Advances in Agronomy, Volume 44

ADVANCES IN

AGRONOMY VOLUME 44

This Page Intentionally Left Blank

ADVANCES IN

AGRONOMY Prepared in Cooperation wirh the AMERICAN SOCIETYOF AGRONOMY

VOLUME 44 Edited by N. C . BRADY Science and Technology Agency for International Development Department of State Washington, D . C .

ADVISORY BOARD N. L. TAYLORH. G . HODGES

E. L. KLEPPERG . L. HORST R. J . KOHEL R. H. MILLER G . E. HAM S . MICHELSON K. H. QUESENBERRY C. W. STUBER G . H . HEICHELD. E. KISSEL

ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishers San Diego New York Boston London Sydney Tokyo Toronto

This book is printed on acid-free paper.

@

Copyright 0 1990 by Academic Press, Inc. All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Academic Press, Inc. San Diego, California 92 10 1 United Kingdom Edition published by Academic Press Limited 24-28 Oval Road, London NW I 7DX

50-5598

Library of Congress Catalog Card Number:

ISBN 0-12-000744-4

(alk. paper)

Printed in the United States of America 9 0 9 1 9 2 9 3 9 8 7 6 5 4

3

2

1

CONTENTS CONTRIBUTORS ..........................................................................

PREPACE ..................................................................................

ix xi

VARIATION IN TIME OF SEEDLING EMERGENCE WITHIN POPULATIONS: A FEATURE THAT DETERMINES INDIVIDUAL GROWTH AND DEVELOPMENT

L. R. Benjamin I. 11. 111. 1V.

.......................................... Introduction ................................ Factors That Influence Time of Seedling Emergence.. ......................... Importance of Variation in Time of Seedling Emergence to Crop Development .............................................................................. Conclusions ............................ References .................................................................................

1 2

9 20 21

FORAGE TREE LEGUMES: THEIR MANAGEMENTAND CONTRIBUTIONTO THE NITROGEN ECONOMY OF WET AND HUMID TROPICAL ENVIRONMENTS

Graeme Blair, David Catchpoole, and Peter Horne 1. 11.

111. 1v. V. VI. VII. VIII.

Introduction.. .......................................................... Species of Useful Tree Legumes. ............................................... Agronomic Performance of Tree Legumes ........................................ Tree Legume Leaf as Animal Feed .................................................. Management of Tree Legumes........................................................ Nitrogen Yields of Material Harvested from Tree Legumes .................. Nitrogen Recycling via Leaf and Excreta.. ...................... Conclusion ................................................................................. References ........................................................................

21 28 29 34 36 45 46 49 50

STATISTICAL ANALYSES OF MULTILOCATIONTRIALS

Jose Crossa I. 11.

Introduction ................................................................................ Conventional Analysis of Variance.. ................................................ V

55

51

vi

CONTENTS

111. Joint Linear Regression................................................................. IV. Crossover Interactions.. ................................................................ V. Multivariate Analyses of Multilocation Trials .................................... VI. AMMl Analysis ........................................................................... VII. Other Methods of Analysis ............................................................ VIII. General Considerations and Conclusions .......................................... References .................................................................................

61

68 70 76 80 81 82

EVALUATION AND DOCUMENTATION OF GENETIC RESOURCES IN CEREALS

A. B. Damania I. 11. 111.

IV. V. VI. VII.

Introduction ..................... ................................................ Evaluation of Cultivated Wheat ...................................................... Evaluation of Cultivated Barley ......................................... Genetic Resources from Ethiopia.. .................................................. Evaluation of Wild and Primitive Forms of Wheat and Barley ............... Documentation of Genetic Resources .............................................. Summary and Conclusions. .................... References .................................................................................

87 90 93 95 96 I02 I05 107

MODELING CROP ROOT GROWTH AND FUNCTION

Betty Klepper and R. W.Rickman 1. 11.

111. IV. V. VI.

Introduction .............................................. Early Models .............................................................................. Desirable Model Features.. ................................................... Model Components ...................................................................... Some Existing Root Growth and Function Models ............................. Limitations to Development of Root Growth Models ..........................

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

I13 114 115

118 I28 130 131

GENETIC MANIPULATION OF THE COWPEA (Vigna unguiculafa [L.] Walp.) FOR ENHANCED RESISTANCE TO FUNGAL PATHOGENS AND INSECT PESTS

A. 0. Latunde-Dada 1. 11. 111.

IV.

Introduction ................................................................................ Insect Pests ................................................................................ Fungal Pathogens.. ....................................................................... Tissue Culture Technology ............................................................

133 139 141 142

vii

CONTENTS V.

Conclusions and Epilogue.. .................................................... ........ References ............................................................ ....................

.

149

I 50

NITROGEN FIXATION BY LEGUMES IN TROPICAL AND SUBTROPICAL AGRICULTURE

Mark B. Peoples and David F. Herridge

I. 11.

111. IV. V. VI .

.

Introduction.. .................................................................. ........... Methods of Assessing N2 Fixation N2 Fixation in Legume Production Systems ...................................... Contribution of Legume N to Plant and Animal Producti Strategies to Enhance N2 Fixation ................................................... Concluding Remarks .................................,............ References ............................. ..............................................

i56 158 177 190 202 216 216

DISTRIBUTION, COLLECTION, AND EVALUATION OF Gossypium

A. Edward Percival and Russell J. Kohel

.

11.

Introduction.. ........ ........ ....................... Distribution ..........,.....................................................................

Ill. IV . V.

Evaluation.. ................................................................................ Concluding Remarks ................................ ..

1.

225 228 235 245 253 253

BREEDING WHEAT FOR RESISTANCE TO Septoria nodorum AND Septoria tritici

Lloyd R. Nelson and David Marshall

I. 11. 111. IV. V. VI .

Introduction.. .............................................................................. Identification of Resistance. ....................... Pathogen Variation.. ........... ......................................................... Genetics of Resistance ............................................. Sources of Resistance ..... ....,....................... Discussion and Conclusions ...........................................................

.

INDEX......................................................................................

257 258 268 270 272 272 214

279

This Page Intentionally Left Blank

CONTRIBUTORS Numbers in parentheses indicate the pages on which the authors' contributions begin.

L. R. BENJAMIN ( l ) , AFRC Institute of Horticultural Research, Wellesbourne, Wanvick CV35 9EF, England GRAEME BLAIR (27), Australian Centre for International Agricultural Research, Department of Agronomy and Soil Science, University of New England, Armidale, New South Wales 2351, Australia DAVID CATCHPOOLE* (27), Australian Centre for International Agricultural Research, Department of Agronomy and Soil Science, University of New England, Armidale, New South Wales 2351, Australia JOSE CROSSA (55), Biometrics and Statistics Unit, International Maize and Wheat Improvement Center (CIMMYT),06600 Mexico D. F., Mexico A. B. DAMANIA (87), Genetic Resources Unit, International Centerfor Agricultural Research in the Dry Areas (ICARDA),Aleppo, Syria DAVID F. HERRIDGE (155), Australian Centre for International Agricultural Research (Project 8800),New South Wales Agriculture and Fisheries, Tamworth, New South Wales 2340, Australia PETER HORNE (27), Australian Centre for International Agricultural Research, Department of Agronomy and Soil Science, university of New England, Armidale, New South Wales 2351, Australia BETTY KLEPPER ( 1 13), United States Department of Agriculture, Agricultural Research Service, Columbia Plateau Conservation Research Center, Pendleton, Oregon 97801 RUSSELL J. KOHEL (229, United States Department of Agriculture, Agricultural Research Service, Southern Crops Research Laboratory, College Station, Texas 77840 A. 0. LATUNDE-DADA (133), Department of Crop Production, College of Agricultural Sciences, Ogun State university, Ago-Iwoye, Ogun State, Nigeria DAVID MARSHALL (257), Texas A&M University Research and Extension Center at Dallas, Dallas, Texas 75252 LLOYD R. NELSON (257), Texas A&M University Agricultural Research and Extension Center at Overton, Overton, Texas 75684

*Present address: Queensland Department of Primary Industries, Ayr. Queensland 4807, Australia.

ix

X

CONTRIBUTORS

MARK B. PEOPLES ( I S ) , Australian Centre for International Agricultural Research (Project SSOO), CSIRO Division of Plant Industry, Canberra, A . C . T . 2601, Australia A. EDWARD PERCIVAL (225), United States Department of Agriculture, Agricultural Research Service, Southern Crops Research Laboratory, College Station, Texas 77840 R. W. RICKMAN ( 1 13), United States Department of Agriculture, Agricultural Research Service, Columbia Plateau Conservation Research Center, Pendleton, Oregon 97801

PREFACE In the nearly 25 years I have had the privilege of editing this serial, it has been an inspiration to witness the desire and willingness of crop and soil scientists to prepare review articles for their compatriots. Scientists throughout the world have taken literally thousands of hours of their own time and energies to prepare articles for Advances in Agronomy gaining little but the satisfaction that they were doing a favor for their fellow scientists. They have indeed furthered the cause of science by writing these reviews. The authors of the nine articles in this volume have followed the tradition of their predecessors. Located at research institutions in six different countries, they maintain the international focus of this serial. The subjects covered vary from genetic resources of cereals and cotton to timing of seedling emergence and modeling of root growth and function. Two articles focus on tropical crops and agriculture as well as agro-forestry, subjects of keen concern as low-income farmers of the tropics struggle to increase food production while maintaining the quality of their environment and especially of their soils. Efforts to increase the host resistance of wheat and cowpeas are covered in two articles. These are evidence of the increasing focus on alternatives to the use of chemical pesticides to control plant pests. Host resistance is one of the characteristics evaluated in multilocational field trials around the world. The statistical analyses of such trials are the subject of another article in this volume. Thanks are due the advisory board of the American Society of Agronomy and the directors-general of three international agricultural research centers for suggesting topics and authors for this serial. To the 15 authors of the important reviews contained herein I express my special gratitude. N. C. BRADY

xi

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY, VOL. 44

VARIATION IN TIME OF SEEDLING EMERGENCE WITHIN POPULATIONS: A FEATURE THAT DETERMINES INDIVIDUAL GROWTH AND DEVELOPMENT L. R. Benjamin AFRC Institute of Horticultural Research Wellesbourne, Warwick CV35 9EF, England

Introduction Factors That Influence Time of Seedling Emergence A. Water B . Temperature C . Sowing Depth D . Seed Attributes E. Conclusions Ill. Importance of Variation in Time of Seedling Emergence to Crop Development A. Total Plant Growth B. Partitioning between Organs C . Organ Morphology and Composition D. Longevity IV. Conclusions References 1. 11.

I. INTRODUCTION The point in time when the growing point of a shoot emerges from the soil into the aerial environment is one of the most easily observed events in crop development. The range of percentage and timing of seedling emergence can be large, even in cultivated species, because emergence is the culmination of a large number of preceding events. For example, FinchSavage (1984) reported in carrots sown at eight different times between 1 1 February and 15 June in the United Kingdom that the number of days between the 10 and 90 percentile points for emergence ranged from 11 to 24. In noncultivated Umbelliferae. and in some cultivated Umbelliferae 1 Copyright Q 1990 by Academic Press. Inc. All rights of reproductionin any form reserved.

2

L. R. BENJAMIN

species, there are also dormancy mechanisms that result in periodicity of emergence times (Roberts, 1979). Much attention has been directed to unraveling the complex interactions between the agronomic and genetic factors that influence seedling emergence. The salient features of these factors and their influence on variation in seedling emergence time will be reviewed here. Most plant communities are characterized by intense competition between individuals for growth resources, which leads to the development of dominance hierarchies (Watkinson, 1985; Benjamin and Hardwick, 1986; Weiner and Thomas, 1986). These hierarchies are important in crops because they contribute to variation in weight per plant, which is undesirable for a market that increasingly requires uniformly sized produce for processing and the fresh market (Anonymous, 1982), and can also contribute to loss of economic yield (Benjamin and Hardwick, 1986). The hierarchies are also important in natural communities because they contribute to size-dependent fecundity (Pacala and Slander, 1985; Watkinson et al., 1989) and to size-dependent mortality (Mithen et ul., 1984; Schmitt et ul., 1987; Thomas and Weiner, 1989). This review will examine the relevance of time of seedling emergence to the development of these hierarchies and seek mechanisms to account for any relationships.

II. FACTORS THAT INFLUENCE TIME OF SEEDLING EMERGENCE Crops have been bred and selected for genetic uniformity and elimination of seed dormancy mechanisms. Engineers have developed sophisticated equipment to produce good seedbed tilths and to sow seeds at uniform depth. So why should there be variation in times of seedling emergence? About a hundred papers are published annually on germination (Lovato, 1981) and this subject has been reviewed extensively (Koller, 1972; Heydecker, 1973; Harper, 1977; Heydecker and Coolbear, 1977; Johnston, 1979; Perry, 1982). The purpose of this review is not to make another exhaustive study but rather to examine the ideas that are implicit in most of these studies and to determine how much is already known about the causes of variation in time of seedling emergence. Although germination is a complex process, it has only three requirements: water, warmth, and a free exchange of gases. Seed dormancy is common in most species, but despite some notable exceptions, it has largely been overcome in commercial crops (Villiers,

SEEDLING EMERGENCE WITHIN POPULATIONS

3

1972; Maguire, 1983). Seeds are usually sown directly into soil, often at varying depths, and the subsequent germination is in conditions of fluctuating temperature and water supply. The effects of these factors on time of seedling emergence will be considered in the next sections. A. WATER Water influences the spread in time of seedling emergence in a number of ways. First of all, water is essential for germination, so any restriction on its supply reduces the rate and final percentage of seed germination. The supply of water to the seed is governed by the conductivity of the soil water (Collis-George and Sands, 1959; Williams and Shaykewich, 1971; Hadas and Russo, 1974a), the degree of seed-soil water contact (Sedgley, 1963; Manohar and Heydecker, 1964; Collis-George and Hector, 1966; Harper and Benton, 1966; Hadas and Russo, 1974a), and the osmotic and matric potentials (Collis-George and Sands, 1959, 1962; Manohar and Heydecker, 1964; Williams and Shaykewich, 1971; Dasberg and Mendel, 1971; El-Sharkawi and Springuel, 1979; Ross and Hegarty, 1979; Willat and Struss, 1979; Tipton, 1988). Increasing the supply of water can restrict the supply of oxygen that is necessary for germination. Oxygen is sparingly soluble and its solubility decreases with increasing temperature, whereas the metabolic demand for this gas increases with temperature. In addition, any complementary buildup of carbon dioxide has a poisoning effect on germination (Heydecker, 1958). Consequently, even slight excesses in the amount of moisture can have large inhibitory effects on germination (Heydecker et al., 1969; Dasberg and Mendel, 1971). Hanks and Thorp (1956) reported that emergence of wheat was restricted when the oxygen diffusion rate g cm-2 min-I. This corresponded to an air (ODR) fell below 75-100 x pore space of 16% in a silt clay and 25% in a fine-sandy loam. Subsequent work shows that the rate of germination is restricted by an ODR as little as 20 x lo-' g cmP2min-l (Dasberg and Mendel, 1971). Insufficient oxygen supply has been alleviated by coating seeds with calcium or zinc peroxides, or by incorporating calcium peroxide in the growing medium, but the beneficial effects varied greatly with species and occurred only when the moisture content of the growing medium was very high. Furthermore, the addition of peroxides in drier media often had a detrimental effect on germination and emergence (Brocklehurst and Dearman, 1983; Langan et af., 1986). Even if germination proceeds rapidly and uniformly, there can be a wide spread in time of seedling emergence due to a restriction of seedling growth

