample area to be collected, and the mean density of animals to be encountered at any given

B. J. BURD ET AL.

196

station, and can be used to determine the number of replicates required for the desired degree of precision. Downing (1979) and Downing and Anderson (1985) discuss the assumptions of the method. In general, to gain acceptable precision, many replicates must be taken when small samplers are used and when the densities of the animals are low. Holme and McIntyre (1984) discuss a similar method of estimating sampling efficiency (from Elliott, 1977), based on the sample mean, the degree of clumping, and the precision required. At an expected density of about 100 animals per replicate (regardless of sampler size), and b = 1.5 (according to Taylor’s power law), a precision of 20% (i.e. 20% uncertainty) would require a total of 12 replicates. At 1000 animals, the same conditions would call for only five

-n

4.0-

Q)

1.0

1.2

1.4

1.6

1.8

2.0

2.2 2.4

2.6

Log,o Mean Density FIG. 3. A comparison of the empirical “universal” density (X) vs variance ( S 2 ) power functions (S2 = aXb)proposed by Downing (1979: slope = 1.5) for freshwater benthic invertebrates and by Vezina (1988: slope = 1.2) for marine invertebrates (for a discussion, see Section 1II.C.I ) . The plots illustrate Vezina’s contention that marine benthic invertebrates tend to be less aggregated than freshwater ones.

BENTHIC MARINE INFAUNAL STUDIES

197

Using Downing’s method, a sampling program which includes high- and low-density stations will have to include enough replicates to sample adequately the low-density stations. Unfortunately, this may mean oversampling high-density stations if an equal number of replicates is taken at all stations, which has implications for the acceptance or rejection of statistical hypotheses about those stations (see Toft and Shea, 1983). Alternatively, for economic reasons, it might be advisable to consider the use of a different number of replicates or a different quadrat size at each station, so that standard error of the estimated mean density would be similar for all stations. It is conceivable that researchers might be uncomfortable with this method although it has an appealing logic. 2. Community structure In the preceding section, the spatial distribution of an assemblage was considered. This section is concerned with the species composition of the assemblage and the univariate approaches to examining community structure. This section discusses the concept of diversity and related issues which have resulted in a host of information indices, distributions and graphical displays. Many simple indices have been proposed and used, particularly for pollution studies (e.g. Satsmadjis, 1985). These are often misleading, or may be no more effective than the visual subjective conclusions of experts. (a) Diversity theories There are numerous definitions of diversity (see Washington, 1984, for a review) but a good general definition commonly used by ecologists is “a measure of the species composition of an ecosystem, in terms of the number and relative abundance of the species” (Legendre and Legendre, 1983). Specific definitions and related algorithms are numerous (see Washington, 1984). Hurlbert (1971) suggests that diversity is an abused “non-concept” in current ecological use. Certainly, there is considerable controversy about the validity of the use of diversity measures at all. The term “species richness” has commonly been used interchangeably with “diversity”, although most authors use species richness to describe only the total number of species in an assemblage. Dominance refers to the degree to which an assemblage is “dominated by” individual species. Therefore, an assemblage with many species of relatively equal abundance has a low degree of dominance. The inverse of dominance is “evenness”, so that an assemblage with low dominance will have high evenness. Diversity, therefore, combines the concepts of dominance, evenness and species richness, and is a measure of the relationship between species richness (number of

198

B. J. BURD ET AL.

species in the community) and the distribution of individuals among the species. Traditional ecological theory from terrestrial biology states that there are fewer numerically dominant species in species-rich communities than in species-poor ones. This seems to apply generally to tropical marine habitats and the deep-sea (for review see Birch, 1981), although there are numerous exceptions. Many general theories attempt to relate diversity to latitudinal and temporal gradients. Pianka (1966) described six such theories which are strongly interrelated. No attempt will be made here to discuss separate theories, but some of the factors believed to be fundamental in the maintenance of diversity will be described briefly in the following sections. (For a more recent review of diversity theories, see Washington, 1984.) Perhaps the most simplistic mechanism thought to control diversity is time. There is a persistent traditional belief among ecologists that, over time, assemblages become more complex, and therefore more diverse. This theory is somewhat difficult to test, and unlikely considering the wide range of other factors that is known to affect species richness and evenness, and which vary over time. Time is therefore not seriously discussed in relation to diversity maintenance in benthic communities. In resource-limited assemblages, intraspecific competition has traditionally been considered the most important mechanism in the evolution and maintenance of species diversity. Ecological theory from Hutchinson’s time states that competing individuals of a given species would be selected according to those traits that reduced intraspecific Competition, thus resulting in diversification and eventually speciation or geographic separation (competitive exclusion) of two closely related species (e.g. island biogeography: Simberloff, 1978). Similar niche separation between species may occur due to competition, although Levinton (1982) points out that there is surprisingly little evidence of niche separation even in intensely competitive species from marine benthic assemblages. Interspecific cooperation can also play an important role in assemblage structure (e.g. in food exploitation by tubificid oligochaetes: Brinkhurst, 1980). There have been opposing schools of thought on the importance of competition and cooperation as factors in community structure. Competition theory is in itself a complex issue beyond the scope of this chapter. Some ecologists state that competition per se is non-existent in certain communities, and that factors such as predation and mutualism are most important for sustaining community diversity and structure (for reviews see Levinton, 1982; Strong, 1983). Commit0 and Ambrose (1985) reviewed the roles of predatory infauna in the control of trophic assemblage structure of infaunal communities. The authors make the interesting observation that botanists and invertebrate biologists seem rather unimpressed by consider-

BENTHIC MARINE INFAUNAL STUDIES

199

ations of interspecific competition, whereas vertebrate zoologists are more inclined to seek such relationships. Perhaps this is because behavioural interactions can be more readily observed in vertebrate communities than in benthic infaunal assemblages. Several theories are briefly mentioned here because of their application to, and influence on, benthic ecological theory. In aquatic ecosystems, the most important aspect of diversity has traditionally been its relationship to stability. Unfortunately, it is not clear from the on-going arguments among ecologists that there is any relationship between diversity and stability, especially since few can agree on a clear-cut definition of stability. Coupled with this problem is the difficulty in measuring or proving the existence of assemblage equilibrium points around which species composition may fluctuate. Stability itself is defined in various ways, but perhaps the most commonly used definition originated with Margalef (cited by Smedes and Hurd, 1981): stability is the resistance of the community to change from stress caused by external disturbance. Such resistance may be different for different components of the assemblage (e.g. meiofauna and macrofauna). DeAngelis and Waterhouse (1987) review the related concepts of community stability and equilibrium. Despite the difficulties in definitions and measurements, theories have been put forward as to the factors which affect stability and diversity. Competition and cooperation have already been mentioned briefly as factors affecting diversity. Such biotic factors are notoriously difficult to measure. Theories describing abiotic (particularly disturbance and perturbation) effects on communities have had a particular influence on benthic infaunal ecology. A brief description of a few of these follows. The Stability-Time hypothesis was introduced by Sanders (1968, 1969), along with a univariate graphical method of depicting diversity known as rarefaction (see Simberloff, 1977, for a review of rarefaction methods in ecology). Rarefaction has since been used extensively to determine sampling efficiency in benthic studies. The theory states that an increase in benthic diversity within a community is established in two ways (paraphrased): ( I ) Short-term, non-equilibrium or transient diversity is induced by a lowlevel, unpredictable physical or biological perturbation or stress, resulting in biological undersaturation. In effect, small-scale disturbances produce empty patches which can be filled temporarily by nonequilibrium species, thus increasing diversity. (2) Long-term, equilibrium or evolutionary diversity is a result of past biological interactions in physically benign and predictable environments. This is seen most clearly in deep-sea habitats in which diversity usually increases with increasing depth (Rowe et al., 1982). Long and

200

B. J. BURD ET AL.

Lewis (1987) disputed the idea that diversity increases offshore in most cases. Pearson (1975) suggested that serial succession in polluted habitats is best explained by the Stability-Time hypothesis. The simplicity of the theory has been disputed by several authors, however. Thistle (1983a,b) and Shin and Thompson (1982) provided contradictory examples from marine benthic ecology. Abele and Walters (1978, 1979) and Josefsen (1981) introduced a different hypothesis which states that a species-area relationship is sufficient to account for observed patterns of species richness (diversity). This “area effect” means that variations in habitat heterogeneity, either physically (Abele and Walters, op. cit.; Kay and Knights, 1975; Probert and Wilson, 1984) or biologically induced (Josefsen, op. cit.), can account for variations in assemblage diversity. The larger the area sampled, the greater the effect. One aspect of Sanders’ (1968) Stability-Time hypothesis has received some attention in recent benthic ecological studies. The concept of maintenance of high diversity in communities by a patchy “mosaic” of disturbed sites with a predominance of short-term, non-equilibrium species, has led to the postulation that species richness in stable communities is maintained by a temporal mosaic of former disasters. Therefore, a community is a collection of relics and recoveries (Johnson, 1973; McCall, 1977). The patchy nature of the benthic environment promotes questions about the efficiency of sampling design in benthic studies (Thistle, 1983a,b). Thistle (1981) has reviewed this concept in the light of recent benthic ecological studies. If disasters occur frequently enough that there is reasonable expectation that one will occur within the range and lifetime of an individual opportunist, this is an important mechanism for maintaining diversity in a community in equilibrium. He pointed out that such patchy disturbances are caused by sources of pollution (see Pearson and Rosenberg, 1978) as well as natural causes (see Maurer et af., 1978). Other theories for modelling transient patches are discussed by DeAngelis and Waterhouse (1987). They point out that if the scale of a stochastic disturbance is smaller than the range of the biological assemblage in question, the disturbance will help to maintain diversity, since the ability to recolonize patches from surrounding areas will offset local extinctions, particularly for those species which are less competitively effective in successionally advanced locales. The quantification of this temporal mosaic phenomenon in benthic studies was discussed by Abugov (1982). He pointed out that the maintenance of competitively inferior (“fugitive”) species, by colonization of disturbed patches in a community, was dependent upon the frequency and spacing of patches. To measure this, he described a model of patch occurrence and developed a “phasing parameter” which represents the spatial and temporal

BENTHIC MARINE INFAUNAL STUDIES

20 1

environmental phasing rate of patches (see also Levin, 1984). He concluded that diversity was maximized at “intermediate” levels of disturbance. Rhoads and German0 (1986) have introduced a method for mapping successional mosaics on the sea-floor. The researcher uses a sediment-profile camera to monitor long-term changes in benthic community structure, or the effect and duration of patches. A more practical approach might be to study specific communities in which detailed information can be obtained from living and dead assemblages. Foraminifera would provide such a diagnostic tool, as their taxonomy is totally dependent on the shell or test (see Smedes and Hurd, 1981). The temporal mosaic theory has raised numerous questions about succession and recolonization in disturbed and normal communities. This has led to a series of laboratory and in situ patch experiments (cf. McCall, 1977, 1978; Leppakoski, 1975; Winiecki, 1986) to examine the time sequence involved in returning a bare patch to the same composition as the surrounding community. There are numerous discussions of these succession patterns and some interesting observations have emerged. For example, McCall ( 1977) pointed out that although the colonization sequence always proceeds with the same “types” of species groupings, the actual colonizer species will vary in any given area. Results of recolonization experiments in intermittently low-oxygen areas indicate that communities may never reach stability in an abiotically controlled environment (Leppakoski, 1975). The aforementioned theories about development and maintenance of diversity in benthic communities are only briefly touched upon here. The original sources should be consulted for detail. Many graphical methods and mathematical models have been developed for depicting diversity and related concepts. Once again, a small selection of these is included because they have strongly influenced the way benthic ecologists interpret community structure. For a historical review of these methods and theories, see Washington ( 1984). The following selection assumes implicitly that diversity can be expressed by any monotonous function having a minimum when all elements belong to the same class and a maximum when all belong to a different class. (b) Rarefaction methods Sanders (1968) introduced the use of rarefaction methods for describing the relationship between species richness and dominance, based on his StabilityTime hypothesis. Rarefaction involves the classification of entities of one hierarchical level into entities of a higher level. For example, one might relate number of species from a series of sample units to the number of genera. In the most common form, often called a species abundance curve, the number of species is plotted against the number of individuals. Each

202

B. J. BURD ET AL.

point in the plot corresponds to a sample unit. Typically, a power function is fitted to the data. This produces a characteristic curve describing the distribution of individuals over species in the set of samples (Fig. 4). In relation to the Stability-Time hypothesis, Sanders (1969) suggested that the flatter the species abundance curve the more the distribution approximates the idealized physically controlled community, whereas steep curves reflect biotically accommodated communities with large numbers of species per unit number of individuals.

Number o f Individuals or Sample Units (A); o r L o g Number o f Individuals or Sample Units ( 6 ) FIG.4. Portrayal of typical species abundance curves used in benthic studies to examine sampling efficiency. The tapering off of curve A represents the area where incremental sampling effort (i.e. more replicates or larger sample size) produces only a marginal increase in the number of species obtained.

Engen (1979) points out the confusion relating to the names of species abundance curves. Often they are called species area curves, although Holme and McIntyre (1984) distinguish species area curves as the cumulative number of species vs number of sampling units (or sampling area). This is still a form of rarefaction. Species abundance and species area curves have been used in the design of sampling programs (see Jumars, 1975; Section 1II.E. 1). It is particularly important to know the expected number of species in a sample drawn from low-diversity habitats or from areas where the density of animals is low (e.g. deep sea). In a species area rarefaction plot of a given assemblage (this can

BENTHIC MARINE INFAUNAL STUDIES

203

be one station or many depending on the assumptions made about the assemblage: see Fig. 4), sampling bias can be reduced by calculating the number of species for one sampling unit using the mean value of all the units sampled, then for two sampling units using the mean of the sum of all the pairs, and so on (see Holme and McIntyre, 1984). This should give some indica’tion of the relative increase in species coverage expected if the number of replicates is increased. Simberloff (1978) points out that a “best fit” power curve derived from rarefaction plots assumes a random spatial distribution of individuals, which is rarely true. The result is that the more clumped the assemblage in a community, the more rarefaction overestimates the number of species expected in a sample. By extrapolation, the larger the individual sample size, the less likely that clumping will affect the sampling results. Birch (1981) and Simberloff (1978) discuss the applicability and limitations of rarefaction methods in marine ecology. Simberloff points out that the calculation of a series of diversity indices (such as mentioned in the next subsection) for each sample unit (or station) is often not specific enough for community descriptions, and is therefore less valuable than plotting the data from a set of sample units as a simple rarefaction curve. (c) Diversity indices Other representations of diversity based on species richness and relative abundance of species have been suggested. Diversity indices can be calculated for each station or replicate, or may be calculated using pooled data from a number of stations grouped for some rational reason. Computation of a diversity index reduces each column of the data matrix (or group of columns) to a single number. Therefore, sites can be compared and sorted according to diversity, as long as the researcher takes into consideration the fundamental limitations in specificity of such measures. Birch (1981) discussed the use of diversity indices in benthic studies and cautioned that they assume that dominance decreases with increasing species richness. There are many marine situations where dominance increases with increasing species richness (see also Hurlbert, 197I , 1984), particularly in tropical areas. This controversy about the positive or negative relationship between species richness and dominance was further discussed by Rejmanek (’1 al. (1985), who concluded that the two are not linearly related except over small intervals of J ’ (evenness, the inverse of dominance), but are related by a quadratic function. At low values of J’ the relation is positive, and at high values of J’ the relation is negative (see Fig. 5). Unfortunately, the fit of this function to the data given by Rejmanek et al. (1985) is not too convincing. Other problems with diversity indices are illustrated by limnological studies. For example, the reduction in diversity and increase in dominance of

204

B. J. BURD E T A L .

certain tolerant “indicator” species below sewage effluents produces obvious shifts in diversity. Similar shifts might be produced by sampling in nearshore, surf-affected sandy bottoms instead of mud habitats just below the surf zone. Diversity indices might therefore fail to distinguish between communities with totally different taxonomic and environmental structures. Environmental deterioration that causes a drastic loss of diversity is readily detectable by diversity indices, as well as any number of other simple methods. More subtle but often seriously damaging factors can cause the substitution of one “suite” of species for another, or shift dominance from one taxonomic group to another with profound ramifications up the food chain. Diversity measures may overlook such changes entirely.

FIG. 5. Quadratic relationship between evenness and species richness (from Rejmanek el al., 1985), showing that at low values of evenness the relationship is positive, and at high values of

evenness the relationship is negative.

Pielou (1 969) discussed the relative merits of different diversity indices, including the most commonly used in benthic ecology, the Shannon-Weiner H’ (described by Shannon and Weaver, 1963) and Margalef‘s “d” (Margalef, 1958). The Spearman rank correlation coefficient is often used to compare dominance (1 - J’,where J‘ is evenness- Pielou, 1966; 1 - J’ = 0 where there are no dominants; and 1 - J‘ = 1, where there is only one species) with species richness ( S ) .There are other diversity indices which are not commonly used in the benthic literature. Washington (1984) provides a discussion of terminology and historical perspective on the development of the many different indices. An alternative way to depict diversity developed from empirical plots of the distribution of individuals among species (grouped in a geometric series

BENTHIC MARINE INFAUNAL STUDIES

205

of abundance classes) vs number of species. The entire plot will normally represent one station (or the combined data from a group of stations assumed to have a common diversity). Such distributions are sometimes called species abundance curves, but should not be confused with rarefaction curves, since they do not involve the classification of one hierarchical level into another, and usually represent only one sample unit or station. A theoretical distribution (such as normal or Gaussian) is often fitted to the aforementioned distribution (which is often referred to as a “diversity distribution”). The fitted functions can then be compared for different stations, times, etc. Fisher et al. (1943) suggested the general applicability of the log series distribution for describing such abundance distributions. Preston (1948, 1962) discussed the utility of the canonical log normal distribution for this purpose. The latter has gained more empirical support in benthic studies than the log series distribution (see Engen 1978, 1979). Gauch and Chase (1974) compare methods for fitting a Gaussian curve to abundance data. Among these is the method of least-squares estimation, which is described in many statistical packages and textbooks. Iterative maximum likelihood methods may be commonly used for this purpose in the future. The log normal distribution is based on the observation that the distribution of the number of species vs individuals per species (in geometric classes) often gives a truncated normal distribution, with more of the normal curve evident as sample size increases (Preston, 1962, 1980). The portion of the curve evident in a given sample is bounded by the veil line (as illustrated in Fig. 6). Plotted on a scale of percentage of total species or cumulative species, the log normal distribution produces a straight line. The slope of this line indicates the comparative degree of dominance vs species richness for the sample. In their search for biological justification of the log normal distribution, Ugland and Gray (1982) concluded that most communities are not actually based on a single log normal distribution. Instead, rare species, moderately common species and abundant species produce a mixture of three or more log normal distributions. In many plots of individuals per species vs number of species, these separate distributions are evident (see Gray and Pearson, 1982; see also Fig. 7). This points out the importance of ensuring that comparisons of log normal distributions are based on the same total abundances in all cases (Hurlbert, 1971). Otherwise, the percentage of rare species obtained will be different and the position of the veil line will not be comparable. Like other diversity indices, the log normal method has been criticized because it assumes that dominance declines as species richness increases. In some tropical ecosystems, there can be many species with high abundances

206

B. J. BURD E T A L .

(see Birch, 1981), producing a severe departure from the log normal curve. Also, the more aggregated an assemblage, the poorer the log normal approximation to the data (as with rarefaction).

Abundance Class (number p e r species) 6 . Conceptual illustration of the canonical log normal distribution of number of species as proposed by Preston (1962). The veil line illustrates the cut-off point beyond which most sampling programs fail to obtain data. FIG.

I’S abundance

Assuming that log normal distributions provide an adequate model for data from some benthic assemblages, changes in a given distribution over time or space may be used to study the effects of organic pollution on benthic community structure. Ugland and Gray (1982) suggested that benthic assemblages are really made up of a mixture of three or more abundance distributions which can be approximated by symmetrical (discrete) binomial functions. The binomial functions are usually closely overlapped in undisturbed or equilibrium communities, so that the resulting overall distribution closely fits a single log normal function (see Ugland and Gray, 1982, Fig. 1). However, these separate binomial functions tend to spread apart in organically enriched communities, thereby causing a deviation from the log normal distribution (see Fig. 7). Gray and Pearson (1982) point out that this traditional method of assessing the fit of a log normal curve to benthic assemblage data (see Fig. 8; see also Ugland and Gray,

BENTHIC MARINE INFAUNAL STUDIES

207

1982, Fig. I ) are insensitive for separation of the component binomial distributions inherent in benthic assemblages, particularly in polluted areas. Gray and Pearson (1982) prefer to use simple plots of number of species vs abundance per species (as in Fig. 8) to delineate clearly the multiple peaks in abundance of benthic assemblage data. They use this latter plotting method to try to select objectively pollution indicator species which are “moderately common” in polluted areas (the second abundance peak: classes IV-VI in Fig. 8). Their rationale for using these species as indicators seems to be mainly that they are not ubiquitous, and are therefore discriminatory (see also Pearson et al., 1983).

VI

I

I1

Ill

IV

v

VII

Vlll

VI

FIG. 7. An illustration of the overall log normal functions calculated for cumulative (or probits) species. from several years of time-series data from Loch Eil, Scotland (from Gray and Pearson, 1982, Fig. 4) 1963, prior to pollution; 1967, 1 year after pollution began; 1971 and 1973, 5 and 7 years after pollution began. The x axis represents geometric classes of abundance per species. The flattening of function lines for cumulative log normal distributions of species per abundance class is postulated to be related to progressive pollution at the site. The break in several function lines suggests that there are at least two separate log normal distributions in the abundance data from 1967 and 1971. Therefore, the use of an overall log normal distribution for this data is invalid.

Rygg (1986) indicated that the log normal distribution is not a valid model for areas of heavy metal pollution, since the abundance of all species tends to drop off as pollution increases, and no dominants emerge. Rygg (1985a) examined the effects of various heavy metals on diversity of communities and suggested the use of the diversity indices E(Sn) (=expected number of species per 100 individuals) and H‘ (Shannon and Weaver, 1963) to identify groups of negative indicator species, or those species groups most likely to disappear along a variable gradient of pollution (Rygg, 1985b). The system apparently works in industrial pollution conditions as well as organic, although the levels of diversity indicative of different levels of pollution must still be arbitrarily selected.

B. J. BURD ~ A L .

208 30

2or

A

1967

v)

x

!! n.

I

11 Ill IV

A

v

V I VII

"[ 10

1973

d

INDIVIDUALS PER SPECIES I N GEOMETRIC CLASSES

FIG.8. Simple plots of number of species YS individuals per abundance class (instead of log normal plots) more clearly delineate the separate (binomial) abundance distributions of rare, moderately abundant and common species groups in the Loch Eil data shown in Fig. 7 (from Gray and Pearson, 1982, Fig. 5). Gray and Pearson (1982) suggest that species from the second abundance group (the peak of moderately abundant species) can be used as pollution indicators.

Some authors have contended that the log normal distribution does not fit the diversity patterns of many soft-bottom assemblages, impacted or otherwise (cf. Platt, 1985; Hughes, 1984). Nelson (1987) examined an extensive number of marine studies and found frequent examples of considerable variation from the log normal and log series distributions. Preston (1980) suggests some potential data problems which may cause deviations from the canonical log normal distribution. Hughes (1984) has suggested that assemblage growth in benthic communities is often arithmetic (instead of geometric as assumed by the log normal method) because of limiting factors such as predation and environmental influences. Therefore, many diversity distributions become increasingly concave rather than linear (as in log series) or truncated (as in log normal), as the number of rare species increases (especially in the tropics). Schmidt and Garbutt (1985) suggested that the Gamma distribution may be useful for fitting concave diversity distributions without a mode (as described by Hughes, 1984). They found that out of 136 marine fouling communities tested, 128 conformed to the Gamma distribution. Other distribution models have been suggested for describing communities, including the Poisson-Inverse Gaussian distribution (Ord and Whitmore, 1986). None of these distributions addresses the basic limitations inherent in a univariate abundance model.

BENTHIC MARINE INFAUNAL STUDIES

209

3. Biomass distributions Benthic assemblage structure can be described in terms of weight instead of abundance data, although there are problems related to the use of biomass as a quantitative measure. This section describes a few such methods. Biomass is used as a general term in benthic studies, but can refer to a wide range of measures of weight (see Section 1II.A). Unfortunately, biomass data are usually more variable than abundance data, both within and between samples. Long-term studies of the relationship between biomass and benthic community structure are relatively rare. One such study was undertaken over a 6-year period by Moller et af. (1985) at 15 different shallow soft-bottom locations on the Swedish west coast. Of increasing interest to benthic ecologists is the use of combined abundance and biomass descriptions of community structure. In plots of biomass concentration as a function of logarithmic intervals of organism size (i.e. no species identifications), three biomass peaks are evident (Schwinghamer, 1981, 1983). It turns out that the minima between peaks effectively separate the grain surface dwellers (bacteria) from the interstitial organisms (meiofauna), and the meiofauna from the macrofauna. These peaks in size distribution are surprisingly stable spatially and temporally, and vary only in relative magnitude under different environmental conditions (particularly sediment type: Schwinghamer, 1983). Gerlach et al., ( 1985) confirmed Schwinghamer’s ( 1 98 1, 1983) results, converting biomass to organic carbon values, and postulating metabolic peaks for the meiofauna and macrofauna. The authors also point out problems with lumping the foraminifera in with meiofauna in such size distributions. Schwinghamer ( 1983) proposed some causal theories to explain interactions among size groups, and between size groups and environmental variables, and later (Schwinghamer, 1988) suggested that there may be some relation between the size structure of benthic communities and their proximity to pollution sources. The size distributions proved to be less consistent than taxonomic ones, and may be useful only for baseline monitoring functions where taxonomic analysis is not feasible. Warwick (1984) differentiated between meiofaunal and macrofaunal groups using plots of log normal frequency distributions of species body size (dry weight), and suggested that the differentiation between bacterial, meio- and macrofauna is the result of evolutionary optimization, so that sizes in between these groups are inefficient. Platt ( 1 985) reviewed the ecological concept of “Trophic Level Formalism” (see also Schwinghamer, 1981), which is a model for describing the relationship between organism size and its position in the trophic structure. Some of the early theories on which the model is based were expressed by

210

B. J. BURD E T A L .

Sheldon et al. (1972, 1977) and the concept was formalized into a descriptive model for pelagic ecosystems by Platt and Denman (1977). This model is best represented by an Eltonian Pyramid of animal abundance vs animal weight, or an inverted pyramid using biomass vs animal weight (described by Sanders, 1960). The model incorporates such variables as size of food and respiration. Assuming that large quantities of organic matter are not tied up in the sediment, so that all available organic material is in use, the total primary productivity of a given community should be equal to the sum of the biomass of all the different trophic groups within that system. This causal analysis approach to the examination of the dynamic interaction of the members of benthic communities should be of value in the study of benthic productivity and its relationship to primary production and fisheries. The development of effective theories and models to relate benthic production to fisheries is a necessary step for the management of fisheries stocks (Mills, 1975). A different approach to abundance and biomass descriptions was introduced by Warwick (1986, 1988a,b; Warwick er al., 1987) for analysing pollution gradients. The method, termed ABC, is based on the assumption that as pollution disturbance increases, the large dominants in the normally stable assemblage decline in biomass and abundance (see Fig. 9). Simultaneously, the smaller opportunists increase in biomass and abundance. In polluted stations, total biomass decreases with respect to total abundance. Plots of the relative biomass and abundance distributions illustrate normal, moderate and grossly polluted stations. However, this method has two important drawbacks. First, the method assumes that the assemblage was initially stable or in equilibrium, a condition rarely true in environmentally controlled habitats, e.g. tidal flats (see Beukema, 1988) or estuaries. Secondly, the moderately polluted stations which are often of greatest interest, produce the most ambiguous results. This is a common problem with most current pollution indices. Warwick (1988a) suggested that the use of taxonomic groupings higher than species may partially alleviate the first problem, since the higher groups are less sensitive to natural habitat factors than individual species. This proposition is arguable, and would require convincing verification. All of the methods mentioned in this section require greater scrutiny. Schwinghamer (1983) points out that size comparisons in benthic assemblages cannot supplant taxonomic descriptions for characterizing communities, but that the two approaches can be complementary. This combined approach is still relatively rare in benthic ecology. In-depth productivity studies are even rarer than simple biomass studies. The determination of production or productivity requires not only information about flow rates of organic biomass between species in the assemblage, but also about the trophic relationships between them. A review of

21 1

BENTHIC MARINE INFAUNAL STUDIES

considerations and methods for analysing energy flow or rates of change of biomass through benthic systems is given by Crisp (1984). 100

50

w o

D 1

I

-

0

t

-

E

G I

5

5

10

1 10

RANK FIG.9. Plots of macrofauna data aggregated at the family level (from Warwick, 1988b. Fig. 3) illustrating the relative changes in abundance (squares and solid lines) and biomass curves (crosses and broken lines) over varying pollution conditions.

212

B. J. BURD E T A L .

D. Statistical Inference 1. Hypothesis testing Historically, there has been little emphasis placed upon hypothesis testing in benthic studies. For the analytical methods described in Sections III.C.2 and III.C.3, statistical hypotheses are not required. The only situation in which statistical inference might be applied is in the estimation of an assemblage or sample parameter (e.g. mean abundance of a given species). In that case, bias, standard errors or confidence limits are important concepts. Also, the “goodness of fit” of the data to a certain model or distribution may be tested. In such a case, the hypothesis being tested is straightforward, and the results are readily interpretable. Methods of estimation are included in all introductory statistical textbooks and will therefore not be discussed further. Statistical hypotheses are rarely used for outlining and designing benthic studies. General and often vague objectives have been the norm in most benthic studies. Ecologists are becoming increasingly conscious of the importance of stating and testing clear and reasonable hypotheses, since the reliability of the sampling design, analysis and results are so dependent on this. This is particularly important in multivariate analyses and sampling designs, as the complexity of results can make subjective interpretations difficult. Recently, several papers have been published (Roughgarden, 1983; Simberloff, 1983; Quinn and Dunham, 1983) which discuss the rigour and application of hypothesis testing in mensurative (survey) ecological studies. In practical terms, it is difficult to test independent and mutually exclusive hypotheses in a multi-factor system, in which each factor may have a proportional or partial effect on the hypothesis being tested. Therefore, strict falsification by independent tests of hypotheses concerning any one factor is impossible. Also definitive statements about causality in a multivariate situation must assume that every alternative hypothesis has been identified and rejected. In many cases, it is impossible even to identify every alternative hypothesis (concerning biotic and abiotic factors). Furthermore, one cannot control environmental fluctuations between replicates in survey studies, so that generalizations beyond the sample are exceptionally difficult. In practice, therefore, most hypotheses in benthic studies involve making predictions about the degree of effect or the probability that factors are affecting a community or organism. The veracity of the test depends on a number of factors, including clarity and discreteness of the hypothesis, adequacy of sampling, conformity of data to the underlying probability distribution (i.e. normal, chi-squared, etc.) and other assumptions on which the test is based. Throughout Section III.E, commonly applied inferential tests will be discussed in connection with multivariate data analyses.