4

L. R. BENJAMIN

by the strength of the soil, which is largely governed by its water content (Arndt, 1965; Collis-George and Williams, 1968; Royle and Hegarty, 1977; Hegarty and Royle, 1978). In addition, slow drying gives closer packing of soil particles, resulting in high soil strengths (Gerard, 1965). However, the more important restriction to seedling emergence is the creation of crusts by surface drying (Hanks and Thorp, 1956,1957; Royle and Hegarty, 1977; Nuttall, 1982). Spread in time of emergence is determined by soil strength because the emergence force that seedlings exert develops dynamically and is a linear function of volumetric soil-water content and the cross-sectional area of the seedlings (Gerard, 1980). Taylor and Broeck (1988) measured the emergence force exerted by nine vegetable species at 25°C in sand at 15% moisture content and showed that the time taken to exert maximum force ranged from 4 hr in red beet to 21 hr in snap bean. Increasing the salinity of soil water caused a reduction in the size of the emergence force exerted by seedlings and increased the time required to exert the maximum force (Sexton and Gerard, 1982). Adding nitrogen fertilizers to soils has reduced percentage emergence, presumably because of osmotic effects (Hegarty, 1976a; Page and Cleaver, 1983). Nearly all studies of the effects of fertilizers on seedling growth have examined only final percentage emergence. However, Henriksen (1978)showed that addition of 75 or 150 kg N ha-' prior to sowing onions increased the standard deviation of emergence times by about half a day as well as reduced the percentage emergence compared with addition of the nitrogen after emergence. The salts that increase the salinity of the soil water can have the opposite effect of stimulating seedling growth by supplying essential mineral nutrients, such as phosphorus (Costigan, 1984). Most studies of seedling emergence have imposed constant soil moisture conditions. In nature, however, seeds are exposed to a fluctuating supply of water. This fluctuation affects variation in mean time of seedling emergence between populations in both weed (Roberts, 1984) and cultivated species (Hegarty, 1976b; Finch-Savage, 1986), but is also liable to be a major determinant of the individual-to-individual variation in time of seedling emergence within a population. B. TEMPERATURE In some species, seeds require exposure to low temperatures to break dormancy (see Roberts, 1972 for a review). All species show a qualitative relationship between germination parameters and temperature. The usual

SEEDLING EMERGENCE WITHIN POPULATIONS

5

responses are an approximately linear increase in rate (reciprocal in time taken to start of or some percentage of germination) with increasing temperature from a threshold to a maximum, with or without a plateau, followed by a linear decline at superoptimal temperatures. As a consequence of this linearity, it is convenient to describe the effect of temperature on the mean time of seedling emergence in terms of temperature sums (often erroneously called heat sums) (Hegarty, 1973; Bierhuizen and Wagenvoort, 1974; Garcia-Huidobro et al., 1982a). Only a few studies have examined the effect of temperature on the variability in the time of germination, but there is evidence in carrots that the spread in time of germination decreases with increase in temperature over the range 5°C to 25°C (Gray, 1979). In nature, seeds are exposed to fluctuating temperatures and, for noncultivated species, this might be a requirement for germination (Thompson, 1974). However, for most cultivated species, fluctuations in temperature have negligible practical effects on time of germination (Wagenvoort and Bierhuizen, 1977; Garcia-Huidobro et al., 1982b). In moist seedbeds, the rate of seedling emergence has a relationship with temperature similar to that described for germination (Muendel, 1986; Finch-Savage, 1986), and temperature sums have been used to describe the effects of temperature on the timing of emergence (Khah et al., 1986; Tenhovuori, 1986). However, the timing of seedling emergence is not governed entirely by the relationship between germination and temperature, because low temperatures also increase the time taken by seedlings to exert maximum force (Gerard, 1980). An example of this interaction between temperature and soil compaction was found in calabrese by Hegarty and Royle (1978). They showed that as temperature decreased from 20°C to 6"C, percentage emergence decreased from 93% to 78% when 0.6 N cm-* pressure had been applied, but the percentage emergence decreased from 90% to 33% when 4.8 N cm-* had been applied. The interaction between temperature and soil-water matric potential was quantified by Tenhovuori (1986), who showed that the temperature sum required for 50% emergence increased linearly above a threshold value as the soil-water matric potential increased. Gummerson ( 1989) examined the influence of seedbed preparation practices on the influence of moisture content, impedance, aeration, and temperature on the emergence of sugar beet. Of all four factors, he reported that temperature was the one that consistently limited rate of seedling emergence. There appear to be no studies that have quantified the relative importance of temperature for the spread in time of seedling emergence.

6

L. R. BENJAMIN

C. SOWING DEPTH The amplitude of diurnal variation in temperature lessens and the time of maximum and minimum daily temperature shifts with increasing depth (Orchard and Wurr, 1977). Hence, deep-sown seeds experience a more uniform temperature than shallow-sown seeds. Similar considerations would apply to moisture content of the soil. However, deep-sown seeds would be expected to take longer to emerge and small seeds often do not have the ability to penetrate through a deep layer of soil (Moore, 1943; Black, 1956; Stickler and Wassom, 1963; Arnott, 1969; Snyder and Filban, 1970; Bedford and MacKay, 1973; Wagenvoort and Bierhuizen, 1977; Abul-Fatih and Bazzaz, 1979; Buckley, 1982;Nuttall, 1982). This inability could be due to insufficient stored materials to generate the osmotic gradient necessary to overcome the pressure exerted by the soil (Black, 1956)or the seedling might be too weak to withstand the forces necessary to overcome the resistance of the soil, with the hypocotyl breaking as it drags the cotyledon through the soil (Rathore et al., 1981). Nuttall (1982) attributed better emergence of rape from small seeds to the requirement for less energy to push small cotyledons through the soil crusts. However, the differences in seed sizes were confounded with differences in cultivar in his experiments and the optimum sowing depths for cabbage, lettuce, carrot, and onion were 1.5-2.5 cm, despite differences between these species in seed size, presence of endosperm, and being mono- or dicotyledons (Heydecker, 1956). D. SEEDATTRIBUTES The previous sections have concentrated on the effects of the external environment on variation in time of seedling germination and emergence, but attributes of the seed also influence the rate of germination and emergence. Even in a favorable, uniform environment, seeds do not germinate synchronously but display a probability of germinating in a unit length of time (Thornley, 1977; Bould and Abrol, 1981).The effect of environmental factors such as temperature and water supply is to influence this probability of seed germination (Harper, 1977; Bould and Abrol, 1981). This stochastic nature might be an inevitable consequence of germination being a chain of many physical, biochemical, and physiological events (Thornley, 1977; Tipton 1984). For example, in carrots, germination was faster in seeds containing large embryos (from primary umbels) than in those containing small embryos (from secondary umbels) (Gray, 1979). Also the

SEEDLING EMERGENCE WITHIN POPULATIONS

7

standard deviation of germination time was less in seed lots with a low seed-to-seed variation in embryo length (mature seed lots) than in those containing variable embryo lengths (immature seed lots). Although dormancy is not considered to be a problem for germination in most cultivated species, it is well known in many natural species. Furthermore, there is a well-known inhibition of germination in some cultivated species by specific environmental stimuli. For example, light can inhibit germination of some cultivars of lettuce, tobacco, and tomato (Pollock, 1972). The corky capsules that surround beet seeds contain a watersoluble inhibitor to germination (but see Morris et al., 1984 for an opposing view). Carrot seeds were considered not to contain such inhibitors, but recent work on seed priming has revealed their presence (Pill and FinchSavage, 1988). Thus, these inhibitors of germination may be more ubiquitous in cultivated crops than previously suspected. Attributes of the seed also interact with environmental factors to determine the rate of germination. For example, Harper and Benton (1966) showed that the germination of all types of seeds was restricted by low matric potential when placed on sintered glass disks, but mucilaginous seeds were least sensitive, spiny reticulate seeds were the most sensitive, and smooth seeds showed a graded response to water tension. Small seeds were less sensitive to water tension than large seeds. Time of seedling emergence is controlled by genetic constitution (Eagles, 1988; Lafond and Baker (1986) and seed size (Lafond and Baker, 1986). However, some studies showed no effect of seed size on time of seedling emergence (Naylor, 1980; Stanton, 1984). These inconsistent results might be attributed to the use of different growing media in the different studies. The optimum seed-soil water contact for germination is achieved when the mean aggregate size of soil particles is one-fifth to one-tenth of the seed’s diameter (Hadas and Russo, 1974b). Adverse soil conditions might be partially overcome by using seeds that are “robust.” Some seed lots have persistently high field emergence over a wide range of soil conditions (Hegarty, 1974). Osmotic priming of seeds often improves seedling establishment, presumably by bringing all seeds to a uniformly mature state (see Bradford, 1986 for a review of this technique). It might be possible to breed for specific seed properties that favor germination, for example, small seeds and cracked testas (Whittington, 1978). However, these factors that favor germination might be detrimental to emergence of seedlings in field conditions. It is interesting to consider whether seed attributes or environmental factors dominate time of seedling emergence. Only a few studies have addressed this question directly, but indirect studies indicate an overriding importance of the environment. For example, varying soil texture has

8

L. R. BENJAMIN

large effects on seedling emergence (Hammerton, 1961 ;Wurr e f al., 1982). There are even large interactions between the method of watering (by capillary action or by surface watering) and soil moisture content on the percentage of seedling emergence (Heydecker, 1961). However, attributes of the seed can influence emergence in unexpected ways. For example, oil seed rape seedlings adapt to high soil impedance by decreasing the time taken to develop their full emergence force. This response to soil impedance was enhanced or inhibited by substances that affected ethylene production or action (Clarke and Moore, 1986). When such subtle interactions occur between seed and environment, it is perhaps naive to determine their relative importance. However, the relative importance of various attributes of seeds for variation in time of seedling emergence has been estimated (Waller, 1985). Waller collected seeds from cleistogamous (self-pollinating) and chasmogamous (cross-pollinating)flowers of jewelweed (Impatiens capensis) and found that between a third and a quarter of the variation in time of seedling emergence was associated with seed weight, seed type, maternal parent, and their interaction. E. CONCLUSIONS

Variation in time of seedling emergence arises because it is the culmination of a large number of preceding processes whose rates can differ between individuals. Differences between seeds in their genetic constitution, development on the mother plant, and exposure to extraneous factors, such as fungal attack, produce variation in time of germination in uniform environments. In nature, an additional source of variation in time of germination occurs from heterogeneity of the soil. Harper et al. (1965) suggested that germination of broadcast seeds depends on available sites of warmth and moisture. This idea is a useful concept also for buried seeds. Hegarty and Royle (1978) noted that there was greater seedling emergence in a dry soil that had been compacted than in a similar soil that had not been compacted. They speculated that compaction had improved the water supply to the seeds, presumably by increased seed-water contact, which effectively increased the number of sites for germination. Dasberg and Mendel (1971) claimed that “the rate of seed-water uptake governs germination. This rate is determined in general by the energy status of the water in the germination medium, by its conductivity, and by the area of contact between seed and medium, which is a function of pore geometry and surface tension.” Seed death is another important aspect of the effects of soil conditions on seedling emergence (Harper, 1955; Hegarty, 1978). In a system as multifaceted as the seed-soil complex, it is inevitable that

SEEDLING EMERGENCE WITHIN POPULATIONS

9

there is a wide spread in seedling emergence times, even in a crop sown synchronously (Hegarty, 1976b). The foregoing studies indicate that manipulation of any one of a number of processes would reduce the spread in time of seedling emergence, but no one treatment would produce synchrony of emergence. Spread in time of seedling emergence can be reduced by improved seed production techniques, by laboratory techniques to bring all seeds to maturity (priming), and by improved engineering to give uniform depth and optimum seed-soil contact in the seedbed. However, the most pragmatic way of reducing the spread in time of seedling emergence is to ensure a continued supply of soil moisture at a level that is optimum for germination during the period of imbibition, radicle emergence, and eruption of the shoots through the soil surface. The purpose of the rest of this review is to examine the importance of this spread in time of seedling emergence to the subsequent development of the plant community.

111. IMPORTANCE OF VARIATION IN TIME OF SEEDLING EMERGENCE TO CROP DEVELOPMENT In the remaining part of this review, I shall examine the importance of seedling-to-seedling variation in time of emergence within populations on the subsequent development of each plant. Although much work has been done to compare the effects of treatments that influence mean time of seedling emergence in separate populations (Hegarty, 1976b; Symonides, 1978; Gummerson, 1989), this is of little value for determining the importance of time of seedling emergence on the interactions between individuals within a population. Therefore, the remainder of the review will be confined to those few studies that examined the development of individual plants. Is the time of seedling emergence truly a boundary in the course of plant development, or is it just an event of no consequence to the plant but easily observed by humans? Certainly there is a switch from carbon for growth provided by the mobilization of seed reserves to carbon provided by photosynthesis. However, presumably some mineral nutrient and water absorption occurs through the roots of a seedling whose shoots have not yet broken the soil surface. Black and Wilkinson (1963) appear to be the only workers to have experimentally distinguished between time of seedling emergence and preemergence relative growth rate. Pregerminated subterranean clover seeds were sown either synchronously or in plots containing a mixture of

10

L. R . BENJAMIN

two sowing dates. In mixed sowing date plots, the plants were sown on a square grid with the early and late sowings alternating in a checkerboard design. The difference in sowing time was either 2, 4, or 8 days and the sowing positions were 1.5 cm apart. Despite sowing seeds carefully at uniform depth in boxes of compost, there was a spread in emergence time of 8 to 14 days for the early-sown plants and 5 to 10 days for the late-sown plants. Thus, there was a wide range of preemergence growth rates, which resulted in a wide overlap in the seedling emergence times from the different sowings. When the plants had a foliage height of 30 cm, the logarithm of dry weight was negatively correlated with time of seedling emergence. The novel point about this work was that the effect of time of sowing on the regression of weight per plant at harvest on seedling emergence time was examined and was found to have only a slight effect. Thus, time of seedling emergence, and not any correlation with preemergence growth rate, was responsible for the subsequent effects on plant weight. In the following sections the effect of time of seedling emergence on the growth, form, and composition of individuals in populations is examined.

A. TOTALPLANT GROWTH

At the point of seedlingemergence, plant weight is still minute compared with the potential and probable weight that the plant will attain. Plant growth at this time is nearly always exponential because there is virtually no self-shading or competition for growth resources from neighbors. Consequently, a difference of a few days in seedling emergence time can result in a manyfold difference in weight between plants. For example, Black and Wilkinson (1963)found that a delay in emergence of 5 days brought about a reduction in subterranean clover weight of about 50%, and a delay of 8 or 9 days produced a reduction in weight of at least 75%. Gray (1976) showed that, in lettuce at 26 days after sowing, seedlings that had emerged at 7 days were five to six times larger than those that had emerged at 15 days. When considering the importance of relative times of seedling emergence for the subsequent dry matter increment of individuals within a population, a number of factors have to be taken into consideration. First, the spread in time of seedling emergence in a population is important, but most studies relied on the natural spread in time of seedling emergence rather than attempting to modify it artificially. The time from seedling emergence to harvest is also relevant. As this time interval increases, there is a greater probability that individual plant growth will be influenced by some extraneous factor, for example, herbivory. Finally, the “state” of the population must be considered, such as the density of plants, or

SEEDLING EMERGENCE WITHIN POPULATIONS

11

whether the plants under consideration are a monoculture or part of a mixed-species population. Table I presents a summary of these factors and the percentage weight variation accounted for by spread in time of seedling emergence in a number of species. The percentage of weight variation associated with spread in time of seedling emergence was usually greatest when fewer than 20 days have elapsed between mean time of seedling emergence and harvest (Table I). (Gray, 1976; Benjamin, 1987). The paper by Benjamin (1982) (Table I) best illustrates how the percentage of weight variation associated with spread in time of seedling emergence increases as spread in time of seedling emergence increases. This trend is also seen in other work (Table I). Benjamin (1982) reduced the spread in time of carrot seedling emergence from 19 to 14 days by using a contact herbicide to kill the first seedlings to emerge. This treatment had no detectable effect on the frequency distribution of storage roots in different diameter grades, suggesting that relatively slight changes in spread of seedling emergence time have no detectable influence on variation in weight. Controlling time of seedling emergence might allow other sources of weight variability to be expressed more strongly. However, the residual sum of squares of root weight (after taking out the effect of emergence time) from plots that had small spreads in emergence times was no greater than the residual sum of squares from plots that had large spreads in emergence times (Benjamin, 1982). Therefore, there is no evidence that other sources of variation are able to express themselves to a greater extent when time of seedling emergence is controlled. The effect of density on the relationship between spread in time of seedling emergence and plant weight is difficult to interpret from Table I because the other factors were varied too. The very high percentage variation in weight associated with spread in time of seedling emergence reported by Ross and Harper (1972) is perhaps because they used extremely high densities (Table I). Benjamin (1982) showed that the increase in coefficient of variation (CV) of carrot shoot and storage root weight at wider spreads in seedling emergence time was greater at 400 than at 25 plants mP2. Ross and Harper (1972) noted that the last cocksfoot seedlings to emerge in a population growing at 30,000 plants m-’ had a weight only a little greater than that of seeds, even after 35 days of growth. The implication is that late-emerging seedlings are denied resources for growth by the earlieremerging plants and these competitive interactions are more intense at high densities. More direct evidence that relative time of seedling emergence determines the relative amount of growth resources that can be captured by an individual comes from a second experiment by Ross and Harper (1972) in

Table I Percentage Shoot Weight Variation Associated with Spread in Time of seedling Emergence

Species e

N

Dactylis glomerata Lactuca sativa

Spread in time of seedling emergence 10

Density (m-*)

Mean days from emergence to harvest

Accounted % variation

in weight

Notes

Reference

30,000

40

95

Ross and Harper

13 13 13 13 13 13 13

17 63 69 17

81 63 49 94

Gray (1976)

(1972) 9 9 9 11 11 6 5-10

Lolium perenne

-a

Impatiens capensis

23

24.48, or%

seedlings in 34 x 12cm tray 500- 1000

20

90

22 17

72 80

180

50

ca. 105

<1

Naylor (1980)

Fecundity measured, plants grouped into six emergence classes

Howell (1981)

Daucus carota

Lactuca sativa Daucus carota

Allium porrurn

Ludwigia leptocarpa Raphanus raphanistrum Beta vulgaris

a

Not stated.

25 and 400 25 and 400 25 and 400 9

ca. 120 ca. 120 ca. 120 ca. 33 ca. 117 ca. 27 ca. 75 ca. 140 ca. 259 110 ca. 50

3 4 72 61 25 11 22 ca. 0

21

ca.250 ca.250 ca.250 ca.250 18 18 1/23-cm pot 10 seeds in 10 x 10-cm tray 16

ca. 56

ca. 0

20 20 16 16 3

92 46 108 54 1/18-cm pot

11 93 8 80 77

44 39 83 58 ca. 0

3-5 6-1 1 15-21 ca. 10 19 19 43 43 11 11 18 3

5-10 7-23 16-48 0-12

Root weight measured Root weight measured Root weight measured Individual plants harvested from within crop when mature Root weight measured Root weight measured Root weight measured Root weight measured

Benjamin (1982) Wurr and Fellows (1983) Benjamin (1984a)

Benjamin (1984b) Dolan (1984) Fecundity measured, plants grouped into two emergence classes

Stanton (1985) Benjamin (1987)

14

L. R. BENJAMIN

which the growth of seedlings that had emerged at different times with and without the presence of neighbors was compared. Where neighbors were not present, then the difference in weights between plants that had emerged at different times was explained by the differences in durations of growth. However, where neighbors were present, then the late-emerging seedlings were smaller than expected from just their reduced period from emergence to harvest. Benjamin (1984~)grew carrots in either pure or mixed sowing date combinations at 400 plants m-*. The sowing dates were separated by a fortnight, thus ensuring different times of seedling emergence. At 35 days after the first sowing, the late-sown plants in mixed sowing date treatments had ceased growing, whereas there was no evidence that the late-sown plants in pure stand had ceased growth at the end of the experiment at 56 days. This experiment provides some of the strongest evidence that the rank each individual occupies in the hierarchy of sizes maintained by competition is largely determined by its time of seedling emergence relative to that of its neighbors. VanBaalen et al. (1984) set up a replacement series experiment with Scrophularia nodosa and Digitalis purpurea, species that have similar seed weights, but the former had a higher relative growth rate when grown in monoculture. The replacement series was repeated with the D. purpurea sown either at the same time as, 7 days earlier, or 7 days later than S . nodosa. They concluded that a delay of 1 day in the emergence of D. purpurea in mixed stand resulted in a 28% reduction in yield per plant, whereas a delay of 1 day in the emergence of S . nodosa gave a reduction in yield per plant of 5%. Therefore, the importance of spread in time of seedling emergence depends on the seedling relative growth rate, and perhaps the effect of given delay in the time of emergence on the growth of an individual should be “scaled” by the mean relative growth rate. The relative times of seedling emergence have been considered to be of importance in studies of competition between species. Nieto et al. (1968) first introduced the idea that there were critical periods in the development of crops during which the presence of weeds will suppress the growth of the crop, these critical periods being defined in terms of time after crop seedling emergence (Hewson and Roberts, 1971; Spitters and vandenBergh, 1982). Elberse and deKruyf (1979) set up replacement series experiments using Hordeum vulgare and Chenopodium album at a range of densities and with sowing the Chenopodium at various times before the Hordeum. They concluded that a difference of 15 days was the critical period; Hordeum outcompeted Chenopodium if Chenopodium was sown less than 15 days before Hordeum, but Chenopodium outcompeted

SEEDLING EMERGENCE WITHIN POPULATIONS

15

Hordeurn if sown move than 15 days before Hordeurn. The relative densities of the two species was of relatively minor importance to the outcome of these competitive interactions. More recently, Cousens et al. (1987) have developed a model to describe the relationship between crop yield and the combined effect of weed density and its time of seedling emergence relative to that of the crop. Some of the studies listed in Table 1 claimed to show that little of the variation in plant weight was associated with time of seedling emergence. How could this occur if the general tenet of this review is correct? Sometimes the lack of an association between plant weight and time of seedling emergence was because of a very low spread in time of seedling emergence (Dolan, 1984; Benjamin, 1987). The spread in time of seedling emergence in carrots in the study by Benjamin (1984a) was 19 days, yet only 3% weight variation at 33 days was associated with time of seedling emergence. However, in this study the vast majority of seedings (82%)emerged over only a 5-day period. The studies of Howell (1981) and Stanton (1985) were unable to show that spread in time of seedling emergence was important perhaps because they used fecundity rather than plant weight as a measure of plant performance, and they also classified seeding emergence date into a restricted number of categories. Sangakkara and Roberts (1986), in a competition study between grass species, argued that they had eliminated differences in time of seedling emergence between ryegrass, prairie grass, and cocksfoot by using transplants. They concluded that “competitive relationships seem to be determined by initial seedling size and the capacity to accumulate greater quantities of dry matter” and that “early emergence had no apparent effect in determining the competitive hierarchy established between these species.” However, they neglected the possibility that the different initial seedling sizes at the time of transplanting were a direct consequence of differences in time of seedling emergence. The bulk of the literature indicates that spread in time of seedling emergence is a major source variation in mature plant weight, probably because differences in emergence time have a large effect on seedling size at the point in time when plants start to compete with one another for growth resources. Time of seedling emergence has its largest effect when spread is large, when seedlings have a high relative growth rate (warm, moist conditions and fast-growing species), when between-plant competition is intense (at high plant densities), and when harvests are made not long after seedling establishment. In the following sections the effect that seedling emergence time might have on the morphology, physiology, and biochemistry of plants will be examined.

16

L. R. BENJAMIN

B. PARTITIONING BETWEEN ORGANS The previous sections have shown that the late-emerging seedlings become the smallest in the population and the ones most likely to be shaded by their neighbors. The position occupied by the foliage influences the temperature, humidity, and quantity and quality of light that shoots experience, which can have a large effect on the partitioning between organs. For example, Acock et al. (1979) showed that with increasing radiation or decreasing temperature, a greater percentage of dry matter is partitioned to the stem at the expense of the roots in chrysanthemums. This study shows that it is difficult to determine a general effect of late emergence on partitioning of dry matter because late-emergers will tend to have foliage both in the shade and at a lower temperature than early-emergers. Studies that examine the competition between species are relevant because between-species differences in plant stature might be analogous to the differences in stature within species caused by differences in time of seedling emergence. These studies also show conflicting effects of stature on partitioning of dry matter between organs. For example, Digitalis purpurea (a rosette-forming plant) grew more slowly and had a lower root : shoot ratio when grown in association with Scrophularia nodosa (a stem-forming plant) than when grown in monoculture (VanBaalen et al., 1984). In contrast, Hawthorn and Cavers (1982) reported that groups of Plantago major and P. rugelii plants, suppressed by the presence of larger neighbors, allocated a greater percentage of dry matter to the roots and caudices, whereas the larger plants devoted a greater proportion of dry matter to reproductive growth. Barnes (1979) presented extensive data from field experiments with carrots to show that at any given time there was a linear relationship between the logarithm of shoot weight and the logarithm of storage root weight, but the intercept of this relationship decreased with time. The variation in plant weight at any one time was largely due to the different imposed density treatments. The implication is that at any one instant, the relative growth rate of the shoots is a fixed proportion of that of the storage roots, but this proportion changes with time. Barnes reasoned that the shift in the relationship between shoot and storage root relative growth might be due to the lower maintenance respiration rate in storage organs. The expectation would be that plants derived from seedlings that had emerged at different times would exhibit different partitioning of photosynthates between organs because the plants were effectively of different ages. In a purely theoretical paper, Moms and Myerscough (1987) considered the effect of partitioning of dry matter between shoots and roots on the competition between plants in even-aged monocultures. They concluded

SEEDLING EMERGENCE WITHIN POPULATIONS

17

that the yield-density relationships could be influenced by shoot-root partitioning of dry matter, which implies that there could be complex feedback processes occurring in crops in which differences in time of emergence lead to differences in amounts of resources available to each plant, which in turn can lead to differences in shoot-root partitioning, which in turn leads to differences in the ability of plants to compete with their neighbors. Until data are presented to support this view, then the more simple assumption should hold, that is, suppressed plants adapt to allow them to scavenge resources not taken up by dominant plants (Robinson, 1986), but this adaptation does not deny resources to the dominant plants. Thus, there appear to be no studies that have examined the partitioning of dry matter between foliage and fibrous roots separately in plants derived from early- and late-emerging seedlings, but the effects of different emergence times on shoot-root partitioning might depend on the light intensity falling on the crop and the mineral status of its soil. C. ORGANMORPHOLOGYA N D COMPOSITION Iwaki (1959) made a large number of interesting observations on the effect of mixed or pure stands and the relative times of sowing on the canopy structure and weight of buckwheat (Fugopyrum esculentum) and green grams (Phaseolus uiridissirnus). In pure stands, buckwheat increases in weight slightly faster, and grows much taller, than green grams. In mixed stands there is a severe suppression of the growth of the green grams. This suppression is lessened by sowing the green grams earlier than the buckwheat, but a difference of 13 days was not sufficient to overcome the dominance of the buckwheat. An interesting feature of this study was that the canopy was harvested separately in 10-cm-deep strata, so that the effects of mixed and pure stand on canopy architecture could be examined. I have estimated the percentage of the photosynthetic system in each 10-cm band of the canopy from his Fig. 1 (Table 11). The plants growing in pure stand deployed over half their photosynthetic dry matter in the top 10-cm layer of the canopy, whereas the plants growing with buckwheat (which grew to a height of 120 cm) deployed their photosynthetic dry matter more uniformly along the length of the plant. However, these suppressed plants were nearly the same height as those in pure stand, but their total weight was less than half that of the pure stand plants (Table 11). Presumably similar morphological responses would occur where the cause of the dominance was a difference in time of seedling emergence. Perry (1985) reviewed the literature on competition in forest stands and

L. R. BENJAMIN

18

Table 11 Effect of Stand Type on the PhotosyntheticDry Matter (Percentage of the Entire Plant) at Various Heights above the Ground in a Crop of Green Grams” Stand type Height (cm) 0-10 10-20 20-30 30-40 40-50 Total (9) a

Sole

with Buckwheat

2

13

5 16

25 52 70

13

38 38 20

From Iwaki (1959).

included the morphological differences between dominant and suppressed trees. He concluded that shade leaves are larger and have more chlorophyll per unit area, and that sometimes the small trees have greater height and radial growth rates than their larger neighbors because of a lower respiration rate and higher specific leaf area. Dominant trees have larger and broader crowns, which allows them to maintain high rates of photosynthesis per tree, but at the expense of high maintenance respiration rate. Perry (1985) also pointed out that suppression of plant growth can influence the chemical composition of trees, with “reduced allocation of photosynthates to those chemical constituents that are energy-expensive.’’ To give an idea of the effects he quoted Chung and Barnes (1977), who determined the number of grams of glucose required to synthesize 1 g of the following substances in Pinus taedu: lipids, 3.02 g; phenolics, 1.92 g; lignin, 1.90 g; nitrogenous compounds, 1.58g; organic acids, 1.43 g; carbohydrates, 1.13 g. The more energy demanding compounds could be involved in deterring herbivory. Where insects and molluscs are the herbivores then the smallest plants in the community are most prone to predation (Perry, 1985; Gange et af., 1989; Thomas and Weiner, 1989), but where there are grazing mammals then the largest plants can be preferentially eaten (Weiner, 1988). Thus, the relative times of seedling emergence are likely to have a large effect on the shape and composition of plant organs. This is likely to be important in terms of the adaptation of plants to competition for growth resources. However, there are also commercial implications because there

SEEDLING EMERGENCE WITHIN POPULATIONS

19

is an increasing emphasis on the quality aspects of harvested produce, which includes its shape, size, and composition. For example, Wurr and Fellows (1983) showed that in studies with two cultivars of lettuce sown at three occasions, the heads of plants derived from late-emerging seedlings matured later than those derived from early-emerging seedlings. The results of these studies that follow the development of individual plants contrast strongly with those of Symonides (1978), who observed the year-to-year difference in the rate of development of five species contrasting in their rate and duration of development. Symonides (1978) found that the rate of development in all five species was faster in those years in which the date time of seedling emergence had been delayed because of climatic conditions. In this latter study, the development of the entire population rather than that of individual plants had been observed. Clearly, within populations, the tendency for accelerated development with late seedling emergence is overridden by the effects of competition with larger plants.

D. LONGEVITY There have been a number of studies that have demonstrated that the plants derived from late-emerging seedlings are more likely to die. For example, Stanton (1985) split a wild radish population into two emergence time cohorts. The early-emerging cohort consisted of larger plants and the probability of mortality before flowering was 0.21, but the corresponding value was 0.55 for the late cohort. vanderToorn and Pons (1988) studied the emergence patterns and survival of two Plantago species in plots in which other species (Poa triuialis, Phleum pratense, Alopecurus prarensis, Trifolium repens, and Ranunculus repens) had either been removed, clipped to a foliage height of 5 cm or 15 cm, or left unclipped. They reported a spread in seedling emergence times of 10 and 12 weeks for the two species. Seedlings that emerged earlier had higher survival rates, and there were greater survival rates in more open environments, with survival being positively correlated with the transmission of light in the different clipping treatments. Shaw and Antonovics (1987) made a population census of Slauia lyrata plants growing for nearly 3 years in pots. They reported that the plants that founded the population had the greatest chance of survival and produced the greatest number of seeds. Westoby and Howell (1986) approached the relationship between emergence time and mortality in a different way. They investigated whether the population structure of a community influenced the course of densitydependent self-thinning in radish. This course was described by graphs of logarithm of biomass per unit area plotted against logarithm of numbers of

20

L. R. BENJAMIN

surviving plants. When plants are self-thinning as a result of between-plant competition, the path of points on the graphs runs from bottom right (high density, low biomass) to top left (low density, high biomass). The slope of this path is claimed to be -4 (see White, 1980; Antonovics and Levin, 1980; Westoby, 1984 for reviews of self-thinning). Different population structures were achieved by making either a single sowing or two sowings separated by between 4 and 23 days in a pot. They concluded that the course of self-thinning in those pots that had a staggered sowing time was not consistently different from the course in the pots that had a single sowing. The implication is that plants derived from late-emerging seedlings are no more likely to die from competition with their neighbors than those derived from early-emerging seedlings. However, the relationship between mean weight of survivors and plant density varied greatly between the single sowing date pots in different experiments. This implies that this relationship is not sufficiently consistent to be used to distinguish between large differences in the probability of mortality of plants derived from early- and late-emerging seedlings. Indeed, Weller (1987) has now shown that many of the sets of data that were thought to support the -4 power law in fact do not have statistically significant correlations between logarithm of weight and logarithm of density.

IV. CONCLUSIONS Clearly, there can be a wide range of seedling emergence times, even in synchronously sown commercial crops. The causes of different times of seedling emergence are numerous, but the limited evidence suggests that in monocultures these preemergence processes are important only because they influence time of seedling emergence. This is an important hypothesis that deserves more attention and further experimental verification. If true, it replaces the plethora of possible interactions between individuals in monocultures by a simple measurable statistic, the relative times of seedling emergence. This statistic can be used as a predictor of future plant growth. If time of seedling emergence is to be used as a predictor of plant performance, there is a need for more consistent intervals of time into which emergence is classified. Where spread of emergence time is very wide, then seedlings that have emerged over several days often have been grouped, but more usually emergence is recorded daily. Consequently, comparisons between experiments have been made difficult because the time intervals into which seedlings were grouped have been inconsistent.