BENTHIC MARINE INFAUNAL STUDIES

213

2. Statistical power There are two types of error ( I and 11) associated with inferential hypothesis testing. The probability of making a type-I error (a), is the probability of mistakenly rejecting a true null hypothesis. In practice, it is usually set at a value between I and 10%. However, the probability of making a type-I1 error (p) is the probability that a false null hypothesis has not been rejected. The complement of p (1 - p) is equivalent to the power of the statistical test, which indicates the reliability of the statistical result. The probability of a type-I1 error is not commonly calculated, and neither are standard acceptable levels (20% is considered reasonable) in use. The power of a statistical test is valuable to know when the value of a is being used to measure the degree of an effect. If p is known to be very low, one can confidently use a as a measure of effect. Also, the unexpected failure to reject a null hypothesis may sometimes be explained by the value of p. An explanation of the use of power in ecological research is given by Toft and Shea (1983). Some contingency tables for p at different levels of a are given by Cohen (1977) for standard univariate statistical tests. Several factors affect the size of p: these include the value of a, the sample size, and the magnitude of the critical effect being measured by a. Obviously, the more stringent the requirement for a (i.e. 1YOor less), the more likely the probability of making a type-I1 error and therefore the lower the power of the statistical test. The larger the sample size, the lower the probability of making either type of error. Fortunately, most benthic studies have reasonably large sample sizes. The “effect size” or magnitude of the effect being measured is important, since the stronger or more obvious the effect to be measured, the greater the power of the test. This also means that with large sample sizes, even trivial effects may be statistically significant. Obviously, what constitutes an important biological effect is somewhat arbitrary, and standard guidelines are not often available. Cohen (1977) has made an attempt to standardize effect sizes for different types of application. The calculation of statistical power is complicated except in very simple inferential tests of univariate hypotheses (e.g. ANOVA, t-test), since the underlying probability distribution of the data must be taken into consideration. In multivariate hypotheses, the calculation is extremely complex and unwieldy. Therefore, such calculations are rarely practical for multivariate tests. There is, however, some hope that bootstrap methods can be used to estimate the power of multivariate tests (see Beran, 1986). Such methods have not yet been applied in benthic community studies, but are appealing because they are distribution-free. The evaluation of the power of a statistical method is the only statistically correct method for determining the proper number of replicates and samples required in order to use correctly statistical inferential methods. For

214

B. J. BURD E T A L .

example, an undersized sample can make it difficult to detect departures from the null hypothesis. On the other hand, very large samples tend to detect even small and inconsequential departures from the null hypothesis. In this way, a priori decisions about “effect size” or the magnitude of departure from the null hypothesis that is worth testing, can help determine sample size. E. Multivariate Data Analyses Each column or sample unit of the faunal data matrix (see Section 1II.A) can be considered as a multivariate dependent variable. The inherent limitations involved in univariate community descriptive indices as discussed in Section 1II.C are:

( I ) The exclusion of information on actual taxonomic content of data, which makes it difficult to compare effectively two different communities, or to monitor temporal changes in a non-polluted community structure. (2) The difficulty associated with fitting multivariate assemblages into preconceived univariate distributions. (3) The difficulty in examining the relationship among more than two variables (faunal or environmental variables) using multiple univariate comparisons (both because of escalating error and interdependence between comparisons). Because of these limitations, benthic marine ecologists have borrowed and adapted multivariate descriptive methods which were introduced in terrestrial ecological studies. Multivariate statistics are not described in introductory textbooks, because they are considered to be advanced topics in statistical studies. Mills (1969) described the two basic groups of statistical analysis that have developed out of the “community” and “continuum” viewpoints, as “classification” (grouping of stations or species according to relative similarity) and “ordination” (distribution of stations or species along a small number of ordinate axes which reduce the dimensionality of the assemblage as much as possible). The distinction between classification and ordination is somewhat artificial, as they are usually just different graphical manipulations of the same data analysis, and many researchers use a combination of methods to show trends in the data. In fact, one is usually most confident in the robustness of results if several descriptive and inferential methods lead to similar conclusions.

BENTHIC MARINE INFAUNAL STUDIES

215

1. Similarity indices

A similarity (or, inversely, distance) coefficient measures the similarity between the community structure of any two sample units. For a faunal data matrix (see Table l), this means that a similarity coefficient can be calculated for each pair of sample units or replicates (columns). The collection of pairwise similarities is typically summarized as a symmetric similarity matrix. This is the most common usage and will be described in Sections III.E.2 and 3. Alternatively, similarity indices can be used to compare each sample unit (or station) with some reference value (site) (see Pontasch and Brusven, 1988). There are many kinds of similarity measures in common use in benthic studies, which include: the Bray-Curtis coefficient (Bray and Curtis, 1957alternatively called index of affinity, percentage similarity or Czekanowski coefficient); the Canberra-Metric (Lance and Williams, 1967); Jaccard’s index (Jaccard, 1908- presence/absence data only); Steinhaus’ coefficient (refer to Motyka et al., 1950); the Zurich-Montpellier index (Kuchler, I967 - presence/absence data only); Normalized Expected Species Shared (NESS), which can be used to study the importance of dominant and rare species and is a generalization of Morisita’s Index (refer to Lopez-Jamar, 1981); Fager’s index of Affinity (Fager, 1967); and Euclidean distance measures, such as Orloci’s index (Orloci, 1975) or the Manhattan metric (refer to Legendre and Legendre, 1983). Reviews of similarity methods and their various advantages and shortcomings include Clifford and Stephenson (1975), Legendre and Legendre (1983), Washington (l984), Cormack (1971), Williams (1971), Goodall (1973) and Green and Vascotto (1978). Unfortunately, it is difficult to identify an appropriate (justifiable) set of criteria by which comparisons can be made. Thus researchers have resorted to comparisons based on the analysis of real data and what the results “should” look like (Pontasch and Brusven, 1988), studies using simulated data, e.g. Bloom, 1981, found that the Canberra-Metric, Morisita’s and Horn’s Information Theory all diverge greatly from the theoretical standard: only the Bray-Curtis ( = Czekanowski) coefficient accurately reflected predicted similarity); or the application of intuitive criteria such as dependence on sample size (e.g. Kobayoshi, 1987). The problem has been compounded by the fact that not all similarity measures have the same range. Nevertheless, ecologists have attempted to identify the “best” similarity measures for specific applications to benthic faunal data. Field et al. (1982) recommend the Bray-Curtis coefficient (after appropriate transformation of the data) as a good all-round similarity measure for abundance data, but suggest that Canberra-Metric may be more useful for biomass data since it weights all species equally, thus

216

B. J. BURD ET AL.

avoiding gross skewing of the data by a few large specimens. In speciesimpoverished areas, the Manhattan metric may be appropriate, since it is a measure sensitive to the total number of species present. The variation in emphasis of different similarity measures points out the difficulty in comparing benthic studies which use different indices. Poore and Rainier (1979) compared the two most commonly used measures, the Canberra-Metric and Bray-Curtis coefficients. Grassle and Smith (1976) compared the BrayCurtis, Canberra-Metric and NESS similarity methods and preferred the last. Popham and Ellis (1971) compared the Jaccard and Zurich-Montpellier indices (presence/absence data only), and decided that the latter provided more information with which to distinguish species and atypical samples. Smith et al. (1988) point out some of the limitations of similarity matrices, particularly at very low and very high levels of similarity between sample units (studies in which a diverse sample set is being examined). For example, dissimilarity indices can be skewed along a strong environmental gradient as dissimilarity approaches 0 and loo%, so that the indices change much more slowly at these extremes than in the middle of the distribution of values. In data sets where these extreme values are never approached, this is not a problem. In cases where the data set is diverse, Smith et al. (1988) recommend several modifications to correct this, including the use of a “stepacross” procedure originally described by Williamson (1 978), which results in a matrix which may include dissimilarities greater than 100%. This suggests that the scale of measurement has been changed, which is problematical since dissimilarities greater than 100% are only meaningful if it can be assumed that some strong environmental factor is affecting assemblage distribution in a relatively simple (i.e. linear) way. Otherwise, how is it possible to interpret dissimilarities greater than loo%? An alternative method for transforming the raw dissimilarity indices to avoid skewness utilizes a dissimilarity measure coined as ZAD (cf. Mahon et al., 1984). Similarity matrices have also been used to examine sampling efficiency. The species abundance (or species area) curves described in Section III.C.2.b) are of limited value for determining sampling coverage because of their univariate nature and the underlying assumption that the distribution is random. To overcome this problem, a method has been developed which uses similarity (presence/absence only) in a similarity/area curve to determine the sampling effort required to obtain an acceptable percentage of species (for discussion of similarity area curves, see Kronberg, 1987). Weinberg ( I 978) compared a qualitative similarity index (presence/absence) with a quantitative one (abundance) to determine community minimal area for sampling.

BENTHIC MARINE INFAUNAL STUDIES

217

2. Classification If the data set comprises sample units from discrete community structures (whether a sampling artefact or not), a classification approach may be used to separate the sample sites into a moderate number of clusters. Cluster analysis is an objective method for grouping the objects (sites or species) according to similarity of community structure. The main advantage of this approach is its simplicity of interpretation (if the clusters are sufficiently distinct). Clustering methods which hierarchically sort dissimilarity matrices fall into four categories: agglomerative, divisive, constructive and direct optimization (for a review, see Gordon, 1987). The most commonly used of these is agglomerative, hierarchical cluster analysis, which involves the combination of a similarity or dissimilarity coefficient matrix with a linkage rule. The linkage rule defines a measure of the similarity between two arbitrary groups of sites, building hierarchically until all groups are linked as one cluster. Boesch (1977) pointed out that divisive methods are theoretically more promising but computationally expensive for large databases, so that methods have not been adequately developed for use in ecology. Standard linkage rules include: single linkage (the chaining property of single linkage is undesirable in most applications); complete linkage; unweighted or weighted pair group mean average sort (Sneath and Sokal, 1973), which joins two groups at the collective average similarity level for each group and produces little distortion of the actual resemblance relationships (Boesch, 1977; see also centroid clustering); flexible sorting (Lance and Williams, 1967; Clifford and Stephenson, 1975; Boesch, 1977); proportional link linkage (Sneath, 1966; see Legendre et al., 1985); and a Euclidean method (Gordon, 1987). Attempts to combine the results from a number of different clustering schemes do not, in general, provide a satisfactory method (cf. Stephenson et al., 1972), unless a “consensus” method is being applied (see Gordon, 1987) to test the robustness of results. (For reviews of linkage rules, see Sneath. 1966, Cormack, 1971 and Gordon, 1987.) Interpretation of a cluster analysis is usually done by optimally rotating the results of the hierarchical linkage and plotting a dendrogram (see Fig. 10). In the plotting of the linkages from a hierarchical classification, optimal rotation simply refers to the rearrangement of stations about the linkage nodes to avoid crossing branches in the resulting dendrogram (this is a geometric rotation which results in no mathematical transformation of data). There is some loss of information using dendrograms derived from a cluster analysis, since the multivariate similarity between sites has been reduced to a single number by the linkage method (i.e. the similarity between any pair of sites cannot be determined from the dendrogram and similarity

218 1.o I

B. J. BURD E T A L .

0.1

0.5 I

I

I

I

I

I

I

I

CRUISE 1

I

1 A1

-2 9 % 2

A4

28 %

-

A2

9%

r

A5

7%

- 11%

tl %

I

l!O

I

I

0!7

I

D2 43% 0!5

I

I

0!3

I

Oll

BRAY -CURTIS DISSIMILARITY INDEX FIG. 10. Cluster dendrogram of benthic faunal abundances from the coastal shelf of Vancouver Island off Barkley Sound (from Brinkhurst, 1987), with significances of clusters calculated using a bootstrap method (Nemec and Brinkhurst, 1988a). Values of 5% or less represent significant clusters, meaning that the stations within that linkage are homogeneous (interchangeable), but the entire group is significantly different from all other groups.

scale alone). Measures are available for assessing the loss of information in a cluster diagram, though this is rarely done in benthic application studies (e.g. “cophenetic correlation coefficient” used in numerical taxonomy: see Sneath and Sokal, 1973; Gordon, 1987). The skewed nature of the extreme

219

BENTHIC MARINE INFAUNAL STUDIES

values in some similarity matrices (discussed by Smith et al., 1988) can be lessened using cluster analysis, since it tends to average the similarity between stations for hierarchical clustering, and because the intermediate levels of comparison are usually of more interest in the interpretation of cluster analyses than the extremes of 0 and 100%. A map in which successive similarity levels are shown as concentric rings is often useful for depicting station groupings (see Fig. 1 1 ) and can be a valuable management tool.

12(r000’

125’00’

FIG. 1 I . Concentric station clusters mapped to illustrate the significant station groupings ( P< So/,) for the follow-up cruise to that shown in dendrogram format in Fig. 10. The time elapsed between cruises was 5 months.

One of the main disadvantages of classification methods has been the lack of an objective criterion for determining the number of legitimate clusters. Until recently, most authors have arbitrarily selected a preferred optimum similarity level to determine the number of groups. The number of clusters must be sufficiently large so that the important differences in community structure are captured by the groupings, but not so large that the differences between clusters are comparable to those seen among the replicates drawn from each site. The availability of replicates is therefore an important factor in making subjective judgements about cluster groups. The relatively new non-parametric “bootstrap” method (Efron, 1982; Diaconis and Efron, 1983; Efron and Gong, 1983; Felsenstein, 1985) has recently been applied to the problem. Nemec and Brinkhurst (1988a,b) describe a bootstrap method that can be used to assess the “statistical significance” of clusters, provided that replicate samples are available. The

220

B. J. BURD ET AL.

method tests the hypothesis that the two station groups that are joined at a particular linkage level are the same (see Figs 10 and 1 I). A similar method has been used to compare two different dendrograms by testing the hypothesis that the two dendrograms are the same (i.e. the sample units from both clusters can collectively be considered replicates from a single community) at any given linkage level (for applications of this method, see Burd and Brinkhurst, 1987; Brinkhurst, 1987; Brinkhurst et al., 1987). Strauss (1982) used a non-parametric approach based on a randomization method, which is suitable for the analysis of presence/absence data and compares the observed linkage levels with the linkage levels for the randomized data matrices to test for significant clusters. Raup and Crick (1979) have also defined a similarity measure for inferential testing of presence/absence data which is applicable in palaeontology studies. They outline a probabilistic (counting) method for comparing two communities and placing a probability on their different structures. Smith et al. (1986) describe a bootstrap method for producing confidence intervals for the similarity between two algal communities. Clarke and Green (1988) discussed a randomization method for testing the significant differences between sites or sets of sites in multispecies data. Unfortunately, these latter two methods are only valid for pair-wise comparisons and may therefore suffer from multiple comparison problems. Once the set of representative clusters has been decided upon, it is often desirable to determine which species, or groups of species, are useful for characterizing the clusters. Indicator species are often selected subjectively. An apparently objective method of identifying indicator species is to compare the mean (relative) abundance of each species across the clusters, using a series of one-way ANOVA F-tests (one for each species) or a set of “pseudo F-tests” (Mirza and Gray, 1981). Those species that exhibit “significant” differences across the clusters are considered useful for discriminating between clusters (cf. Shin, 1982; Shin and Thompson, 1982; Field et al., 1982). Some studies mention the potential pitfalls of such an approach (e.g. violation of the underlying normal assumptions, multiple comparisons problem using univariate tests). Field et al. (1982) also mention the use of an information statistic and chi-square analysis for distinguishing important species. R-mode (inverse) cluster analyses are sometimes used in concert with a Q-mode (station) cluster analysis. The species matrix is then physically rearranged so that the sites are aligned according to the results of the Qmode analysis, and the species according to the R-mode analysis. This is commonly referred to as a “Two-way coincidence table” or “Nodal analysis” (Boesch, 1977), and can be useful for spotting misclassifications, which can occur during cluster analysis. An example of this is given in Smith et al. (1988). The rearranged data are visually examined for trends (cf. Hughes

BENTHIC MARINE INFAUNAL STUDIES

22 1

and Thomas, 1971a,b; Smith and Greene, 1976; Flint and Holland, 1980), but clear results are rare.

3. Ordination Each column (sample unit or station) of the data matrix corresponds to a point in a space with dimensions equal to the number of rows (species). If the columns can be represented as points in a space with considerably fewer dimensions, the representation is loosely referred to as an ordination. Each dimension in the reduced space then represents an “ordination co-ordinate”. An ordination co-ordinate for a given sample unit may or may not be a simple (linear) transformation of the species abundances, depending on the method of ordination used. Like cluster analysis, ordination resulis in some information 1 0 s which is directly related to the extent to which the original number of dimcilsions is reduced. Hence the community structure of the sites will be described to a greater or lesser degree by the ordination coordinates. If the number of ordination co-ordinates is two or three, the relationship among the sites can be seen by plotting the sires in a plane or in a three-dimensional space. If the community structure of the sites can be classified into discrete classes, the plot is expected to show a clustering of the sites (Fig. 11). If the data form a continuum, the sites will be spread out over some region defined by the ordination. Ordination can be used to examine data which is clustered into discrete communities, though the results are often more difficult to interpret than classification methods. However, if species data do not form discrete clusters but conform more to a continuous distribution, the data should be examined using ordination rather than classification methods. This occurs most commonly in nearshore and estuarine areas and some polluted areas with dominant physical factors. Ordination methods include a variety of different types of analysis, some of which have been in common use in terrestrial ecological studies for many years (see Whittaker, 1967). There are two broad classes of ordination: metric multidimensional scaling and non-metric multidimensional (or ordinal) scaling. Metric dimensional scaling refers to any method in which the distance between data points can be approximated by Euclidean co-ordinates, i.e. it is assumed that the dissimilarities between any pair of objects can be approximated by a metric distance (in a lower dimensional space). It does not necessarily require that the distance or dissimilarity index itself be metric (for assumptions, see Chatfield and Collins, 1980). Non-metric methods do not attempt to approximate the actual dissimilarities between pairs of objects, but rather preserve only the agreement between the rank order of the pair-wise dissimilarities.

222

B. J. BURD

Er AL.

Many of the common ordination methods utilize one of the standard dissimilarity measures described in the previous section. Shin (1982) points out that a “continuum” method such as ordination should only be applied to data sets that are relatively homogeneous (i.e. few zero records). Many authors use data reductions or primary transformations to satisfy this requirement (see Section IILA), but in so doing may introduce problems of interpretation. Mills (1969) indicated that more effort in sampling and analysis is required to perform properly gradient or ordination than is required for classification analyses. Also, some ordination methods assume monotonicity of the response curve of the species with respect to environmental factors, or may even assume linearity. Multidimensional scaling methods include: Principal Components Analysis; Principal Co-ordinates Analysis or Gower’s method of principal coordinates (Gower, 1966), which was originally formulated by Torgerson (1952, 1958); non-metric multidimensional scaling; and Correspondence Analysis. Legendre and Legendre (1983) describe all of the above ordination methods, and provide examples. Also, a little-used somewhat subjective ordination method was described by Bray and Curtis (1957) and has been discussed further by Shepard (1980) and Beals (1984). Factor Analysis refers to a specific but different technique which is arguably a form of ordination. Unfortunately, “factor analysis” is often used in the benthic literature as a catch-all term for ordination. Principal Components Analysis (PcpA) is a specific application of “metric” multidimensional scaling (for assumptions of metric distance measures, see Legendre and Legendre, 1983). PcpA preserves the multivariate (Mahalanobis) Euclidean distances between the sites (Q-mode analysis) with the restriction that because this transformation is linear, the distances between sites may be distorted if the faunal data are not linear. In PcpA, the ordination co-ordinates are linear combinations of individual species abundances. In Q-mode analysis, each set of co-ordinates represents a sample unit. The percentage of the variance that is explained by the first few principal components (or dimensions) is often used to assess the quality of the representation. If two or three components are sufficient to account for 50-90% of the variance, the representation may be acceptable. Legendre and Legendre (1983) discuss a number of misuses of PcpA, including the need to reduce or roll-up data (see Section III.A.1) to avoid distortions of Euclidean distances that may result when there is a large number of “double-zero’’ pairs. Some decisions made during PcpA are subjective, such as the selection of a data transformation (see Section III.A.2) or factor rotation. There are an infinite number of orthogonal and oblique factor rotations. The efficacy of rotation can best be illustrated by Fig. 12, which shows various distributions

223

BENTHIC MARINE INFAUNAL STUDIES

t-

; :

0

-,'

' + '

FIG. 12. Conceptual illustration of the effects of different types of rotational schemes for interpreting ordination analyses (see Section III.E.3 for a description). The dotted lines represent rotated axes. (A) Oblique rotation (orientation of axes with respect to each other is changed, making the axes dependent on each other) best explains the groupings of points. (B) No rotation is required for three groups but the presence of a fourth group suggests that another dimension or axis is required to explain sufficient variance in the analysis. (C) The scatter of points cannot be explained by ordination, and no rotation will improve the depiction. (D) A simple orthogonal (orientation of axes with respect to each other maintained) rotation best explains the structure of the groups.

of sample points (co-ordinates) on a pair of component axes, and the effects of rotating those axes. Such rotations do not change the amount of variance in the data explained by the ordination, but can simplify the interpretation of the structure of data points. This may be desirable if the component axes are to be interpreted in some manner, but is probably not necessary if the only requirement is to determine group membership of data points. If the solid lines in Fig. 12 represent the first two unrotated component axes identified during the analysis, the various patterns of points show the effects of different types of scenarios and rotations. In (B), the variance in the majority of sample points can be easily interpreted using the original axes. However, there is an extra cloud of points (circled by a broken line) for which the variance may only be explained by a third component axis (not shown). In (D), the points are not easily interpreted by the original axes, but an orthogonal rotation of the axes (so that the two axes remain perpendicular or independent of each other, as illustrated by the broken lines) does

224

B. J. BURD E T A L .

provide simple structure. In (A), there is no orthogonal rotation of axes that would adequately allow interpretation of the variance in the data. Therefore, an oblique rotation which does not preserve the original orientation of axes with respect to each other may be required. This also means that the axes are no longer independent, producing interpretation problems in the final analysis. In (C), the points show a random pattern which cannot be effectively reduced in dimensionality by ordination. Therefore, rotation is not an issue. There are various arguments for and against the use of different types of rotations, which are described by most multivariate statistical textbooks. A good example of the application of PcpA is given in Lie ( 1 974), who used both normal and inverse analyses. Some unusual but questionable applications have been evident in the literature. For example, Chester et al. (1983) applied a PcpA analysis to a data set combining assemblage data as well as environmental variables. Long and Lewis (1987) used a step-wise combined classification and ordination method which involved removing stable station groups at each step. The approach is not easy to follow, partly due to lack of clarity about the identification of the “stable groups” removed and confusion about how replicates were handled. The analytical method is also difficult to follow in the re-analysis of Petersen’s original community data by Stephenson et al. (1972), partly because of problems with the original data set. Whereas PcpA attempts to explain the variances in the species abundances by producing weighted linear composites of observed variables, common factor analysis attempts to explain the co-variances among the species abundances in terms of a small number of unobserved factors (which are often given some physical interpretation). The main problem with factor analysis is that it is indeterminate, i.e. there is not a unique solution for a particular model. For a short but enlightening discussion of the differences between principal components and factor analysis (as well as a useful critique), refer to Wilkinson (1988, p. 408). For a more mathematical and in-depth discussion of these differences, refer to Reyment (1963). Gower’s (1966) discussion of the relationship between PcpA, PcdA and f x t o r analysis is particularly clear. A more general form of Metric Multidimensional Scaling (MDS) is Principal Coordinates Analysis (PcdA), introduced by Torgerson ( 1 952, 1958) and described in a more generalized form by G3wer (1966). PcdA differs from PcpA and factor analysis in that it does not necessarily use a Euclidean distance measure. PcdA can use distances given by dimensionless similarity indices (such as Bray-Curtis, Canberra-Metric, etc.). The ordination co-ordinates are therefore dimensionless. Note that the use of a metric distance measure is sufficient to produce a solution, but not necessary (e.g. Bray-Curtis is non-metric). An equivalent solution to PcpA is found if the

BENTHIC MARINE INFAUNAL STUDIES

225

distance measure in Gower’s method is Euclidean. Otherwise, PcdA suffers from the same interpretation problems as PcpA. The aforementioned methods are all linear ordinations, which assume that the faunal data are linear with respect to the ordination dimensions imposed. Unfortunately, most species abundance and biomass data show a decline to 0 for both low and high values along any environmental or biological gradient. This non-monotonicity in assemblage data produces a “horseshoe”-shaped distribution of stations when plotted using linear ordination axes (see Pielou, 1977). Problems of non-linearity in ordinations applied to species assemblage data are discussed by Austin and Noy-Meir (1971). Williamson (1978) and Smith et al. (1988) approach the problem from the perspective of the dissimilarity matrices used in ordination. In cases where the data set is diverse (i.e. dissimilarities approach 0 and loo%), the skewness in the dissimilarity distribution at these extremes produces the “horseshoe effect” on the ordination axes. The recommended “step-across” procedure (see Section III.E.2) addresses this problem but can produce dissimilarities > 100%. Therefore, Smith et al. (1988) recommend the use of the rank of the samples for the step-across dissimilarity matrix, rather than actual values. This produces the equivalent of a non-metric multidimensional scaling ordination, which may have the effect of eliminating the “horseshoe effect”. Non-linearity in the data can also be dealt with by the application of nonlinear ordination methods. Non-linear ordinations such as non-metric Multi-Dimensional scaling (for review see Kruskal and Wish, 1978; Shepard, 1980; Field et al., 1982; Ramsay, 1982; Kirkwood and Burton, 1988) and Gaussian ordination may be useful for situations when continuous species data are being compared with environmental gradient data and it is not realistic to assume that the relationship is linear (for a discussion, see Green and Vascotto, 1978). Non-metric MDS (proximity analysis) is a generalization of metric MDS. This has been used in benthic applications recently (see Field et a/., 1982; Clarke and Green, 1988 and Smith et al., 1988). but in practice there is no a priori reason to select a particular nonlinear transformation, so the linear approach may be as reasonable as anything else. Gauch et al. (1974) discuss the use of Gaussian ordination as an alternative to linear and ranking forms. It assumes that species are distributed in bell-shaped patterns along environmental gradients, a pattern which is often obvious in direct gradient analysis. Ordination is sometimes carried out in an R-mode fashion so that species are plotted as points in the ordination space. This causes problems of interpretation (especially with 200-300 species), as well as computational problems and is not commonly used. Smith et al. (1988) provide examples of such an analysis using data standardizations and transformations.

226

B. J. BURD E T A L .

Hill and Gauch (1980) have recently pointed out problems with many commonly used ordination methods. They introduced a new method called Detrended Correspondence Analysis (DCA), which is an extension of ordination but which uses standard deviations of species abundances across ordination distances, instead of abundances. This method has not appeared often in the benthic literature, so that objective discussion of the advantages and disadvantages is unavailable. The method may serve to reduce the scale of disparity in abundances between rare and common species (see Section III.A.2). Alongi (1986) presented an example of DCA Q- and R-mode analyses sorted and analysed simultaneously, producing some interpretative challenges.

4. Multivariate comparisons with habitat factors Benthic community survey studies are usually confined to discussions of the effects of environmental factors (abiotic) on community structure. The effects of specific biotic factors (other than general characterizations such as dispersion and diversity: see Section 1II.C) on community structure are usually examined experimentally (exclusion and recolonization experiments) and are therefore outside the scope of this review. There are a few examples of survey studies of biotic factors which utilize multivariate analyses. A good example is given by Smith (1981), who examined the influence of sand dollars on community structure in ten different beach areas. Smith’s study is appealing because of the well-defined objective, which allowed an effective and controlled delineation of sample sites and other environmental factors. Unfortunately, studies of the effects of abiotic factors on community structure are rarely so well-defined. Commonly, the stated study objective is “to describe the effect of environmental factors on invertebrate infaunal community structure.” Field et al. (1982) stress the importance of examining the biotic (species matrix) and abiotic data separately, before comparisons are made, rather than performing complex manipulations on mixed data sets (see also Green and Vascotto, 1978). Because of the myriad of environmental factors potentially influencing an assemblage, and their interdependence (e.g. depth and sediment type), causal conclusions are difficult to make except where the effects are so profound and consistent that the resultant faunal patterns can be attributed to a single factor. Even then, the unambiguous naming of that causal factor is problematical (e.g. is an observed difference in community structure between sand and silt stations due to sediment texture, organic content, microbial composition, turbulence or some unknown combination of factors?). One approach to the problem is to select an assemblage in which the environmental factors affecting benthos can be reduced with reasonable

BENTHIC MARINE INFAUNAL STUDIES

227

confidence to one or two. An example of this is a rocky intertidal study in South Africa by Field and McFarlane (1968), in which wave exposure was the overwhelmingly dominant environmental factor, with all sample sites otherwise fairly uniform. The results were not statistically tested, but were so obvious as to require no further testing. Other examples of such a situation include hypoxic marine habitats, where the overwhelming dominance of oxygen makes all other factors irrelevant (see Burd and Brinkhurst, 1984), or in estuaries, where salinity overrides all other factors (see Chapman and Brinkhurst, 1981). Such situations are not common, particularly in softbottom areas (see Rosenberg, 1980). Several factors stand out as being important in most undisturbed (nonpolluted) benthic community distributions. These include a range of sediment characteristics such as mean particle size, sediment structural complexity, organic content, ATP or chlorophylls in sediment, clay size fractions or percentages, sulphide content, presence of macroflora (e.g. eelgrass) or microfloral composition. Sediment factors are most commonly cited as the dominant abiotic factors in benthic assemblage structure (for a review, see Gray, 1974). Other important factors in benthic community distribution include depth, or depth-associated factors, which in many cases cannot be considered independent of sediment characteristics. Water characteristics such as natural or stagnation-induced hypoxia are often related to depth (Burd and Brinkhurst, 1984). Rosenberg (1980) has suggested that natural stagnation can cause much more widespread defamation than pollution-induced stagnation. This is illustrated by conditions in the Baltic Sea. Salinity in nearshore and estuarine habitats and shallow fjord areas is usually depthrelated, but is extreme only in nearshore or estuarine situations, or brackish seas such as the Baltic (for a review of brackish water studies, see Hedgpeth, 1983). Surprisingly, factors operating within the sediment are usually ignored in soft-bottom studies, possibly because of the difficulties involved in sampling and analysis of interstitial waters. These factors include interstitial salinity (see Chapman and Brinkhurst, 198l), sediment oxygen/ sulphide balance, sediment structure and microflora. Sediment factors are probably of the utmost importance, because the water chemistry overlying the sediments may not fluctuate to the same degree or in synchrony with the interstitial water. Important but rarely cited factors related to benthic community structure include the geographic and temporal separation of sampling stations, which may explain a drift in the species composition of communities, e.g. there may be geographic and temporal variations in recruitment or larval dispersion (Burd and Brinkhurst, 1987). Pollution factors include a wide variety of organic and inorganic

228

B. J. BURD ET AL.

contaminants from sewage and pulp waste, as well as chemical toxins and heavy metals. Other forms of pollution are mainly related to disturbance, such as dredging. Gray et al. (1988) emphasize that the most common problem with most univariate and multivariate methods in benthic ecological surveys is that of separating natural and unnatural (pollution) environmental effects. The analytical approach to examining the relationship between environmental factors and community structure depends on the discreteness of the community groups. If the data form discrete clusters, there are several effective statistical approaches for examining environmental effect. Analysis of variance (ANOVA) and multiple discriminant analysis are the two most commonly used methods. ANOVA is used to determine whether or not the mean value of some environmental variable varies significantly across the clusters (cf. Jones, 1986). For example, the mean sediment particle size might be computed for each cluster. If there is a significant difference in these means it might be concluded that the differences in community structure may be due, in part, to particle size differences. Many studies seem to test several environmental variables using a series of single-factor ANOVA tests. If the number of tests is large, it is advisable to use a multivariate ANOVA or adjust for the multiple comparisons (e.g. by using a studentized NewmanKeuls test), since the overall confidence level of a series of tests is usually less than that of an individual test. If a large enough number of independent tests is performed at the 0.05 (P)level of significance, approximately 5% will reject the null hypothesis even when it is true. ANOVA is a parametric test which assumes normality and independence of the sample units, and a homogeneous variance for each variable, although the test is fairly robust with respect to certain departures from these assumptions. The Kruskal-Wallis test is a non-parametric version of the univariate one-way ANOVA F-test, which can be used if the environmental variable does not have a normal distribution (it does not help if the independence or homogeneous variance assumptions fail, but will help if the distribution is skewed or otherwise non-normal). Another non-parametric test for comparing an assemblage cluster analysis with an independently determined “covariate” dendrogram (based on one or more environmental factors) is described by Nemec and Brinkhurst (1988b). It uses a FowlkesMallows (Fowlkes and Mallows, 1983) test statistic (which is a measure of the degree of similarity between two dendrograms) to test the null hypothesis that the two dendrograms are unrelated, (for examples of this method, see Burd and Brinkhurst, 1987; Brinkhurst, 1987; Brinkhurst et al., 1987). Green and Vascotto (1978) describe the use of multiple discriminant analysis ( = canonical variate analysis) for examining multifactor effects on assemblages. Discriminant analysis can be thought of as a procedure for reducing the dimensionality of the environmental space (rather than of the

BENTHIC MARINE INFAUNAL STUDIES

229

species data matrix as in ordination). Multiple (clusters, not variables) discriminant analysis is used to construct a set of orthogonal “environmental axes” which correspond to linear combinations of the environmental variables (i.e. discriminant functions), such that when the sites are plotted using these axes (using the values of the environmental variables at each site) there is a maximum separation of the clusters along each axis (Fig. 13).