SEEDLING EMERGENCE WITHIN POPULATIONS

21

Strictly speaking, the relative weights of plants when establishment is complete are probably better predictors of future plant performance (Sangakkara and Roberts, 1986), but differences in weight between plants that emerge on the same day are likely to be small compared with the differences in weight between plants emerging on different days. An implicit assumption of this hypothesis is that the growth of plants after seedling emergence is largely under conditions in which individuals are competing with one another for growth resources, and the outcome of this contest for resources is determined by their relative weights. It is tempting to consider that this hypothesis can be extended to mixedspecies stands, but the work of Iwaki (1959) on competition between buckwheat and green grams shows that the faster rates of shoot elongation of buckwheat allowed this species to outcompete green grams even when the green grams had emerged several days prior to the buckwheat. Recent work by Ballare et ul. (1987) shows that the red:far red light ratio perceived by a seedling is modified by the presence of neighbors long before there is any between-seedling shading. This modification in the light regime was perhaps responsible for longer internode growth and lower leaf: stem dry weight ratio is Sinupsis seedlings grown on the sunlit side of a grass “fence” compared with seedlings grown on the sunlit side of a bleached grass “fence.” This implies that seedlings can detect and respond to the presence of neighbors before there is competition for resources. Thus, seedlings might modify their growth to allow them to partially overcome the disadvantage of late emergence. Relative time of seedling emergence is an important criterion for the future growth of plants, and its importance to plants has perhaps not been fully recognized. Further work is required to determine whether this importance is ubiquitous, generally true of all plant communities, true just of monocultures, or present just in special experimental conditions. Most evidence to date implies an importance in most monocultures and probably in many mixed-species stands. ACKNOWLEDGMENTS

I thank colleagues at IHR Wellesbourne for their criticisms of this manuscript.

REFERENCES Abul-Fatih, H. A., and Bazzaz, F. A. 1979. New Phytol. 83,817-827. Acock, B., Charles-EdwardsD. and Sawyer, S. 1979. Ann. Bot. 44,289-300. Anonymous. 1982. ‘‘Report on the Foot Industry and Technology, Cabinet Office Advisory Council for Applied Research and Development.” HMSO, London.

22

L. R. BENJAMIN

Antonovics, J., and Levin, D. A. 1980. Annu. Rev. Ecol. Syst. 11,411-452. Arndt, W., 1965. Ausr. J . Soil. Res. 3,55-68. Arnott, R. A. 1969. J . Er. Grussl. SOC.24, 104-110. Ballare, C. L., Sanchez, R. A,, Scopel, A. L., Casal, J. J. and Ghersa, C. M. 1987. Plunr Cell Environ. 10,551-557. Barnes, A., 1979. Ann. Eor. 43,487-499. Bedford, L. V., and MacKay, D. B. 1973. J . Nurl. Insr. Agric. Eor. (U.K.) 13,50-62. Benjamin, L. R. 1982. J . Agric. Sci. 98,537-545. Benjamin, L. R. 1984a. J . Agric. Sci. 102,69-77. Benjamin, L. R. 1984b. J . Agric. Sci. 103,527-537. Benjamin, L. R. 1984~.Ann. Eor. 53,549-557. Benjamin, L. R. 1987. J . Agric. Sci. 108,221-230. Benjamin, L. R.,and Hardwick, R. C. 1986. Ann. Eof.58,757-778. Bierhuizen, J. F., and Wagenvoort, W. A. 1974. Sci. Horfic. (Amsterdam) 2,213-219. Black, J. N. 1956. Ausr. J . Agric. Res. 7,98-109. Black, J. N., and Wilkinson, G. N. 1963. Ausf. J . Agric. Res. 14,628-638. Bould, A., and Abrol, B. K. 1981. Seed Sci. Technol. 9,601-61 I . Bradford, K. J. 1986. HortScience 21, 1105-1 112. Brocklehurst, P. A., and Dearman, J. 1983. Seed Sci. Technol. 11,293-299. Buckley, R. C. 1982. Eiorropicu 14,314-315. Chung, H., and Barnes, R. L. 1977. Can. J . For. Res. 7, 106-1 11. Clarke, C., and Moore, K. G. 1986. Ann. Eor. 58,363-369. Collis-George, N., and Hector, J. B. 1966. Ausr. J . Soil Res. 4, 145-164. Collis-George, N., and Sands, J. E. 1959. Aust. J . Agric. Res. 10,628-637. Collis-George, N., and Sands, J. E. 1962. Ausr. J . Agric. Res. W, 575-585. Collis-George, N., and Williams, J. 1968. Ausr. J . SoilRes. 6, 179-192. Cnstigan, P. A. 1984. Plunr Soil 79, 19!-201. Cousens, R., Brain, P., O’Donovan J. T., and O’Sullivan, P. A. 1987. Weed Sci. 35,720-725. Dasberg, S, and Mendel, K. 1971. J . Exp. Eor. 22,992-998. Dolan, R.W. 1984. Am. J . Eor. 71, 1302-1307. Eagles, H. A. 1988. Theor. Appl. Genet. 76,459-464. Elberse, W. T., and deKkruyf, H. N. 1979. Nerh. J . Agric. Sci. 27, 13-26. El-Sharkawi, H. M., and Springuel, I. V. 1979. Seed Sci. Technol. 7,27-37. Finch-Savage, W. E. 1984. J . Horfic. Sci.59,403-410. Finch-Savage, W. E. 1986. J . Agric. Sci. 107,249-256. Gange, A. C., Brown, V. K., Evans, I. M., and Storr, A. L. 1989. J . Ecol. 77,537-551. Garcia-Huidobro, J., Monteith, J. L., and Squire, G. R. 1982a. J . Exp. Eor. 33,288-296. Garcia-Huidobro, J., Monteith, J. L., and Squire, G. R. 1982b. J . Exp. Eor. 33,297-302. Gerard, C. J. 1%5. Soil Sci. SOC.A m . Proc. 29,641-645. Gerard, C . J. 1980. Agron. J . 72,473-476. Gray, D. 1976. Ann. Appl. Eiol. 82,569-575. Gray, D. 1979. Seed Sci. Technol. 7, 169-178. Gurnmerson, R. J. 1989. J. Agric. Sci. 112, 159-169. Hadas, A., and Russo, D. 1974a. Agron. J . 66,643-647. Hadas. A., and Russo, D. 1974b. Agron. J . 66,647-652. Hammerton, J. L. 1961. J . Agric. Sci.56,213-228. Hanks, R.J., andThorp, F. C. 1956. SoilSci. SOC.A m . Proc. 20,307-310. Hanks, R.J., andThorp, F. C. 1957. SoilSci. Soc. Am. Proc. 21,357-359. Harper, J. L. 1955. Ann. Appl. Eiol. 43,696-708.

SEEDLING EMERGENCE WITHIN POPULATIONS

23

Harper, J. L. 1977. “Population Biology of Plants,” pp. 111-147. Academic Press, London. Harper, J. L., and Benton, R. A. 1966. J . Ecol. 54, 151-166. Harper, J. L . , Williams, J. T., and Sagar, G. R. 1965. J . Ecol. 53,273-286. Hawthorn, W. R., and Cavers, P. B. 1982. Can. J. Bot. 60,2424-2439. Hegarty, T. W. 1973. I n “Seed Ecology” (W. Heydecker, ed.), pp. 411-432. Butterworth, London. Hegarty, T. W. 1974. J . Hortic. Sci. 49, 189-1%. Hegarty, T. W. 1976a. J . Sci. Food Agric. 27, %2-968. Hegarty, T. W. 1976b. J . Hortic. Sci. 51, 133-146. Hegarty, T. W. 1978. Acta Hortic. 83,297-308. Hegarty, T. W., and Royle, S. M. 1978. J. Appl. Ecol. 15,897-904. Henriksen, K . 1978. Acta Hortic. 72, 69-78. Hewson, R. T., and Roberts H. A. 1971. J. Hortic. Sci. 46,471-483. Heydecker, W. 1956. J . Hortic. Sci. 31,76-88. Heydecker, W. 1958. Nature (London) 181, 1140-1 141. Heydecker, W. 1%1. I n “Advances in Horticultural Science and Their Applications, Proceedings of the XVth International Horticultural Congress, Nice, 1958” (J.-C. Garnaud, ed.), Vol. 1, pp. 381-392. Pergamon, Oxford. Heydecker, W. 1973. Univ. Nottingham Sch. Agric. Rep. 1973/1974 50-67. Heydecker, W., and Coolbear, P. 1977. Seed Sci. Technol. 5,353-425. Heydecker, W., Orphanos, P. I., and Chetram, R. S. 1969. Proc. I n t . Seed Test. Assoc. 34, 297-304. Howell, N. 1981. A m . Mdl. Nat. 105, 312-320. Iwaki, H. 1959. Jpn. J . Bot. 17, 120-138. Johnston, M. E. H. 1979. I n “Advances in Research and Technology of Seeds” (J. R. Thompson, ed.), Part 4, pp. 43-83. PUDOC, Wageningen, Netherlands. Khah, E. M., Ellis, R. H., and Roberts, E. H. 1986. J . Agric. Sci. 107,431-438. Koller, D. 1972. I n “Seed Biology” (T. T. Kozlowski, ed.), pp. 1-101. Academic Press, New York. Lafond, G. P., and Baker, R. J. 1986. Crop Sci. 26,341-346. Langan, T. D., Pendleton, J. W., and Oplinger, E. S. 1986. Agron. J . 78,769-772. Lovato, A. 1981. I n “Advances in Research and Technology of Seeds” ( J . R. Thompson, ed.), Part 6, pp. 86-120. PUDOC, Wageningen, Netherlands. Maguire, J. D. 1984. I n “Advances in Research and Technology of Seeds” (J. R. Thompson, ed.), Part 9, pp. 25-60. PUDOC, Wageningen, Netherlands. Manohar, M. S., and Heydecker, W. 1964. Nature (London) u12,22-24. Mithen, R., Harper, J. L., and Weiner, J. 1984. Oecologia 62,57-60. Moore, R. P. 1943. J. A m . SOC.Agron. 35,370-381. Moms, P. C., Grierson, D., and Whittington, W. J. 1984. J . Exp. Bot. 35, 994-1002. Moms, E. C., and Myerscough, P. J. 1987. Ann. Bot. 59,629-633. Muendel, H.-H. 1986. Agron. J. 78,765-769. Naylor, R. E. L. 1980. New Phytol. 84,313-318. Nieto, J. H., Brondon, M. A., and Gonzalez, J. T. 1968. PANS C14, 159-166. Nuttall, W. F. 1982. Agron J . 74, 1018-1022. Orchard, T. J., and Wurr, D. C. E. 1977. Sci. Hortic. 6, 179-184. Pacala, S . W., and Slander, J. A. 1985. A m . Nat. 125,385-41 1. Page, E. R., and Cleaver, T. J. 1983. J. Sci. Food Agric. 34, 13-22. Perry, D. A. 1982. Sci. Hortic (Canterbury, Engl.) 33, 67-75.

24

L. R. BENJAMIN

Perry D. A. 1985. In “Attributes of Trees as Crop Plants” (M. G. R. Cannell and J. E. Jackson, eds.), pp. 481-506. Institute of Terrestrial Ecology, Huntingdon, England. Pill, W. G., and Finch-Savage, W. E. 1988. Ann. Appl. Biol. 113,383-389. Pollock, B. M. 1972. In “Viability of Seeds” (E. H. Roberts, ed.), pp. 164-165. Chapman & Hall, London. Rathore, T. R., Ghildyal. B. P., and Sachan, R. S. 1981. PI. Soil62,97-105. Roberts, E. H. 1972. In “Viability ofSeeds” (E. H. Roberts, ed.), pp. 321-359. Chapman and Mull, London. Roberts, H. A. 1979. J. Appl. Ecol. 16, 195-201. Roberts, H. A. 1984. Ann. Appl. Biol. 105,263-275. Robinson, D. 1986. Ann. Bot. 58,841-848. Ross, H. A., and Hegarty, T. W. 1979. Ann. Boi. 43,241-243. Ross, M. A., and Harper, J. L. 1972. J . Ecol. 60,77-88. Royle, S . M., and Hegarty, T. W. 1977. J . Hortic. Sci. 52,535-543. Sangakkara, U. R., and Roberts, E. 1986. J . Agron. Crop Sci. 156,279-284. Schmitt, J., Eccleston, J., and Ehrhardt, D. W. 1987. J. Ecol. 75,651-665. Sedgley, R. H. 1%3. Ausi. J. Agric. Res. 14,646-653. Sexton, P. D., and Gerard, C. J. 1982. Agron. J. 74,699-702. Shaw, R. G . , and Antonovics, J. 1987. New Phyiol. 107,415-426. Snyder, F. W., and Filban, C. 1970. J . A m . SOC.Sugar Beet Technol. 15,703-708. Spitters, C. J. T., and vandenBergh, J. P. 1982. In “Biology and Ecology of Weeds’’ (W. Holzner and N. Numata, eds.), pp. 137-148. Junk, The Hague. Stanton, M. L. 1984. Ecology 65, 1105-1 112. Stanton, M. L. 1985. Oecologia 67,524-531. Stickler, F. C., and Wassom, C. E. 1%3. Agron. J . 55,78. Symonides, E. 1978. Ekol. Pol. 26,273-286. Taylor, A. G., and Broeck, C. W. T. 1988. HoriScience 23,367-369. Tenhovuori, M. 1986. J. Agric. Sci. Finl. 58, 185-192. Thomas, H. 1972. In “Viability of Seeds” (E. H. Roberts, ed.), pp. 370-372. Chapman & Hall, London. Thomas, S. C., and Weiner, J. 1989. J . Ecol. 77,524-536. Thompson, P. A. 1974. J. Exp. Boi. 25, 156-163. Thornley, J. H. M. 1977. Ann. Boi. 41, 1363-1365. Tipton, J. L. 1984. J . A m . SOC.Horiic. Sci. 109,451-454. Tipton, J. L. 1988. J . A m . SOC. Hortic. Sci. 1l3, 129-133. VanBaalen, J., Kuiters, A. T. and vanderwoude, C. S. C. 1984. Oecol. Plant. 5,279-290. vanderToorn, J., and Pons, T. L. 1988. Oecologia 76,341-347. Villiers, T. A. 1972. In “Seed Biology” (T. T. Kozlowski, ed.), pp. 220-281. Academic Press, New York. Wagenvoort, W. A., and Bierhuizen, J. F. 1977. Sci. Horiic. 6,259-270. Waller, D. M. 1985. New Phytol. 100,243-260. Watkinson, A. R. 1985. In “Studies on Plant Demography: A Festschrift for John L. Harper” (J. White, ed.), pp. 275-289. Academic Press, London. Watkinson, A. R., Lonsdale, W. M., and Andrew, M.H. 1989. J. Ecol. 77, 162-181. Weiner, J. 1988. In “Plant Population Biology” (A. J. Davy, R. J. Hutchings, and A. R. Watkinson, eds.), pp. 59-81. Blackwell, Oxford. Weiner, J., and Thomas, S. C. 1986. Oikos 47,211-222. Weller, D. E. 1987. Ecol. Monogr. 57,23-43. Westoby, M. 1984. Adv. Ecol. Res. 14, 167-225. Westoby, M., and Howell, J. 1986. J . Ecol. 74,343-359.

SEEDLING EMERGENCE WITHIN POPULATIONS

25

White, J. 1980. In “Demography and Evolution in Plant Populations” (0.T. Solbrig,ed.), pp. 21-48. Blackwell, Oxford. Whittington, W. J. 1978. Acra Horric. 72, 39-47. Willat, S. T., and Struss, R. G. 1979. Ann. Bor. 43,415-422. Williams, J., and Shaykewich, C. F. 1971. J . Exp. Bor. 22,586-597. Wurr, D . C. E., and Fellows, J. R . 1983. J . Horric. Sci. 58, 561-566. Wurr, D. C. E., Fellows, J. R., and Gray, D. 1982. J . Agric. Sci. 99, 123-129.

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY, VOL. 44

FORAGE TREE LEGUMES: THEIR MANAGEMENT AND CONTRIBUTION TO THE NITROGEN ECONOMY OF WET AND HUMID TROPICAL ENVIRONMENTS Graeme Blair, David Catchpoole,’ and Peter Horne Australian Centre for International Agricultural Research Department of Agronomy and Soil Science University of New England Armidale, New South Wales 2351, Australia

1. 11. 111. IV. V.

Introduction Species of Useful Tree Legumes Agronomic Performance of Tree Legumes Tree Legume Leaf as Animal Feed Management of Tree Legumes A. Age of First Cutting B. Cutting Height C. Cutting Frequency D. Density E. Interaction between Density and Cutting Management VI. Nitrogen Yields of Material Harvested from Tree Legumes VII. Nitrogen Recycling via Leaf and Excreta VIII. Conclusion References

I. INTRODUCTION Although herbaceous legumes usually form the basis of temperate pasture, tropical herbaceous legumes are renowned for their inability to persist under heavy grazing; in this context, tree and shrub legumes have been

’

Present address: Queensland Department of Primary Industries, Ayr. Queensland 4807, Australia. 21 Copyright 0 1990 by Academic Press. Inc. All rights of reproduction in any form reserved.

28

GRAEME BLAIR ET AL.

increasingly proposed as a more suitable alternative (Anning, 1980; Jones et al., 1982; Jones and Jones, 1982; Brewbaker, 1986). In many areas of the wet and humid tropics, ruminant animals are not permitted to graze freely, but instead are tethered or kept in pens and cages, and feed is brought to the animals. Generally, feed for these animals is cut from areas unsuitable or unavailable for cropping, for example, hillsides or roadside verges and other public “common” land, and as such is often of poor nutritive value. Rarely is there any effort made by farmers to deliberately plant forage species, because of the perception that any improvement in draft power and animal production does not compensate for a reduction in the area available for cropping. In this context, leguminous trees can play a vital role in improving the nutritive value of the animal feed.

II. SPECIES OF USEFUL TREE LEGUMES There has been increasing recognition by scientists within the past few decades of the potential of tree legumes to increase agricultural and silvicultural production (Felker and Bandurski, 1979), especially in Third World countries where nitrogenous fertilizers are often too expensive for purchase by farmers. It is particularly in the poorer countries of the wet and humid tropics that increased adoption of management systems incorporating N-fixing tree legumes can result in improved human nutrition and living standards. Brewbaker and Styles (1982) listed 37 leguminous tree species they considered to be of major importance and 26 of minor importance. Halliday and Nakao (1982) prepared a list of more than lo00 woody species (legume and nonlegume) thought to have the potential for nitrogen fixation. Tropical leguminous trees have been used to various degrees as sources of fodder, fuelwood, or timber for many years in Africa, Asia, Australia, and the Americas. The increasing attention given them is reflected in the number of institutions conducting research on their use in those areas (Van den Beldt and Huxley, 1982). Some species are now widely distributed and used outside their country of origin. Notable examples include Leucaena spp. and Gliricidia sepium, which originated from Central America and are now important components of agricultural systems in many countries of Asia and Africa (Brewbaker and Styles, 1982; Sumberg, 1985). Growth of species introduced into new areas is often better than in the area of origin because of the absence of native pests and diseases (Whiteman, 1980). This quarantine can last for varying periods, but, once an outbreak of disease or insect attack occurs, it can have a serious effect on

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

29

plant performance. This has occurred recently in many parts of the Pacific and Southeast Asia, where the leaf psyllid (Heteropsylfa cubana) has attacked Leucaena leucocephala and caused severe yield reductions. Leucaena leucocephala has received the most attention, probably because of its popularity and multiplicity of uses (forage, human food, green manure, fuelwood, timber) (Brewbaker and Hutton, 1979; CSIRO, 1979; Delpeche et al., 1979; Mendoza and Javier, 1980; Chandrasekaran, 1981; Dutt, 1981; Mendoza et al., 1981; Chee and Devendra, 1983; Pathak and Patil, 1983; Pound and Martinez-Cairo, 1983; Prussner, 1983; Bray, 1984), but there is still relatively little information available on methods of cultivation and harvesting. In an attempt to help researchers become aware of other work currently being undertaken with Leucaena, since 1980 the Nitrogen Fixing Tree Association (NFTA) has published preliminary research results in annual volumes of “Leucuenu Research Reports.” Since 1983, NFTA has also published annually the research communications on other species in “Nitrogen Fixing Tree Research Reports.” The U.S. National Academy of Sciences has published general information on Leucaena (NAS, 1977, 1984), Calliandra (NAS, 1983c),Acacia (NAS, 1983b),and fuelwood trees (NAS, 1980, 1983a). Detailed species descriptions of many tropical forage tree legumes are contained in Skerman (1977). Evans and Rotar (1987a,b) reviewed the role of Sesbania spp. in agricultural systems, particularly in Third World countries. In addition to supplying forage, fuelwood, and timber, sometimes the foliage, flowers, and pods of tree legumes are used as a vegetable (NAS, 1977),and for some species the high concentration of certain chemicals has warranted consideration as sources of pharmaceuticals and other chemicals (NAS, 1977; Brewbaker and Styles, 1982; Romeo, 1984). Leguminous trees are often planted to provide shade for plantation crops. Leucaena is widely used in this role for crops such as coffee, cacao, tea, cinchona, citrus, and nutmeg (NAS, 1977). Sometimes the tree legume also functions as a support for climbing crops, for example, Gliricidia sepiurn has been used as a support tree for cultivation of vanilla (Alconero ef af., 1973). Tree legumes have the greatest potential to increase the quality and level of protein in the diet of people in Third World countries when it is used as a source of feed for animals.

Ill. AGRONOMIC PERFORMANCE OF TREE LEGUMES Leucaena has been shown to be extremely high yielding in a range of locations, with annual forage dry matter yields of 26.8 t ha-’ in Fiji

30

GRAEME BLAIR ET A L .

(Partridge and Ranacou, 1973), 11.5 t ha-' in Japan (Kitamura, 1986), 12.9 t ha-' in Australia (Ferraris, 1979), 14.3 t ha-' in Hawaii (Evensen, 1984), 14.2 t ha-' in Thailand (Topark-Ngarm, 1983), 8.4 t ha-' in the first year of growth in Indonesia (Ella et al., 1989), and 15.6 t ha-' ha-' over a 2.5-year period in Indonesia (Catchpoole and Blair, 1990a). There are very few reports of forage yields from other species. Holm (1972) obtained 7.2 t DM ha-' year-' from Sesbania grandijlora in Thailand. Chadhokar and Lecamwasam (1982) reported Gliricidia leaf fresh weight yield in Sri Lanka in excess of 40 t ha-', with dry matter percentage of leaf ranging between 20% and 23%. Gutteridge and Akkasaeng (1985) compared the growth of 15 tree species during the first 6 months after planting. Sesbania formosa and Sesbania sesban gave the best yields, whereas Sesbania grandijlora was comparable to Leucaena leucocephala and Calliandra calothyrsus. Evans and Rotar (1987b) also obtained comparable fodder yields from various perennial Sesbania species, Calliandra calothyrsus, and Leucaena leucocephala, during the first year after establishment. Little information is available on the best means of establishment and subsequent management practices for most species (Maasdorp and Gutteridge, 1986), although researchers with Leucaena have addressed such topics as rhizobiology (Diatloff, 1973; Jarvis, 1981; Bushby, 1982; Manjunath et al., 1984; Sanginga et al., 1986), response to pH (Ahmad and Ng, 1981; Norani and Ng, 1981),effect of CaC03 (Hutton and Andrew, 1978; Tham et al., 1977),nutrition and symptoms of nutrient deficiency (Gonzales et al., 1980; Haag and Mitidieri, 1980),and methods of establishment (Pound et al., 1980a,b;Pound and Santana, 1980; Falvey, 1981; Cooksley, 1981; 1982; Oakes, 1984; Olvera and Blue, 1985; Olvera and West, 1985). In an evaluation of 16 tree legumes conducted over a 4-year period in South Sumatra, Indonesia, Blair et al. (1988) found that after a 2-year establishment period, leaf yields were highest in Acacia mangium; this declined with time such that the highest total leaf production was recorded in Cassia siamea (Fig. 1). For another species comparison experiment conducted in South Sulawesi, Indonesia, Catchpoole and Blair ( 1990a) recorded leaf yields of 18.2, 19.2, 21.8, and 6.7 t DM ha-' in C. calothyrsus, G. sepium, L . leucocephala, and S . grandi3ora over a 14-month period from establishment. The high yield in Leucaena was recorded prior to the arrival of psyllids in this part of Indonesia. Forage yields per hectare are likely to be maximized if the shrub legumes are in a mixture with an understory of herbaceous pasture species, through increased efficiency of utilization of light and soil moisture (Gray, 1970). In an experiment conducted in Indonesia, Catchpoole and Blair (199Oa)

31

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS 1800r

1600

0 Feb. '82 to Ian. '84

-

a JM.

'84 to Fcb. '85

69 Feb. '85 to Apr. '86

1400 -

e .

1200

-

t . 4 d .

n

1000-

M

Y

U

Ix

800-

L

3

600

-

400

-

200 -

-

FIG. 1. Leaf production (g DM per tree) of the ten most productive species for three growth periods at Nakau, South Sumatra.

Table I

Dry Matter Yields (t ha-') over a 14-Month Period of TreelGrass Mixtures and Monocultures at Gowa, South Sulawesi, Indonesia"

Leaf DM Stem DM Total legume DM Grass DM Edible DM' Total DM

Tree monoculture

Treelgrass mixture

Grass monoculture

18 .74""

16.92" 9.18' 26. 10" 7.82" 24.74" 33.92"

12.00" 12.00" 12.00'

10.26" 29.00" 18.74" 29.00" ~

" From Catchpoole and Blair (]!Ma).

" Means within a row with the same letter are not significantly different ( p < .05). Edible DM = legume leaf DM and grass DM.

32

GRAEME BLAIR ET AL.

found that undersowing leguminous trees (Leucaena, Calliandra, Glirici&a) with Panicum maximum cv. Riverdale increased the dry matter yield of edible forage from 18.7 to 24.7 t ha-' over a 14-month period (Table I). In another Indonesian experiment, Horne and Blair (1990) found that although the quantity of light penetrating to 20 cm above the soil surface was similar in mixtures of Leucaena cut to 100 cm with Setaria sphacelata (HLSET) or Pennisetum purpureum (HLPEN) and Leucaena cut to 30 cm with Setaria (LLSET) or Pennisetum (LLPEN), the shape of the light profile was markedly different (Fig. 2) and the canopy within each stratum had a very different grass-legume composition.

b

a

-8 v

& 8

210

170

Leucaena Leaf

130 90

60

60 40 20 0 Relallve Irradlance (100 I/Io)

100

80

60

1

I

1

I

I

4 0 20 0 Reletlve lrradlanae (100 1/10) 60

80

I

I

1

I

I

I

I

I

I

20

40

60

0

20

40

60

80

?!z5l d

C

210

170 h

8

130

2M

90

.r.

4

t

100

0

210

v

Light proflle

p---o

90

50

GO

10100

80

60

20

40

Relallve lrradiance (100 I

1

I

0 1/10)

I

10100

80

60

20

40

0

Relative lrradiance (100 I/lo) f

I

I

I

20 40 60 0 20 40 60 C Leaf D.M. (kg ha-ld-') C Leaf D.M. (kg lia-'d-') FIG. 2. Mean light profiles of the mixtures and downward cumulative leaf dry matter production. (Z leaf) in the canopies of (a) HLSET, (b) LLSET, (c)HLPEN, and (d) LLPEN. Horizontal bars indicate standard errors of relative irradiance means. 0

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

33

The poor performance of low-cut Leucaena in both mixtures resulted largely from competitive interference for light. Pennisefum,although slow to recover from cutting, was able to produce a tall canopy that severely shaded low-cut Leucaena for most of each growth period. Examination of the mixture and monoculture leaf yields in the experiment of Home and Blair (1990) illustrates the “trade-off” that occurs between the legume and the grass in mixtures. The data presented in Table I1 compare the total leaf yields obtained from the mixtures over the 1-year experimental period with the potential yields obtained when the same land area was subdivided evenly between monocultures of the component species (the 50 : 50 monoculture option). Forage from all four monoculture options consisted of approximately 30% Leucaena leaf, a percentage widely considered to be optimal for ruminant production (Pound and Martinez-Cairo, 1983). Although total forage yields were higher from the low-cut Leucaena mixtures than from their monoculture options, Leucaena leaf made up no more than 12% of the total forage. The reason for this low percentage contribution, compared with the monoculture options, is the unexpectedly good performance of low-cut Leucaena in monoculture and its poor performance in mixtures. There are many reports in the literature that low cutting has an adverse effect on Leucaena leaf and stem yields and these have been summarized by Home e f al. (1986). Although successful incorporation of shrub legumes into an existing herbaceous pasture might prove difficult (Shaw, 1965), herbaceous species could be added to an established stand of shrub legumes by oversowing (Jones et al., 1982). A mixture of shrub legumes and grass also helps to provide an appropriate mixed diet for grazing ruminants (Jones and Bray, 1983) and would help reduce erosion on sloping land. In an experiment of Catchpoole and Blair (1990a), the undersowing of tree legumes with Panicum maximum did not increase the edible N yield of

Table 11 Total Leaf Dry Matter Yields (t ba-’ year-’) from the Mivtures and Potential Yields from the 50 :50 Monoculture Option Mixture

LLPEN HLPEN LLSET HLSET

50 : 50 monoculture

Leucaena

Grass

Total

Leucaena

Grass

Total

0.8 4.2 2.2 8.2

15.2 10.7 16.3 6.2

16.0 14.9 18.5 14.4

4.2 4.3 4.2 4.3

9.3 9.3 10.4 10.4

13.5 13.6 14.6 14.7

34

GRAEME BLAIR ET AL.

the system (652 kg ha-' year-'). Where grass was grown as a monoculture, the edible N yield was only 88 kg ha-'. Another approach is to fence off stands of shrub legumes and to control stocking of the shrub legume pastures and larger areas of herbaceous pasture in a rotation system, to enable optimum use of each pasture resource (Shaw, 1968; Blunt and Jones, 1977; Anonymous, 1981; Paterson et al., 1982; Foster and Blight, 1983). Judicious use of the shrub legume component would enable high-quality feed to be accumulated for use during dry periods (Jones and Bray, 1983). Another grazing system proposed for Leucaena (Wildin, 1980)is to grow the legume as tall trees and to allow grazing of the lower branches and seedlings under the canopy by stock. However, much research is needed to work out the best management system for animals grazing shrub legume pastures, such as size and arrangement of paddocks, time and frequency of grazing, stocking rates, tree spacing, and fertilizer requirements. The ultimate measure of pasture productivity is animal production itself.

IV. TREE LEGUME LEAF AS ANIMAL FEED Many plant species probably form a part of the diet of domestic animals at some time, but Table 111 can be used as a guide to the number of leguminous shrub and tree species used regularly as a source of feed. Often trees and shrubs are only used for feed during drought, or other times when herbaceous material is unavailable (Gray, 1970). A notable exception is Leucaena leucocephala, which can be harvested year-round to provide forage for penned animals in a cut-and-carry feed system (NAS, 1984).

Table 111 Nitrogen-Fixing Tree and Shrub Legumes Used for Forage" Mimosoideae

No. spp.

Papilionoideae

No. spp.

Caesalpinoideae

No. spp.

Acacia Paraserianthes Desmodium Leucaena Pithecellobium Pterocarpus Prosopis

12 1

Cajanus Chamaecytisus Desmanthus Gliricidia Medicago Pangamia Robinia

1 I I I 1

Parkinsonia Calliandra Erythrina Sesbania

1 1 2

I 5 1

2 4

After Brewbaker (1986).