FIG. 13. A conceptual illustration of the operation and purpose of multiple discriminant functions for maximally separating groups along environmental variables or gradients (after Green and Vascotto, 1978) using the minimum number of axes. In this case, the discriminant function ( 3 ) is some combination of the properties of the original variables 1 and 2.

The first axis accounts for the greatest separation of axes, and therefore represents the environmental space which accounts for that maximal separation. Each subsequent axis represents increasingly minor separations of the clusters. Since each axis, or discriminant function, is a combination of the environmental variables, it is often possible to determine which environmental variables are contributing the most to the cluster pattern. Green and Vascotto (1978) point out that the advantage of this method is that it does not assume a linear relationship between the biotic variables (assemblage structure) and the environmental variables, but rather a linear relationship among the environmental variables (e.g. salinity vs temperature). In practice, certain environmental variables may require transformations (such as particle size percentages and pH) to produce linearity. If there are many environmental factors to consider simultaneously, a discriminant analysis might be easier to interpret than a multivariate ANOVA (since the data can

230

B. J. BURD ET AL.

be plotted more easily: Fig. 13), provided the procedure is effective in reducing the dimensionality (i.e. a small number of discriminant functions result in good separation of the groups). The drawback, of course, is that it applies only when there are distinct clusters of samples. Several examples of multiple discriminant analysis are given by Smith et a/. (1988). Shin (1982) also described an application of the multiple discriminant analysis approach. His example is not ideal, however, because the clusters do not appear to be well-defined. The examples in Shin and Thompson (1 982) and Green and Vascotto (1978) are more convincing. This type of analysis may be inferentially tested by a method such as chi-square analysis (Shin, 1982), although Green and Vascotto (op. cit.) suggest that the simple descriptive approach is more conservative. If the data do not cluster into distinct groups or communities, a continuum method is appropriate. A standard approach for examining the role of environmental variables is to use multiple regression or correlation analysis to related ordination co-ordinates derived from the species data to the environmental variables. The linearity assumption should always be assessed using regression diagnostics, such as scatter plots and residual plots. Simple linear regression is often used to investigate each environmental variable separately. The multiple comparisons problem arises here as well as with univariate ANOVA. A multiple regression approach is helpful for looking at the variables simultaneously. Smith and Green (1976) use ridge regression to overcome the problem of the distortions that can arise when there are intercorrelations between the explanatory variables. A reference for ridge regression is given as Marquadt and Snee (1975), although this method has not been widely used in benthic ecological research. Smith et al. (1988) discuss several examples using multiple regression analysis to examine the relationship between combinations of environmental factors and assemblage factors. Canonical correlation is sometimes used to examine the relationship between the biotic and environmental variables. Canonical correlation has been described by Legendre and Legendre (1983) as a generalization of multiple (linear) regression, which can be thought of as “a double principal components analysis followed by a rotation of the canonical axes in order to make them superimpose”. Canonical correlation examines the (linear) relationship between two sets of variables. In the typical benthic application, the two sets of variables are the species data set and the environmental data set (see Chester et al., 1983; Penas and Gonzales, 1983; Smith et al., 1988). A test of significance of the canonical variates is Bartlett’s test (Penas and Gonzales, 1983), which is a non-parametric method using contingency tables. Smith er al. (1988) suggest that canonical correlation provides no more useful data than multiple regression analyses and can be more complex

BENTHIC MARINE INFAUNAL STUDIES

23 1

and difficult to interpret. Both assume linearity between community structure and environmental variables. A univariate non-parametric method for examining the effects of environmental factors on assemblage structure has been proposed and involves comparison of ordination station loadings on a given axis and a chosen environmental factor using Spearman’s rank correlation (described by Legendre and Legendre, 1983). This ranking measure is similar in interpretation to a simple Pearson’s r, but it is useful for examining the degree of monotonic relationship between two variables when the data are nonnormal, or the relationship is non-linear. Still, it must be accepted that a rank order transformation involves some loss of information. Such a method may be suitable when the ordination co-ordinates are suspected of being non-linearly related to the environmental variables (see Hughes and Thomas, 1971a).

F. Time-series Analysis Specific analytical methods have been used to compare large sets of temporal data. These methods are based on the same basic theory as the spatial surveys discussed in this chapter. Many benthic studies cover a long time period, but are not planned as detailed time-series analyses per se (see Chester et al., 1983; Govaere et a / . , 1980). An example of a relatively longterm benthic study is given by Beukema and Essink (1986), who correlated the abundance fluctuations of a series of tidal flat species over a 17-year period in order to separate global patterns of natural fluctuation from localized disturbances such as pollution. They found that 50% of the fluctuations were synchronized and correlated over wide areas. Williams and Stephenson (1973) discuss the three basic methods used for analysing timeseries benthic data, the oldest and most common being to obtain a series of data matrices over time and compare them subjectively or in some inferential manner. They also suggest certain methods of combining species abundance, station and time data in two- or three-dimensional comparisons, the latter of which is problematical but may command more attention as consistent, long-term studies become more common. Legendre et a/. (1985) discuss the problem of mapping successional events in ecological communities. They propose a method that uses a “chronological clustering” of samples from a single station, which are replicated over time, to identify discrete successional steps in the species composition. A non-parametric (randomization) procedure is used as a fusion criterion for the groups. An important component of the method is the exclusion of erratic or random singleton measurements which do not fit into the

232

B. J. BURD ET AL.

successional pattern. Unfortunately, the method requires some subjective decisions which affect the power of the test, and it is not particularly effective for data sets in which a large number of the pair-wise linkages join highly dissimilar groups. In their examples, Legendre et af. (1985) use relatively large significance levels (about 20%) to get viable results. This seems unacceptably high. Nemec and Brinkhurst’s method (1988a) would provide a similar type of test, without the problems mentioned above. Other examples of time-series analyses include the computation of spatial autocorrelation (Pielou, 1977) and the methods described by Barnard et al. (1986) which have potentially important applications in long-term environmental impact studies. However, the effective use of such methods usually requires a large number of identically treated temporal replicates (25 or more depending on sample design), which is rare indeed in benthic community studies. Time-series studies, and their specific analytical problems, are rapidly becoming the current hot topic in benthic ecology. There will undoubtedly be a greater emphasis on long-term analyses of assemblage structure in the future.

IV. Summary The major problems addressed here have not changed since the early 1900s. The methodology has changed dramatically. In sampling the diffuse literature on this topic, one gets the impression that progress in generalizing, simplifying and explaining the processes involved in benthic community organization, is painfully slow in comparison with more clearly defined areas of scientific study. The key to further progress in this field is to develop sampling and analytical techniques in concert. This presupposes that research aims and study hypotheses can be more clearly stated in benthic studies than has often been the case in the past. We have tried to provide a perspective on the development of analytical and sampling methods in benthic survey studies. Methods have evolved from the original intuitive approach based on the indicator species (or group) concept, to objective univariate indices which provide a useful initial characterization of a community or spatial pattern. These univariate methods are being progressively replaced or enhanced by more rigorous descriptive and inferential multivariate methods. Researchers of benthic systems can be simply overwhelmed by the variety and complexity of analytical methods available. In this chapter, we have not attempted to provide handbook (or cookbook) procedures. In all honesty, we cannot define right or wrong methodology, though this has been attempted to some degree (cf. Green and Vascotto, 1978; Field et al., 1982;

BENTHIC MARINE INFAUNAL STUDIES

23 3

Smith et al., 1988; Gray et al., 1988). Instead, we have attempted to provide the prospective researcher with a feel for the assumptions involved with certain methods, or the particular prejudices that have led to the emergence of currently “popular” methods. If anything can be concluded from the often contradictory opinions of different ecological authorities, it is that the analytical approach should be straightforward, avoiding the common trap of using a whole suite of complex (often uninterpretable) methods when one or two would be sufficient. If the data set is “robust”, the results will probably not be seriously affected by the use of questionable statistical methods. Unfortunately, the validity of results cannot be judged by other readers without an adequate description of the sampling design and the analytical methods. This is probably the most common shortcoming in many applied papers in benthic ecology. In addition, data management is a pressing problem in many studies. Much more attention should be paid to the biases and consistency of sampling methods and how they relate to the assumptions inherent in the statistical tests, to the validity of data transformations or reductions, and to the taxonomic adequacy. In most cases, the researcher may have reasonable confidence in the power of statistical inferential tests, due to the large sample sizes characteristic of most benthic studies. On the other hand, large sample sizes can lead to the problem of over-interpretation of “trivial” but statistically significant results. Therefore, the effect size of interest should be carefully considered. Because of the limitations inherent in species identifications, the use of biomass/size spectra to analyse trophic relationships and population structure in benthos is an alternative approach which seems to be receiving some attention. The potential usefulness of combining taxonomically based and biomass-based studies has not yet been fully explored, partly because of the additional costs and labour required for accurate descriptions. The appeal of non-parametric simulation or randomization methods for hypothesis testing is expected to increase, since these methods eliminate many of the problems encountered when attempting to fit aggregated, multispecies data to parametric models. Most researchers would agree that the introduction of reliable and flexible methods to simplify the often confusing and frustrating process of data analysis would be welcome. To this end, more effort should be made to invent clear graphical methods for depicting complex statistical results, for the benefit of managers, public groups and political agencies with policy decisions to make (see Figs 10 and 11). As methods and understanding of basic mechanisms affecting benthic communities improve, researchers seem to be attempting more ambitious studies. Long-term data are now available in many areas, as well as the type of widespread sampling coverage and data-handling methods which may

234

B. J. BURD ET AL.

eventually encourage researchers t o readdress broad community issues of the type raised originally by Petersen (1913-191 5 ) and later by Thorson ( 1957, 1966).

V. References Abele, L. G. and Walters, K. (1978). The stability-time hypothesis: Reevaluation of the data. American Naturalist 114, 559-567. Abele, L. G. and Walters, K. (1979). Marine benthic diversity: A critique and alternative explanation. Journal of Biogeography 6, 115-126. Abugov, R. (1982). Species diversity and phasing of disturbance. Ecology 63, 289293. Allee, W. C., Emerson, A. E., Park, O., Park, T. and Schmidt, K. (1949). “Principles of animal ecology” W. B. Saunders, Philadelphia. Alongi, D. M. (1986). Population structure and trophic composition of the freeliving nematodes inhabiting carbonate sands of Davies Reef, Great Barrier Reef, Australia. Australian Journal of Marine and Freshwater Research 37, 609-619. Anger, K. (1975). On the influence of sewage pollution on inshore benthic communities in the south of Kiel Bay. Helgolander wissenschaftliche Meeresuntersuchungen 27, 408-438. Austin, M. P. and Noy-Meir, I. (1971). The problem of non-linearity in ordination: Experiments with two gradient models. Journal of Ecology 59, 763-773. Baker, J. M. and Wolff, W. J. (1987). “Biological Surveys of Estuaries and Coasts”. Estuarine and Brackish-water Sciences Association Handbook. Cambridge University Press, Cambridge. Barnard, J. W., Myers, J., Pearce, F., Ramsey, F., Sissenwine, M. and Smith, W. (1986). Surveys for monitoring changes and trends in renewable resources: Forests and marine fisheries. American Statistician 39, 363-373. Beals, E. W. (1984). Bray-Curtis ordination: An effective strategy for analysis of multivariate ecological data. Advances in Ecological Research 14, 1-52. Becker, D. S. and Armstrong, J. W. (1988). Development of regionally standardized protocols for marine environmental studies. Marine Pollution Bulletin 19, 3 1C313. Beran, R. (1986). Simulated power functions. Annuls of Statistics 14, 151-173. Beukema, J. J. (1988). An evaluation of the ABC-method (abundance/biomass comparison) as applied to macrozoobenthic communities living on tidal flats in the Dutch Wadden Sea. Marine Biology 99,425-433. Beukema, J. J. and Essink, K. (1986). Common patterns in the fluctuations of macrozoobenthic species living at different places on tidal flats in the Wadden Sea Hydrobiologia 142, 199-207. Birch, D. W. (1981). Dominance in marine ecosystems. The American Naturalist 118(2), 262-274. Bloom, S. A. (1981). Similarity indices in community studies: potential pitfalls. Marine Ecology Progress Series 5, 125-128. Boesch, D. F. (1973). Classification and community structure of Macrobenthos in the Hampton Roads area, Virginia. Marine Biology 21, 226-244. Boesch, D. F. (1977). Application of numerical classification in ecological investigations of water pollution. Special ScientiJic Report No. 77, Virginia Institute of Marine Science. U.S. Department of Commerce National Technical Information Service PB-269 604.

BENTHIC MARINE INFAUNAL STUDIES

235

Botton, M. L. (1979). Effects of sewage sludge on the benthic invertebrate community of the inshore New York Bight. Estuarine and Coastal Marine Science 8, 169-180. Bray, J. R. and Curtis, J. T. (1957). An ordination of thc upland forest communities of southwestern Wisconsin. Ecological Monographs 27, 325-349. Brey, T., Rumohr, H. and Ankar, S. (1988). Energy content of macrobenthic invertebrates: General conversion factors from weight to energy. Journal of Experimental Marine Biology and Ecology 117, 271-278. Brinkhurst, R. 0. (1974). “The Benthos of Lakes”. Macmillan, London, 190 pp. Brinkhurst, R. 0. ( 1980). Production biology of the Tubificidae (Oligochaeta) In “Aquatic Oligochate Biology” (R. 0. Brinkhurst and D. G . Cook, eds), pp. 205209. Plenum Press, New York. Brinkhurst, R. 0. (1987). Distribution and abundance of macrobenthic infauna from the continental shelf off southwestern Vancouver Island, British Columbia. Canadian Technical Report on Hydrography and Ocean Sciences 85, 86 pp. Brinkhurst, R. O., Burd, B. J. and Kathman, R. D. (1987). A review of benthic studies in Alice Arm, B.C. during and following cessation of mine tailings disposal, 1982 to 1986. Canadian Technical Report on Hydrography and Ocean Sciences 89, 45 pp. Brown, J. H. (1984). On the relationship between abundance and distribution of species. American Naturalist 124, 255-279. Buchanan, J. B. (1963). The bottom fauna communities and their sediment relationships off the coast of Northumberland. Oikos 14, 154-175. Buchanan, J. B. and Moore, J. J. (1986). Long-term studies at a benthic station off the coast of Northumberland. Hydrobiologia 142, 121-127. Burbanck, W. D., Pierce, M. E. and Whiteley, G . C. (1956). A study of the bottom fauna of Rand’s Harbor, Massachusetts: An application of the ecotone concept. Ecological Monographs 26, 236243. Burd, B. J. and Brinkhurst, R. 0. (1984). The distribution of the galatheid crab (Munida quadrispina) (Benedict 1902) in relation to oxygen concentrations in British Columbia fjords. Journal of Experimental Marine Biology and Ecology 81, 1-20.

Burd, B. J. and Brinkhurst, R. 0. (1987). Macrobenthic infauna from Hecate Strait, British Columbia. Canadian Technical Report on Hydrography and Ocean Sciences 88, 123 pp. Burns, G . R. (1978). Distribution of benthos in a stagnant estuarine bay. Canadian Journal of Zoology 56, 1664-1668. Chang, B. and Winnell, M. H. (1981). Comment on the fourth-root transformation. Canadian Journal of Fisheries and Aquatic Sciences 38, 126-127. Chapman, P. M. and Brinkhurst, R. 0. (1981). Seasonal changes in interstitial salinities and seasonal movements of subtidal benthic invertebrates in the Fraser River estuary, B.C. Estuarine and Coastal Shelf Science 12, 49-66. Chapman, P. M., Dexter, R. N. and Goldstein, L. (1987). Development of monitoring programmes to assess the long-term health of aquatic ecosystems: A model from Puget Sound, U.S.A. Marine Pollution Bulletin 18, 521-527. Chatfield, C. and Collins, A. J. (1 980). “Introduction to Multivariate Analysis”, pp. 189-211. Chapman and Hall, London. Chester, A. J., Ferguson, R. L. and Thayer, G . W. (1983). Environmental gradients and benthic macroinvertebrate distributions in a shallow North Carolina estuary. Bulletin of Marine Science 33, 282-295.

236

B. J. BURD ET

AL.

Christie, N. D. (1975). Relationship between sediment texture, species richness and volume of sediment sampled by a grab. Marine Biology 30, 89-96. Clarke, R. B. and Milne, R B. (1955). The sublittoral fauna of two sandy bays on the Isle of Cumbrae, Firth of Clyde. Journal of the Marine Biological Association of the United Kingdom 34, 161-180. Clarke, K. R.and Green, R. H. (1988). Statistical design and analysis for a biological effects study. Marine Ecology Progress Series 46, 2 13-226. Cliff, A. D. and Ord, J. K. (1981). “Spatial Processes: Models and Applications”, Pion, London, 266 pp. Clifford, H. T. and Stephenson, W. (1975). “An Introduction to Numerical Classification”. Academic Press, London, 229 pp. Cochran, W. G. (1963). “Sampling Techniques”, 2nd edn. John Wiley, New York. Cohen, J. (1977). “Statistical Power Analysis for the Behavioural Sciences”. Academic Press, London. Commito, J. A. and Ambrose, W. G., Jr (1985). Multiple trophic levels in softbottom communities. Marine Ecology Progress Series 26, 289-293. Cormack, R. M. (1971). A review of classification. Royal Statistical Society Journal Series A 134, 321-367. Crisp, D. J. (1984). Energy flow measurements. In “Methods for the Study of Marine Benthos” (N. A. Holme and A. D. McIntyre, eds), pp.284373. Blackwell Scientific, Oxford. Cuff, W. and Coleman, N. (1979). Optimal survey design: Lessons from a stratified random sample of macrobenthos. Journal of the Fisheries Research Board of Canada 36, 35 1-36 I . Dayton, P. K. and Hessler, R. R. (1972). The role of disturbance in the maintenance of deep-sea diversity. Deep Sea Research 19, 199-208. DeAngelis, D. L. and Waterhouse, J. C. (1987). Equilibrium and nonequilibrium concepts in ecological models. Ecological Monographs 57, 1-21. Devore, J. L. (1987). “Probability and Statistics for Engineering and the Sciences”, 2nd edn. Brooks-Cole, Monterey, Calif., 672 pp. Diaconis, P. and Efron, B. (1983). Computer-intensive methods in statistics. Scientijc American 248, 96108. Downing, J. A. (1979). Aggregation, transformation and the design of benthos sampling programs. Journal of the Fisheries Research Board of Canada 36, 14541463. Downing, J.A. (1980). Precision vs. generality: A reply. Canadian Journal of Fisheries and Aquatic Sciences 37, 1329-1330. Downing, J.A. (1981). How well does the fourth-root transformation work? Canadian Journal of Fisheries and Aquatic Sciences 38, 127-129. Downing, J . A. (1986). Spatial heterogeneity: Evolved behaviour or mathematical artefact? Nature. London 323, 255-257. Downing, J.A. and Anderson, M. R. (1985). Estimating the standing biomass of aquatic macrophytes. Canadian Journal of Fisheries and Aquatic Sciences 42, 1860-1 869. Dybern, B. I., Ackefors, H. and Elmgren, R. (1976). Recommendations on methods for marine biological studies in the Baltic Sea. Baltic Marine Biologists 1, 98 pp. Edmonson, W. T. and Winberg, G . G . (1971). “A manual on Methods for Assessment of Limnological Sampling: Productivity in Freshwaters”. IBP Handbook No. 17. Blackwell Scientific, Oxford. Efron, B. (1982). “The Jackknife, the Bootstrap and Other Resampling Plans”. SIAM, Philadelphia, 92 pp.

BENTHIC MARINE INFAUNAL STUDIES

237

Efron, 9 . and Gong. G. (1983). A leisurely look at the bootstrap, jackknife and cross-validation. American Statistician 37, 36-48. Eleftheriou, A. and Holme, N. A. (1984). Macrofauna techniques. In “Methods for the Study of Marine Benthos” (N. A. Holme and A. D. McIntyre, eds), 2nd edn. Blackwell Scientific, Oxford. Elliott, J. M . (1977). ‘Some Methods for the Statistical Analysis of Samples of Benthic Invertebrates’, 2nd edn. Scientific Publications of the Freshwater Biological Association, No. 25. Ellis, D. V. (1971). A review of marine infaunal community studies in the Strait of Georgia and adjacent inlets. Syesis 4, 3-9. Ellis, D. V. (1985). Taxonomic sufficiency in pollution assessment. Marine Pollution Bulletin 16, 459. Ellis, D. V. (1987). Biomass loss in wet-preserved reference collections. Collection Forum 3, ( 1 and 2). Elmgren, R. (1978). Structure and dynamics of Baltic benthos communities, with particular reference to the relationship between macro- and meiofauna. Kieler Meeresforschungen, Sonderheft 4, 1-22. Engen, S. (1978). “Stochastic Abundance Models”. Chapman and Hall, London/ Halsted Press, John Wiley, New York. Engen, S. (1979). Stochastic abundance models in ecology. Biornetrics 35, 331-338. Fager, E. W. (1957). Determination and analysis of recurrent groups. Ecology 38, 586595. Felsenstein, J. (1985). Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39, 783-791. Field, J. G. and McFarlane, G. (1968). Numerical methods in marine ecology. 1. A quantitative similarity analysis of rocky shore samples in False Bay, South Africa. Zoologica A,fricana 3, 119-137. Field, J. G., Clarke, K. R. and Warwick, R. M. (1982). A practical strategy for analysing multispecies distribution patterns. Marine Ecology Progress Series 8, 37-52. Fisher, R. A., Corbet, A. S. and Williams, C. B. (1943). The relation between the number of individuals and the number of species in a random sample of an animal population. Journal qf Animal Ecology 12, 42-58. Flint, R. W. and Holland, J. S. (1980). Benthic infaunal variability on a transect in the Gulf of Mexico. Estuarine and Coastal Marine Science 10, 1-14. Fowlkes, E. 9 . and Mallows, C. L. (1983). A method for comparing two hierarchical clusterings (with comments). American Statistician Association 78, 553-584. Gage, J. and Geekie, A. D. (1973a). Community structure of the benthos in Scottish sea-lochs. 11. Spatial pattern. Marine Biology 19, 41-53. Gage, J. and Geekie, A. D. (1973b). Community structure of the benthos in Scottish sea-lochs. 111. Further studies on patchiness. Marine Biology 20, 89-100. Gage, J. D., Lightfoot, R. H., Pearson, M. and Tyler, P. A. (1980). An introduction to a sample time-series of abyssal macrobenthos: Methods and principal sources of variability. Oceanologica Actu 3, 169-176. Gauch, H. G . , Jr and Chase, G. E. (1974). Fitting the Gaussian curve to ecological data. Ecology 55, 1377-1381. Gauch, H. G., Jr, Chase, G . 9. and Whittaker, R. H. (1974). Ordination of vegetation samples by Gaussian species distribution. Ecology 55, 1382-1 390. Gerlach, S. A., Hahn, A. E. and Schrage, M. (1985). Size spectra of benthic biomass and metabolism. Marine Ecology Progress Series 26, 161-173.

238

B. J. BURD E T A L .

Gillis, G. P. (1978). The effect of the spring freshet on selected biota in the St. John River Estuary. MSc thesis, University of New Brunswick, 121 pp. Gleason, H. A. (1926). The individualistic concept of the plant association. Bulletin of the Torrey Botanical Club 53, 7-26. Goodall, D. W. (1973). Sample similarity and species correlation. In “Handbook of Vegetation Science” (R. Y. Whittaker, ed.), Part V, pp. 107-156. Junk, The Hague. Gordon, A. D. (1987). A review of hierarchical classification. Journal of Royal Statistical Society of America 150, 119-137. Govaere, J. C. R., Van Damme, D., Heip, C. and De Coninck, L. A. P. (1980). Benthic communities in the southern bight of the North Sea and their use in ecological monitoring. Helgolander wissenschaftliche Meeresuntersuchungen 3, 507-52 1. Gower, J. C. (1966). Some distance properties of latent root and vector methods used in multivariate analysis. Biometrika 53, 325-338. Grassle, J. F. and Smith, W. (1976). A similarity measure sensitive to the contribution of rare species and its use in investigation of variation in marine benthic communities. Oecologia (Berlin) 25, 13-22. Grassle, J. P. and Grassle, J. F. (1976). Sibling species in the marine pollution indicator Capitella (Polychaeta). Science 192, 567-569. Gray, J. S. (1974). Animal-sediment relationships. Oceanography and Marine Biology Annual Review 12, 223-261. Gray, J. S. (1976). The fauna of the polluted River Tees estuary. Estuarine and Coastal Marine Science 4, 653-676. Gray, J. S. (1985). Ecological theory and marine pollution monitoring. Marine Pollution Bulletin 16, 224-227. Gray, J. S. and Christie, H. (1983). Review: Predicting long-term changes in marine benthic communities. Marine Ecology Progress Series 13, 87-94. Gray, J. S. and Pearson, T. H. (1982). Objective selection of sensitive species indicative of pollution-induced change in benthic communities. I. Comparative methodology. Marine Ecology Progress Series 9, 11 1-1 19. Gray, J. S., Aschan, M., Carr, M. R., Clarke, K. R., Green, R. H., Pearson, T. H., Rosenberg, R. and Warwick, R. M. (1988). Analysis of community attributes of the benthic macrofauna of Frierfjord/Langesundfjord and in a mesocosm experiment. Marine Ecology Progress Series 46, 151-165. Green, R. H. (1979). “Sampling Design and Statistical Methods for Environmental Biologists”. John Wiley, New York. Green, R. H. and Vascotto, G. L. (1978). A method for the analysis of environmental factors controlling patterns of species composition in aquatic communities. Water Research 12, 583-590. Greig-Smith, P. (1964). “Quantitative Plant Ecology”. 2nd edn. Butterworths, London. Hailstone, T. S. (1976). Delimitation of subtidal macrobenthos associations at the mouth of the Brisbane River. Australian Journal of Marine and Freshwater Research 21, 2 17-238. Harrison, A. D. and Hynes, H. B. N. (1988). Ethiopian mountain streams and rivers. Archivum Hydrobiologia 81 (suppl.), 1-36. Hartley, J. P. (1982). Methods for monitoring offshore macrobenthos. Marine Pollution Bulletin 13, 150-154. Hartley, J. P. and Dicks, B. (1987). Sampling methods in estuaries. In “Biological Surveys of Estuaries and Coasts” (J. M. Baker and W. J. Wolff, eds), pp. 106130.

BENTHIC MARINE INFAUNAL STUDIES

239

Estuarine and Brackish-Water Sciences Association Handbook. Cambridge University Press, Cambridge. Hedgpeth, J. W. (1983). Coastal ecosystems: Brackish waters, estuaries and lagoons. In Marine Ecology, Vol. V, Part 2 (0.Kinne ed.). John Wiley, New York. Hellawell, J. M. (1986). “Biological Indicators of Freshwater Pollution and Environmental Management”. Elsevier, London. Herman, P. M. J. and Heip, C. (1988). On the use of meiofauna in ecological monitoring: Who needs taxonomy? Marine Pollution Bulletin 19, 665-668. Hill, M. 0. and Gauch, H. G. (1980). Detrended correspondence analysis, an improved ordination technique. Vegetatio 42, 47-58. Hoaglin, D. C . , Mosteller, F. and Tukey, J. W. (1983). “Understanding Robust and Exploratory Data Analysis”. John Wiley, New York. Holme, N. A. (1950). Population dispersion in Tellina tenuis da Costa. Journal of the Marine Biological Association of the United Kingdom 29, 267-280. Holme, N. A. (1964). Methods of sampling the benthos. Advances in Marine Biology 2, 171-260. Holme, N. A. and McIntyre, A. D. (1984). “Methods for the Study of Marine Benthos”, 2nd edn. Blackwell Scientific, Oxford, 386 pp. Hopkins T. L. (1964). A survey of marine bottom samplers. Progressive Oceanographer 2, 213-256. Horikoshi, M. (1970). Quantitative studies on the smaller macrobenthos inhabiting various topographical environments around the Sagami Bank in the deep-sea system of Sagami Bay. Journal of the Oceanographical Society of Japan 26, 159182. Howard, J. D. and Frey, R. (1975). Estuaries of the Georgia coast, U S A . : Sedimentology and biology. 11. Regional animal-sediment characteristics of Georgia estuaries. Senckenbergiana Maritirna 7 , 33-103. Hoyle, M. H. (1973). Transformations- an introduction and a bibliography. International Statistical Reviews 41, 203-223. Hughes, R. G. (1984). A model of the structure and dynamics of the benthic marine invertebrate communities. Marine Ecology Progress Series 15 1-1 I . Hughes, R. N. and Thomas, M. L. H. (1971a). Classification and ordination of benthic samples from Bedeque Bay, an estuary in Prince Edward Island, Canada. Marine Biology 10, 227-235. Hughes, R. N. and Thomas, M. L. H. (1971b). The classification and ordination of shallow-water benthic samples from Prince Edward Island, Canada. Journal of Experimental Marine Biology and Ecology 7 , 1-39. Hurlbert, S. H. (1971). The nonconcept of species diversity: A critique and alternative parameters. Ecology 52, 577-586. Hurlbert, S. H. (1984). Pseudoreplication and the design of ecological field experiments. Ecological Monographs 54, 187-2 1 I . Hynes, H. B. N. (1960). “The Biology of Polluted Waters”. Liverpool University Press, Liverpool. Illies, J. and Botosanneau, I. (1963). Problemes et methods de la classification et de la zonation ecologique des eaux courantes, considerees surtout du point de vue faunistique. Mitteilungen internationale Vereinigung fur theoretische und angewandte Lirnnologie 12, 57 pp. Jaccard, P. (1908). Nouvelles recherches sur la distribution florale. Bulletin Societe vaudoise Science Nature1 44,223-270. Johnson, R. G . (1973). Conceptual models of benthic communities. In “Models in Paleobiology” (T. J. M. Schopf, ed.). Freeman Cooper, San Francisco.

240

B. J. BURD E T A L .

Jones, A. R. (1986). The effects of dredging and spoil disposal on macrobenthos, Hawkesbury Estuary, N.S.W. Marine Pollution Bulletin 17, 17-20. Jones, N. S. (1950). Marine bottom communities. Biological Reviews 25, 283-3 13. Josefson, A. B. (1981). Persistence and structure of two deep macrobenthic communities in the Skagerrak (West coast of Sweden). Journal of Experimental Marine Biology and Ecology 50, 63-97. Jumars, P. A. (1975). Methods for measurement of community structure in deep-sea macrobenthos. Marine Biology 30,245-252. Kay, D. G. and Knights, R. D. (1975). The macro-invertebrate fauna of the intertidal soft sediments of south east England. Journal of the Marine Biological Association of the United Kingdom 55, 81 1-832. Kershaw, K. A. (1973). “Quantitative and Dynamic Plant Ecology”, 2nd edn. Arnold, London. Kirkwood, J. M. and Burton, H. R. (1988). Macrobenthic species assemblages in Ellis Fjord, Vestfold Hills, Antarctica. Marine Biology 97, 445-457. Knox, G. A. and Fenwick, G. D. (1981). Zonation of inshore benthos off a sewage outfall in Hawke Bay, New Zealand. New Zealand Journal of Marine and Freshwater Research 15, 417-435.

Kobayoshi, S. (1987). Heterogeneity ratio: A measure of beta-diversity and its use in community classification. Ecological Research 2, 101-1 1 I . Kolkwitz, R. and Marsson, W. A. (1908). Ecology of plant saprobia. Berichte der Deutschen Botanischen Gesellschaft 26, 505-5 19.

Kronberg, I. (1987). Accuracy of species and abundance minimal areas determined by similarity area curves. Marine Biology 96, 555-561. Kruskal, J. B. and Wish, M. (1978). “Multidimensional Scaling”. Sage, Beverley Hills, Calif. Kuchler, A. W. (1967). “Vegetation Mapping”. Ronald, New York. Kuenzler, E. J. (1961). Structure and energy flow of a mussel population in a Georgia salt marsh. Limnology and Oceanography 6, 191-204. Lance, G. N. and Williams, W. T. (1967). A general theory of classificatory sorting strategies. I. Hierarchical systems. Computing Journal 9, 373-380. Legendre, L. and Legendre, P. (1983). “Numerical Ecology”. Elsevier, Amsterdam, 419 pp. Legendre, P., Dallot, S. and Legendre, L. (1985). Succession of species within a community: Chronological clustering, with applications to marine and freshwater zooplankton. American Naturalist 125, 257-288. Leppakoski, E. (1975). Macrobenthic fauna as indicator of oceanization in the southern Baltic. Merentutkimuslait. JuIklHavsforskningsinst. Skr. 239, 28&288. Leppakoski, E. (1977). Monitoring the benthic environment of organically polluted river mouths. In “Biological Monitoring of Inland Fisheries” (J. S. Alabaster, ed.). Elsevier, Barking. Leppakoski, E. (1979). The use of zoobenthos in evaluating effects of pollution in brackish-water environments. In “The Use of Ecological Variables in Environmental Monitoring”, pp. 15 1-1 58. National Swedish Environmental Protection Board PM1151. Levin, L. A. (1984). Life history and dispersal patterns in a dense infaunal polychaete assemblage: Community structure and response to disturbance. Ecology 65, 11 851200.

Levings, C . D., Foreman, R. E. and Tunnicliffe, V. J. (1983). Review of the benthos of the Strait of Georgia and contiguous fjords. Canadian Journal ofFisheries and Aquatic Sciences 40, 112&1141.

BENTHIC MARINE INFAUNAL STUDIES

24 1

Levinton, J. S. (1982). “Marine Ecology”. Prentice-Hall, Englewood Cliffs, N.J., 526 pp. Lie, U. (1968). A quantitative study of benthic infauna in Puget Sound, Washington, USA, 1963-1 964. Fiskeridirektoratets Skrifter, Serie Havundersskelser 14, 556 pp. Lie, U. (1974). Distribution and structure of benthic assemblages in Puget Sound, Washington, USA. Marine Biology 26, 203-223. Lilliefors, H. W. (1967). On the Kolmogorov-Smirnov test for normality with mean and variance unknown. American Statistical Association Journal 62, 399402. Long, B. and Lewis, J. B. (1987). Distribution and community structure of the benthic fauna of the north shore of the Gulf of St. Lawrence described by numerical methods of classification and ordination. Marine Biology 95, 93-101. Lopez, G. R. (1988). Comparative ecology of the macrofauna of freshwater and marine muds. Limnology and Oceanography 33, 946-962. Lopez-Jamar, E. (1981). Spatial distribution of the infaunal benthic communities of the Ria de Muros, North-West Spain. Marine Biology 63, 29-37. Mahon, R., Smith, R. W., Bernstein, B. B. and Scott, J. S. (1984). Spatial and temporal patterns of groundfish distribution on the Scotian Shelf and in the Bay of Fundy. 1970-1 98 I . Canadian Technical Report on Fisheries and Aquatic Sciences 1300, 164 pp. Margalef, D. R. (1958). Information theory in ecology. General Systematics 3,36-71. Margalef, R. (1969). Diversity and stability: A practical proposal and model of interdependence. Brookhaven Symposia in Biology 22, 25-37. Marquadt, D. W. and Snee, R. D. (1975). Ridge regression in practice. American Statistician 29, 3 . Masse, H. (1972). Quantitative investigations of sand-bottom macrofauna along the Mediterranean north-west coast. Marine Biology 15, 209-220. Maurer, D., Watling, L., Kinner, P., Leathern, W. and Wethe, C. (1978). Benthic invertebrate assemblages of Delaware Bay. Marine Biology 45, 65-78. McCall, P. L. (1977). Community patterns and adaptive strategies of the infaunal benthos of Long Island Sound. Journal of Marine Research 35, 221-266. McCall, P. L. (1978). Spatial-temporal distributions of Long Island Sound infauna: The role of bottom disturbance in the nearshore marine habitat. In “Estuarine Interactions” ( M . L. Wiley, ed.), pp. 191-219. Academic Press, London. McIntyre, A. D. (1970). Bibliography on methods of studying the marine benthos. F A 0 Fisheries Technical Paper 98, 96 pp. Mills. E. L. (1969). The community concept in marine zoology, with comments on continua and instability in some marine communities: A review. Journal of the Fisheries Research Board of Canada 26, 1415-1428. Mills, E. L. (1975). Benthic organisms and the structure of marine ecosystems. Journal of the Fisheries Research Board of Canada 32, 1657-1663. Mirza, F. B. and Gray, J. S. (1981). The fauna of benthic sediments from the organically enriched Oslofjord, Norway. Journal of Experimental Marine Biology and Ecology 54, 18 1-207. Moller, P., Pihl, L. and Rosenberg, R. (1985). Benthic faunal energy flow and biological interaction in some shallow marine soft-bottom habitats. Marine Ecology Progress Series 27, 109-1 2 1. Morisita, M. ( 1959). Measuring of interspecific association and similarity between communities. Memoirs of the Faculty of Science, Kyushu University Series E (Biology) 3, 65-80. Motyka, J., Dobrzanski, B. and Zawadzki, S. (1950). Wstepny badania nad lakami poludniowo swhodniej Lubelszczyzny. (Preliminary studies on meadows in the

242

B. J. BURD E T A L .

south-east of the province Lublin.) Annales Universitatis Mariae Curie-Sklodowska Sectio E: Agricultura 5, 367447. Nelson, W. G. (1987). An evaluation of deviation from the lognormal distribution among species as a pollution indicator in marine benthic communities. Journal of Experimental Marine Biology and Ecology 113, 181-206. Nemec, A. and Brinkhurst, R. 0. (1988a). Using the bootstrap to assess statistical significance in the cluster analysis of species abundance data. Canadian Journal of Fisheries and Aquatic Sciences 45, 965-970. Nemec, A. and Brinkhurst, R. 0. 1988b. The Fowlkes-Mallows statistic and pattern recognition in cluster analysis. Canadian Journal of Fisheries and Aquatic Sciences 45, 971-975. Nichols, J. A. (1977). Benthic community structure near the Woods Hole sewage outfall. Internationale yevue der gesamten Hydrobiologie 62, 235-244. Ord, J. K. and Whitmore, G. A. (1986). The Poisson-inverse Gaussian distribution as a model for species abundance. Communications in Statistics, Theory and Methods 15, 853-871, Orloci, L. (1975). “Multivariate Analysis in Vegetation Research”. Junk, The Hague. Pearson, T. H. (1975). The benthic ecology of Loch Linnhe and Loch Eil, a sea-loch system on the west coast of Scotland. IV: Changes in the benthic fauna attributable to organic enrichment. Journal of Experimental Marine Biology and Ecology 20, 1 4 1 . Pearson, T. H. and Rosenberg, R. (1978). Macrobenthic succession in relation to organic enrichment and pollution of the marine environment. Oceanography and Marine Biology Annual Reviews 16, 229-3 1 1. Pearson, T. H., Gray, J. S. and Johannessen, P. J. (1983). Objective selection of sensitive species indicative of pollution-induced change in benthic communities 2. Data analyses. Marine Ecology Progress Series 12, 237-255. Penas, E. and Gonzales, G. (1983). Relationships between benthic infauna and environmental factors in three beaches of the Ria de Arosa embayment (Spain) using canonical correlation analysis. Journal of Experimental Marine Biology and Ecology 68, 245-256. Petersen, C. G. J. (1913, 1914). Valuation of the sea: 11. The animal communities of the sea-bottom and their importance for marine zoogeography. Report of the Danish Biological Station 21, 1 4 8 . Petersen, C. G. J. (1915a). On the animal communities of the sea bottom in the Skagerrak, the Christiania fjord and the Danish waters. Report of the Danish Biological Station 23, 1-8. Petersen, C. G. J. (1915b). A preliminary result of the investigations on the valuation of the sea. Report of the Danish Biological Station 23, 29-32. Phillips, D. J. H. and Segar, D. A. (1986). Use of bio-indicators in monitoring conservative contaminants: Programme design imperatives. Marine Pollution Bulletin 17, 10-17. Pianka, E. R. (1966). Latitudinal gradients in species diversity: A review of concepts. American Naturalist 100, 33-46. Pielou, E. C. (1966). Shannon’s formula as a measure of species diversity: Its use and misuse. American Naturalist 100, 463-465. Pielou, E. C. (1969). An introduction to mathematical ecology. Wiley Interscience, New York. Pielou, E. C. (1977). “Mathematical Ecology”. John Wiley, New York, 385 pp. Platt, T. (1985). Structure of the marine ecosystem: Its allometric basis. Canadian Bulletin of Fisheries and Aquatic Sciences 213, 55-64.

BENTHIC MARINE INFAUNAL STUDIES

243

Platt, T. and Denman, K. (1977). Organisation in the pelagic ecosystem. Helgolunder wissenschaftliche Meeresuntersuchungen 30, 575-58 I . Pontasch, K. W. and Brusven, M. A. (1988). Diversity and community comparison indices: Assessing macroinvertebrate recovery following a gasoline spill. Water Research 22, 619-626. Poore, G. C. B. and Kudenov, J. D. (1978a). Benthos of the Port of Melbourne: The Yarra River and Hobsons Bay, Victoria. Australian Journal of Marine and Freshwater Research 29, 141-155. Poore, G. C. B. and Kudenov, J. D. (1978b). Benthos around an outfall of the Werribee sewage treatment farm, Port Phillip Bay, Victoria. Australian Journal of Marine and Freshwater Research 29, 157-167. Poore, G. C. B. and Rainer, S . (1979). A three-year study of benthos of muddy environments in Port Phillip Bay, Victoria. Estuarine and Coastal Marine Science 9,477-497. Popham, J. D. and Ellis, D. V. (1971). A comparison of traditional, cluster and Zurich-Montpellier analyses of infaunal pelecypod associations from two adjacent sediment beds. Marine Biology 8, 26CL266. Preston, F. W. (1948). The commonness and rarity of species. Ecology 29, 254-283. Preston, F. W. (1962). The canonical distribution of commonness and rarity. Part I. Ecology 43, 185-21 5 . Preston, F. W. (1980). Noncanonical distributions of commonness and rarity. Ecology 61, 88-97. Probert, P. K. and Wilson, J. B. (1984). Continental shelf benthos off Otago Peninsula, New Zealand. Estuarine and Coastal Sherf Science 19, 373-39 1. Quinn, J. F. and Dunham, A. E. (1983). O n hypothesis testing in ecology and evolution. American Naturalist 122, 602-617. Raffaelli, D. (1987). The behaviour of the nematode/copepod ratio in organic pollution studies. Marine Environmental Research 23, 135-152. Raffaelli, D. G. and Mason, C. F. (1981). Pollution monitoring with meiofauna, using the ratio of nematodes to copepods. Marine Pollution Bulletin 12, 158-163. Ramsay, J. 0. ( 1 982). Some statistical approaches to multidimensional scaling. Journal of the Royal Statistical Society A145,285-312. Raup, D. M. and Crick, R. E. (1979). Measurement of faunal similarity in paleontology. Journal of Paleontology 53, 12 13-1 227. Rees, H. L. (1984). A note on mesh selection and sampling efficiency in benthic studies. Marine Pollution Bulletin 15, 225-229. Reish, D. J. (1959). A discussion of the importance of the screen size in washing quantitative marine bottom samples. Ecology 40, 307-309. Rejmanek, M., Parton, E., Sikora, J. P. and Sikora, W. B. (1985). Dominance in marine ecosystems: A comment. American Naturalist 126, 724727. Reyment, R. A. (1963). Multivariate analytical treatment of quantitative species associations: An example from palaeoecology. Journal qf Animal Ecology 32,535547. Rhoads, D. C. and Germano, J. D. (1986). Interpreting long-term changes in benthic community structure: A new protocol. Hydrobiologia 142, 291-308. Ripley, B. D. (1981). “Spatial Statistics”. John Wiley, New York. Ristich, S . S., Crandall, M. and Fortier, J. (1977). Benthic and epibenthic macroinvertebrates of the Hudson River. Estuarine and Coastal Marine Science 5, 255266. Rosenberg, R. (1973). Succession in benthic macrofauna in a Swedish fjord subsequent to the closure of a sulphite pulp mill. Oikos 24, 244-258.

244

B. J. BURD E T A L .

Rosenberg, R. (1980). Effect of oxygen deficiency on benthic macrofauna in fjords. In “Fjord Oceanography” (H. J. Freeland, D. M. Farmer and C. D. Levings, eds), pp. 569-602. Nato Conference Series IV: Marine Sciences 4. Plenum Press, New York. Roughgarden, J. (1983). Competition and theory in community ecology. American Naturalist 122, 583-601. Rowe, G . T., Polloni, P. T. and Haedrich, R.L. (1982). The deep-sea macrobenthos on the continental margin of the northwest Atlantic Ocean. Deep-sea Research 29, 257-278. Rutledge, P. A. and Fleeger, J. W. (1988). Laboratory studies on core sampling with application to subtidal meiobenthos collection. Limnology and Oceanography 33, 274-280. Rygg, B. (1985a). Distribution of species along pollution-induced gradients in benthic communities in Norwegian fjords. Marine Pollution Bulletin 16, 469474. Rygg, B. (1985b). Effect of sediment copper on benthic fauna. Marine Ecology Progress Series 25, 83-89. Rygg, B. (1986). Heavy-metal pollution and log-normal distribution of individuals among species in benthic communities. Marine Pollution Bulletin 17, 3 1-36. Saila, S. B., Pikanowski, R. A. and Vaughan, D. S. (1976). Optimum allocation strategies for sampling benthos in New York Bight. Estuarine and Coastal Marine Science 4, 119-128. Sanders, H. L. (1960). Benthic studies in Buzzard’s Bay. 111. The structure of the softbottom community. Limnology and Oceanography 5, 138-1 53. Sanders, H. L. (1968). Marine benthic diversity: A comparative study. American Naturalist 102, 243-282. Sanders, H. L. (1969). Benthic marine diversity and the stability-time hypothesis. In “Diversity and Stability in Ecological Systems”. Brookhaven Symposium in Biology No. 22, Upton, New York. Satsmadjis, J. (1985). Comparison of indicators of pollution in the Mediterranean. Marine Pollution Bulletin 16, 395-400. Schmidt, G . H. and Garbutt, D. J. (1985). Species abundance data from fouling communities conform to the gamma distribution. Marine Ecology Progress Series 23, 287-290. Schwinghamer, P. (1981). Characteristic size distributions of integral benthic communities. Canadian Journal of Fisheries and Aquatic Sciences 38, 1255-1263. Schwinghamer, P. ( 1 983). Generating ecological hypotheses from biomass spectra using causal analysis: A benthic example. Marine Ecology Progress Series 13, 151166. Schwinghamer, P. (1988). Influence of pollution along a natural gradient and in a mesocosm experiment on biomass-size spectra of benthic communities. Marine Ecology Progress Series 46, 199-206. Shackley, S. E. and Collins, M. (1984). Variations in sublittoral sediments and their associated macro-infauna in response to inner shelf processes; Swansea Bay, U.K. Sedimentology 31, 793-804. Shannon, C. E. and Weaver, W. (1963). “The Mathematical Theory of Communication”. University of Illinois Press, Urbana. Sheldon, R. W., Prakash, A. and Sutcliffe, W. H., Jr (1972). The size distribution of particles in the ocean. Limnology and Oceanography 17, 327-340. Sheldon, R. W., Sutcliffe, W. H. and Paranjape, M. A. (1977). Structure of pelagic food chain and relationship between plankton and fish production. Journal of Fisheries Research Board of Canada 34, 2344-2353.

BENTHIC MARINE INFAUNAL STUDIES

245

Shelford, V. E. (1935). Some marine biotic communities of the Pacific coast of North America. Part I: General survey of the communities - their extent and dynamics. Ecological Monographs 5, 249-324. Shepard, R. N. (1 980). Multidimensional scaling, tree-fitting and clustering. Science 10. 390-398. Shin. P. K . S. (1982). Multiple discriminant analysis of macrobenthic infaunal communities. Journal of Experimental Marine Biology and Ecology 59, 39-50. Shin, P. K. S. and Thompson, G . B. (1982). Spatial distribution of the infaunal benthos of Hong Kong. Marine Ecology Progress Series 10, 37-47. Simberloff, D. (1977). Use of rarefaction and related methods in ecology. American Society for Testing and Materials, Special Technical Publication 652, 15&165. Simberloff, D. ( 1 978). Using island biogeographic distributions to determine if colonization is stochastic. American Naturalist 112, 7 13-726. Simberloff, D. (1983). Competition theory, hypothesis testing, and other community ecological buzzwords. American Naturalist 122, 62G635. Smedes, G. W. and Hurd, L. E. (1981). An empirical test of community stability: Resistance of a fouling community to a biological patch-forming disturbance. Ecology 62, 1561-1572. Smith, A. L. (1981). Comparison of macrofaunal invertebrates in sand dollar (Dendraster excentricus) beds and in adjacent areas free of sand dollars. Marine Biology 65, 191-198. Smith, E. P., Genter, R. B. and Cairns, J., Jr (1986). Confidence intervals for the similarity between algal communities. Hydrobiologia 139, 237-245. Smith, R. W. and Greene, C. S. (1976). Biological communities near submarine outfall. Journal of the Water Pollution Control Federation 48, 1894-1912. Smith, R. W., Bernstein, B. B. and Cimberg, R. L. (1988). Community+nvironmental relationships in the benthos: Applications of multivariate analytical techniques. In “Marine Organisms as Indicators” (D. F. Soule and G . S. Kleppel, eds), pp. 247-326. Springer-Verlag, New York. Smith, W. (1978). Environmental survey design: A time series approach. Estuarine and Coastal Marine Science 6, 2 17-224. Sneath, P. H. A. (1966). A comparison of different clustering methods as applied to randomly-spaced points. Classification Society Bulletin 1, 2-1 8 . Sneath, P. H. A. and Sokal, R. R. (1973). “Numerical Taxonomy: The Principles and Practice of Numerical Classification”. W. H. Freeman, San Francisco, 573 pp. Soule, D. F. and Keppel, G . S. (1988). Marine Organisms as Indicators. SpringerVerlag, New York, 342 pp. Stephenson, W. and Cook, S. D. (1980). Skewness of data in the analysis of speciesin-sites-in-times. Proceedings of the Royal Society, Queensland 91, 37-52. Stephenson, W., Williams, W. T. and Cook, S. D. (1972). Computer analyses of Petersen’s original data on bottom communities. Ecological Monographs 42, 387415. Strauss, R. E. (1982). Statistical significance of species clusters in association analysis. Ecology 63, 634-639. Strong, D. R. Jr (1983). Natural variability and the manifold mechanisms of ecological communities. American Naturalist 122, 636660. Taylor, L. R.(1961). Aggregation, variance and the mean. Nature, London 189, 732735. Taylor, L. R. (1980). New light on the variance/mean view of aggregation and transformation: Comment. Canadian Journal of Fisheries and Aquatic Sciences 37, 1330-1332.

246

B. J. BURD ET AL.

Taylor, L. R., Woiwod, I. P. and Perry, J. N. (1978). The density-dependence of spatial behaviour and the rarity of randomness. Journal of Animal Ecology 47, 383406. Taylor, W. D. (1980). Comment on aggregation, transformations and the design of benthos sampling. Canadian Journal of Fisheries and Aquatic Sciences 37, 13281329. Tetra Tech, Inc. (1986). Recommended protocols for sampling and analyzing subtidal benthic macroinvertebrate assemblages in Puget Sound. Draft Reporf TC-3991-04. U.S. Environmental Protection Agency. Thistle, D. (198 I). Review: Natural physical disturbances and communities of marine soft bottoms. Marine Ecology Progress Series 6, 223-228. Thistle, D. (1983a). The role of biologically produced habitat heterogeneity in deepsea diversity maintenance. Deep-sea Research 30, 1235-1 245. Thistle, D. (1983b). The stability-time hypothesis as a predictor of diversity in deepsea soft-bottom communities: A test. Deep-sea Research 30, 267-277. Thorson, G . (1957). Bottom communities (sublittoral or shallow shelf). In “Treatise on Marine Ecology and Paleoecology” (J. W. Hedgpeth, ed.), Vol. 1, pp. 461534. Memoirs of the Geological Society of America, 67. Thorson, G . (1966). Some factors influencing the recruitment and establishment of marine benthic communities. Netherlands Journal of Sea Research 3, 267-293. Toft, C. A. and Shea, P. J. (1983). Detecting community-wide patterns: Estimating power strengthens statistical inference. American Naturalist 122, 618-625. Torgerson, W. S. (1952). Multidimensional scaling: I. Theory and method. Psychometrika 17, 401419. Torgerson, W. S. (1958). “Theory and Methods of Scaling”. John Wiley, New York. Tukey, J. (1977). “Exploratory Data Analysis”. Addison-Wesley, Reading, Mass., 688 pp. Tyler, P. and Shackley S. E. (1978). Comparative efficiency of the Day and SmithMcIntyre grabs. Estuarine and Coastal Marine Science 6, 4 3 9 4 5 . Ugland, K . I. and Gray, J. S. (1982). Log-normal distributions and the concept of community equilibrium. Oikos 39, 171-1 78. Verner, J., Pastorok, R., OConnor, J., Severinghaus, W., Glass, N. and Swindel, B. (1 985). Ecological community structure analysis in the formulation and enforcement of law and policy. American Statistician 39, 393402. Vezina, A. F. (1988). Sampling variance and the design of quantitative surveys of the marine benthos. Marine Biology 97, 151-155. Warwick, R. M. (1984). Species size distributions in marine benthic communities. Oecologia 61, 3241. Warwick, R. M. (1986). A new method for detecting pollution effects on marine macrobenthic communities. Marine Biology 92, 557-562. Warwick, R. M. (1988a). The level of taxonomic determination required to detect pollution effects on marine benthic communities. Marine Pollution Bulletin 19, 259-268. Warwick, R. M. (1988b). Analysis of community attributes of the macrobenthos of Frierfjord/Langesundfjord at taxonomic levels higher than species. Marine Ecology Progress Series 46, 167-170. Warwick, R. M. and Davies, R. M . (1977). The distribution of sublittoral macrofauna communities in the Bristol Channel in relation to the substrate. Estuarine and Coastal Marine Science 5 , 267-288.

BENTHIC MARINE INFAUNAL STUDIES

247

Warwick, R. M., Pearson, T. H. and Ruswahyuni (1987). Detection of pollution effects on marine macrobenthos: Further evaluation of the species abundance/ biomass method. Marine Biology 95, 193-200. Washington, H. G. (1984). Diversity, biotic and similarity indices: A review with special relevance to aquatic ecosystems. Water Research 18, 653-694. Weinberg, S. (1978). The minimal area problem in invertebrate communities of Mediterranean rocky substrata. Marine Biology 49, 3 3 4 0 . Whittaker, R. H. (1967). Gradient analysis of vegetation. Biological Reviews 49, 207264. Whittaker, R. H . (1970). “Communities and Ecosystems”, 2nd edn. Macmillan, New York, 385 pp. Widbom, B. and Elmgren, R. (1988). Response of benthic meiofauna to nutrient enrichment of experimental marine ecosystems. Marine Ecology Progress Series 42, 257-268. Wigley, R. L. ( I 967). Comparative efficiencies of Van Veen and Smith-McIntyre grab samples as revealed by motion pictures. Ecology 48, 168-169. Wilkinson, L. (1988). “SYSTAT: The System for Statistics”. Systat Inc., Evanston, Ill. Williams, W. T. (1971). Principles of clustering. Annual Reviews in Ecology and Systematics 2, 303-326. Williams, W. T. and Stephenson, W. (1973). The analysis of three-dimensional data (sites x species x times) in marine ecology. Journal of Experimental Marine Biology and Ecology 11, 207-227. Williamson, M. H. (1978). The ordination of incidence data. Journal qf Ecology 66, 91 1-920. Wilson, J. B., Holme, N. A. and Barrett, R. L. (1977). Population dispersal in the brittle-star Ophiocomina nigra (Abildgaard) (Echinodermata: Ophiuroidea). Journal of the Marine Biological Assocation of the United Kingdom 57, 405-439. Winiecki, C . I. (1986). Benthic community development in Boca de Quadra, Alaska. MSc thesis, University of Alaska, Fairbanks, Alaska.

This Page Intentionally Left Blank

Plankton Production and Year-class Strength in Fish Populations: an Update of the Match/ Mismatch Hypothesis D. H. Cushing 198 Yurmouth Road, Lowestoft, Suflolk, U.K. NR32 4AB

Introduction ., .. .. .. .. .. .. .. .. The Original Hypothesis .. .. .. .. .. .. .. I l l . The MemberiVagrant Hypothesis .. .. .. .. .. .. IV. Extension to Waters Equatorward of 40" Latitude .. .. .. .. A. Background . . , . .. .. .. .. .. .. .. B. Events in upwelling areas . . .. .. .. .. .. .. C. Lasker events .. .. .. .. .. .. .. .. D. Spawningand food . . . . .. .. .. .. .. .. V. Testing the Hypothesis . . . . .. .. .. .. .. .. VI. Some Examples . . . . .. .. .. .. .. .. .. A. The original approach .. .. .. .. .. .. .. B. The spawning distributions of seven species in the Southern Bight of the North Sea . , .. .. .. .. .. .. .. .. C. The link between recruitment and larval abundance .. .. .. D. The Baltic , . .. .. .. .. .. .. .. .. E. The gadoid outburst .. .. .. .. .. .. .. F. The great salinity anomaly of the 1970s . , .. .. .. .. G. Catches of spiny lobsters and the time of onset of the spring outburst off Tasmania ., .. .. .. .. .. .. .. .. H. Larvae of the Arcto-Norwegian cod .. .. .. .. .. VII. Discussion .. .. .. .. .. .. .. .. .. VIII. Acknowledgements .. .. .. .. .. .. .. .. IX. References .. .. .. .. .. .. .. .. ..

I.

11.

250 25 1 256 259 259 259 26 1 262 264 265 265 270 27 1 273 274 276 278 279 28 1 285 285

Copyrighc Q /Po0 by Academic Press Limited A / / rights of reproduction in any form reserved

ADVANCES IN MARINE BIOLOGY VOLUME 26 ISBN 0-12-026126-X

249

250

D. H. CUSHING

1. Introduction The degree of match and mismatch in the time of larval production and production of their food has been put forward as an explanation of part of the variability in recruitment to a stock of fish (Cushing, 1974, 1975, 1982). The magnitude of recruitment is not completely determined until the yearclass finally joins the adult stock, and the processes involved probably begin early in the life-history of the fish when both their growth and mortality rates are high (Ricker, 1954). Hjort (1914) thought the level of recruitment was established during this period between hatching and first-feeding (the critical period according to Marr, 1956; May, 1974). However, I extended the match/mismatch hypothesis to cover the subsequent development through larval life up to metamorphosis, and possibly just beyond. An essential part of my hypothesis was the suggestion of Ricker and Foerster ( 1948) that under adventitious predation, i.e. without aggregation, well-fed larvae grow quickly and experience less predatory mortality than poorly fed ones at a given stage of development (Cushing and Harris, 1973; Shepherd and Cushing, 1980). Hence, one would expect growth and mortality to be inversely related. Almost as a consequence of match and mismatch, fish in temperate waters should release their larvae during the spring or autumn peaks in the production cycle, when more food is available. Anderson (1988) listed a number of hypotheses regarding the survival of pre-recruits, and he came to the conclusion that the growth/mortality hypothesis (as he named it) was most significant. As well as extending my match/mismatch hypothesis to cover a longer period in the life-history of fish, it was also more explicitly related to climatic factors and to the Sverdrup (1958) model than Hjort's first hypothesis. Hjort's (1914) second hypothesis, i.e. that recruitment may be determined by the loss of larvae through advective processes, was revived by Bailey (1981), who found recruitment to the hake stock off the west coast of the U.S.A. to be negatively correlated with an index of upwelling. I decided to examine the effect of upwelling on recruitment in waters equatorward of 40" latitude and, subsequently, I re-examined the data in support of my original hypothesis together with more recent information, including the member/ vagrarlt hypothesis. The match/mismatch hypothesis has developed over a number of years in several publications, and I thought it desirable to bring all of the evidence together here. However, it should be noted that although my original hypothesis applied only to recruitment to fish stocks, I now also include spiny lobsters and Dungeness crabs.