I 1

4

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

35

When animals are maintained in low-quality roughage such as rice straw, the inclusion of high-protein tree legume leaf can have a significant effect on animal performance. An obstacle to the widespread acceptance of tree legumes as forage sources has been low quality or palatability (Jones, 1979; Minson and Wilson, 1980; D'Mello, 1982; Brewbaker, 1986). Chemical analyses and digestibility data of leaf of Leucaena leucocephala, Gliricidia sepium, and Sesbania grand;Jlora have been reported by Holm (l973), Adenaye (1979), Minson and Wilson (1980), and Ekpenyong (1986). Vercoe (1989) has presented data on 39 Australian tree and shrub species. Palatability and acceptance by stock are generally not a problem with Leucaena (Bray, 1984), but problems of stock acceptance of other tree legume species have been reported (Skerman, 1977). Gliricidia has been used as a feed supplement for sheep (Chadhokar and Kantharaju, 1980) and dairy cattle (Chadhokar and Lecamwasam, 1982). Carew (1983) showed that Gfiricidiaas the sole dry season diet for dwarf sheep was able to maintain them for a continuous period of 21 weeks, despite an initial drop in liveweight. There is a paucity of published information on the feed quality of other species, despite their potential as sources of high-protein feed for ruminant animals. Most of the literature reporting liveweight gains of animals grazing shrub legumes documents research predominantly undertaken in Queensland, Australia, with Leucaena. Jones and Bray (1983) have summarized research in Australia on animal production from grazed Leucuena pastures. Results from these experiments were obtained before Jones and Lowry (1984) showed that the mimosine toxicity problem in Australian ruminants could be overcome by infusion of rumen fluid from Indonesian ruminants. Jones and Bray (1983) surmised that elimination of Leucaena toxicity would result in a marked improvement in animal production. Despite the toxicity problem, Falvey (1976) reported that a Leucaena-Cynodon dactylon pasture in the Northern Territory of Australia resulted in higher liveweight gains than were obtained from heifers grazing Townsville stylo-Panicum maximum pastures. Jones and Jones (1982) reported annual liveweight gains of cattle grazing Leucaena-Setaria pastures of 3 1 1 kg ha-' compared t o 200 kg ha-' from Siratro-Setariu pastures. In Malaysia, Wong et af. (1983) reported that liveweight gains of cattle rotationally grazing a Leucaena-Brachiaria decumbens mixture were comparable to liveweight gains obtainable from nitrogen-fertilized pastures. In the Philippines, Moog (1983) reported that leaf production from Leucaena-Zmperata cylindrica pastures is three times that from native Imperata grasslands, and liveweight gains from Leucaena-Panicum maximum pastures have been of the order of 440 kg ha-'. However, nonru-

36

GRAEME BLAIR ET AL.

minants are susceptible to mimosine and this has led to problems when Leucaena makes up too high a proportion of the diet of poultry, rabbits, or humans (NAS, 1977). There appears to be no published information on performance of animals grazing other tree or shrub legume species, although other species are known to have potential. In Australia several species of Acacia, notably A. aneura, are grazed by animals, particularly during times of drought (Gray, 1970). Shrubby species belonging to the genus Desmodium were observed by Bryan (1966) to persist under grazing for 5 years, and thrived on soils unsuitable for Leucaena.

V. MANAGEMENT OF TREE LEGUMES The large number of latent growing points in the stem cambium of tree legumes, which are protected from damage by grazing, enables rapid recovery after being cropped by the grazing (or browsing) animal (Gray, 1970). The more extensive permanent root system of shrub legumes compared to that of herbaceous legumes confers an ability to retain highquality forage under conditions of moisture stress. This, together with reserves of nonstructural carbohydrates stored in roots and stem, enables rapid resumption of growth after the return of favorable conditions (NAS, 1979). However, a serious limitation to the widespread use of shrub legumes is the length of time and difficultiesinvolved in establishment (Maasdorp and Gutteridge, 1986; Jones and Bray, 1983). Another problem that can be encountered is that shrub legumes tend to grow too tall, unless some form of management can be imposed on the stand to keep the production of edible material within the reach of grazing ruminant animals (Gray, 1970). Most experiments measure yield by cutting material from trees, thus simulating cut-and-carry utilization of tree legumes. Although such information can sometimes indicate the potential of leguminoustrees to provide large quantities of high-protein animal feed, inappropriate choice of height or harvest interval for a given environment might equally result in yields well below potential. Accordingly, there has been some effort to determine cutting management strategies, primarily for the species Leucaena, given its widespread use as a forage and fuelwood tree in developing countries of tropical regions. Although the general usefulness of Leucaena is well known and high yields have been reported, there is relatively little information on how to manage it for maximum or sustainable forage production. Results from

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

37

experiments investigating the effect of cutting height and/or cutting frequency on Leucaena yield are summarized in Table IV. A major problem with many research reports is the use of incorrect or ill-defined yield terminology (e.g., Mendoza and Javier, 1980). Many reports are of limited utility because of a failure to accurately define “herbage” without making it clear whether this refers to edible material only or to total biomass yield. Most of the yield figures in Table IV conform with the assertion of Blom (1980) that annual edible dry matter production from Leucaena will generally range form 6 to 18 t DM ha-’ year-’. Comparison of yields and the effects of cutting height and frequency on Leucaena is made difficult because of the different cultivars used by the various researchers. The findings of cutting management experiments will be dependent on whether arboreal (“Salvador”), shrubby (“Hawaiian”), or lowbranching (“Peru”) cultivars have been used. Although the Salvador types with high leaf: stem ratios have often outyielded the Hawaiian and Peru types, the results have been confounded by the method of harvesting Table IV Cutting Height and Cutting Frequency Effects on Yield of LeucaendPb Cutting height (cm) 5 1 1SOd 150”

100 (5) (10)

Cutting interval (days) ( 120)

( 150) (60)

70d 70d (40) 110“

90

120“ 30/60/90 nsp 40 30 90

(100)

110

75 30-90 ns 120d

60 42-84 ns (30)

(30) 30“ 50d

Yield (t ha-’ year-‘)

Reference

50.6 FW total 28.7 FW total 39.4 FW total 28.0 FW total 113.0 FW total 13.0 DW leaf 9.0 DW leaf 8.2 DW leaf 5.4 DW leaf 14.6 DW total 12.5 DW total I 1 .O DW leaf 14.2 DW leaf 15.2 DW leaf 4.0 DW leaf

Takahashi and Ripperton (1949)r Castillo el al. (1979) Krishnamurthy and Munegowda (1982a) Krishnamurthy and Munegowda (l982b) Siregar (1983) Guevarra et al. (1978) Ferraris (1979) Savory and Breen (1979) Pathak et a / . (1980) Perez and Melendez (1980) Osman (1981a.b) Semali et al. (1983) Topark-Ngarm (1983) Evensen (1984) Isarasanee et al. (1985)

“ After Horne et al. (1986).

’

Figures not in parentheses are the best height/frequency used. Figures in parentheses are the only heightlfrequency used. Cited by Evensen (1984). Maximum height or longest frequency used. No significant differences.

”

38

GRAEME BLAIR ET AL.

(Evensen, 1984). Results from experiments investigating effects of cutting height and cutting frequency are often conflicting and will be considered separately here to clarify general trends emerging from the literature. A. AGEOF FIRST CUTTING Little information exists on the effect of age of tree legumes at first cutting on their productivity and persistence. Bhumibhaman er al. (1984) assessed the coppicing of 1-year-old Leucaena leucocephala trees and concluded that the amount of regrowth was positively related to stump size. Similar results were reported by Pathak and Patil(l983). Ella (1987) reported an increase in both leaf and wood production in L . leucocephala, C. calothyrsus, S . grandiflora, and G . sepium as age to first cutting increased from 13 to 21 months (Table V). Subsequent yields recorded over a 12-month period revealed that the yields of S . grandiflora and C . calofhyrsus were not related to age of first cutting, whereas in L . leucocephala and G . sepium, the highest yields were recorded from the trees that were first cut at 21 months. These results have prompted research to investigate the role of fastgrowing, weakly perennial species in supplying forage to animals (Bray er al., 1989). These studies have recorded high leaf yields from Cajanus cajan, Codariocalyx gyroides, and Sesbania sesban at sites in Indonesia and Australia. Planting of these species would allow farmers to leave the more perennial tree species uncut for a longer period and hence increase overall productivity. Table V Leaf and Wood Yield (mg DM Tree-’) of Tree Legumes“ Cut to 100 cm at Various Ages from Sowing Age (months)

Leaf

Wood

13 I5 17 19 21

0.35 a*” 0.40 a 0.71 b 1.35 c 2.11 d

0.47 a 0.60 a 1.53 a 3.59 b 6.79 c

a Mean of L. leucocephala, C. calorhyrsus, S . grandgora, and G. sepium.

Numbers followed by the same letter within a column are not significantly different.

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

39

B. CUTTINGHEIGHT Recommendations from cutting height experiments in the literature deal mostly with Leucaena and show wide variability, ranging from extremes as low as 5 cm up to 3 m. Takahashi and Ripperton (1949, cited by Evensen, 1984), in an early investigation of the effect of cutting height on Leucaena, compared total production when trees were cut at 5,38, or 76 cm and found that a cutting height of 5 cm gave maximum yields. By contrast, other workers have often found that higher cutting heights give better yields. Mendoza et al. (1975) found that yields of Leucaena cut at 300 cm were more thar double that of Leucaena cut at a height of 15 cm. Unfortunately, they c id not include an intermediate range of cutting heights in their experiment. Herrera (1967, cited by Evensen, 1984), Pathak et al. (1980), and Perez and Melendez (1980) all found that the highest yields were obtained from the highest cutting height under investigation, but as these were 75 cm or below, it might be that their experiments did not cover a wide enough range to properly investigate this effect. Horne and Blair (1991) found no significant difference in leaf or wood yields when L . leucocephala cv. Cunningham was cut at 30 or 100 cm. Krishnamurthy and Munegowda (1982a) found that 150 cm gave higher yields than 15-cm and 75-cm cutting height treatments, and in an additional comparison (Krishnamurthy and Munegowda, 1982b), the highest cutting height treatment (150 cm) again resulted in the highest yields compared with cutting heights of 60,90, or 120cm. Most other workers have reported optimum cutting heights in the range of 90-120 cm (Osman, 1981c; Siregar, 1983; Isarasanee et al., 1985). Lazier (1981) imposed three cutting frequency treatments (14-, 40- and 56-day intervals) with three heights of cutting ( 5 , 25, and 50 cm) on Codariocalyx gyroides and found that the 56-day cutting interval and the two higher cutting heights gave the best yields (2.7 t leaf DM ha-' year-'). Siregar (1983) imposed a range of cutting height treatments (up to 150 cm) on Calliandru calothyrsus and Flemingia congesta and reported that cutting height at 1 m gave the best yield. C. CUTTINGFREQUENCY Cutting frequency treatments imposed by researchers range from harvest intervals of 4 weeks (Perez and Melendez, 1980; Osman, 1981a) up to 16-17 weeks (Takahashi and Ripperton, 1949, cited by Evensen, 1984; Ferraris, 1979). Most experiments have set fixed cutting intervals, but it has been agreed that in areas with large variability in climatic conditions

40

GRAEME BLAIR ET A L .

(e.g., with distinct wet and dry seasons), optimum production is more likely if trees are cut on the basis of stage of regrowth (Evensen, 1984). Guevarra et al. (1978), Herrera (1967, cited by Evensen, 1984),Isarasanee et al. (1985), and Ferraris (1979) reported experiments in which cutting frequency treatments were imposed on the basis of attainment of some predetermined height of regrowth. In a more recent experiment, Catchpoole and Blair (1990a) found that leaf production in L . leucocephala, G . sepium, C calothyrsus, and S . grandifIora was unaffected by cutting to a height of 100 cm when the trees reached 150,200,or 250 cm. By contrast, stem production averaged over species increased from 10.16 to 21.45 t ha-' as attainment height was increased from 150 to 250 cm. Data from cutting experiments that use predetermined stages of regrowth to indicate the time of harvesting are more meaningful and more readily applicable to other locations than if only fixed cutting intervals are used (Hegde, 1983). Results of cutting frequency experiments with Leucaena reported in the literature vary considerably, and the variation in cutting heights makes it difficult for any general conclusions to be drawn. For example, whereas Takahashi and Ripperton (1949, cited by Evensen, 1984) and Ferraris (1979) obtained the best yields from infrequently cut trees at low cutting heights, Pathak et al. (1980) and Perez and Melendez (1980) found that yields were maximized at high cutting heights with shorter cutting intervals. The situation becomes a little clearer when the results from investigations reporting the effect of cutting frequency on total yield are considered separately from studies on the effects of cutting interval on yields of leaf or edible forage from Leucaena. Most workers investigating the effects of cutting frequency on total Leucaena yields have generally found that infrequent harvesting gives the highest total dry matter yields. Osman (1981a) compared cutring intervals of 30, 60, 90,and 120 days, with the 90-day treatment resulting in the highest total yield. The higher total DM production under infrequent cutting is generally attributed to the higher proportion of the total yield made up of stem material (Osman, 1981b; Horne et a / . , 1986). Ferraris (1979) found that leaf made up 54% of the total Leucaena yield after 60 days of regrowth, but only 31% after 120days. Similarly, Guevarra el al. (1978) reported a decline in the edible fraction from 81% after 70 days to 60% after 110 days of regrowth, and Topark-Ngarm (1983) from 69% at 40 days to 60% after 60 days. Thus it appears that the higher total yields of Leucaena obtained with infrequent cutting are due to a relative increase in the stem component (a lower leaf stem ratio) with lengthened harvest interval. However, it is difficult to draw general conclusions from experiments

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

41

investigating the effects of cutting frequency on leaf yields, as the results are much less consistent. Pathak et al. (1980) and Das and Dalvi (1981) found that the most frequent cutting treatments studied (40-day and 60-day intervals, respectively) gave the best yields of Leucaena leaf. In contrast, Semali et al. (1983) found that infrequent cutting (1 10 days) resulted in the highest leaf yields. Topark-Ngarm (1983) obtained higher leaf yields with a 60-day harvest interval than from a more frequently cut 40-day regrowth treatment. Evensen (1984) reported no significant differences in leaf yield across cutting heights of 30,60, and 90 cm, with harvest intervals of 40,60, and 120 days, despite the lower leaf fraction in the total yield at 120 days (44%) compared to at 40 days (69%). These results of Evensen (1984) have particular relevance to management recommendations for Leucaena production, as his findings imply that infrequently cut Leucaena can be used as a source of fuelwood without adversely affecting forage yield. However, such a practice is dependent on the nature of the farmer's tree legume resource and forage requirements-a smallholder with a limited number of trees will have to cut each tree more often to obtain a continuous supply of forage (Home et al., 1986). Although the available literature on the effects of cutting frequency on leaf dry matter yields is confusing, it is well known that the forage quality of Leucaena leaf declines with age, and therefore material from infrequently cut trees with a higher proportion of older leaf has a lower nutritive value than material harvested from frequently cut trees (Takahashi and Ripperton, 1949, cited by Evensen, 1984;Tangendjaja et al., 1986). Semali et al. (1983) showed that calcium, phosphorus, and nitrogen levels in leaf declined as cutting interval was lengthened beyond 60 days. An exception is that of Evensen (1984), who found no significant effect of cutting frequency on crude protein content, although this could be a reflection of the highly fertile soil on which his experiments were conducted. There is very little information concerning the effects of cutting height and frequency on yields of tree legume species other than Leucaena. Gore and Joshi (1976, cited by Evans and Rotar, 1987a) obtained the highest yields of N,P,K-fertilized Sesbania sesban with infrequent cutting (every 9-10 weeks) than with frequent cutting (5-6 weeks), over a 31-week growing period (14.2 t leaf DM ha-').