THE MATCH/MISMATCH HYPOTHESIS UPDATED

25 1

II. The Original Hypothesis The original match/mismatch hypothesis consisted of two parts: first, that fish in temperate waters spawned at a fixed time and, secondly, the larvae were released during the spring or autumn peaks in plankton production. The variation in production of larval food (e.g. copepod nauplii and copepodites for many fish; the appendicularian Oikopleura for the plaice, Pleuronectes platessa L. in the southern North Sea) then depended on the variation in the time of onset and the duration of primary production (Cushing, 1974, 1975). The hypothesis assumed that during the spring peak, plankton production followed Sverdrup’s ( 1 953) model, and that during the autumn peak the same principles applied, but in reverse - spring and autumn are the periods of mixing or weak stratification. The hypothesis was based on two observations. First, the spring and autumn spawning herring stocks (Clupea harengus L.) in the northeast Atlantic did in fact release their larvae at a time when they would grow in the spring and autumn peaks in plankton production (Cushing, 1967). The seasonal estimates of “greenness” for each spawning area were averaged over a number of years using the Continuous Plankton Recorder Network (Cushing, 1967). The plankton recorder has a coarse mesh through which the smaller algal cells are able to escape when present in small quantities; however, when present in large quantities, they are retained on the mesh. During peaks of phytoplankton, the production of all algae is governed by the critical depth and the production ratio (i.e. compensation depth/depth of mixing). The estimate of “greenness” as a relative index of the quantity of phytoplankton was validated by Gieskes and Kraay (1977). I also reported each stock’s peak time for spawning, the spread in time over which spawning took place and the number of days to the point of no return (i.e. the time at which the larvae must have food in order to survive). The autumn spawners hatch in the western North Sea and the larvae then drift and migrate as young juveniles to the nursery grounds in the east (Cushing, 1962). On this eastward movement, they feed on zooplankton which grows in a pronounced autumn outburst of “greenness” (Cushing, 1967). The areas from which these “greenness” data were taken, together with the spawning grounds of the herring (Harden Jones, 1968), are shown in Fig. 1. Buckmann (1 942) was able to present some evidence for this drift across the North Sea by larvae from the Downs and Dogger spawning groups. Hainbucher et al. (1986) have provided average trajectories of water particles (in “Sverdrups”) under wind stress between 1969 and 1982, starting from different points. The relevant courses from the Thames and Rhine (the Downs spawning group), the Humber (the Dogger spawning group) and an

252

D. H. CUSHING

area off the Moray Firth (the Buchan spawning group, which includes the West of Scotland) are plotted in Fig. 1 (Saville and Bailey, 1979). Note that the term “spawning group” has been used to avoid confusion with the present unit of management, i.e. the North Sea stock. Bartsch and Backhaus (1987) have calculated similar trajectories with different forms of vertical migration made by herring larvae. They suggest that year-to-year differences in advection may contribute to the annual variability of recruitment. Subsequently, Bartsch et al. (1989) presented evidence to show that herring larvae cross the North Sea from their spawning grounds in the west to their nursery grounds in the east -and that this passage is wind-influenced. The second observation (Cushing, 1969) was that four temperate species herring, plaice, sockeye salmon (Oncorhynchus nerka Walbaum) and cod (Gadus rnorhua L.)-spawn at a fixed season, the peak date of which has a standard deviation of about 1 week, although the fish may spawn for a period of 2-3 months. Two standard deviations amount to about 2 weeks, but the standard error of the mean date is only 1-2 days; Rothschild (1986) has noted that the first statistic shows the spread as the average of evolutionary processes, and that the second indicates the most important time for spawning. Subsequent analysis suggested that the spawning of the Arcto-Norwegian cod did, in fact, occur 1 week or so later over an extended period between 1891 and 1931 (Cushing and Dickson, 1976). The material was based on the weekly distribution of catches in the Vestfjord for many decades. Using evidence of maturation, (Pedersen, 1984) found that the peak date for spawning is somewhat later than the date of peak catch. Ellertsen et al. (1987), taking the dates of 50% spawning at Lofoten between 1976 and 1983, have shown that the real date of peak spawning is 14 days later than that of peak catch. The main conclusion was that the aforementioned four species spawned at about the same time each year in temperate or high latitudes. The time of onset of the spring peak in plankton production in the North Sea and north east Atlantic can vary from year to year by as much as 6 weeks, as shown by the Continuous Plankton Recorder (CPR) data (Colebrook, 1965; Colebrook and Robinson, 1965; Robinson, 1970). Colebrook (1979) has extended the analysis of the CPR observations and has illustrated spatial and temporal differences. Dickson et al. (1988a) found that the time of onset of the spring peak in plankton production in the western North Sea shifted 4-6 weeks later between the 1950s and the 197Os, shifting back again in the 1980s. Colebrook (1982) also showed that the spring peak in the northeast Atlantic occurred before the water column was fully stabilized, i.e.

THE MATCH/MISMATCH HYPOTHESIS UPDATED

253

FIG. I . The autumn spawning grounds of the North Sea herring (black areas: from Harden Jones, 1968) and the climatological mean transports (arrows: Hainbucher et a / . , 1986) from the spawning grounds (Buchan and Dogger, autumn transport; Downs, winter transport) to the nursery grounds in the eastern North Sea. The hatched areas include the small squares used for sampling the “greenness” in each of the presumed larval drifts.

254

D. H. CUSHING

the period of weak stratification. There are two main ecosystems in the sea, fully stratified and weakly stratified water (Cushing, 1989). The original match/mismatch hypothesis was based on the contrast between the relatively fixed time of spawning and the highly variable time of peak production. However, to this should be added the possibility of the start of production being brought forward or delayed. The recognition of the extent of thermal fronts in tidal seas (Simpson and Hunter, 1974; Pingree et al., 1975, Le Fevre, 1986) has changed our view of the production cycle in two ways. First, in protected coastal waters shoreward of the fronts, there is a continuous production cycle which peaks in June or July, (Grall, quoted by Le Fevre, 1986; Kremer and Nixon, 1978; Diwan, quoted in Raymont, 1980; Franz and Gieskes, 1984). In relatively shallow and protected water with the constant attenuation of irradiance, production starts when the critical depth exceeds the depth of water, and the subsequent course of production is governed by the production ratio. In other words, with a constant coefficient of attenuation, the time of onset of production and its subsequent development depends only on irradiance. Production commences in shallow water and develops later in deeper water (Cushing, 1972) to a depth, where stratification begins, i.e. where the fronts develop in summer. However, such areas are often susceptible to strong winds in spring and the coefficient of attentuation is decreased (Lee and Folkard, 1969; Dickson and Reid, 1983). The time of onset is then delayed as a function of strong winds, as is the rate of development of peak production. Thus, in enclosed tidal areas, the time of onset of production is controlled solely by irradiance, whereas in open tidal areas it is also controlled by attenuation which is governed by the presence or absence of strong winds. It is interesting to note that herring spawn in relatively shallow waters and that their larvae grow in mixed water regimes. The second consequence of the recognition of thermal fronts is that phytoplankton production in temperate waters has been shown to be concentrated within the thermocline during the summer (Pingree et al., 1977; Holligan and Harbour, 1977). Production continues above the thermocline (Holligan et al., 1984). The result is spring and autumn peaks with production at the thermocline and in the euphotic layer in between. Because there is apparently no variation in the timing of the production in the summer in stratified waters, the match/mismatch hypothesis will not apply in that season. The matchjmismatch hypothesis is illustrated conceptually in Fig. 2. Because fish in temperate waters may spawn at a fixed season, production of fish larvae and that of their food may be matched or mismatched in time. Larval development is an inverse power function of temperature, and is extended in cold water. (Note that the error bars shown above the distri-

THE MATCH/MISMATCH HYPOTHESIS UPDATED

255

bution of production in time in Fig. 2 are asymmetrical. The variability in time of the production of larval food is represented by the much broader symmetrical error bands above the distribution of larval food.)

Eggs

H Frequency

Larval Food

Larvae

HMATCH

MISMATCH

FIG.2. The match/mismatch hypothesis. The production of eggs, larvae and larval food are shown as distributions in time. The match or mismatch is represented by the overlap in time between the production of fish larvae and that of their food. In temperate waters, fish may spawn at a fixed season and the production of early stage eggs reflects this. The production of larvae, like all egg stages, is inversely related to temperature by a power law; this is expressed in the asymmetrical error bars, based on inter-annual differences. The production of larval food depends upon the timing of the production cycle and the long error bar expresses the annual variation. With a low stock, there is a greater chance of mismatch than with a high stock (recruitment is usually more variable with a low stock).

Full match and full mismatch are equally unlikely, but they may correspond to uncommonly rich and poor year-classes. The question arises as to why fish have not evolved a mechanism to match their larval production to that of their food. As will be shown below, such mechanisms exist where the interval between batch spawnings depends upon food (at least in waters equatorward of 40" latitude). In temperate waters, such a device may well exist but may not be effective. Fish that spawn before the spring peak in plankton production have to forecast its time of onset, which is difficult taking into account the highly variable stress caused by the weather. One way of overcoming this problem is to spread spawning activity over a period of time (as pointed out by Dr John Shepherd, pers. comm.). The hypothesis of match/mismatch is based on the belief that some of the

256

D. H. CUSHING

variability of recruitment to temperate fish stocks is affected by climatic change. Williamson (1961) analysed a change from a neritic to an oceanic planktonic regime in the northern North Sea by Principal Component Analysis and showed that half of the variance was correlated with a change in the degree of vertical mixing. This was not a response by a single population, but one by a community. Dickson et al. (1988a) have shown that the delay of about 1 month in the time of onset of the spring peak and the decline in number of Pseudocalanus by a factor of five in the western North Sea, was associated with the increase in northerly winds and gales between the 1950s and the 1970s; both delay and decline were reversed in the early 1980s. Thus, in addition to the variation in time of onset of the spring peak, there are secular trends based on climatic change. Many correlations between recruitment and climatic factors, such as wind strength and temperature, have been published. Though their failure may have been due to improper statistical manipulation, it was more likely due to the highly non-linear properties of a system responding to climatic change (Rothschild, 1986). Cushing (1982) discussed the possible relations between recruitment and climatic change in some detail. Such relations may originate in the year-to-year differences in the time of onset of primary production, which is the simplest way in which climatic factors can affect recruitment. Shepherd et al. (1984) found a correlation between the first principal component of the recruitments of the stocks of nine fish species in the North Sea and the first principal component of temperature in that region. A similar conclusion emerged from Hollowed et al.’s ( 1 987) analysis of recruitment in the Alaska gyral; they found that recruitments were correlated within, but not between, regions several hundred kilometres across, the same size as depressions and anticyclones. Also, within such a region, some prominent year-classes were common to a number of stocks. Such are the current bases of the belief that climatic factors or climatic change affect the generation of recruitment at an early stage in the life-history of fish.

111. The Member/Vagrant Hypothesis In considering the member/vagrant hypothesis here, I discuss only the bearing of this hypothesis on that of match/mismatch and not its more general consequences. An essential part of the member/vagrant hypothesis is the idea of a “larval retention area”. For example, herring larvae are supposed to be retained by behavioural mechanisms close to thermal fronts for several months at places around the British Isles and on Georges’ Bank off the east coast of the U.S.A. They are said to grow there at a constant rate irrespective of the different types of food available (Iles and Sinclair, 1982).

THE MATCH/MISMATCH HYPOTHESIS UPDATED

257

More precisely: larvae of a discrete herring population develop within, and are thus adapted to the specific oceanographic conditions of their larval retention areas.. . larvae are not retained passively by the circulation features of the larval distributional area, but rather by an active behavioral response to the physical regime. . . populations are observed to maintain a discrete distribution in a relatively fixed location for several months” (Sinclair and Tremblay, 1984).

In the shelf waters around the British Isles, the fronts collapse during equinoctial gales and the larvae are not necessarily retained within the thermal fronts but, aided by winds, drift eastwards across the North Sea, as noted by Biickmann (1942; see also Cushing, 1986). More developed methods have confirmed this conclusion. The larvae and juveniles of capelin (Mallotus villosus L.), yellowtail flounder (Limanda ferruginea (Storer)) and American plaice (Hippoglossoides platessoides (Fabricius)) are retained on the southern Grand Bank (Frank et al., 1989) and a proportion of young pelagic cod are retained within a swirl on their spawning grounds on Brown’s Bank, the remainder drifting to the richer food grounds off southwest Nova Scotia (Suthers and Frank, 1989). Though retention areas d o exist, like that off southwest Nova Scotia for herring larvae (originally proposed by Iles and Sinclair, 1982, and confirmed by Stephenson and Power, 1989), Campana et al. ( 1 989) have found that during their passage from Brown’s Bank to southwest Nova Scotia, cod larvae are retained along the course of the larval drift before being dispersed over Brown’s Bank again. Furthermore, Heath and MacLachlan (1987), Heath et al. (l987), Kimboe and Johansen (1986) and Kiorboe et al. (1988) have all shown that patches of early autumn-spawned herring larvae drift considerable distances off the Scottish coasts before the fronts collapse. There is thus a question as to whether herring larvae are retained in the fronts around the Scottish coasts; indeed, those from the West coast of Scotland move quickly eastward into the North Sea (Heath, 1989). I conclude that retention areas do exist (usually for short periods), and where they exist they form part of the larval drift between the spawning grounds and nursery grounds. The crucial point is whether retention areas are necessary for growth and survival. Except off southwest Nova Scotia, they do not persist for very long. There are also places where retention areas do not exist at all, and one should beware of attributing properties of retention to hydrographic structures (see Sinclair, 1988, on the Southern Bight of the North Sea). A retention area may be described physically (as a swirl) or geographically (over a bank), and because it is really part of the larval drift or migration from the nursery ground to the spawning ground, one might imagine interannual differences arising in the course of that migration. Thus, different proportions of larvae might then reach a specified nursery ground (as

258

D. H. CUSHING

suggested by Bartsch et a[., 1989). Retention areas have stimulated a lot of research at sea, and if herring spawn in tidally mixed waters (to a depth of 100 m in the waters around the British Isles), their larvae must be found at the fronts or where the fronts had once been. The second question concerns the effect of food, and whether fish larvae grow at a constant rate and whether they are distributed independently of their food. It has already been noted that the young pelagic cod on Brown’s Bank drift towards richer feeding grounds. Frank (1988) showed that Pseudocalanus, the preferred food of the herring larvae off southwest Nova Scotia, was more abundant where the larvae lived. Munk et al. (1986), K i ~ r b o eand Johansen (1986) and Kimboe et al. (1988) have all shown that herring larvae growth is limited by food. From published observations at sea, Fortier and Gagne (in press) demonstrated that herring larvae grew at about half the maximal rate. They also found that spring-spawned larvae in mixed water in the St Lawrence hatched 2 weeks before the peak in plankton production, and that they had a good survival rate. In contrast, the autumnspawned larvae suffered heavy mortality after the food stocks ran out. I conclude that the growth of fish is modified by the availability of food along the course of the larval drift. Thirdly, Sinclair’s member/vagrant hypothesis states that the time of spawning and the time of onset of production should be decoupled (Sinclair and Tremblay, 1984). This is based on the idea that the time at which spawning takes place is governed by the need to metamorphose in summer, between April and October. Spring-spawning herring proceed to metamorphosis quickly, whereas autumn spawners do so slowly throughout the winter. Sinclair and Tremblay (1984) suggest (without evidence) that conditions in the spring retention areas are better than those in the autumn ones. Spring- and autumn-spawning herring do spawn before the spring and autumn peaks in plankton production (Cushing, 1967). The time at which seven species in the Southern Bight of the North Sea spawn is coupled to the time of onset of the spring peak in plankton production (see Fig. 5). The simplest explanation is that they do so in order to exploit the available food. It is possible that autumn-spawned herring larvae survive on the overwintering populations of Pseudocalanus and metamorphose in the following summer. The suggested decoupling of the time at which spawning takes place from the onset of production is not supported by the facts. An anonymous reviewer suggests that the two hypotheses be reconciled. Sinclair and Tremblay (1984) state that each spawning group must be adapted to “the specific oceanographic conditions of their larval retention area”. If and where there are retention areas along the passage of the larval drift, such conditions might include the availability of food. The spawning date of each population has evolved in response to the average timing of the onset of production of plankton that exists in a retention area. However,

THE MATCH/MISMATCH HYPOTHESIS UPDATED

259

such production cycles can exist independently of such areas, e.g. the seven species in the Southern Bight of the North Sea (see Fig. 5). Thus, the basis of the match/mismatch hypothesis holds for the variation of climatic factors, as expressed by the critical depth and the production ratio.

IV. Extension to Waters Equatorward of 40" Latitude A.

Background

In 1974, there was no direct evidence to show that recruitment was in fact linked to match or mismatch (Cushing, 1974). Later, it was noted that the Californian sardine (Sardinops caerulea (Girard)) spawned during many months of the year (Cushing, 1969). Similarly, in the oceans of the tropics and subtropics, it was shown that tuna and tuna-like species tended to spawn right across these waters for most seasons of the year. Hence, the form of mechanism displayed in Fig. 2 was not then considered to apply equatorward of 40" latitude. This conclusion is now reconsidered. I have noted (Cushing, 1989) that there are two primary ecosystems in the open ocean: one based on very small grazers in the microbial food loop, and the other based on rather larger ones ( > 5 pm). The first is found in the oligotrophic ocean and in the well-stratified summer waters of lakes and seas. (It will be recalled that the period of the summer thermocline in temperate seas was excluded from the match/mismatch hypothesis.) The second ecosystem is the traditional food chain - diatoms, copepods and fish-which is found where the water is weakly stratified, in the early and middle stages of a spring peak, in the later stages of an autumn peak and in an upwelling area or an oceanic divergence. Let us now examine events in tropical and subtropical seas outside the oligotrophic ocean. B. Events in Upwelling Areas In the last decade or so, many links have been established between indices of upwelling and recruitment. Because the food for fish larvae is produced in upwelling areas, positive correlations between recruitment and upwelling are to be expected. In recent years, negative correlations have been associated with offshore drift, and hence with the advective loss of larvae. However, I pointed out that the greatest production would derive from moderate or slow upwelling (Cushing, 197 I ) . Hence the simple distinction between positive and negative correlations for indicating the effect of food or offshore drift may be misleading.

260

D. H. CUSHING

Marr (1960) found a positive relationship between the recruitment of the Californian sardine and the inshore temperature at the Scripps pier, an inverse relation with upwelling (Marr did not attempt correlation because there was one “outlier” in the relationship, the 1939 year-class, for which he gave a reasonable explanation). Dickson et al. (1988a) also demonstrated an inverse relationship between upwelling (averaged over the 3 years preceding the catch) and the sardine (Sardinia pilchardus L.) catch off Portugal. There are also relationships between upwelling and sardine (Bakun and Parrish, 198l), mackerel (Scombev juponicus Houttuyn: Parrish and McCall, 1978), Dungeness crab (Cancer magister Dana: Peterson, 1973; Botsford and Wickham, 1975) and hake (Merluccius productus: Bailey, 1981) year-classes, some of which are direct, some inverse, some simple and others complex, some with catches lagged by 3 years and others with recruitment. Upwelling is a complex process and it has many forms, any of which may change its nature in the course of time. A negative relation between recruitment and an index of upwelling might indicate, among other things, that a proportion of the eggs and larvae is swept offshore. Bailey (1981) showed that recruitment to the stock of Pacific hake was inversely related to Ekman transport. Indeed, that best year-classes were recorded in years of downwelling (which, incidentally, would have brought the larvae inshore to their nursery ground). Such losses must exist, because no inherently variable current system can deliver all survivors to the nursery grounds. Indeed, the gonad must include a proportion devoted to such loss. Losses may be considered to be a fraction of subsequent recruitment with a constant predatory mortality; alternatively, losses may be compensated for by subsequent processes. In this respect, the larger the initial loss, the greater the subsequent survival, either because more food becomes available or because predators lose interest. The first alternative explicitly denies any dependence of mortality on food during larval drift; indeed, the larvae should grow at a constant or even a maximal rate, and the effects of predation should be independent of density. In the second, the processes of growth and mortality continue during the larval drift, in a similar manner to that indicated earlier as an essential part of the match/ mismatch hypothesis. I have suggested (Cushing, 1971) that the greatest production can only be achieved by a slow rate of upwelling because a fast rate will sweep eggs and larvae away before peak production is reached in the rising water and they become vulnerable to grazing. Bakun and Parrish (1981) and Husby and Nelson (1982) both suggest that pelagic fish tend to spawn where there is a relatively slow rate of upwelling. Cury and Roy (1989) distinguish upwellings linked to turbulence from those that are not. They suggest that recruitment depends directly on upwelling if turbulence is absent, as off the

THE MATCH/MISMATCH HYPOTHESIS UPDATED

26 1

Ivory Coast. In turbulent upwellings (Morocco, Peru, Senegal) with wind stress < 5 m/s, recruitment might become reduced. Hence, both direct and inverse relationships are expected as turbulence increases. Dickson et al. (1988a) found that the Pacific sardine, with its high growth rate, probably needs a relatively low but consistent rate of upwelling; the northern anchovy, which grows at half the rate of the sardine, tolerates a faster rate of upwelling (data from Ahlstrom, 1966). Thus positive and inverse links between recruitment and indices of upwelling can arise from turbulence and also from biological factors, such as differences in growth rates. Therefore, the dichotomy of positive/negative links between upwelling and recruitment do not indicate dependence upon food or the lack of it. A negative relationship between recruitment and upwelling can, of course, be the result of offshore drift, but there are also other explanations. C. Lasker Events Hunter (1972) reported that the larvae of the northern anchovy require food densities of 1 nauplius/l or 30 Gymnodinium splendensll. Lasker (1975) thought that such densities of food only existed in the chlorophyll layers of the thermocline off southern California. When a storm upset the stability of the water column off San Onofre, there was not enough food for the larvae (Lasker, 1978). Furthermore, when most production in 1975 was diverted into diatoms (which the anchovies did not eat) and a red tide, the ensuing year-class was the worst for 15 years (Lasker, 1981). The use of daily growth ring counts on larval otoliths (Brothers et al., 1976) yields a distribution of birth dates (Lasker, 1981), thus allowing the train of survival to be ascertained. Such distributions vary in time from year to year, in contrast to the fixed spawning season in temperate and high-latitude waters. Peterman and Bradford (1987) have demonstrated an inverse relationship between larval death rate and the number of calm periods per month (i.e. 4 days of wind stress < 5 m/s, - a Lasker event). Peterman et al. (1988) showed that there was no relationship between the number of 19-day-old larvae and subsequent recruitment in the northern anchovy (perhaps the time-series of recruitment estimates was too short). Mendelsohn and Mendo (1987) reported that there were few Lasker events off Peru. Cury and Roy (1989), however, noted an inverse relationship between turbulence and the number of Lasker events off Peru, thereby confirming Lasker’s hypothesis. Lasker’s work suggested that the processes which determine the magnitude of recruitment begin during larval life, and that they may depend upon the larval food supply. Upwelling off southern California is intermittent and northern anchovy larvae may need Lasker events to survive the first feeding stage, but their slow growth rate also allows them to survive more upwelling.

262

D. H. CUSHING

D. Spawning and Food Many marine species produce several batches of eggs in a year. The northern anchovy completes many cycles in a season, from primary oocyte to ovulation, each of which lasts about 1 month (Hunter and Leong, 1980). Hunter and Leong also suggested that the number of batches per spawning season in the northern anchovy varied with ration. The growth of oocytes is governed by factors that affect body growth, i.e. food and metabolic hormones. Wootton (1977) showed that egg production per spawning batch in the stickleback (Gasterosteus aculateus L.) did not depend on ration, but that the number of spawnings did, so that the interspawning interval was inversely related to ration. Therefore, upwelling areas or divergences present the chance of spawning in a food-rich environment. As periods of upwelling last for several months, there is every chance that such an opportunity will arise. This reproductive strategy may be characteristic of waters equatorward of 40" latitude. In the North Pacific subtropical anticyclone, Yamanaka ( 1 978) established an association between year-class strength and temperature in four tuna stocks (Fig. 3). The direct effect of temperature on the metabolism of the fish or on the rate of development of the eggs and larvae (which must be very rapid in the warm water, i.e. 2-3 days at the most), is unlikely to be more than slight. The wind stress across the ocean tends to be greater in the peripheral regions of the subtropical anticyclone where the tuna live and where the divergences are common (Hidaka and Ogawa, 1958). Divergences bring cool water to the surface and hence there is local plankton production. The tuna spawn for 9 months of the year and distribute their larvae across the whole Pacific (Matsumoto, 1966; Nishikawa et al., 1985) thereby spawning as they feed. High recruitment tends to occur when the wind stress is relatively great and presumably where the divergences are more extensive and perhaps more intense. But there may be an additional mechanism, in that the fish may be able to average the effects of their food supply during the interspawning interval merely because they are large, so that hormonal control responds to averaged changes in food concentration. When the Peruvian anchoveta (Engraulis ringens Jenyn) feeds on phytoplankton, it obtains a maintenance ration only, which is inadequate for growth and reproduction (Cushing, 1978, using the data of de Mendiola, 1971). In other words, they eat algae, apparently to their disadvantage. They spawn towards the end of the strong upwelling period and their larvae grow during the following period of weak upwelling when copepods comprise most of their food (Walsh, 1978). In an evolutionary sense, anchovies may adopt the reproductive strategy of spawning at the end of the period of strong upwelling, with the result that their larvae have food in the form of

263

THE MATCH/MISMATCH HYPOTHESIS UPDATED

nauplii and copepodites during the period of weak upwelling. However, the period of weak upwelling may, in fact, produce more food than the period of strong upwelling. Yellowfin

Yellowfin

Cold

Warm NStrong

Warm

Cold

Weak

FIG.3. Strong year-classes of tuna in cold water and poor year-classes in warm water in the Pacific - yellowfin in the West Pacific and West Indian Oceans, bigeye in the East Pacific and albacore in the Coral Sea (Yamanaka, 1978).

Thus, it appears that in these waters, fish - as exemplified by the anchovy, the sardine and the tuna may well match their reproductive strategy and physiology to feeding. In other words, they spawn as they feed and their larvae grow in the food patches and the mismatch is minimized. Poleward of 40" latitude, the same objective is achieved by minimizing the spread in time of the onset of production by spawning at a fixed season. Thus, we apply the match/mismatch hypothesis in its original form to waters poleward of 40" latitude, and add to it the thesis that reproduction is more nearly matched to the production of larval food in upwelling areas and oceanic divergences equatorward of 40" latitude. ~

264

D. H. CUSHING

V. Testing the Hypothesis Sinclair (1988) gives five tests of the match/mismatch hypothesis. The first contrasts the spring and autumn peaks in the Baie des Chaleurs in the Gulf of St Lawrence and the midsummer peak in the lower St Lawrence; the latter has a protected tidal production cycle. The spawning distributions in time of seven species are about the same in the two areas, but the larvae peak in May and June in the lower St Lawrence and in June and July in the Baie des Chaleurs. According to de la Fontaine et af. (1984a,b), there is a clear relationship between the number of larvae and the quantity of zooplankton in the Baie, but not in the lower St Lawrence, where the animals drift away from the sampling stations. This test does not deny the match/mismatch hypothesis, but it may well support it. Methot (1983) provided the second test based on the survival of the larvae of the northern anchovy (data derived from the relative ratio of birth dates to that of the larval distribution by months). He found that survival was greater in a year of low upwelling than in one of normal upwelling, and concluded that low upwelling prevented loss of larvae in offshore flow. As noted above, this might be too simple an explanation and, indeed, there may have been more food available in the year of low upwelling. For the third test, O’Boyle et af. (1984) found that eggs and larvae of four demersal fish species were present all year round on the Scotian shelf, so that there may have been no opportunity for match or mismatch. The material is presented as log n numbers on time so that we should look more closely at the peaks rather than the tails of the distributions. It can then be seen that they correspond reasonably well with the spring and autumn peaks in plankton production, as presented in Sinclair et af. (1984). The fourth test, based on Koslow (1984) and Koslow et af. (1987), showed that the recruitments of stocks over broad areas were correlated within species. Sinclair believed that the timing of spawning events and plankton blooms could not coincide over such distances. As noted above, Shepherd et af. (1984) and Hollowed et af. (1987) show that recruitments are correlated within large regions but not between them. The last test of the hypothesis is the link between the high survival rate of Pacific mackerel off Baja California and high sea level records from coastal tidal stations (Sinclair et af., 1985); the survival rates were negatively correlated with a number of estimates of primary production and microzooplankton abundance between Point Conception and San Diego. The high sea level occurs in the El Niiio years, and therefore the survival index is linked to northerly transport associated with the Kelvin wave possibly generated by El Niiio. But Parrish and McCall (1978) found positive correlations between log n recruitment of Pacific mackerel and sea-surface

THE MATCH/MISMATCH HYPOTHESIS UPDATED

265

temperatures off La Jolla and in Marsden Square 102 (between Point Conception and San Diego) as a result of warm water flowing from the south. They also found positive correlations with upwelling and divergence. The water rising to the surface in the El Niiio years would have been warm, with animals present not normally recorded from the region between Point Conception and San Diego. Hence the argument based on the negative correlation between survival index and food, further north, fails. Pacific mackerel depend on southerly transport, and therefore the argument for lack of advection for year-class success also fails. The match/mismatch hypothesis has not been tested and the interpretations of data are not convincing. If the match/mismatch hypothesis were true, we would expect that: ( 1 ) Fish would release their larvae into peaks of plankton production or

upwelling areas. (2) Because the essential processes depend upon food availability, larval growth rate would be expected to vary from year to year and to be nearly always less than the maximum, and that mortality due to constant predation should vary inversely with growth rate. (3) The earliest and crudest link between recruitment and food should be found during the larval stage.

VI. Some Examples In addition to the original approach, a number of other events will be considered. These are the spawning distributions of demersal fish in the Southern Bight of the North Sea, the link between recruitment and larval abundance, the increase in cod and herring stocks in the Baltic, the “gadoid outburst” in the North Sea, the great salinity anomaly of the 1970s, the link between spiny lobster catches off Tasmania and New South Wales, and the time of onset of the spring peak in plankton production and the larval drift of the Arcto-Norwegian cod. A.

The Original Approach

Figure 4 shows the extension of the original seasonal distributions of “greenness” (1948-62) to more recent years (1963-83). The early 1980s, during which the decline in the zooplankton since the 1950s (Glover et al., 1972) was reversed (Colebrook et af., 1984), is not considered because it was so short a period of time compared with 1948-62 and 1963-83. In the future, it would be desirable to use in addition and separately the decade 1983-93

266

D. H. CUSHING

for the following reason. The recovery of the northeast Atlantic communities in the early 1980s was most marked in 1983, when the Atlanto-Scandian stock of herring started to recover as well as herring stocks in the northern North Sea.

Norwegian Sea

1948-62 II It

0 Gu

1963-83

1948-62 4 1963- 83 $

(a)

Central Buchan

THE MATCH/MISMATCH HYPOTHESIS UPDATED

J

F 1

M 1

A 1

1948-62

M 1

J 1

J 1

A 1

S 1

O 1

N 1

D 1

1

-

1963-78

267

Jw Kv

1948-62

/*\

t Lu @Mu *Nu

1963-83

1948-62

1963-83

A

LV

Dogger

FIG.4. Seasonal cycles of production estimated by “greenness” in selected small squares of the Continuous Plankton Recorder Network, comparing 194842 with 1963-83. (a) Norwegian Sea to Central Buchan; (b) East Buchan to Central Dogger.