D. DENSITY The variation in tree spacing and planting configuration between published investigations into Leucaena cutting management strategy is a further confounding factor in the interpretation of the effects of cutting height and frequency on yield. The density employed in cutting height and fre-

42

GRAEME BLAIR ET A L .

quency experiments is usually determined with a view to the intended practical application. Rarely has equidistant spacing been used, but rather Leucaena has been planted in rows to simulate hedgerow or alley cropping management systems and, as such, density expressed as number of trees per unit area is less meaningful than specifications of inter- and intrarow tree spacing. Generally most researchers have found that Leucaena planted at higher densities gives better yields than at lower densities. Pathak et al. (1980) reported higher leaf dry matter yields (5.4 t ha-' year-') from trees at a density of 40,000 trees ha-' than at a density of 15,000 trees ha-'. (Both of these density treatments were cut at a height of 30 cm every 40 days.) Castillo et al. (1979) compared four densities (3000, 5000, 60oO, and 10,000 trees ha-') and obtained significantly higher yields from the two highest densities. Three densities (lO,OOO, 30,000, and 60,000 trees ha-') were compared by Savory and Breen (1979), who reported that 60,000 trees ha-' gave the highest yields. Van den Beldt (1983) reported that wood yields were higher at wider spacings. Ella (1987) studied tree legumes in mixtures with grasses in a distinctly monsoonal climate in South Sulawesi, Indonesia, where soil moisture deficits limit plant growth in the dry season and incident light is potentially limiting in the wet season. Four tree legume species (Sesbania grandijlora, Calliandra calothyrsus, Leucaena leucocephala, and Gliricidia sepium) were grown at four densities (5000, 10,OOO, 20,000, and 40,000 trees ha-') with an underplanted grass (Panicum maximum). Cutting intervals of 6 and 12 weeks were applied to the trees (cut at 1 m) and the grass (cut at 12 cm). Although there was some variability in results between species, there was a general trend toward increased grass leaf dry matter yields and reduced tree leaf dry matter yields with reduced tree density in the wet season. In the dry season, however, grass yields were positively related to tree density. Ella (1987) attributed this seasonal reversal of the tree density effect on grass yields to competition for light, which inhibited grass growth at the higher densities and longer cutting intervals, in the wet season and to reduced direct sunlight and transpiration benefiting grass growth in the same treatments in the dry season. A further confounding factor influencing the effect of tree density on grass yields is nitrogen fixation. Kitamura (1986) investigated the longterm effects of varying tree densities and regrowth intervals on the productivity of low-cut Leucaena (30 cm) interplanted with tall tropical grasses (Pennisetum purpureum and Panicum maximum). Total dry matter yields of Leucaena were reduced to only 2 1% of monoculture yields by Penniseturn, compared with 5% in the experiment of Horne and Blair (1990). This is probably because of the very high densities employed (10-cm intrarow spacings and 15-,30-, 60-, and 100-cm interrow spacings). Kitamura (1986) found that edible dry matter yields of both Leucaena and Penniserum were

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

43

maximized at the highest densities and longest growth intervals (attainment of 200-cm height by the Leucaena canopy). The explanation given for this was substantial transfer of nitrogen from the Leucaena to the grass in these treatments. The effects of interrow and intrarow spacing on yield have been investigated by Guevarra et al. (1978), Ferraris (1979), and Savory and Breen (1979). Ferraris (1979) found no differences between interrow spacings of 30 or 90 cm. Savory and Breen (1979) also found no significant differences between Leucaena leaf yields of the two interrow treatments (60-cm and 100-cm spacing between rows), but there was a significant effect of intrarow spacing, with significantly higher yields (7 t leaf DW ha-’ year-’) from trees 30 cm apart within a row compared to the 7.5- and 15-cm intrarow treatments. This is in contrast to the findings of Guevarra et al. (1978), who reported that wider intrarow spacing (45 cm) gave higher yields than closer spacings (15 or 30 cm). Thus the general observation that high densities result in higher yields must, however, take into account the planting configuration of trees, as the evidence from Guevarra et al. (1978) indicates that the closer spacing within rows can result in lower yields.

E. INTERACTION BETWEEN DENSITYA N D CUTTING MANAGEMENT One possible factor giving rise to the often conflicting findings of experiments investigating the relationship between height and frequency on leaf yields might be that of density. For example, the variation in cutting height recommendations could be due to the effect of density on rooting volumes. At high densities, which limit the soil volume that can be exploited by a single tree, a low cutting height will more severely impair regrowth than at lower densities. This shows the importance of the size of stumps remaining, from which reserves can be mobilized for regrowth, and explains the better yields obtained by Pathak et al. (1980), Perez and Melendez (1980), and Mendoza et al. (1975) at higher cutting heights. The literature in not consistent concerning cutting frequency effects on leaf yield and might be due partly to the different densities adopted by the various experimenters. There is a strong interaction between cutting frequency and density (Horne et al., 1986). The ideal time to harvest the woody species is after canopy closure and attainment of maximum Leaf Area Index, but before any shedding of leaves due to shading has begun. Harvesting after this stage will result in lower yields due to leaf abscission, whereas harvesting before this optimum stage will increase the relative proportion of the regrowth period taken up by the initial “lag phase” before regrowth commences (Evensen, 1984). Thus, at low densities, infrequent cutting will give the best yields, whereas at high densities, the

44

GRAEME BLAIR ET A L .

leaf canopy will close earlier and therefore shorter cutting intervals will result in higher yields. Planting arrangement will also affect the optimum time to harvest. Trees growing in widely spaced rows will not be subject to the same degree of light competition as trees equidistantly spaced, because of lateral irradiation. Consequently, one might expect longer cutting intervals to give the highest yields with trees planted in rows, whereas more frequent cutting is required with equidistantly spaced trees at the same density. At this stage, however, there is not yet a clear indication of the effect of density and/or plant arrangement on leaf yields or the leaf : stem ratio. Another possible reason for the conflicting findings concerning optimum cutting frequency is a potential interaction with height of cutting. Low cutting heights could limit the number of buds from which regrowth can arise. Evidence for this comes from the work of Perez and Melendez (1980), who showed that Leucaena cut at 30 cm formed fewer buds (89) after a harvest than trees cut at 50 cm ( 1 12 buds). This might explain why Guevarra et al. (1978) and Ferraris (1979), with very low cutting heights of 5- 10 cm, found that longer harvest intervals resulted in higher leaf yields, whereas those workers who reported that frequent cutting gave maximum leaf yields had higher cutting heights. Evensen (1984), however, did not find any interaction between cutting heights (30,60, and 90 cm) and cutting interval (40, 60, and 120 days). There is little information available on the long-term effect of cutting management on growth habit and form of each tree. The work of Savory and Breen (1979) suggested that frequent cutting ultimately results in a greater number of shoots regrowing from the stump than occurs with less frequent cutting. Consequently, it is possible that the optimum harvest interval might change with time. However, to date no cutting frequency experiment has been conducted long enough for this to become apparent. There are many other cutting management factors besides height and frequency that might be important determinants of yield and persistence but that have not yet been addressed by researchers. For example, it is likely that regrowth will be more rapid from shoots remaining on the tree after cutting than if regrowth arises solely from new buds, but no information is yet available to compare the effect of total versus partial removal of leaf on long-term yields. Another factor that has received little attention is the effect of age or size of the tree at the time of the first harvest on subsequent yields. Ella (1987) has found that an additional consideration that has not received any attention is whether initial cuts should be made at a lower height than subsequent harvests in order to promote a multibranched stem with theoretically more potential bud sites from which regrowth can arise.

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

45

VI. NITROGEN YIELDS OF MATERIAL HARVESTED FROM TREE LEGUMES Yields of nitrogen in material cut from leguminous trees are presented in Table VI. Information about possible nitrogen yields is scanty, even for Leucaena, because many of the experimenters who have reported dry matter yields did not take samples of leaf or stem material for analysis of nitrogen content. Generally, most researchers have found a mean nitrogen concentration in leaf tissue within the range of 3.5% to 4.5% N , although Endrinal and Mendoza (1979, cited by Mendoza, 1984) showed the N content of young leaves to be as high as 6% N , whereas for mature leaves the nitrogen concentration was found to drop to as low as 2.6% N. Table VI Reported Nitrogen Yields in Material Harvested from Tree Legumes Species Leucaena leucocephala

Sesbania sesban Sesbania grandgora Sesbania cannabina Calliandra calothyrsus Gliricidia sepium

Nitrogen yield (kg N ha-' year - I )

Reference

614 (leaf)

Evensen (1984)

414 (leaf) 581 (leaf) 195 (leaf) 485 (leaf, 9 months only) 240 (total) 751 (total) 613 (total) 715 (total) 630 (total) 450 (total) 731 (leaf) 848 (total) 160-195 (leaf, 221 days only) 120 (leaf) 160 (total) 76 (total, 70 days only) 474 (leaf) 581 (total) 586 (leaf) 700 (total)

Ferraris (1979) Guevarra et al. (1978) Hutton and Beattie (1976) Hutton and Bonner (1960) Anslow (1957)" Eriksen and Whitney (1982) Ferraris (1979) Guevarra et al. (1978) Kitamura (1986) Takahashi and Ripperton (1949)" Catchpoole and Blair (1990a)

" Cited by Hutton and Bonner (1960). Cited by Evans and Rotar (1987a).

Gore and Joshi (1976)b Catchpoole and Blair (1990a) Yost et al. (1985) Catchpoole and Blair (1990a) Catchpoole and Blair (1990a)

46

GRAEME BLAIR E T A L .

The nitrogen yield of 715 kg N ha-' year-', obtained by Guevarra et al. (1978), has often been quoted by other researchers in the past decade as evidence of the capacity of Leucaena to fix large quantities of atmospheric nitrogen. However, as Halliday and Somasegaran (1983) pointed out, this figure cannot be attributed solely to biological nitrogen fixation because of the contribution from soil nitrogen. Eriksen and Whitney (1982), in estimating nitrogen fixation by Leucaena, attempted to allow for nitrogen that had been taken up from the soil. They subtracted the average N uptake of three grasses grown adjacent to the tree legume plots from the total nitrogen accumulation in the harvested Leucaena material and obtained an estimate of nitrogen fixation of 656 kg N ha-' year-'. The limitation of this approach is that soil N uptake by the relatively shallow-rooted grass species might not provide a good indication of soil N uptake by deeprooted perennial Leucaena. This and other difficulties associated with techniques to estimate N fixation by tree legumes are discussed later. High-protein forage can, in theory, be continually cut from three legumes fertilized with all nutrients except nitrogen without adversely affecting nitrogen yields, because of the capacity for symbiotic nitrogen fixation. However, it is most unlikely that tree legumes used as sources of cut-and-carry forage throughout developing countries ever receive any fertilizer, and consequently yields will decline with nutrient depletion. Humphreys (1978) noted that where herbaceous pastures form part of a cut-and-carry animal feeding system, the reduction in pasture soil fertility is very rapid if the dung, urine, and reject forage are not carefully collected and returned to the pasture.

VII. NITROGEN RECYCLING VIA LEAF AND EXCRETA Although cut-and-carry animal feeding systems are the norm in densely populated areas of the tropics, in other areas where population pressure is not as high, or land is unsuitable for cropping, ruminant animals are permitted to graze freely. There are a number of advantages with grazing compared to cut-and-carry systems. The labor input is generally less, and nutrient rundown of the pasture is likely to be greatly reduced as a result of nutrient recycling through the animal. Disadvantages include trampling and the need for fencing and provision of water. In both systems there is potential for large amounts of nitrogen to be lost through ammonia volatilization. There are few reports of the direct contribution of tree legumes to an associated nonlegume. In a field experiment conducted by Catchpoole and

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

47

Blair (1990a),901 kg N ha-' year-' was removed in leaf and stem material of L . leucocephala, C. calothyrsus, G . sepium, and S . grandijlora when grown as a monoculture. No difference in grass N yield was recorded between the grass in monoculture and where it was planted under the tree legume (Table VII). In a subsequent glasshouse experiment using a split-root technique with "N, Catchpoole and Blair (1990b)found that the direct transfer of N from L . leucocephala and G . sepium was only 9% of the total harvested N yield. Most workers in the field of nitrogen cycling believe that transfer of biologically fixed nitrogen from legume to soil predominantly occurs as a result of leaf fall and subsequent decomposition, or via passage through grazing animals (Vallis, 1978, 1979). Substantial leaf fall has been shown to occur in a Leucaena-Gliricidia plantation by Sajise et al., cited by Van den Beldt (1983),who recorded an annual litter fall of 1 1.3t DM ha-'. Van den Beldt (1983)found leaf fall in a "Hawaiian Giant" Leucaena plantation to be 8.5 t DM ha-' year-', with an average nitrogen content of 1.2%. The relatively low N content of leaf litter compared to leaf on the plant is due to mobilization of nitrogen from leaf before abscission occurs. Thus the nitrogen contribution from tree legumes as shade trees might not be as high as in an alley cropping situation. There is a strong correlation between rate of decomposition and the nitrogen content of organic matter (Stevenson, 1986). Consequently, leaf cut from hedgerows with a low C/N ratio will break down and release N more quickly than abscised leaves with a high C/N ratio. In addition, it is possible that there will be a greater input of

Table VII Nitrogen Yields (kg ha-') over a 14-Month Period of Tree/Grass Mixtures and Monocultures at Gowa, South Sulawesi, Indonesia"

Leaf N yield Stern N yield Total legume N yield Grass (%N) Grass N Yield Total edible N Yield Total N yield ~~

~~

a

Tree monoculture

Treelgrass mixture

Grass monoculture

751 ab 150 a 901 a 752 a 901 a

669 b 132 b 801 b 1.3 a 100 a 769 a 901 a

0.9 b 103 a 103 b 103 b

~

From Catchpole and Blair (1990a). Means with the same letter are not significantly different ( p < .05).

48

GRAEME BLAIR ET AL.

nitrogen from nitrogen fixation into an alley cropping system because, according to one theory of symbiotic nitrogen fixation (Bethlenfalvay and Phillips, 1977; Bethlenfalvay et al., 1978), removal of large quantities of nitrogen from the legume will increase the demand of the plant for available nitrogen, thus stimulating the rate of N2 fixation. In addition to increasing the total level of soil nitrogen, leguminous trees also can increase the availability of nitrogen and other nutrients for uptake by the companion crop. Tiedemann and Klemmedson (1973) found the relative yield of grass grown in soil taken from under mesquite (Prosopis glandulensis) to be 15 times that of intratree soil, whereas the soils differed in total N by a factor of only 3. Tree legumes are known to have deep rooting systems, up to 50 m in some instances (Phillips, 1963), thus nutrients in low concentration throughout the soil profile can be taken up by the tree roots and become concentrated near the soil surface from litter fall and decomposition (Virginia, 1986). This redistribution of nutrients from deep soil to the surface increases the supply of nutrients to a shallow-rooted companion crop but, unlike nitrogen, the total quantity of the nutrient in the system does not increase. Tree roots can result in a loosening of the soil and improvement in porosity, and wind-blown soil particles can be intercepted by the tree canopy and washed to the ground beneath (Sanchez et al., 1985). Shade and the protection of the litter layer result in a moderation of soil temperature and dampening of temperature fluctuations, which can lead to more favorable conditions for microbiological breakdown and release of nutrients from organic matter (Virginia, 1986). The idea of incorporating tree legumes in fencelines as “living fences” or on the bunds between rice paddies has won widespread acceptance in many areas of the tropics. Piggin and Parera (1985) reported that in Timor large areas of hilly country have been planted to Leucaena leucocephala, and forage material is cut from these slopes and carried to animals tethered to horticultural trees (e.g., coffee or coconuts) in backyard gardens. This has resulted in an apparent improvement in horticultural yields since the adoption of this practice. It is likely that much of the increased growth is a result of nitrogen cycled through the animal and returned in feces and urine to the soil, given the low maintenance requirement of adult ruminants for nitrogen (Henzell and Ross, 1973). In Java, feces and uneaten leaf falling through the slatted floors of animal pens is periodically collected and spread onto neighboring vegetable gardens as fertilizer. Thus, improvement of animal feed quality through the use of tree legumes has ramifications not only for increased animal production, but also for greater crop yields. However, to date no research to quantify this avenue of nitrogen contribution from tree legumes in a cut-and-carry management system has been reported.

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

49

The use of material cut from tree legumes as fertilizer for nonlegume crops has long been a practice in many areas of Asia (Brewbaker and Glover, 1987; Evans and Rotar, 1987a). Traditional “green manuring” techniques, however, usually involved carrying branches or leaves cut from trees grown in one area to the intended area for cropping (Bin, 1983), whereas in an alley cropping system the trimmed foliage does not have to be camed as far (NAS, 1979). There has been very little research conducted comparing yields of nonlegume crops within an alley cropping system with yields obtainable from a monoculture of the crop on a per hectare basis. Hoekstra (1983)conducted an economic analysis of a simulated alley cropping system in a semiarid environment based on the assumptions of Torres (1983) that introduction of Leucaena hedgerows would halve the area available for maize, but at a rate of increase of 13.5 kg maize ha-’ for every kilogram of organic N added from Leucaena prunings. Research conducted at the International Institute of Tropical Agriculture in Nigeria showed that the material cut from Leucaena hedgerows was able to sustain maize grain yield for two consecutive years, but with no addition of prunings, maize yield declined (Kang et al., 1981).