268

D. H. CUSHING

J

F M A M I

I

I

I

J

J 1

1

A 1

S 1

O

N

1

1

D I

I

I

1948-62

0 Lx

1

1948-62

0

ov

I

1963-83

0

ov

Lw

~~

I I

Downs

1948-62 A R S

PIymout h

1957-62

&a 1963-78

v-vx

Dunmore

x pq v Pr

THE MATCH/MISMATCH HYPOTHESIS UPDATED F M A M J

J I

I

I

I

J

A

I

I

I

S

O I

N I

269

D I

I

~

1948-62

4

Dk

1963-83

4

Dk

c

DI

b

48-62

Kq a63-78

Hebrides

FIG.4. (c) East Dogger, Downs and Plymouth; (d) Iceland and Hebrides. Such distributions were used in Cushing (1967) to show that the herring released their larvae at a favourable time in the production cycle.

In the western North Sea and in the southern Norwegian Sea, the spring peak in plankton production was delayed in 1963-83. In the eastern North Sea, there was n o delay, although the magnitude of both the spring and autumn peaks was probably reduced. As noted above, Dickson et al. (1988a) showed that the delay in the western North Sea was due to an increase in northerly winds and gales during the spring. The Atlanto-Scandian stock of herring did not recover in the later period but did recover in 1983. However, the Continuous Plankton Recorder observations in the southern Norwegian Sea had been curtailed before that date. There was no change in the time of onset of the autumn peak in the North Sea, although it may have been reduced in magnitude. However, Reid (1978) showed that the abundance of diatoms was much reduced during the autumn peak in the central North Sea during the 1970s, which may have been caused by the diminished numbers of Pseudocalanus, such copepods depending on cells > 5 pm in diameter. It is possible that the decline in North Sea herring recruitment in the 1970s was partly associated with this event.

270

D. H. CUSHING

B. The Spawning Distributions of Seven Species in the Southern Bight of the North Sea

Using egg distributions, Fig. 5 (adapted from Harding et al., 1978) shows the spawning periods of seven species in the Southern Bight of the North Sea: plaice, (Pleuronectes platessa L.), cod (Gadus morhua L.), sandeel (Ammodytes spp.), whiting (Merlangius merlangus L.), bib or pout (Trisopterus luscus L.), flounder (Pleuronectes flesus L.) and dab (Limanda limanda L.) (Cushing, 1990). Figure 5 also shows the increase in “greenness” averaged for the two periods 1948-62 and 1963-83, and for the two squares Ov and Ow of the Continuous Plankton Recorder Network, as an indication of the onset of the spring peak in plankton production. The seven species have timed their spawning periods so that their larvae grow during the spring peak in plankton production where their food develops. The plaice and sandeels feed on Oikopleura. The small cells taken by this appendicularian are produced under the same conditions as the larger phytoplankton sampled by the plankton recorder, i.e. production begins when the critical depth exceeds the depth of water; and the compensation depth as a fraction of the depth of water controls the rate of development of production.

Greenness

D

J

F

M

A

M

J

FIG.5. The spawning distributions in time of seven fish species in the Southern Bight of the North Sea. Also, the development of “greenness” in the Southern Bight for the two periods 194842 and 1963-83, from the Continuous Plankton Recorder Network (this method was validated by Gieskes and Kraay, 1977).

271

THE MATCH/MISMATCH HYPOTHESIS UPDATED

C.

The Link Between Recruitment and Larval Abundance

Brander and Thompson (1989) have shown that the night/day ratio of herring larval catches by high-speed nets (Gulf 111, V or their derivatives) increases with the sizes of the larvae. The mean night/day ratio increases from about 1.5 when the larvae are 10 mm in length to 2.4 when they are 20 mm in length, as the proportion escaping increases with size. Hence the correlation between recruitment and numbers of larvae might be expected to decrease with size (although no such evidence appears from Table 1). An important general point to be made here is that unless the larvae are properly sampled at all ages, the measures of growth and mortality needed for understanding the processes will be difficult to establish. Moreover, the high-speed nets used in the North Sea are more efficient than many others that are in general use. Burd and Parnell (1973), working with the Downs stock of herring, and MacCall (l979), working with the Californian sardine, have shown that in these two pelagic species at least, the magnitude of recruitment is linked in a rough way to the densities of their larvae. Van der Veer (1986) found that an index of recruitment to the southern North Sea plaice depends (only six observations) on the number of larval plaice in the flood waters of the Wadden Sea where the nursery ground lies. Parmanne and Sjerblom (1984, 1987) have sampled herring larvae and their food at seven stations in four areas off the coast of Finland. Using a Gulf V net, samples were taken each week during May-August between 1974 and 1986. They established a link between the abundance of herring larvae and subsequent recruitment (as 0-groups from virtual population estimates), as shown in Table 1. CORRELATION BETWEENNUMBERS OF HERRING LARVAEOFF FINLAND IN JUNEAND SUBSEQUENTRECRUITMENT, BY SIZE OF LARVAE AND BY AREAOF SAMPLING (After Parmanne and Sjerblom, 1984, 1987)"

1. TABLE

Area

Size

< IOmm 10-15 mm

>15mm > IOmm

29

30

32

0.714** 0.644* 0.632* 0.649*

0.492* 0.480 0.240 0.464*

0.209 0.798* 0.792** 0.828**

Area 31 was excluded from the data in this table because the number of herring larvae caught there were so low. * P < 0.05; ** P =z 0.02.

"

272

D. H. CUSHING

Parmanne and Sjsblom also established correlations between recruitment (as 0-groups in the virtual population analysis) and zooplankton biomass per herring larva, particularly in July and August. The value of this work lies in the fact that such correlations can be established within a relatively short time, about 1-2 months after hatching. Leggett and his colleagues have made a thorough study of the emergence of capelin larvae (from the seabed) and of their survival in the bays of eastern Newfoundland. The larvae emerge in relatively warm water during periods of onshore winds (Frank and Leggett, 1981, 1982). Offshore winds with local upwelling inhibit emergence because the upwelled water is cold and not turbulent. There was more zooplankton in the relatively warmer water. Finally, it was found that the continued emergence of larvae depended on the frequency of onshore winds, and that their subsequent growth depended on sustained temperatures, expressed as degree-days. Leggett et al. (1984) related year-class strength of capelin (as 0-groups from virtual population analysis) to the frequency of onshore winds and degreedays as follows: In R = 16.10 - 0.19 Wind

+ 0.19 Tempsum (r2 = 0.58)

where R is recruitment, Wind is the frequency of onshore winds and Tempsum is in degree-days. The importance of this model is that it was derived step-by-step from detailed observations of ( I ) the development of the eggs with temperature, (2) the effects of the onshore winds and (3) the presence of food in the water. It is a realistic model of the structure of events that lead to the establishment of a year-class. The survival of Arcto-Norwegian cod larvae will be discussed in more detail below. Figure 6 shows the dependence of numbers of 0-group cod on the post-larval index, or index of early juveniles, for 1979-85 (Sundby et ul., 1989). The juveniles, 2-3 months after hatching, still feed on Culunus. Recruitment indices at 3 years of age are correlated with the 0-group indices for 1966-81 (r2 = 0.67; but the 1985 and 1986 year-classes were probably poorly estimated by the 0-group survey: Anon. 1988). Thus we have evidence from capelin, herring, plaice and cod that the magnitude of recruitment can be detected up to 2 months after hatching. It has been believed that the magnitude of recruitment cannot be established until perhaps 9 months after hatching. There is now enough evidence to show that the processes by which recruitment is determined do begin during the larval stages. There is thus every reason to continue studies on the growth and mortality of fish larvae up to metamorphosis and perhaps just beyond.

THE MATCH/MISMATCH HYPOTHESIS UPDATED

0

, 20

I

I

40

60

273

Post larval index FIG. 6. Dependence of 0-group cod on the post-larval index of early juveniles (or post-larvae) for the Arcto-Norwegian cod stock (Sundby ei al., 1989). The 0-group index is used routinely for the prediction of recruitment in I.C.E.S. stock assessments.

D. The Baltic In the 1920s and early 1930s, catches of the Baltic cod amounted to less than 10,000 tonnes. Meyer and Kalle (1950) noted that the catches increased during the late 1930s along the southern shore of the sea from Lubeck to Gdansk. Alander (1948) first associated this increase in catches with the recruitments of 1941-44 and 1947, and with the high salinity inflows into the Baltic which occurred then. The increase in salinity would cause the eggs to float, rather than sink, as they had done frequently when the salinity was low. Further, as the authors note, “a rise of 10m or so in the halocline, where the fry hatch, gives better access to suitable food”. Beverton and Holt (1964) presented tables in which the yield per recruit is given in terms of the estimated vital parameters:

R,

I

80

= Y/W,Y’

where R, is the recruitment in the first year of life, Y is the catch in weight (kg, millions), W , is the weight at infinite age and Y‘ is the yield per recruit

274

D. H. CUSHING

per unit of W,. By periods of years (1925-28, 1935-39 and 1951-57) recruitment had increased by a factor of 17 between the 1920s and the mid1950s (Cushing, 1982). In the 1950s and subsequently, catches remained high at an annual level of more than 200,000 tonnes. Figure 7 shows the increase in size of catches of the cod in the Baltic since the 1930s. Catches of herring and sprat also rose by the same order during the period. Graumann (1966, 1974) showed that the larger and more buoyant eggs rose more quickly through the halocline to the region where the larvae fed and where the larval guts were full; below the halocline, the guts of the larvae were empty. Fonselius (1969) showed that between the 1930s and 1960s, the halocline rose from 80 to 60 m, which probably affected each yearclass as it was generated, from the early 1940s onwards.

I

1920

I

1930

I

I

I

1940

I

1950

I

I

1960

-1

1970

1980

Yea I

FIG. 7. Increase in annual cod catches (tonnes) in the Baltic between 1921 and 1976 (Cushing, 1982).

E.

The Gadoid Outburst

From 1962 onwards, the catches and stocks of five gadoid species in the North Sea increased sharply and remained high; the event has been termed the “gadoid outburst” (Cushing, 1984). In the northeastern North Sea, to which most larval cod drift in spring, the monthly abundance of Calanus (the preferred food of the larval cod: Marak, 1974), was set back by 1 month if we compare 1949-61 and 1962-78 (Fig. 8). An annual index of delay was

275

THE MATCH/MISMATCH HYPOTHESIS UPDATED

5

4

au

3

m 0

0.

'

J

F

M

A

M

I

I

J

J

Month

FIG.8. The seasonal abundance of Calanus (as loge N ) in the northeastern North Sea (area

B1 of the Continuous Plankton Recorder Network) for 1948-61 and 1962-67 (Cushing, 1984).

prepared, showing the number of months between March and July at which the stock of Calanus peaked in any 1 year (after July, the little cod probably start to eat small fish: see Cushing, 1980). Cod recruitment was positively related to the index of delay between 1957 and 1977, i.e. before the increase in stock in the first year of the gadoid outburst and subsequently. There is thus a little evidence that cod recruitment tends to be greater when Calanus is produced later. Dickson et al. (1974) reported a spatial and temporal correlation between cod recruitment and temperature. A stepwise multiple regression of cod recruitment on the following factors was calculated ( r 2 = 0.75): Calanus abundance -0.0013* Delay in production 0.4537** March temperature 0.1496** Calanus abundance x delay 0.0007* Temperature x delay 0.0854** Where

* P < 0.05 and ** P < 0.01

276

D. H. CUSHING

Cod recruitment was correlated with the index of delay and it was modified by the abundance of Calanus and by temperature. The delay occurred during larval and post-larval life and it affected the magnitude of recruitment. Such might be expected if the match/mismatch hypothesis were true.

F. The Great Salinity Anomaly of the 1970s Dickson et al. (1984), Dickson and Blindheim (1984) and Dickson et al. (1988b) have described the passage of the great salinity anomaly of the 1970s from the East Greenland Current to Iceland, West Greenland, the Grand Bank, across the Atlantic to the waters off the British Isles, the Norwegian Sea and the Barents Sea, in that order, between 1965 and 1982. The water mass, which was generated off east Greenland by cool northerly winds in winter in the early 1960s, was cooler and fresher than the surrounding sea; it was several hundred kilometres across and extended to a depth of 400m, retaining its identity for a period of 17 years. An example was made of the year-classes of 15 “deepwater” stocks that might have been affected by the passage of the anomalous water mass (Cushing, 1988). In waters such as the North Sea, mixed by tidal streams, the effect of the water mass upon recruitment might have been attenuated. Figure 9 shows some examples. Recruitment to the Iceland summer spawning herring was very low during the anomalous years 1965-71, and recruitments to a number of other stocks were low during other anomalous periods. Year-classes in the anomalous years were tested with a Wilcoxon rank test with respect to each stock series and, of the 15 stocks examined, 11 were significantly low. North of Iceland, primary production was reduced by half (Thoradottir, 1977) and zooplankton by a factor of three (Astthorsson et al., 1983) throughout the anomalous years 1965-71. Similarly, on the Grand Bank, zooplankton was reduced to the same degree for a period of 3 years (Robinson et al., 1975). The reduced production of plankton in the cooler and fresher water is possibly linked to the low year-classes. Production must have been delayed because it takes longer to heat the cooler water, but the magnitude of the expected delay is as yet unknown. The cause is either a very reduced rate of development due to low food supplies, or a reduction in food FIG.9. Recruitment to the Icelandic summer spawning herring, West Greenland cod. Grand Bank cod, West Scotland saithe, northeast Arctic cod, northeast Arctic haddock (and the 0groups of the two latter stocks), together with the years of the great salinity anomaly of the 1970s as it passed through the waters of the northeast Atlantic (Cushing, 1988). Significant relationships were established for I 1 out of 15 “deepwater” stocks in the North Atlantic.

277

THE MATCH/MISMATCH HYPOTHESIS UPDATED Iceland Summer Spawning n e r r m g Anomalous y e a r s s 1 9 6 5 - 7 1

West Greenlond Cod

-

8 0 - A

^#^^^^

Anomalous y e a r s . 1 9 6 9 - ? 2

-

oann

Norlh-Earl Arclac H a d d o c k Anomalous years, 1 9 7 8 - 8 1

7nnn

..

Norlh Sea Sailhe Anomalous yaars, 1 9 7 5 - 7 7

-

-- -.... -_--

- -- _.__North- Earl Arcttc Haddock Anomalous yaars. 1 9 7 8 - 8 1

-----

North-Earl Arclic Cod Anomalous years, 1 9 7 8 - 8 1

..

North-Earl Arctic Cod Anomalous yaors. 1 9 7 8 - 8 1

278

D. H. CUSHING

due to the delayed production during the period of transient thermoclines, before the seasonal thermocline becomes fully established. Whatever the proximate cause, the reduction in recruitment was possibly the consequence of the mismatch of the production of larvae to that of their food.

rn >r Y

3 3

N

150,

-100-50

-

I

I

I

1940

I

1950

I

I

1

I

I

I

I

1960

I

I

I

1970

I

1

I

1

1980

1

1

1990

FIG. 10. The link between the west winds and catches of spiny lobsters off Tasmania. The changes in zonal westerly winds over Tasmania between 1945 and 1985. The date of onset of the spring bloom off Maria Island (off western Tasmania) is given in (b). Note the extraordinary variation in time of onset. (c) The time-series of catches of spiny lobsters off Tasmania, lagged by 7 years to indicate the timing of recruitment.

G. Catches of Spiny Lobsters and the Time of Onset of the Spring Outburst o f Tasmania

Although fishes are my prime concern here, spiny lobsters share their habit of dispersing larvae into the pelagic environment and, as will be shown, their

THE MATCH/MISMATCH HYPOTHESIS UPDATED

279

recruitments may be matched or mismatched to the time of onset of the spring peak in plankton production. Figure 10a shows the time-series of zonal westerly winds over Tasmania between 1945 and 1985; the variation is determined by the position of the atmospheric high over Australia (which is linked to the El Niiio/Southern Oscillation). As a consequence, the time of onset of the spring peak in plankton production varies by as much as 3; months between September and early January (Fig. lob). Catches of spiny lobsters off Tasmania (lagged by 7 years to indicate the year of recruitment) were correlated with high sea temperatures off the east coast, i.e. with a late spring peak (in differenced observations in time series) as shown in Fig. 1Oc. Catches of spiny lobsters in New South Wales were correlated with an early spring peak (Harris et al., 1988). This study was successful partly because the time-series was a long one, but also because there was a wide variation in the time of onset of the spring peak in plankton production. The implication, which cannot be shown, is that the recruitment to the stocks of spiny lobsters depends on the match or mismatch of larval production to that of their food during the spring peak.

H. Larvae of the Arcto-Norwegian Cod The survival of cod larvae in the Atlantic current north of the Vestfjord in northern Norway has been described by Ellertsen et al. (1987, 1989). They have shown that the number of Calanus nauplii per larval gut depends upon the numbers of nauplii per litre (in the manner of an Ivlev curve). Figure 11 shows the inverse relationship between temperature and the date at which the stock of Calanus copepodite stage 1 reaches a maximum off Skrova in the Lofoten Islands; if the water is cold, production is late, and vice versa. For a reduction in temperature of 2"C, there is a delay of about 6 weeks, very much more than the few days expected by an extended time of development. It may originate in a delayed production cycle, much as production may have been delayed during the passage of the salinity anomaly of the 1970s. Ellertsen and his colleagues also show that the length/ weight relationship of the cod larvae varies from year to year. For 1982-85, the dry weights of larvae 6 mm long were estimated; they reach this length about 15 days after first-feeding (Yin and Blaxter, 1987). The dates of the peak stock of Calanus copepodite stage I for the 4 years were calculated from the dependence of the Julian date of peak stock on temperature. Table 2 shows that the cod larvae 6 mm long achieve a greater weight when their food is produced early.

280

D. H. CUSHING

Temperature FIG. 11. The inverse relationship between the date of maximal production of Calmus and temperature in the Barents Sea (Ellertsen el a / . , 1987). The range of 60 days cannot be accounted for by differences in the rate of development at the temperatures quoted.

TABLE 2. THELINKBETWEENTHE WEIGHTOF CODLARVAE WHEN 6 mm IN LENGTH A N D THE TIMING OF THE PEAKOF ABUNDANCE OF CALANUS COPEPODITE STAGEI (Based on Ellertsen et af., 1987, 1989)

1982 1984 I985 1983

Dry weight at 6 mm

Temperature

(PSI

("C)

Peak of abundance of Calanus (Julian days)

98.75 107.35 120.00 146.33

2.10 3.05 3.50 3.60

149 123 114 109

The earlier the production of Culunus, the greater the dry weight of larvae 6 m m long. Figure 12 (Ellertsen et ul., 1989) shows the match (1980, 1983, 1984, 1985) or mismatch (1960, 1981) of the production of cod larvae with their food (nauplii and copepodite stage l), i.e. Culanusfinmurchicus. Above a density of 5 nauplii/l, the cod larvae are well fed. The 1983 year-class was a

THE MATCH/MISMATCH HYPOTHESIS UPDATED

28 1

strong one, yet from Figs 12 d, e and f (1 983, 1984 and 1985), the matches appear to be the same. But the 1983 year-class individuals put on much more dry weight by the time they were 6 mm long and that year class was twice as strong as the other three year-classes (Anon., 1988). The two mismatched year-classes, (1960 and 1981) turned out to be poor ones (Rothschild, 1986; Anon., 1988); the 1980 year-class was well matched, but it probably suffered during the passage of the salinity anomaly. The production of larvae may be matched to that of their food, but the processes of growth and mortality during larval life must also be examined. A good match may initiate good growth, but it has to be sustained. Jones (1973) suggested that larval haddock grew with a cohort of Culanus on which they fed. The initial match or mismatch and the evolution of growth and mortality through the later stages of larval life are distinct but essential parts of the match/mismatch hypothesis. Though cod spawn a t the same time each year, the onset of production of larval food varies from year to year by as much as 6 weeks. Figure 6 shows how the numbers of post-larval cod (or early juveniles at about 60 days after hatching) are linked to the number of 0-groups and hence to recruitment. It remains to establish the relationship between the growing larvae at 15 days after first-feeding and the post-larvae (or early juveniles) which still feed on Culunus. Thus, recruitment depends upon the match of larval production to that of their food and probably upon growth and mortality during the larval stages.

VII. Discussion The match/mismatch hypothesis has now been extended to the upwelling areas and oceanic divergences equatorward of 40" latitude on the basis that fish in these regions release batches of eggs more frequently when they are well fed and, more generally, that pelagic fish may modify their reproductive strategies so that they can feed and spawn at the same time. However, the growth and mortality of fish larvae and post-larvae have not been covered, either by direct observation or by modelling. The main reason for this is that because the sampling of larvae has often been poor, the growth and mortality have not been well estimated. There are, however, exceptions, e.g. cod post-larvae or early juveniles (Sundby et al., 1989). However, Jones and Henderson (1988), using a simulation model with a fixed date for the spring peak in plankton production, showed that fish larvae should be hatched at the right place and at the right time, i.e. at the start of growth of their food cohort. Because the spring peak is short, the fish larvae need to grow fast, and there is an optimum period for start-

D. H. CUSHING

282

feeding. This is why spawning is spread over time. Such are the essential structures which provide match and mismatch when the date of the spring peak in plankton production varies. It also provides an explanation of why spawning is spread over time.

30-

nauplii

1.3

20.3

9.4

29.4

19.5

30 1980

nauplii 1-1

I20

20.

1.3

O/O

4030-

( c

20.3

9.4

I-'

29.4

19.5

1 1981

nauptii 1-1

THE MATCH/MISMATCH HYPOTHESIS UPDATED

283

30-

1.3

1.3

20.3

9.4

29.4

19.5

FIG. 12. The matches (1980, 1983, 1984 and 1985) and mismatches (1960 and 1981) of the production of cod larvae to that of their food (nauplii and copepodites stage I of Calanus finmarchicus: Ellertsen et al., 1989). Above a density of 5 nauplii/l, the cod larvae are well fed. Percentage distribution of newly spawned cod eggs ( 0 )and first-feeding larvae (0); and the production of copepod nauplii (A).

Nival ei af. (1988) modelled the effect of food on the transfer rate from stage to stage of crustacean larvae. Starvation (representing a physical event such as mixing) produces the expected delay in transfer and such delays have

284

D. H. CUSHING

disproportionate effects under predation. Therefore, a delay in predation is of great importance, particularly when production peaks in early development. This model illustrates the difficulties that occur when growth and mortality are allowed to interact. One of the disappointments of the last decade has been the attention paid to either starvation or predation separately, although Ricker and Foerster (1948) noted that they were probably linked. There are three consequences of the match/mismatch hypothesis. The first is that fish should release their larvae during the spring and autumn peaks in plankton production in temperate waters or into upwelling areas or divergences in low latitude seas. Spring and autumn spawning herring release their larvae during the spring or autumn peaks. In the southern North Sea, seven demersal species release their eggs during the spring peak. More generally, fish that spawn in spring or autumn temperate waters probably do so for the same reason. In waters equatorward of 40" latitude, fish spawn in areas of upwelling and offshore divergence. By spawning more frequently where food is rich, and by adapting their reproductive strategies to feed and spawn at the same time, fish can match their spawning to the production of food for their larvae, as they try to do in temperate waters. This is an extended form of the match/mismatch hypothesis. The response of fishes to upwelling is more complex than has been thought, particularly in that more food may be produced where upwelling is moderate, which is where the pelagic fish tend to spawn. Furthermore, the effects of turbulence appear to be critical; without it, there is a simple direct response of recruitment to an index of upwelling, but if it is too strong recruitment is reduced. The second consequence of the match/mismatch hypothesis is that correlations (if necessarily low) between larval numbers and recruitment are expected. Many such relationships have been established with indices of upwelling or divergence. This can probably only be effective during the larval stages, because the index of upwelling represents production, if in a somewhat complex manner, and hence the effects on larval food. Recruitment to the Baltic herring stock, the Newfoundland capelin stock and the Arcto-Norwegian cod stock can be roughly forecast 1-2 months after hatching. By its nature, the hypothesis is hard to demonstrate. Yet there is evidence in the study of the gadoid outburst in the North Sea, the spiny lobster off Tasmania and New South Wales and the Arcto-Norwegian cod larvae that there are grounds to support it. The reproductive strategy of some fishes begins the processes which tend to determine recruitment in larval life. The interleaved processes of growth and mortality from larval life onwards are then the prime agents in the gener-

THE MATCH/MISMATCH HYPOTHESIS UPDATED

285

a t i o n of recruitment. They m a y be modified b y other factors and there m a y be effects independent of them, such as advection. B u t the limited conclusion now a n d , as proposed by Rothschild (1986), is t h a t investigations of fish larvae should continue to be a p a r t o f t h e study of population dynamics of

fishes.

VIII. Acknowledgements I am grateful to the editors and an anonymous reviewer for detailed and rewarding comments on an earlier draft of the manuscript.

IX. References Ahlstrom, E. H. (1966). Distribution of sardine and anchovy larvae in the California current region off California and Baja California 1951-64: A summary. Special Scientijic Report of the United States Fish and Wildlife Service 534, 1-71. Alander, H . (1948). Swedish trawling in the Southern Baltic. Annales Biologiques du Conseil International pour I'Exploration de la Mer, 3, 1 11-1 13. Anderson, J. T. (1988). A review of size dependent survival during pre-recruit stages of fishes in relation to recruitment. Journalof North West Atlantic Fisheries. 8, 5566. Anon. (1988). Report of the North east Arctic Working Group. International Council f o r the Exploration of the Sea Assessment Report, 5, 1-142. Astthorsson, 0. A., Hallgrimsson, I. and Jonsson, G. S. (1983). Variations in zooplankton in Icelandic waters in spring during the years 1961-82. Rit Fiskideildar, 7 , 73-1 13. Bailey, K. M . (1981). Larval transport and recruitment of Pacific hake. Marine Ecology Progress Series, 6 , 1-9. Bakun, A. and Parrish, R. R. (1981). Environmental inputs to fishery population models for Eastern Boundary Current regions. International Oceanographic Commission Workshop Report 28, 1-37. Bartsch, J. and Backhaus, J. (1987). Numerical simulation of the advection of vertically migrating herring larvae in the North Sea. International Council,for the Exploration of the Sea. C.M. 1987/L:8 (mimeo). Bartsch, J., Brander, K. M., Heath, M., Munk, P., Richardson, K . and Svendsen, E. (1989). Modelling the advection of herring larvae in the North Sea. Nature, London 340,632-636. Beverton, R. J. H. and Holt, S. J. (1964). Tables of yield assessment for fishery assessment. Food and Agricultural Organization of the United Nations Fisheries Technical Paper 38, 1-49. Botsford, L. W. and Wickham, D. E. (1975). Correlation of upwelling indices and Dungeness crab catch. Fisheries Bulletin of the United States Department of Commerce 73, 901-907. Brander, K. and Thompson, A. B. (1989). Die1 differences in avoidance of three vertical profile sampling gears by herring larvae. Journal of Plankton Research 11, 775-784.

286

D. H. CUSHING

Brothers, E. B., Matthew, C. P. and Lasker, R. (1976). Daily growth increments in otoliths from larval and adult fishes. Fisheries Bulletin of the United States Department of Commerce 14, 1-8. Buckmann, A. (1 942). Die Untersuchungen der biologischen Anstalt uber die Okologie der Heringsbrut in der siidlichen Nordsee. Helgolunder wissenschaftliche Meeresuntersuchungen 3, 1-1 7. Burd, A. C and Parnell, W. G (1973). The relationship between larval abundance and stock in the North Sea herring. Rapports et Procis- Verbaux des RPunions, Conseil International pour I’Exploration de la Mer 164, 30-36. Campana, S. E., Smith, S. J. and Hurley, P. C. F. (1989). An age structured index of larval drift and retention in the waters of south west Nova Scotia. Rapports et Procis- Verbaux des RPunions. Conseil International pour I’Exploration de la Mer 191, 50-62. Colebrook, J. M. (1965). On the analysis of variation in the plankton, the environment and the fisheries. Special Publication of the International Commission for North West Atlantic Fisheries 6, 291-302. Colebrook, J. M. (1979). Continuous plankton records: Seasonal cycles of phytoplankton and copepods in th.’ North Atlantic Ocean and the North Sea. Marine Biology 51, 23-32. Colebrook, J. M. (1982). Continuous plankton records: Seasonal variations in the distribution and abundance of plankton in the North Atlantic Ocean and the North Sea. Journal of Plankton Research 4, 435-462. Colebrook, J. M. and Robinson, G. A. (1965). Continuous plankton records: Seasonal cycles of phytoplankton and copepods in the North east Atlantic and the North Sea. Bulletin of Marine Ecology 6 , 123-129. Colebrook, J. M., Robinson, G. A., Hunt, H. G . , John, A. W. G., Bottrell, H. H., Lindley, J. A., Collins, N. R. and Halliday, N. C. (1984). Continuous plankton records: A possible reversal in the downward trend in the abundance of plankton in the North Sea and the North east Atlantic. Journal du Conseil, Conseil International pour I’Exploration de la Mer 41, 301-306. Cury, P. and Roy, C. (1989). Optimal environmental window and pelagic fish recruitment success in upwelling areas. Canadian Journal of Fisheries and Aquatic Sciences 46, 670-680. Cushing, D. H. (1962). Recruitment to the North Sea herring stocks. Fisheries Investigations, London Ser. 2 23, 43-71. Cushing, D. H. (1967). The grouping of herring populations. Journal of the Marine Biological Association of the United Kingdom 41, 193-208. Cushing, D. H. (1969). The regularity of the spawning season in some fishes. Journal du Conseil. Conseil International pour I’Exploration de la Mer 33, 81-92. Cushing, D. H. (1971). Upwelling and the production of fish. Advances in Marine Biology 9, 255-334. Cushing, D. H. (1972). The production cycle and the numbers of marine fish. In “Conservation and Productivity of Natural Waters” (R. W. Edwards and D. J. Garrod, eds), pp. 213-232. Academic Press, London. Cushing, D. H. (1974). The natural regulation of fish populations. In “Sea Fisheries Research” (F. R. Harden Jones, ed.), pp. 399-412. Elek Science, London. Cushing, D. H. (1975). “Marine Ecology and Fisheries”. Cambridge University Press, Cambridge. Cushing, D. H. (1978). The upper trophic levels in upwelling areas. In “Upwelling Systems” (B. Boje and M. Tomczak, eds), pp. 101-1 10. Springer-Verlag, Berlin.

THE MATCH/MISMATCH HYPOTHESIS UPDATED

287

Cushing, D. H. (1980). The decline of the herring stocks and the gadoid outburst. Journal du Conseil. Conseil International pour I’Exploration de la Mer 39, 74-85. Cushing, D. H. (1982). “Climate and Fisheries”. Academic Press, London. Cushing, D. H. (1984). The gadoid outburst in the North Sea. Journal du Conseil, Conseil International pour I’Exploration de la Mer 41, 159-166. Cushing, D. H. (1986). The migration of larval and juvenile fish from spawning ground to nursery ground. Journal du Conseil, Conseil International pour I’ExpIoration de la Mer 43, 4 3 4 9 . Cushing, D. H. (1988). The northerly wind. In “Toward a Theory of BiologicalPhysical Interactions in the World Ocean” (B. J. Rothschild, ed.), pp. 235-244. Kluwer, Dordrecht. Cushing, D. H. (1989). A difference in structure between ecosystems in strongly stratified waters and in those that are only weakly stratified. Journal of Plankton Research 11, 1-13. Cushing, D. H. (1990). Hydrographic containment of a spawning group of plaice in the Southern Bight of the North Sea. Marine Ecology Progress Series 58,287-297. Cushing, D. H. and Dickson, R. R. (1976). The biological response in the sea to climatic changes. Advances in Marine Biology 14, 1-122. Cushing, D. H . and Harris, J. G. K. (1973). Stock and recruitment and the problem of density dependence. Rapports et Proces- Verbaux des RPunions, Conseil International pour 1’E.xploration de la Mer 164, 142-155. Dickson, R. R. and Blindheim, J. (1984). On the abnormal hydrographic conditions in the European Arctic during the 1970’s. Rapports et Proces-Verbaux des RPunions, Conseil International pour I’Exploration de la Mer 185, 201-213. Dickson, R. R. and Reid, P. C. (1983). Local effects of wind speed and direction on the phytoplankton of the Southern Bight. Journal of Plankton Research 5, 4 4 456. Dickson, R. R., Pope, J. G . and Holden, M. J. (1974). Environmental influences on the survival of North Sea cod. In “The Early Life History of Fish” (J. H. S. Blaxter, ed.), pp. 69-80. Springer-Verlag, Berlin. Dickson, R. R., Malmberg, S.-A., Jones, S. R. and Lee, A. J. (1984). An investigation of the earlier great salinity anomaly of 191Ck1914 in waters west of the British Isles. International Council for the Exploration of the Sea C.M. 1984/Gen:4 (mimeo). Dickson, R. R., Kelly, P. M., Colebrook, J. M., Wooster, W. S. and Cushing, D. H. (1988a). North winds and production in the eastern North Atlantic. Journal of Plankton Research 10, 151-169. Dickson, R. R., Meincke, J., Malmberg, S.-A and Lee, A. J. (1988b). The great salinity anomaly in the northern North Atlantic, 1968-1982. Progress in Oceanography 20, 103-1 5 I . Ellertsen, B., Fossum, P., Sundby, S. and Tilseth, S. (1987). The effect of biological and physical factors on the survival of Arcto-Norwegian cod and the influence on recruitment variability. In “Proceedings of the Third Soviet-Norwegian Symposium” (H. Loeng, ed.), pp. 101-126, Murmansk. Marine Science, Bergen. Ellertsen, B., Fossum, P., Solemdal, P. and Sundby, S. (1989). Relations between temperature and survival of eggs and first feeding larvae of the Arcto-Norwegian cod (Gadus morhua). Rapports et Procis- Verbaux des RPunions, Conseil International pour I’Exploration de la Mer 191, 209-219. Fonselius, S. (1969). Hydrography of the Baltic deep basins 111. Fisheries Board of Sweden Series Hydrografi 23, 1-97.

288

D. H. CUSHING

de la Fontaine, Y., Sinclair, M., El Sabh, M. I., Lassus, C. and Fournier, R., (1984a). Temporal occurrence of ichthyoplankton in relation to hydrographic and biological variables at a fixed station in the St Lawrence estuary. Estuarine, Coastal and Shelf Seas Science 18, 177-190. de La Fontaine, Y., El Sabh, M.I., Sinclair, M., Messieh, S. N. and Lambert, J.-D. (1984b). Structure ockanographique et distribution spatio-temporelle d’oeufs et de larves de poissons dans I’estuaire maritime et la partie ouest du Golfe Saint Laurent. Sciences et Techniques de I’Eau 17, 43-50. Fortier, L. and Gagnk, J. (in press). Larval herring dispersion, growth and survival in the St Lawrence estuary: Match/mismatch or membership/vagrancy. Canadian Journal of Fisheries and Aquatic Sciences. Frank, K. T. (1988). Independent distributions of fish larvae and their prey: Natural paradox or sampling artifact. Canadian Journal of Fisheries and Aquatic Sciences 45, 48-59. Frank, K . T. and Leggett, W. C (1981). Wind regulation of emergence times and early larval survival in capelin (Mallotus villosus). Canadian Journal of Fisheries and Aquatic Sciences 38, 2 15-223. Frank, K. T. and Leggett, W. C . (1982). Environmental regulation of growth rate, efficiency, and swimming performance in larval capelin (Mallotus villosus) and its application to the match/mismatch hypothesis. Canadian Journal of Fisheries and Aquatic Sciences 39, 691-699. Frank, K. T. and Leggett, W. C. (1989). Survival value of an opportunistic life stage transition in capelin (Mallotus villosus). Canadian Journal of Fisheries and Aquatic Sciences 40, 1442-1448. Frank, K. T., Carscadden, J. E. and Leggett, W. C. (1989). Comparative analysis of factors underlying retention of capelin and flatfish larvae on the Southern Grand Banks. Rapports et Procis- Verbaux des Reunions, Conseil International pour I’Exploration de la Mer 191, 450. Franz, H. and Gieskes, W. W. C. (1984). The unbalance of phytoplankton and copepods in the North Sea. Rapports et Procis-Verbaux des Reunions. Conseil International pour I’Exploration de la Mer 183, 218-225. Gieskes, W. W. C . and Kraay, G. W. (1977). Continuous plankton records: Changes in the plankton of the North Sea and its eutrophic Southern Bight from 1974 to 1975. Netherlands Journal of Sea Research 11, 334-364. Glover, R. S., Robinson, G . A. and Colebrook, J. M. (1972). Plankton in the North Atlantic - an example of the problems of analyzing variability in the environment. In “Marine Pollution and Sealife” (M. Ruivo, ed.), pp. 4 3 9 4 5 . Fishing News (Books), London. Graumann, G . B. (1966). Some data on the spawning of cod in the Southern Baltic in 1964. Annales Biologiques du Conseil International pour I’Exploration de la Mer 21, 83-84. Graumann, G. B. (1974). Investigations on the spawning of Baltic cod in 1968-70. Rapports et Procis- Verbaux des Reunions, Conseil International pour I’Exploration de la Mer 166, 7-9. Hainbucher, D., Backhaus, J. 0. and Pohlmann, T. (1986). Atlas of climatological and actual seasonal patterns in the North Sea and adjacent shelf regions. Technical Report of the Institut fur Meereskunde of the University of Hamburg, 86 pp. Unpublished manuscript. Harden Jones, F. R. (1968). “Fish Migration”. Edward Arnold, London.

THE MATCH/MISMATCH HYPOTHESIS UPDATED

289

Harding, D.. Nicholls, J. H . and Tungate, D . S. (1978). The spawning of plaice (Pleuronectes platessa L.) in the Southern North Sea and English Channel. Rapports et Procis- Verbaux des Rtunions, Conseil International pour I'Exploration de la Mer 172, 102-113. Harris, G. P., Davies, P., Nunez, M. and Meyers, G. (1988). Interannual variability in climate and fisheries. Nature 333, 754-757. Heath, M. R. (1989). Transport of larval herring (Clupea harengus L.) by the Scottish coastal current. Rapports et Proces- Verbaux des Reunions, Conseil International pour I'Exploration de la Mer 191, 85-91. Heath, M. R. and MacLachlan, P. (1987). Dispersion and mortality of yolk-sac herring (Clupea harengus L.) larvae from a spawning ground to the west of the Outer Hebrides. Journal of Plankton Research 9, 613-630. Heath, M. R., MacLachlan, P. and Martin, J. H. A. (1987). Inshore circulation and transport of herring larvae off the North coast of Scotland. Marine Ecology Progress Series 40, 1 1-23. Hidaka, K and Ogawa, K. (1958). On the seasonal variations of surface divergences of the ocean currents in terms of wind stress over the oceans. Records of Oceanographical Works in Japan 4, 1 2 4 169. Hjort, J. (1914). Fluctuations in the Great Fisheries of Northern Europe viewed in the light of biological research. Rapports et Procis- Verbaux des RPunions, Conseil International pour I'Exploration de la Mer 20, 1-228. Holligan, P. M. and Harbour, D. S. (1977). The vertical distributions and succession of phytoplankton in the western English Channel. Journal of the Marine Biological Association of the United Kingdom 57, 1075-1093. Holligan, P. M., Harris, R. P., Newell, R. C., Harbour, D. S., Head. R. N., Linley. E. A. S., Lucas, A. I., Tranter, P. R. G. and Weekley, C. M. (1984). Vertical distribution and partitioning of organic carbon in mixed, frontal and stratified waters of the English Channel. Marine Ecology Progress Series 14, 1 1 1-127. Hollowed, A. B., Bailey, K. H. and Wooster, W. S. (1987). Patterns of recruitment of marine fishes in the North east Pacific Ocean. Biological Oceanography 5,99-132. Hunter, J. R.(1972). Behaviour and survival of northern anchovy, Engraulis mordax larvae. Reports of the California Cooperative Oceanic Fisheries Investigations 19, 138-146. Hunter, J. R. and Leong, R. (1980). The spawning energetics of the female northern anchovy, Engraulis mordax. Fisheries Bulletin of the United States Department of Commerce 79, 21 5-230. Husby, D. M. and Nelson, C. S. (1982). Turbulence and vertical stability in the California Current. California Cooperative Oceanic Fisheries Investigations 23, 113-129. Iles, T. D. and Sinclair, M. (1982). Atlantic herring: Stock discreteness and abundance. Science 215, 627433. Jones, R. (1973). Density dependent regulations of the numbers of cod and haddock. Rapports et Proces- Verbaux des Reunions. Conseil International pour I'E.xploration de la Mer 164, 166173. Jones, R. and Henderson, E. W. (1988). Simulation studies of fish larval survival. In "Towards a Theory on Biological-Physical Interactions in the World Ocean" (B. J. Rothschild, ed.). pp. 343-372. Kluwer, Dordrecht. Kiorboe, T. and Johansen, K. (1986). Studies of a larval herring (Clupea harengus L.) patch in the Buchan area. IV. Zooplankton distribution and productivity in relation to hydrographic features. Dana 6 , 37-5 1.

290

D. H. CUSHING

Kierrboe, T., Munk, P., Richardson, K., Christensen, V. and Paulsen, H. (1988). Plankton dynamics and larval herring growth, drift and survival in a frontal area. Marine Ecology Progress Series 44, 205-219. Koslow, J. A. (1984). Recruitment patterns in north west Atlantic fish stocks. Canadian Journal of Fisheries and Aquatic Sciences 41, 1722-1729. Koslow, J. A., Thompson, K. R. and Silvert, W. (1987). Recruitment to North west Atlantic cod (Gadus morhua) on George’s Bank and haddock (Aeglejnus melanograrnmus) stocks: Influence of stock size and climate. Canadian Journal of Fisheries and Aquatic Science 44, 2 6 3 9 . Kremer, J. N. and Nixon, S. W. (1978) “A Coastal Marine Ecosystem”. SpringerVerlag, Berlin. Lasker, R. (1975). Field criteria for survival of anchovy larvae: The relation between inshore chlorophyll maximum layers and successful first feeding. Fisheries Bulletin of the United States Department of Commerce 13, 453462. Lasker, R. (1978). The relation between oceanographic conditions and larval anchovy food in the California Current: Identification of factors contributing to recruitment failure. Rapports et Procts- Verbaux des Reunions, Conseil International pour I’Exploration de la Mer 113, 2 12-230. Lasker, R. (1981). Factors contributing to variable recruitment of the northern anchovy (Engraulis mordax) in the Californian Current: Contrasting years 1975 through 1978. Rapports et Procts- Verbaux des RPunions, Conseil International pour I*Exploration de la Mer 178, 375-388. Lee, A. J. and Folkard, A. R. (1969). Factors affecting turbidity in the Southern North Sea. Journal du Conseil, Conseil International pour I’Exploration de la Mer 32, 291-302. Le Fivre (1986). Aspects of the biology of frontal systems. Advances in Marine Biology 25, 163-299. Leggett, W. C., Frank, K. T. and Carscadden, J. E. (1984). Meteorological and hydrographic regulation of year class strength in capelin (Mallorus villosus). Canadian Journal of Fisheries and Aquatic Science 41, 1193-1201. MacCall, A. D. (1979). Population estimates for the waning years of the Pacific sardine fishery. Californian Cooperative Oceanic Fishery Investigation Reports 20, 72-82. Marak, R. R. (1974). Food habits of larval cod, haddock and coalfish in the Gulf of Maine and George’s Bank area. Journal du Conseil, Conseil International pour I’Exploration de la Mer 25, 147-157. Marr, J. C. (1956). The critical period in early life history of fishes. Journal du Conseil. Conseil International pour [’Exploration de la Mer 21, 160-170. Marr, J. C. (1960). The causes of major variations in the catch of Pacific sardine, Sardinops caerulea (Girard). In “Proceedings of the World Scientific Meeting on the Biology of Sardines and related species, (Food and Agricultural Organization), Vol. 3 pp. 667-679. Matsumoto, W. (1966). Distribution and abundance of tuna larvae in the Pacific Ocean. In “Proceedings of the Governor’s Conference on Central Pacific Research, 1966” (T. A. Manier, ed.), pp. 221-230. May, R. C. (1974). Larval mortality in marine fishes. In “The Early Life History of Fish” (J. H. S. Blaxter, ed.) pp. 3-19, Springer-Verlag, Berlin. Mendelsohn, R. and Mendo, J. (1987). Exploratory analysis of anchoveta recruitment off Peru and related environmental series. In “The Peruvian Anchoveta and its Upwellng Ecosystem: Three Decades of Change” (D. Pauly and I. Tsukuayama, eds), pp. 294306 ICLARM, Manila.

THE MATCHIMISMATCH HYPOTHESIS UPDATED

29 1

Mendiola, B. R. de (1971). Some observations on the feeding of the Peruvian anchoveta, Engraulis ringens J., in two regions off the Peruvian coast. In “Fertility of the Sea” (J. D. Costlow, ed.), Vol. 2, pp. 4 1 7 4 0 . Gordon and Breach, London. Methot, R. D. (1983). Seasonal variation in survival of larval northern anchovy, Engraulis mordax, estimated from the age distributions of the juveniles. Fisheries Bulletin of the United States Department of Commerce 81, 741-750. Meyer, P. F. and Kalle, K. (1950). Die biologisch Unstimmung der Ostsee in den letzten Jahrzehnten-eine Folge hydrographischer Wasserum-schichtungen. Archiv fur Fischereiwissenchaft 11, 1-9. Munk, P., Christensen, V. and Paulsen, H. (1986). Studies of a larval herring (Clupea harengus L.) patch in the Buchan area. 11. Growth, mortality and drift of larvae. Dana 6 , 1 1-24. Nishikawa, Y., Homma, M., Ueyanagi, S. and Kikawa, S . (1985). Average distribution of oceanic species of scombroid fishes, 19561981. Far Seas Fisheries Research Laboratory, Ser. 12. 1-99. Nival, P., Carlotti, F. and Sciandra, A. (1988). Modelling the recruitment of marine species. In “Towards a Theory of Biological-Physical Interactions in the World Ocean” (B. J. Rothschild, ed.), pp. 321-342, Kluwer, Dordrecht. O’Boyle, R. N., Sinclair, M., Conover, R. J. and Kohler, A. C. (1984). Temporal and spatial distribution of ichthyoplankton communities of the Scotian Shelf in relation to biological, hydrological and physiographic features. Rapports et ProcPs- Verbaux des RPunions, Conseil International pour I’Exploration de la Mer 183, 27-40. Parmanne, R. and Sjrablom, V. (1984). The abundance of spring spawning herring larvae around Finland in 1982 and 1983 and the correlation between zooplankton abundance and year class strength. International Council for the Exploration of the Sea C.M. 1987/J: 18 (mimeo). Parmanne, R. and Sjrablom, V. (1987). Possibility of using larval and zooplankton data in assessing the herring year class strength off the coast of Finland in 197486. International Counci1,for the Exploration of the Sea C.M. 1987/J: 19 (mimeo). Parrish, R. R. and McCall, A. D. (1978). Climatic variations and exploitation in the Pacific mackerel fishery. Califbrnian Fish and Game Bulletin 167, 1-109. Pedersen, T. (1984). Variation of peak spawning of Arcto-Norwegian cod (Gadus morhua L.) during the time period 1929-82 based on indices estimated from fishery statistics. In “The Propagation of Cod, Gadus morhua L.” (E. Dahl, D. S. Danielsson, E. Moksness and P. Solemdal, eds), Part I . Flradevigen Biological Station. Peterman, R. and Bradford, M. J. (1987). Wind speed and mortality rate of a marine fish, the northern anchovy (Engraulis mordax). Science 235,4786, 354-356. Peterman, R., Bradford, M. J., Lo, N. H. C. and Methot, R. (1988). Contribution of early life stages to interannual recruitment of northern anchovy (Engraulis mordax). Canadian Journal of Fisheries and Aquatic Sciences 45, 8-176. Peterson, W. T. (1973). Upwelling indices and annual catches of Dungeness crab, Cancer magister, along the west coast of the United States. Fisheries Bulletin of the United States Department of Commerce 71, 902-910. Pingree, R. D., Pugh, P. R., Holligan, P. M. and Forster, G . R. (1975). Summer phytoplankton blooms and red tides along tidal fronts in the approaches to the English Channel. Nature 258, 612-677. Pingree, R. D., Maddock, L. and Butler, E. I. (1977). The influence of biological activity and physical stability in determining the chemical distributions of inorganic phosphate, silicate and nitrate. Journal of the Marine Biological Association of the United Kingdom 57, 1065-1073.

292

D. H. CUSHING

Raymont, J. E. G. (1980). “Plankton and Productivity in the Oceans”, 2nd edn, Vol. I . Pergamon Press, Oxford. Reid, P. C. (1978). Continuous plankton records: Long scale changes in the abundance of phytoplankton in the North Sea from 1958 to 1973. Rapports et ProcPs- Verbaux des Reunions, Conseil International pour I’Exploration de la Mer 172, 384-389. Ricker, W. E. (1954). Stock and recruitment. Journal ofthe Fisheries Research Board of Canada 11, 559-623. Ricker, W.E. and Foerster, R. E. (1948). Computation of fish production. Bulletin of the Bingham Oceanographic Collection. 11, 173-21 1. Robinson, G. A. (1970). Continuous plankton records: Variations in the seasonal cycles of phytoplankton in the North Atlantic. Bulletins of Marine Ecology 6, 333-345. Robinson, G. A,, Colebrook, J. M. and Cooper, G . A. (1975). The Continuous Plankton Recorder Survey: Plankton in the I.C.N.A.F. area in 1972. Special Publication of the International Commission of North West Atlantic Fisheries 10, 95-103. Rothschild, B. J. (1986). “Dynamics of Marine Fish Populations”. Harvard University Press, Cambridge, Mass. Saville, A. and Bailey, R. S. (1979). The assessment and management of the herring stocks in the North Sea and to the west of Scotland. Rapports et Procis-Verbaux des RPunions, Conseil International pour I’Exploration de la Mer 177, 112-142. Shepherd, J. G. and Cushing, D. H. (1980). A mechanism for density dependent survival of larval fish as the basis for a stock-recruitment relationship. Journal du Conseil, Conseil International pour I’Exploration de la Mer 39, 160-1 67. Shepherd, J. G., Pope, J. G. and Cousens, R. G. (1984). Variations in fish stocks and hypotheses concerning their links with climate. Rapports et Proc2s- Verbaux des Reunions, Conseil International pour I’Exploration de la Mer 185, 258-267. Simpson, J. H. and Hunter, J. R. (1974). Fronts in the Irish Sea. Nature, 250, 4 0 4 406. Sinclair, M. (1988). “Marine Populations”. Washington Sea Grant Program, Seattle. Sinclair, M. and Tremblay, M. J. (1984). Timing of spawning of Atlantic herring (Clupea harengus harengus) and the match/mismatch theory. Canadian Journal of Fisheries and Aquatic Sciences 41, 10541065. Sinclair, M., Maguire, J. J., Koeller, P. and Scott, J. S. (1984). Trophic dynamic models in light of current resource inventory data and stock assessment results. Rapports et Procis- Verbaux des RPunions, Conseil International pour [‘Exploration de la Mer 184, 269-284. Sinclair, M., Tremblay, M. J. and Bernal, P. (1985). El Niiio events and variability in Pacific mackerel (Scomber japonicus) survival index: Support for Hjort’s second hypothesis. Canadian Journal for Fisheries and Aquatic Sciences, 42, 602-608. Stephenson, R. L. and Power, M. J. (1989). Observations on herring larvae retained in the Bay of Fundy: Variability in vertical movement and position of the patch edge. Rapports et Procis- Verbaux des Reunions, Conseil International pour I’Exploration de la Mer 191, 177-183. Sundby, S., Bjorke, H., Soldal, A. V. and Olsen, S. (1989). Mortality rates during the early life stages and year class strength of the Arcto-Norwegian cod (Gadus morhua L.). Rapports et ProcPs- Verbaux des Reunions, Conseil International pour I’Exploration de la Mer 191, 351-358.

THE MATCH/MISMATCH HYPOTHESIS UPDATED

293

Suthers, I. M. and Frank, K. T. (1989). Dispersal, growth and condition ofjuvenile pelagic cod (Gadus morhua) in southwestern Nova Scotia. Rapports et ProcesVerbaux des RPunions. Conseil International pour I’Exploration de la Mer 191, 46-67. Sverdrup, H. U. (1953). On conditions for vernal blooming in phytoplankton. Journal du Conseil, Conseil International pour I’Exploration de la Mer 18,287-295. Thoradottir, T. (1977). Primary production in north Icelandic waters in relation to recent climatic changes. In “Polar Oceans” ( M . J. Dunbar, ed.) pp. 655-665. Arctic Institute of North America. van der Veer, H. W. (1986). Immigration, settlement and density dependent mortality of a larval and post larval 0-group plaice (Pleuronectes platessa) population in the Wadden Sea. Marine Ecology Progress Series 29, 223-236. Walsh, J. J. (1978). The biological consequences of interaction of the climatic El Nirio, and event scales of variability in the eastern tropical Pacific. Rapports et Procis- Verbaux des Reunions, Conseil International pour I’Exploration de la Mer 173, 182-192. Williamson, M. H. (1961). An ecological survey of a Scottish herring fishery. Part IV: Changes in the plankton during the period 1949-1959. Bulletins of Marine Ecology 5, 207-229. Wootton, R. J . (1977) Effect of food limitation during the breeding season on the size, body components and egg production of female sticklebacks (Gasterosteus aculeatus). Journal of Animal Ecology 46, 823-834. Yamanaka, I . (1978). Oceanography in tuna research Rapports et ProcPs- Verbaux des Reunions, Conseil International pour I’ExpIoration de la Mer 173, 203-21 1. Yin, M . C. and Blaxter, J. H. S. (1987). Temperature, salinity tolerance and buoyancy during early development and starvation of Clyde and North Sea herring cod and flounder larvae. Journal of Experimental Marine Biology and Ecology. 107, 279-290.

This Page Intentionally Left Blank

Taxonomic Index Figures in italic; tables in bold

C

A Abramis brama 83, 84, 120, 126 Acanthocottus bubalis 138 Acanthoglobiusflavimanus 102 Acartia 135 Acesta lopezi 179 Acesta neosuecica 179 Acipenser queldenstaedti 120 Aglaophamus malmgreni 179 Agonus cataphractus 121 Alburnus alburnus 120, 126 ANia quadrilobata 179 Alosa pseudoharengus 120 Ammodytes sp. 27, 120, 133, 149, 210 Amtnodytes americanus 133 Ammodytes marinus 121, 133 Ammodytes personatus 121, 149 Ampharete acutifrons 179 Ampharete jnmarchica 179 Ampharetidae 179 Amphitritinae 179 Amphiura jliformis 181 Ancistrossyllis groenlandica 179 A nguilla anguilla 141 Anguilla rostrata 121, 130, 142 Anguilla sp. I49 Anisakis spp. 38 Anobothrus gracilis 179 Aricidea 179 Arius ,felis 139 Asychis 179 Aurelia sp. 145 Autolytus 179

Calanoida 131 Calanus 212, 274-6, 279, 281 Calanus Jinnmarchicus 280, 283 Caligus sp. 131 Cancer magister 260 Capitella capitata 190 Carassius auratus 128, 129 Ciiridean sp. 131 Centronotus gunnellus 121 Ceratomyxa drepanopsettae 38 Chaenogobius isaza 120, 129 Chone ecaudata 179 Chrysaora melanaster 145 Ciliata mustela 138 Cirrophorus branchiatus 179 Citharichthys arctifions 146 Clarias macrocephalus 92 Clupea harengus 33, 75, 78, 92, 120, 121, 122, 251 Clupea harengus pallasi 73, 92 Clymenura columbiana 179 Contracaecum aduncum 38 Coregonus albula 72, 79, 83, 121, 146 Coregonus artedii 147 Coryphoblennius galerita 131, 139 Cossura longocirrata 179 Cossura modica 179 Cottus extensus 136 Cyanea sp. 145 Cyclopterus lumpus 76 Cyprinus carpio 72, 74, 150

Benthosema glaciale 120, 130 Blennius pholis 137. 142 Brada villosa 179 Brevoortia patronus 121, 131 Brevoortia tyrannus 121 Brosme brosme 33-4

Daphnia magna 135 Decamastus gracilis 179 Derogenes varicus 38 Dicentrarus labrax 85 Diogenichthys laternatus 120 Dorosoma cepedianum 120, 121

295

296

TAXONOMIC INDEX

E Echinorhynchus gadi 38 Engraulis japonicus 137 Engraulis mordax 137, 144 Engraulis ringens Jenyn 262 Entobdella hippoglossi 38, 40, 41, 5 4 5 Etheostoma 80 Euchone incolor 179 Euclymene 179 Euclymene zonalis 179 Euclymeninae 179 Eupagurus bernhardus 27 Eusyllis assimilis 179 Exogone verugera 179

F Fundulus heteroclitus 128

G Gadus esmarkii 27 Gadus macrocephalus 121 Gadus merlangus 145 Gadus minutus 27 Gadus morhua 73, 75, 77, 86, 92, 120, 121, 126, 131,132, 141, 252, 270 Gadus poutassou 33 Galathowenia oculata 179 Gasterosteus aculateus L. 262 Gattyana cirrosa 179 Gilbertidia sigalutes 153 Glycera capitata 179 Glycinde armigera 179 Clyphanostomum pallescens 179 Gobius microps 121 Gobius scorpioides 121 Goniada annulata 179 Grillotia erinaceus 38, 40, 41 Gymnellis 33 Gymnodinium splendens 261 Gyptis 179

H Halichoeres poecilopterus 129 Harmothoe 179 Harmothoe lunalata 179

Hatschekia hippoglossi 38 Heteropneustes 103 Hippoglossus americanus 3 Hippoglossus hippoglossus 2, 3, 5, 76, 78, 86, 92 Hippoglossus hippoglossus stenolepis 2 Hippoglossus platessoides 257 Hippoglossus stenolepis 2 Hippoglossus vulgaris 3 Hyas 27

I Icichthys lockingtoni 145

J Jasmineira pacijica 179

K Kefersteinia cirrata 179

L Lagodon rhomboides 121 Lampenus lampetriformis 121 Leiostomus xanthurus 121, 140 Lepeophtheirus hippoglossi 38, 41 Lepomis sp. 121 Lepomis gibbosus 129 Lepomis macrochirus 127 Lesueurigobius friesii 127 Leuroglossus stilibius 120, 121 Limacina sp. 131 Limanda ferruginea 116, 257 Limanda limanda 270 Limanda Yokohama 73,75, 92, 95, 102 Limnothrissa miodon 130 Liparis sp. 121 Lota Iota 129 Lucioperca lucioperca 83 Lycodes 33

M Mallotus sp. 120 Mallotus villosus 33, 121, 130, 257

297

TAXONOMIC INDEX

Melanogramnrus aegle$nus 120, 126, 131, 132, 148 Menidia beryllina 130 Merlangius merlangus 86, 210 Merluccius bilinearis 120 Merluccius gayi 120 Merluccius merluccius 33 Merluccius productus 260 Merluccius novaezelandiae 130 Micropogonias undulatus 121, 140 Morone chrysnps 120 Mugil cephalus 103 Munida 27 Mylio berda 103

Pndurkeopsis brevipaipa 179 Pomatomus saltatrh 146 Pomatoschistus minutus 138, 139 Porichthys notatus 150 Portunus 27 Psenes sp. 145 Pseudocalanus 256, 258, 269 Pseudnpleuronectes americanus 91

Reinhardtius hippoglnssoides 43 Roccus saxatalis 92 Rutilus rutilus 83, 97, 120, 121, 126

N Neocalanus plumchrus 136 Nentnysis americana 131, 143 Nephrops norvegicus 34 Notnscnpelus elongatus 1 I6

0 Oiknpleura 25 I , 270 Oligocottus maculosus 138, 142, 148 Oncorhynchus nerka 144, 152, 252 Oryzias latipes 103 Osmerus mordax 120, 140

P

S Sagitta elegans 131 Salmn gairdneri 74, 79, 92 Saimo salar 74, 79, 130 Salmn trutta 92 Salvelinus alpinus 92 Sardinia pilchardus 260 Sardinops caerulea 259 Sardinops melanosticta 131 Scardinius erythrophthalmus 120, 121, . I26 Scnlex pleuronectis 38 Scomber japonicus 260 Scomber scombrus 149 Scophthalmus maximus 73, 74, 92 Scyliorhinus caniculus 141 Sebustes sp. 120, 123, 132 Sebustes diploproa 147 Sebastesjavidus 132 Sebustes marinus 33, 34 Snlea solea 74, 76, 141 Snlea vulgaris 121 Somniosus microcephalus 4 1 Squalus acanthias I44 Stnmias boa ferox 120

Pagellus bogaraveo 73, 75 Pagellus centrodontus 33 Pagrus major 101. 121 Pagurus larvae 131 Pagurus sp. 131 Pandalus 27 Paralichthys sp. 121, 140, 141 Parnphyrus vetulus I42 Pelagia nnctiluca 145 Perca.flavescens 120, 141 Phnlas 27 T Platichthys.flesus 87, 139 Plecnglossus altivelis 74, 92, 103 Pleuronectes jesus 121, 210 Themisto sp. 131 Pleuronectes hippoglossus 3 Theragra chalcogramma 129 Pleuronectes platessa 121, 133, 138, 141, Tilapia mossambica 152 152, 25 1, 210 Tinca tinca 84

298

TAXONOMIC INDEX

Tisbe sp. 131 Trachurus mediterraneus 120 Trachurus mediterraneus ponticus 121 Trachurus sp. 145 Trachurus symmetricus 75, 145 Trinectes maculatus 138 Trisopterus esmarkii 121 Trisopterus luscus 210

U urophycis chuss 147

X Xenodermichthys copei 120 Xiphias gladius 124, 125, 135

Subject Index Figures in ifulic; tables in bold

A Abiotic habitat factors in benthic studies 226-7 Abundance of Atlantic halibut eggs 10 and environmental gradients 189 in faunal data matrices 178, 180-1 and spatial distribution 193 Abundance distribution and biomass 210 Age and maturity of Atlantic halibut 7-8 and ontogenetic variation in fish 150 and size of Atlantic halibut 23, 24, 25, 29-31 Albacore and match/mismatch hypothesis 263 Algae, and plankton 251 Amino acids in fish eggs 83 Analytical methods in benthic studies 169-247 Anaphases, delayed in fish eggs 85 Anchovy, spawning 262-3 ANOVA see Variance, Analysis of Anthropogenic impact on community in benthic studies 190-2 Aquaculture of Atlantic halibut 54-63 Aquatic ecosystems, diversity in 199 Artemiu nauplii 53 Ascorbic acid in fish eggs 81, 82 Assemblage models 194 Atlantic croaker, vertical migration of 121, 141 Atlantic halibut aquaculture of 54-63 biology of 1-70 diseases of 4 1-2 distribution of 5, 6, 7 egg development of 12 exploitation of 42-53 fisheries of 43 identity of 3-5

299

immature phase of 22, 23, 24-7 larvae of 14, 15-16, 17-18 mature phase of 27-38 parasites 3 8 4 1 pelagic phase of 11-22, 20-1 and pollution 42 rearing of 53 reproduction 7-1 1 and salinity 53, 55 sexual maturity of 8 species status 2 4 Averaged General Quality Indicator (QIC) 79, 84 AYU egg quality 74, 98-9 egg viability 92

B Bacterial growth in fish eggs 76 Baltic Sea recruitment of cod in 2 7 3 4 and vertical migration 138-9 Barents Sea, halibut fisheries in 43, 45-6 Bartlett’s test in benthic studies 230 Benthic assemblage, and biomass distribution 209 Benthic marine infaunal studies, analytical methods in 169-247 Bigeye and match/mismatch hypothesis 263 Bioenergetic advantage, as zeitgeber 136-7 Biology of Atlantic halibut 1-70 Biomass and abundance 181 distribution in benthic studies 209-12 Bootstrap methods in classification in benthic studies 219-20 Bray-Curtis similarity coefficients in benthic studies 182, 215-16, 218

300

SUBJECT INDEX

Bream egg quality 74, 77, 84, 98-9, 101 vertical migration in 120-1 Brightness, as zeitgeber 1 2 3 4 Broodstock of Atlantic halibut 54-5 management of fish 96-103 and hormonal treatment 102-3 and induced spawning 102-3 and nutrition 97-102, 98-9 Buchan spawning area 252 Buoyancy in fish eggs 76, 77 and feed quality 98-9

C Calunus in match/mismatch hypothesis 272, 274, 275, 276, 280, 281 Canada, halibut fisheries in 43, 45-6, 48-9 Canberra-Metric measures in benthic studies 215-16 Canonical correlation in benthic studies 230-1 Canonical variate analysis see multiple discriminant analysis Capelin and Atlantic halibut 33 in match/mismatch hypothesis 257, 272 vertical migration of 121, 130 Caridean sp. 131 Carotenoids in fish eggs 80, 81 Carp, egg quality 74 Catch of Atlantic halibut 42-52 by area 45, 46, 47 by nation 47, 48,49 development of 42-3 44, 45 management of 51-2 and over-exploitation 49-50 and recruitment and mortality 50-1 Chemical content and egg quality in fish 79-85 Chi-squared goodness of fit models 194 Chorion hardening in fish eggs 12, 13, 76 Chromosomal aberrations and egg quality in fish 85-6 Chronobiological theory and vertical migration 148-50, 152

Circadian rhythms 116-17, 119, 150 Circatidal activity 137-9 short persistence of 138 Classification in benthic marine infaunal studies 217-21 Climate change and recruitment in fish populations 256 and fish stock recruitment 256 Cluster analysis and classification in benthic studies 217, 219, 220, 229-30 Coastal water and plankton production 254 Cod die1 vertical migration of 120-1, 126, 141 egg quality 73-4, 77 egg viability 90, 92 and gadoid outburst 274-6 larvae of and plankton production 279-8 1 and match/mismatch hypothesis 252, 270, 282-3 prey composition of 130, 131, 132 recruitment of in Baltic Sea 2 7 2 4 and salinity anomaly 277 Commensal species and vertical migration 145 Community anthropogenic impact on 190-2 and biomass distribution in benthic studies 209-12 concept of benthic studies 171, 1 8 6 92 vs continuum in benthic marine infaunal studies 18690 descriptive univariate indices 192-2 12 structure in benthic marine infaunal studies 197-209 Competition, in communities 198-9 Continuous Plankton Recorder Network 251-2, 269-70 Continuum, vs community in benthic marine infaunal studies 1 8 6 90 Cooperation in communities 198-9 Copepodites 251 Copepod nauplii 25 1, 283 Copepods

30 1

SUBJECT INDEX

in benthic studies 191 in match/mismatch hypothesis 25 1, 262-3, 269, 280 Correspondence Analysis in benthic studies 222 Cortical reactions in fish eggs 73 Crustacea as food 26-7 as parasites 38 Czekanowski coefficient in benthic studies 215

Divergence see upwelling Diversity in communities 197-201 indices of 203-8 Dogger spawning area 251 Dominance in communities 197, 201 Dominants and biomass distribution 210 Downs spawning area 251, 271 Downwelling, and fish larvae recruitment 260

D Data analysis of in benthic marine infaunal studies 178-232, 179 collection of in benthic marine infaunal studies 172-8 multivariate analysis 214-31 nature of in benthic studies 178-86 reductions 181-2 transformations 182-6 and statistical inference 212-14 time series analysis 23 1-2 Dendograms of Atlantic halibut 4 in benthic studies 217-18, 228 Descriptive univariate indices in benthic studies 192-212 Detrended Correspondence Analysis in benthic studies 226 Die1 vertical migration in fish 115-68 and endogenous rhythms 122-50 as facultative process 154-5 local environmental factors for 145-8 and ontogenetic variation 148-50 types of 119, 120-1 zeitgeber for 122-45 Digestive rates and vertical migration 136 Diseases of Atlantic halibut 41-2 Dispersion measures and spatial distribution 195-7 Dissimilarity coefficient matrix in classification 217 indices in benthic studies 216 Distribution of Atlantic halibut areas 5, 6 differentiation 6 7

E Ecosystems in open oceans 259 water stratification as 254, 259 Ecotone 189 Ectoparasites see Parasites Eels, vertical migration of 121, 130, 141-2. 149 Eggs of Atlantic halibut 11, 12 incubation and hatching 55-7 in rearing experiments 53 quality in fish 71-113 broodstock management 96-103, 98-9

characteristics of 72-86 chemical content 79-85 chromosomal aberrations 85-6 criteria of 74-5 factors of importance 91-103 and fertilization 72, 73, 74-5 morphology 74-5, 7 6 8 overripening and storage 74-5, 9 1-6 physical and physiological properties 76 and size 74-5, 78-9 in wild fish 86-9 1, 88-9 viability of 92 Ekman transport and fish larvae recruitment 260 Eltonian Pyramids 210 Embryo of Atlantic halibut 11-13 development of in fish eggs 7 6 7 Endogenous rhythmicity 11617 categories of in fish migration 119-22

302

SUBJECT INDEX

evidence for 122-50 importance of 118 Energy economics of Atlantic halibut 37-8 Environment gradients and abundance 189 habitat factors in benthic studies 2268 variations of in vertical migration 145-8 Essential fatty acids (EFA) in fish eggs 98-9, 100 Estuaries as ecotone 189-90 and tidal influence on vertical migration 139-41 Euclidian methods in benthic studies 215, 217 Evenness in communities 197, 204 Exploitation of Atlantic halibut 42-53

F Factor Analysis in benthic studies 222, 224 Fager’s index of Affinity in benthic studies 215 Fat content in fish eggs 83 Faunal data matrices 178-8 1 Feed/feeding of Atlantic halibut 26, 27, 33, 34 level and egg quality 97, 100 quality and buoyancy in fish eggs 98-9 windows and vertical migration 133, 134 Fertilization and egg quality 72-5 and overripening of fish eggs 9 3 4 Fish die1 vertical migrations of 115-68 egg quality in 71-1 13 plankton production and year class strength in 249-93 Fisheries of Atlantic halibut 43 Flounder egg quality 74, 87, 88-9 egg viability 92-3, 95 vertical migration of 116, 121. 13940, 146

Food for Atlantic halibut 33 availability of in vertical migration I47 as zeitgeber 129-36 see also nutrition Fowlkes-Mallows test in benthic studies 228 Frequency distribution models 193-5

G Gadoid outburst and recruitment 274-6 Gamma distribution 208 Genetic identity of Atlantic halibut 4 Gonads in Atlantic halibut development of 9 and energy economics 37-8 Gower’s method in benthic studies 222 Greenland, halibut fisheries in 43, 4 6 7 ‘Greenness’ as measure of plankton production 251, 265,266-9 and spawning distribution of species 270 Grillotia and Atlantic halibut 40-1 Growth of Atlantic halibut 25-6, 31-2, 33 experiments on 62, 63 of larvae 28 I , 284

H Habitat factors, comparisons of in benthic studies 226-31 Haddock prey composition of 130, 131, 132 vertical migration of 120, 126, 147-8 Halibut eggs development of 78 viability 92 stripping of eggs 91 see also Atlantic halibut Hatching of Atlantic halibut 55-7 rate of and egg quality 98-9 Herring egg quality 74

303

SUBJECT INDEX

egg viability 92 light as zeitgeber for 122-3 and match/mismatch hypothesis 265-9, 271-2, 276, 277 in retention areas 257-8 spawning of 25 1-2, 253, 256, 27 I vertical migration patterns 120-1. 122, 151-2 Hippoglossus, biology of 1-70 Hormonal treatment and broodstock management 102-3 Horn’s Information Theory in benthic studies 2 15 Horse mackerel, vertical migration of 120, 127, 145 Horseshoe effect in ordination in benthic studies 225 Hydrographic conditions and spawning of Atlantic halibut 10-1 I Hyperiidae I3 1 Hypothesis testing in benthic marine infaunal studies 2 12

I Iceland, halibut fisheries in 43, 45-6, 48 Immature phase of Atlantic halibut 22-7 age/size composition 23-5 distribution 6 7 feeding 2 6 7 growth 25-6 occurrence and migration 22-3 sex ratios 23-5 Incubation of Atlantic halibut 55, 56-7, 59 Inorganic components of fish eggs 82-3 Irradiance, attenuation of and plankton production 254 Isolumes 125-7

J Jaccard’s index in benthic studies 21516

Kruskal-Wallis test. in benthic studies 228

L Larvae abundance and recruitment 271-3 of Atlantic halibut growth of 15-16 occurrence 11-12, 13 and post-larval stage 18-20, 20-1 rearing 57-60 in rearing experiments 53 start-feeding 6&2 yolk-sac 14-18, 5 7 4 0 of cod and plankton production 27981, 282-3 growth of and food supply 258 and members/vagrant hypothesis 2568 size of and egg size 78-9 survival of 264 and egg quality 98-9, 101 time of release of 251-2, 255 and upwelling areas 259-61 Lasker events and plankton production 26 1 Latitude and circadian rhythms in fish 151 Light as zeitgeber 122-9 Linkage rules and classification in benthic studies 217 Lipids composition of in Atlantic halibut 35, 36-7 content in fish eggs 83, 84, 85 and overripening 96 Liver, in Atlantic halibut 39 Locomotor activity and circatidal activity 137, 139 and vertical migration 135 Log normal distribution 205-6, 207-8 Log normal frequency, and biomass distribution 209 Longevity, of Atlantic halibut 3 1-3

M

K Kolmogorov-Smirnov test 194

Mackerel egg quality 74

304

SUBJECT INDEX

and match/mismatch hypothesis 260 survival rates 2 6 4 5 vertical migration of 149 Macrofauna and biomass distribution in benthic studies 209, 211 Management of Atlantic halibut 51-2 Manhattan metric in benthic studies 215 Match/mismatch hypothesis and fish populations 250-6, 255, 284 in low latitude waters 259-63 testing of 2 6 4 5 Mature phase of Atlantic halibut 27-38 age/size composition 29-3 1 body composition 35-7 distribution 6-7 energy economics 37-8 feeding 33-4 growth 31-3 longevity 31-3 occurrence and migration 27-9 sex ratios 29-31 Maturity, sexual, in Atlantic halibut 7-8,51 Meiofauna and biomass distribution 209 in polluted sites 191-2 Member/vagrant hypothesis on fish spawning 256-9 Metabolic rates, and vertical migration 137 Migration of Atlantic halibut immature phase 22, 23 mature phase 27-9, 28 die1 vertical in fish 115-68 environmental factors 145-8 model of 1 2 3 4 Minerals, and feed quality in fish eggs 98-9, 101 Models of assemblage 194 frequency distribution 193-5 Morisita’s Index in benthic studies 215 Morphology and egg quality in fish 74-5, 76-8 malformations 88-9 Mortality of Atlantic halibut 50-1 of larvae 281, 284

Multidimensional scaling methods in benthic studies 221-6 and non-metric scaling 221, 225 and ordinal scaling 221 and Principal Component Analysis 2224 and Principal Co-ordinates Analysis 222, 224-5 Multiple discriminant analysis in benthic studies 228, 229, 230 Multivariate community analysis 172 Multivariate data analysis in benthic marine infaunal studies 21431 classification 2 17-2 1 habitat factors, comparisons 226-3 1 ordination 221-6 similarity indices 215-16

N Negative binomial distribution model 1934 Nematodes in benthic studies 191 Newman-Keuls test in benthic studies 228 Non-linearity in ordination in benthic studies 225 Normalized Expected Species Shared (NESS) in benthic studies 215-16 North Sea halibut fisheries in 43, 45-6 time of spawning in 251-2, 253, 254 Norway, halibut fisheries in 43, 45-6, 47-8 Nutrition, and broodstock management Of fish 97-102 Nyquist criterion for time-series analysis 177

0 Ontogenetic variation and vertical migration 148-50 Ordinal scaling in benthic studies 221 Ordination in benthic studies 221-6, 223 non-linearity problems in 225

305

SUBJECT INDEX

Organic components of fish eggs 83-5 Orloci’s index in benthic studies 215 Osmosis in fish eggs 76 Over-exploitation of Atlantic halibut 49-50 Overripening of eggs and chemical changes 94, 96 and lipid content in fish eggs 96 and storage 74-5, 91-6, 93 and stripping 91-3 Ovulated eggs, viability of 92 Oxygen, concentration of in vertical migration 147

P Parallel communities hypothesis 1 8 6 7 Parasites in aquaculture 54-5 of Atlantic halibut 3 8 4 1 Pelagic phase of Atlantic halibut 11-22 Perturbation effect on community diversity 199 Phases of migration in fish 150-1 Physical and physiological properties and egg quality in fish 76 Phytoplankton see Plankton Pigments and quality in eggs 80-1, 98-9, 102 Plaice and match/mismatch hypothesis 251-2, 251, 270 vertical migration of 121, 133, 138, 141, 152 Plankton production and year class strength in fish populations 249-93 Poisson distribution model 1 9 3 4 Poisson-Inverse Gaussian distribution 208 Pollution and Atlantic halibut 42 and community stability 200, 2 0 6 7 and diversity 2 0 6 7 and egg quality 87-9 as habitat factor in benthic studies 227-8 indicators of 190-1 Positive binomial distribution model 193

Post-larval stage of Atlantic halibut 18-2 1 Predation avoidance as zeitgeber 1 4 3 4 and die1 vertical migration 127 and match/mismatch hypothesis 284 Prey availability of 127, 129-30 movements of 1 3 3 4 Principal Component Analysis in benthic studies 2 2 2 4 and plankton production 256 Principal Co-ordinate Analysis in benthic studies 222, 224-5 Productivity and biomass 210 Protein in Atlantic halibut 35-6 and feed quality in fish eggs 98-9, 100-1

Q Q-mode cluster analysis in benthic studies 220

R Rarefaction and dominance 201-3, 206 Rearing of Atlantic halibut 53, 57-60 Recruitment of Atlantic halibut 50, 51 in Baltic Sea 2 7 3 4 and climatic change 256 in fish populations 250, 284 and gadoid outburst 274-6 and larval abundance 271-3 in low latitude waters 259-63, 284 and salinity anomaly of 1970s 276-8 and tests of match/mismatch hypothesis 264 and upwellings 250, 259-61, 284 Redfish 3 3 4 vertical migration of 120, 123, 132 Relativization of data 184 Reproduction, of Atlantic halibut 7-1 1 Resource-limited assemblages 198 Retention areas and fish spawning 257-8 of larvae and tides 13940

306

SUBJECT INDEX

R-mode cluster analysis in benthic studies 184, 220, 225-6 Roach, vertical migration of 120-1, 126 Rudd, vertical migration of 120-1, 126

S Salinity anomaly of 1970s 276, 277, 278, 281 and Atlantic halibut 53, 55 Salmon egg quality 74, 98-9 spawning of 252 vertical migration of 144, 152 Sampling in benthic studies 171 design and spatial distribution 193-7 devices in benthic studies 173 effort 175-7 precision of 195 sieving 174 size and statistical power 213-14 timing of in benthic studies 177-8 Sandeel in match/mismatch hypothesis 270 vertical migration of 121, 133 Sardines, recruitment in California 260- 1 Screen size in sampling methods 174 Sediment habitat factors in benthic studies 227 Sex ratios in Atlantic halibut 23-5, 2931 Sexual maturity in Atlantic halibut 7, 8 Shannon-Weiner H' 204, 207 Similarity indices in benthic studies 215-16

Size and age of Atlantic halibut 2 3 4 , 25, 30, 31-3 and egg quality in fish 74-5 and larvae 78-9 of embryo of Atlantic halibut 1 1 and ontogenetic variation in fish 149 Smelt, vertical migration of 120, 13940, 143 Sole egg quality 74 vertical migration of 141-2

Spatial dispersion patterns in benthic studies 176 Spatial distribution and sampling design 193-7 Spawning of Atlantic halibut 9 areas 10-11 time of 9, 10 distributions of species 264, 270 and food 262-3, 284 induced, and broodstock management of fish 102-3 spread over time 282-3 time of and plankton production 258 in temperate waters 251-5 and upwelling 262-3, 284 Spearman rank correlation 204, 23 1 Species abundance curve 201-2, 205 Species area curves 202 Species groups, discreteness of 188 Species richness and diversity 197-8, 200-1, 203, 204, 205 Species status of Atlantic halibut 2 4 Spiny lobsters and plankton production 278, 279 Spot, vertical migration of 121, 141 Stability in aquatic communities 199 Stability-Time hypothesis, and diversity in communities 199-202 Standardization of data 184 Start-feeding period of Atlantic halibut 19 Starvation, and match/mismatch hypothesis 2 8 3 4 Statistical inference in benthic studies 212-14 Statistical power in benthic studies 21314 Steinhaus' coefficient in benthic studies 215 Storage of fish eggs see overripening and storage Stress and Atlantic halibut 53, 54, 57-8 Stripping of Atlantic halibut in aquaculture 54 overripening and storage of fish eggs 91, 92-3 Survival rates of larvae of Atlantic halibut 60

307

SUBJECT INDEX

and tests of match/mismatch hypothesis 264-5 Swordfish, vertical migration of 124, 125, 135

T Tasmania, spiny lobsters in 278-9 Taylor’s power law 184 and variance 194-5 Teleost, die1 vertical migrations of 1 1568 Temperature and die1 vertical migration 125 as environmental factor in vertical migration 145-7 and incubation of Atlantic halibut 12 Temperature-independence, of zeitgeber in fish migration 118 Temporal sampling design 177-8 Territoriality and distribution of species 193-4 Thermal fronts in tidal seas 254, 257-8 Thermoclines, and vertical migration 1 4 6 7 , 148 Tides intensity of 138-9 and offshore activity 139-41 and retention of fish larvae 139-40 stream transport and vertical migration 142 as zeitgeber 13743 Time and diversity in communities 198-9 of plankton production 255, 258-9 of spawning of Atlantic halibut 9-10 in temperate waters 251-5 Time series analysis in benthic studies 231-2 Trematodes and Atlantic halibut 38, 41 Trophic Level Formalism 209 Trout egg quality 74, 81-2, 98-9 egg viability 92-5, 93-4 ?’-test 213 Tuna, spawning 262, 263 Turbot egg quality 74 egg viability 92

U United Kingdom, halibut fisheries in 48 Univariate community analysis 171-2 Upwelling and fish recruitment 250, 259-61, 284 and Lasker events 261 and spawning 262-3, 284 and survival rates 264-5 USSR, halibut fisheries in 48-9

v Variance, Analysis of in benthic studies 182. 185, 213, 220 F-test 228 and Taylor’s power law 194-5 Veil line 205, 206 Vitamin C see ascorbic acid Vitamins in Atlantic halibut 36-7 and feed quality in fish eggs 98-9, 101-2

Water content in Atlantic halibut 35 stratification as ecosystem 254, 259 transparency and die1 vertical migration 127 Weather and vertical migration 123 Whiting chromosomal aberrations in 86 in match/mismatch hypothesis 270

Y Yellowfin and match/mismatch hypothesis 263 Yolk-sac larvae of Atlantic halibut 1 4 18, 57-60

Z Zeitgeber cyclic events as 122-45 bioenergetic advantage 136-7

308

SUBJECT INDEX

food 129-36 light 122-9 predator avoidance 143-4 tides 137-43 and endogenous rhythms in fish 117, 118 multiple 152-3, 154 removal of 127-8

Zooplankton and commensal species 145 in rearing experiments of Atlantic halibut 58, 62 seasonal distribution of 265-9 and vertical migration 134-6 Zurich-Montpellier index in benthic studies 215-16

Cumulative Index of Titles Alimentary canal and digestion in teleosts, 13, 109 Antarctic benthos, 10, 1 Artificial propagation of marine fish, 2, 1 Aspects of stress in the tropical marine environment, 10, 217 Aspects of the biology of frontal systems, 23, 163 Aspects of the biology of seaweeds of economic importance, 3, 105 Assessing the effects of “stress” on reef corals, 22, 1 Association of copepods with marine invertebrates, 16, 1 Behaviour and physiology of herring and other clupeids, 1, 262 Biological response in the sea to climatic changes, 14, 1 Biology of ascidians, 9, 1 Biology of clupeoid fishes, 10, 1 Biology of coral reefs, 1, 209 Biology of euphausiids, 7, I ; 18, 373 Biology of mysids, 18, 1 Biology of pelagic shrimps in the ocean, 12, 223 Biology of Phoronida, 19, 1 Biology of Pseudomonas, 15, 1 Biology of Pycnogonida, 24, I, Biology of the Atlantic Halibut, Hippoglossus hippoglossus (L.,1758), 26, 1 Biology of wood-boring teredinid molluscs, 9, 336 Blood groups of marine animals, 2, 85 Breeding of the North Atlantic freshwater eels, 1, 137 Circadian periodicities in natural populations of marine phytoplankton, 12, 326 Comparative physiology of Antarctic Fishes, 24, 321 Competition between fisheries and seabird communities, 20, 225 Coral communities and their modification relative to past and present prospective Central American seaways, 19, 91 Diseases of marine fishes, 4, I Development and application of analytical methods in benthic marine infaunal studies, 26, 169 Die1 vertical migrations of marine fishes: an obligate or facultative process? 26, 115 Ecology and taxonomy of Halimeda: primary producer of coral reefs, 17, 1 Ecology of deep-sea hydrothermal vent communities, 23, 301 Ecology of intertidal gastropods, 16, 1 1 1 Egg quality in fishes, 26, 71 Effects of environmental stress on marine bivalve molluscs, 22, 101 Effects of heated effluents upon marine and estuarine organisms, 3, 63

309

310

CUMULATIVE INDEX OF TITLES

Environmental simulation experiments upon marine and estuarine animals, 19, I33 Estuarine fish farming, 8, 119 Fish nutrition, 10, 383 Flotation mechanisms in modern and fossil cephalopods, 11, 197 General account of the fauna and flora of mangrove swamps and forests in the IndoWest Pacific region, 6, 74 Growth in barnacles, 22, 199 Gustatory system in fish, 13, 53 Habitat selection by aquatic invertebrates, 10, 271 History of migratory salmon acclimatization experiments in parts of the Southern Hemisphere and the possible effects of oceanic currents and gyres upon their outcome, 17, 397 Influence of temperature on the maintenance of metabolic energy balance in marine invertebrates, 17, 329 Interactions of algal-invertebrate symbiosis, 11, 1 Laboratory culture of marine holozooplankton and its contribution to studies of marine planktonic food webs, 16, 21 1 Learning by marine invertebrates, 3, 1 Management of fishery resources, 6, 1 Marine biology and human affairs, 15, 233 Marine molluscs as hosts for symbioses, 5, 1 Marine toxins and venomous and poisonous marine animals, 3, 256 Marine toxins and venomous and poisonous marine plants and animals, 21, 59 Methods of sampling the benthos, 2, 171 Natural variations in 5Nin the marine environment, 24, 389 Nutrition of sea anemones, 22, 65 Nutritional ecology of ctenophores, 15, 249 Parasites and fishes in a deep-sea environment, 11, 121 Parasitology of marine zooplankton, 25, 117 Particulate and organic matter in sea water, 8, 1 Petroleum hydrocarbons and related compounds, 15, 289 Photosensitivity of echinoids, 13, 1 Physiological mechanisms in the migration of marine and amphihaline fish, 13, 248 Physiology and ecology of marine bryozoans, 14, 285 Physiology of ascidians, 12, 2

CUMULATIVE INDEX OF TITLES

31 1

Pigments of marine invertebrates, 16, 309 Plankton as a factor in the nitrogen and phosphorus cycles in the sea, 9, 102 Plankton production and year class strength in fish populations: An update of the match/mismatch hypothesis, 26, 249 Pollution studies with marine plankton, Part 1 : Petroleum hydrocarbons and related compounds, 15, 289 Pollution studies with marine plankton, Part 2: Heavy metals, 15, 381 Population and community ecology of seaweeds, 23, 1 Population biology of blue whiting in the North Atlantic, 19, 257 Predation on eggs and larvae of marine fishes and the recruitment problem, 25, 1 Present status of some aspects of marine microbiology, 2, 133 Problems of oil pollution of the seas, 8, 215 Rearing of bivalve mollusks, 1, 1 Recent advances in research on the marine alga Acetabularia, 14, 123 Recent developments in the Japanese oyster culture industry, 21, I Recent studies on spawning, embryonic development, and hatching in the Cephalopoda, 25, 85 Relationships between the herring, Clupea harengus L., and its parasites, 24, 263 Respiration and feeding in copepods, 11, 57 Review of the systematics and ecology of oceanic squids, 4, 93 Sandy-beach bivalves and Gastropods: A comparison between Donax serra and Bullia digitalis, 25, 179 Scallop industry in Japan, 20, 309 Scatological studies of the Bivalvia (Mollusca), 8, 307 Siphonophore biology, 24, 97 Some aspects of the biology of the chaetognaths, 6, 271 Some aspects of neoplasia in marine animals, 12, 151 Some aspects of photoreception and vision in fishes, 1, 171 Speciation in living oysters, 13, 357 Study in erratic distribution: the occurrence of the medusa Gonionemus in relation to the distribution of oysters, 14, 251 Taurine in marine invertebrates, 9, 205 Upwelling and production of fish, 9, 255

This Page Intentionally Left Blank

Cumulative Index of Authors Glynn, P. W., 19, 91 Goodbody, I., 12, 2 Gotto, R. V., 16, 1 Grassle, J. F., 23, 301 Gulland, J. A., 6, 1 Harris, R. P., 16, 211 Haug, T., 26, 1 Hickling, C. F., 8, 119 Hillis-Colinvaux, L., 17, 1 Holliday, F. G. T., 1, 262 Holme, N. A., 2, 171 Holmefjord, I., 26, 71 Houde, E. D., 25, 1 Howard, L. S., 22, 1 Hunter, J. R., 20, 1 Kapoor, B. G., 13, 53, 109 Kennedy, G. Y., 16, 309 Kjerrsvik, E., 26, 71 Le Fevre, J., 23, 163 Loosanoff, V. L., 1, 1 Lurquin, P., 14, 123 Macdonald, J. A., 24, 321 MacKenzie, K., 24, 263 Mackie, G. O., 24, 97 McLaren, I. A., 15, I Macnae, W., 6, 74 Mangor-Jensen, A,, 26, 71 Marshall, S. M., 11, 57 Mauchline, J., 7, 1; 18, 1 Mawdesley-Thomas, L. E., 12, 151 Mazza, A., 14, 123 Meadows, P. S.,10, 271 Millar, R. H., 9, 1 Millott, N., 13, 1 Montgomery, J. C., 24, 321 Moore, H. B., 10, 217 Naylor, E., 3, 63 Neilson, J. D., 26, 115 Nelson-Smith, A., 8, 215 Nemec, A., 26, 169 Newell, R. C., 17, 329 Nicol, J. A. C., 1, 171 Noble, E. R., 11, 121 ,Omori, M., 12, 233 ,Owens, N. J. P., 24, 389 Paffenhofer, G-A., 16, 21 1

Akberali, H. B., 22, 102 Allen, J. A., 9, 205 Ahmed, M., 13, 357 Arakawa, K. Y., 8, 307 Arnaud, F., 24, 1 Bailey, K. M., 25, 1 Bailey, R. S., 19, 257 Balakrishnan Nair, N., 9, 336 Bamber, R. N.. 24, I Blaxter, J. H. S.,1, 262, 20, 1 Boletzky, S.V.,25, 85 Boney, A. D., 3, 105 Bonotto, S., 14, 123 Bourget, E., 22. 200 Branch, G. M.. 17, 329 Brinkhurst, R. 0..26, 169 Brown, A. C., 25, 179 Brown, B. E., 22, 1 Bruun, A. F., 1, 137 Burd, B. J., 26, 169 Campbell, J. I., 10, 271 Carroz, J. E., 6, 1 Chapman, A. R. O., 23, 1 Cheng, T. C., 5, I Clarke, M. R., 4, 93 Corkett, C. J., 15, 1 Corner, E. D. S., 9, 102; 15, 289 Cowey, C. B., 10, 383 Crisp, D. J., 22, 200 Cushing, D. H., 9, 255; 14, 1; 26, 249 Cushing, J. E., 2, 85 Davenport, J., 19, 133 Davies, A. G., 9, 102; 15, 381 Davies, H. C., 1, 1 Dell. R. K., 10, 1 Denton, E. J., 11, 197 Dickson, R. R., 14, I Edwards, C., 14, 251 Emig, C. C., 19, 1 Evans, H. E., 13, 53 Fisher, L. R., 7. 1 Fontaine, M., 13, 248 Furness, R. W.. 20, 225 Garrett, M. P., 9, 205 Ghirardelli, E., 6, 271 Gilpin-Brown, J. C., 11, 197

313

314

CUMULATIVE INDEX OF AUTHORS

Perry, R. I., 26, 115 Pevzner, R. A., 13, 53 Pugh, P. R., 24, 97 Purcell, J. E., 24, 97 Reeve, M. R., 15, 249 Riley, G. A., 8, 1 Russell, F. E., 3, 256; 21, 60 Russell, F. S., 15, 233 Ryland, J. S., 14, 285 Saraswathy, M., 9, 336 Sargent, J. R., 10, 383 Scholes, R. B., 2, 133 Shelbourne, J. E., 2, 1 Shewan, J. M., 2, 133 Sindermann, C. J., 4, 1

Smit, H., 13, 109 Sournia, A., 12, 236 Stenton-Dozey, J. M. E., 25, 179 Stewart, L., 17, 397 Taylor, D. L., 11, 1 Theodoridts, J., 25, 117 Trueman, E. R., 22, 102; 25, 179 Underwood, A. J., 16, 111 Van-Praet, M., 22, 66 Ventilla, R. F., 20, 309; 21, 1 Verighina, I. A., 13, 109 Walters, M. A., 15, 249 Wells, M . J., 3, 1 Wells, R. M. G., 24, 321 Yonge, C. M., 1, 209