VIII. CONCLUSION The growing body of literature concerning tree legumes is indicative of the increasing realization of their potential role in farming systems and their multiplicity of uses in developing countries. However, much of the evidence of the high productivity of leguminous trees in the tropics is indirect and inferential, with a paucity of information from controlled experimentation and research. There is a need for rigorous information on management options that can best exploit the potential of various species of tree legumes. Although much attention has been focused on Leucaena leucocephala, there has been very little investigative work done on the potential of other perennial legume species, particularly in areas where soil and climatic conditions are unsuitable for Leucaena. There are few published reports of long-term comparisons of productivity of several species at a single site, and virtually no information available on management strategies for cut-and-carry forage production for genera other than Leucaena. Much of the literature on cutting management of Leucaena for forage is conflicting, and there are very few reports of the yields of nitrogen that can be obtained when forage is cut from Leucaena, and little published information exists on yields of N from other species. Leguminous trees are already widely used in tropical Third World countries as sources of timber, fuelwood, forage, and shade for plantation

50

GRAEME BLAIR E T A L .

crops. However, if their role as a source of symbiotically fixed nitrogen is to be exploited fully, then it is important to have knowledge of their nitrogen input in the various agricultural systems in which tree legumes are employed. Developing countries in the wet and humid tropics have been drawn into the trap of using a single species of tree legume (Leucaena) by enthusiastic scientists and agriculturalists. They are now paying the penalty for this as the psyllid insect devastates large areas. Scientists must show the way with alternative species and information on their management to obtain sustainable production. REFERENCES Adenaye, J. A. 1979. Anim. Feed Sci. Techno!. 4(3), 221-225. Ahmad, N., and Ng, F. S. P. 1981. Malays. For. 44(4), 516-523. Alconero, R., Stone, E. G., and Cairns, J. R. 1973. Agron. J . 65(1), 44-46. Anning, P. 1980. Queensl. Aqric. J . 106(2), 148-171. Anonymous 1981. “Queensl. Dep. Primary Ind. Annu. Rep. 1981,” pp. 9-34. QDPI, Brisbane, Australia. Anslow, R. C., 1957. Rev. Agric. Suer. Ile Maurice 36,39-49. Bethlenfalvay, G . J., and Phillips, D. A. 1977. Plant Physiol. 60,419-421. Bethlenfalvay, G. J., Abu-Shakra, S. S., and Phillips, D. A. 1978. Plant Physiol. 62,127-130. Bhumibhaman, S . , Boonarutee, P., Luangjama, L.. and Thavorn, W. 1984. Leucaena Res. Rep. 5,70-71. Bin, J. 1983. Soil Sci. 135(1), 65-69. Blair, G. J., Panjaitan, M., Ivory, D. A., Palmer, B., and Sudjadi, M. 1988. J . Agric. Sci. 111, 435-441. Blom, P. S. 1980. Abstr. Trop. Agric. 6(3), 9-17. Blunt, C. G., and Jones, R. J. 1977. Trop. Grasslands 11(2), 159-164. Bray,, R. A. 1984. Aust. J . Exp. Agric. Anim. Husb. 24(126), 379-385. Bray, R. A., Palmer, B., and Ibrahim, T. 1989. Nirrogen Fixing Tree Rep. 7,7-8. Brewbaker, J. L. 1986. In “Forages in Southeast Asian and the South Pacific Agriculture” ( G . J. Blair, D. A. Ivory, and T. R. Evans, eds.), ACIAR Proceedings No. 12, pp. 43-50. ACIAR, Canberra. Brewbaker, J. L., and Glover, N . 1987. “Woody Species for Green Manuring in Rice-Based Cropping Systems,” Paper presented to the Workshop on Sustainable Agriculture, May 24-29, 1987. IRRI, Los Banos, Philippines. Brewbaker, J. L., and Hutton, E. M. 1979. In “New Agricultural Crops” ( G . M. Ritchie, ed.), pp. 207-257. AAAS Westview Press, Colorado. Brewbaker, J. L., and Styles, B. 1982. “Economically Important Nitrogen Fixing Tree Species” Nitrogen Fixing Tree Association miscellaneous publication 82-104. NFTA, Waimanalo, Hawaii. Bryan, W. W. 1966. Proc. Int. Grassland Congr. 1966 10,311-315. Bushby, H. V. A. 1982. Aust. J. Exp. Agric. Anim. Husb. 22(117), 293-298. CSIRO, Division of Tropical Crops and Pastures. 1979. “Leucaena,” CSIRO Tropical Crops and Pastures Divisional Report 1977-1978. CSIRO, Brisbane. Carew, B. A. R. 1983. Trop. Grasslands 17(4), 181-184.

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

51

Castillo, A. C., Moog, F. A., and Avante, D. C. 1979. “Effects of Row Arrangement and Plant Density on Herbage Production and Growth of Ipil-ipil.” Paper presented at the 16th Annual Convention of the Philippine Society of Animal Science, Manila. Catchpoole, D. W., and Blair, G. 1990a. Aust. J. Agric. Res. (in press). Catchpoole, D. W., and Blair, G. 1990b. Ausr. J . Agric. Res. (in press). Chadhokar, P. A., and Kantharaju, H. R. 1980. Trop. Grasslands 14(2), 78-82. Chadhokar, P. A., and Lecamwasam, A. 1982. Trop. Grasslands 16(1), 46-48. Chandrasekaran, N. R. 1981. Leucaena Res. Rep. 2, 19-20. Chee, W. S., and Devendra, C. 1983. I n “Leucaena Research in the Asian-Pacific Region,” Proceedings of a workshop held in Singapore, November 1982, pp. 55-60. IDRC, Ottawa, Canada. Cooksley, D. G. 1981. Queensl. J . Agric. Anim. Sci. 38(2), 109-116. Cooksley, D. G . 1982. Queensl. J . Agric. Anim. Sci. 39(1), 47-54. Das, R. B., and Dalvi, G. S. 1981. Leucaena Res. Rep. 2,21. Delpeche, M., Marie, D., and Preston, T. R. 1979. Trop. Anim. Prod. 4, (abstract) 90. Diatloff, A. 1973. Queensl. Agric. J . 99(12), 642-644. D’Mello, J. P. F. 1982. World Rev. Anim. Prod. 18(4), 41-46. Dutt, A. K. 1981. Leucaena Res. Rep. 2,22-23. Ekpenyong, T. E. 1986. Anim. Feed Sci. Tech. 15(3), 183-188. Ella, A. 1987. Unpublished M. Rur. Sci. Thesis, Univ. of New England, Armidale, New South Wales. Ella, A., Jacobson, C., Stur, W. W., and Blair, G. 1989. Trop. Grasslands 23,28-34. Eriksen, F. I., and Whitney, A. S. 1982. Agron. J . 74(4), 703-709. Evans, D. O., and Rotar, P. P. 1987a. “Westfield Trop. Agric. Ser. No. 8.” Westview Press, Colarado. Evans, D. O., and Rotar, P. P. 1987b. Trop. Agric. 64, 193-200. Evensen, C. L. I. 1984. M.Sc. Thesis, Univ. of Hawaii, Manoa, Hawaii. Everist, S. L . 1969. Queensl. Dep. Primary Ind. Div. Plant Ind. Advisory Leajet no. 1024. Falvey, L. 1976. Trop. Grasslands 10, 117-122. Falvey, I. L. 1981. Inr. Tree Crops J . 1(4), 237-244. Felker, P., and Bandurski, R. S. 1979. Econ. Eor. 33(2), 172-184. Ferraris, R. 1979. Trop. Grasslands 13,20-27. Foster, A. H., and Blight, G. W. 1983. Trop. Grasslands 17(4), 170-177. Gonzales, A., Andrew, C. S., and Pieters, W. H. J. 1980. Tech. Pap. Div. Trop. Crops Pastures, CSIRO, 1980 No. 21. Gore, S. B., and Joshi, R. N., 1976. Indian J . Agron. 21(1), 39-42. Gray, S. G. 1970. Trop. Grasslands 4,57-62. Guevarra, A. B., Whitney, A. S., and Thompson, J. R. 1978. Agron. J. 70(6), 1033-1037. Gutteridge, R. C . , and Akkasaeng, R. 1985. Nitrogen-Fixing Tree Res. Rep. 3,60-61. Haag, H. P., and Mitidieri, J. 1980. Leucaena Newsl. 1 , 6 . Halliday, J., and Nakao, P. 1982. “The Symbiotic Affinities of Woody Species Under Consideration as Nitrogen-Fixing Trees.” Univ. of Hawaii NiFTAL Project, Honolulu, Hawaii. Halliday, J, and Somasegaran, P. 1983. In “Leucaena Research in the Asian-Pacific Region,” pp. 27-32. Proceedings of a workshop held in Singapore, November 1982, pp. 27-32. IDRC, Ottawa. Hegde, N. 1983. In “Leucaena Research in the Asian-Pacific Region,” Proceedings of a workshop held in Singapore, 23-26 November, 1982, pp. 73-78. IDRC, Ottawa. Henzell, E. F., and Ross, P. J. 1973. I n “Chemistry and Biochemistry of Herbage” (G. W. Butler and R. W. Bailey, eds.), Vol. 2, pp. 227-245. Academic Press, London.

52

GRAEME BLAIR E T A L .

Herrera, P. G., 1967. Agric. Trop. 23,34-42. Hoekstra, D. A. 1983. Agrofor. S y s t . 1(4), 335-345. Holm, J. 1972. Thailand J . Agric. Sci. 5,227-236. Holm, J. 1973. Thailand J. Agric. Sci. 6,257-266. Home, P., and Blair, G. 1991. Aust. J . Agric. Res. (in press). Home, P. M., Catchpoole, D. W., and Ella, A. 1986. “Forages in SE Asian and South Pacific Agriculture”, (G. J. Blair, D. A. Ivory, and T. R. Evans, eds.), ACIAR Proceedings No. 12, pp. 164-169. ACIAR, Canberra. Humphreys, L. R. 1978. “Tropical Pastures and Fodder Crops.” Longmans, Essex, England. Hutton, E. M., and Andrew, C. S. 1978. Aust. J . Exp. Agric. Anim. Husb. 18(90), 81-88. Hutton, E. M., and Beattie, W. M. 1976. Trop. Grasslands 10(3), 187-194. Hutton, E. M., and Bonner, I. A. 1960. J. Aust. Inst. Agric. Sci. 26,276-277. Isarasanee, A., Shelton, H. M., Jones, R. M., and Bunch, G. A. 1985. Leucaena Res. Rep. 5 , 3-4. Jarvis, B. D. W. 1981. (A. H. Gibson and W. E. Newton, eds.), “Current Perspectives in Nitrogen Fixation,” p. 459. Aust. Acad. Sci., Canberra, Australia. Jones, R. J. 1979. World Anim. Rev. 31, 13-23. Jones, R. J., and Bray, R. A. 1983. I n “Leucaena Research in the Asian-Pacific Region,” pp. 41-49. Proceedings of a workshop held in Singapore, November 1982, IDRC, Ottawa. Jones, R. J., and Jones, R. M. 1982. Trop. Grasslands 16(1), 24-29. Jones, R. J., and Lowry, J. B. 1984. Experimentia 40, 1435-1436. Jones, R. J., Jones, R. M., and Cooksley, D. G. 1982. “Agronomy of Leucaena leucocephala,” Division of Tropical Crops and Pastures, Information Service Sheet 41-44. CSIRO, Australia. Kang, B. T., Wilson, G. F., and Sipkens, L. 1981. Plant Soil 63(2), 165-179. Kitamura, Y. 1986. J . Agric. Sci. 106,623-624. Krishnamurthy, K., and Munegowda, M. K. 1982a. Leucaena Res. Rep. 3,25. Krishnamurthy, K., and Munegowda, M. K. 1982b. Leucaena Res. Rep. 3,31-32. Lazier, J. R., 1981. Trop. Grassl. 15(1), 10-16. Maasdorp, B. V., and Gutteridge, R. C. 1986. Trop. Grasslands 20, 127-134. Manjunath, A., Bagyaraj, D. J., and Gopala Gowda, H. S. 1984. Plant Soil 78,445-448. Mendoza, R. C. 1984. I n “Asian Pastures,” pp. 114-134. FFTC Book Series No. 25. pp. 114-134. Mendoza, R. C., and Javier, E. Q. 1980. Leucaena Newsl. 1,26. Mendoza, R. C., Altamarino, T. P., and Javier, E. Q. 1975. Philipp. J . Crop Sci. 1(3), 149- 153. Mendoza, R. C., Escano, L. R., and Javier, E. Q . 1981. Leucaena Res. Rep. 2,. 42. Minson, D. J., and Wilson, J. R. 1980. J . Aust. Inst. Agric. Sci. 46,247-249. Moog, F. A. 1983. I n “Leucaena Research in the Asian-Pacific Region,” pp. 69-72. Proceedings of a workshop held in Singapore, November 1982, IDRC, Ottawa. NAS 1977. “Leucaena: Promising Forage and Tree Crop for the Tropics.” National Academy of Sciences, Washington, D.C. NAS 1979. “Tropical Legumes: Resources for the Future.” National Academy of Sciences, Washington, D.C. NAS 1980. “Firewood Crops: Shrub and Tree Species for Energy Production.” National Academy of Sciences, Washington, D.C. NAS 1983a. “Firewood Crops: Shrub and Tree Species for Energy Production, Volume 11.” National Academy of Sciences, Washington, D.C. NAS 1983b. “Mangium and Other Fast Growing Acacias for the Humid Tropics.” National Academy of Science, Washington, D.C.

FORAGE TREE LEGUMES IN TROPICAL ENVIRONMENTS

53

NAS 1983~.“Calliandra: A Versatile Small Tree for the Humid Tropics.” National Academy of Sciences, Washington, D.C. NAS 1984. “Leucaena: Promising Forage and Tree Crop for the Tropics,” 2nd Ed. National Academy of Sciences, Washington, D.C. Norani, A., and Ng, F. S. P. 1981. Malays. For. 44(4), 516-523. Oakes, A. J. 1984. Trop. Agric. 61(2), 125-127. Olvera, E., and Blue, W. G. 1985. Trop. Agric. 62(1), 73-76. Olvera, E., and West, S. H. 1985. Trop. Agriculture 62(1), 68-72. Osman, A. M. 1981a. Leucaena Res. Rep. 2,37-38. Osman, A. M. 1981b. Leucaena Res. Rep. 2,35-36. Osman, A. M. 1981~.Leucaena Res. Rep. 2,37-38. Patridge, I. J., and Ranacou, E. 1973. Trop. Grasslands, 7,327-329. Paterson, R. T., Samur, G. S., and Sauma, 0. 1982. Trop. Anim. Prod. 7(1), 9-13. Pathak, P. S., and Patil, B. D. 1983. In “Leucaena Research in the Asian-Pacific Region,” Proceedings of a workshop held in Singapore, November 1982, pp. 83-88. IDRC, Ottawa. Pathak, P. S., Rai, P., and Roy, R. D. 1980. Forage Res. 6(1), 83-90. Perez, P., and Melendez, F. 1980. Trop. Anim. Prod. 5,282. Phillips, W. S. 1963. Ecology 44,424. Piggin, C. M., and Parera, V. 1985. In “Shrub Legume Researchin Indonesiaand Australia,” Proceedings of a workshop held at Balai Penelitian Ternak, Ciawi, Indonesia, Feb. 2, 1984, pp. 19-27. (E. T. Craswell and B. Tangendjaja, eds.), ACIAR, Canberra. Pound, B., and Martinez-Cairo, L. 1983. “Leucaena: Its Cultivation and Uses.” Overseas Development Administration, London. Pound, B., and Santana, A. 1980. Trop. Anim. Prod. 5 , 280. Pound, B., Ruiz, G., and Santana, A. 1980a. Trop. Anim. Prod. 5(1), 91. Pound, B., Santana, A., and Ruiz, G. 1980b. Trop. Anim. Prod. 5(3), 228-231. Prussner, K. 1983. In “Leucaena Research in the Asian-Pacific Region,” Proceedings of a workshop held in Singapore, November 1982, pp. 161-168. IDRC, Ottawa. Romeo, J. T., 1984. Biochem. Sysr. Ecol. l2(3), 293-298. Sanchez, P. A., Palm, C. A., Davey, C. B., Scott, L. T., and Russell, C. E. 1985. Tree crops as soil improvers in the humid tropics. In “Attributes ofTrees as crop. plants” (M. C. R. Cannell and J. E. Jackson, eds.), pp. 327-358. Titus Wilson & Son, Cambria, Great Britain. Sanginga, N., Mulongoy, K., and Ayanaba, A. 1986. Biol.Agric. Hortic. 3(4), 347-352. Savory, R., and Breen, J. 1979. UNDP/FAO Project MLW/75/020, Working Paper 17. Semali, A., Syamsimar, D., and Manurung, T. 1983. “Proceedings 5th World Animal Production Conference” pp. 61 1-612. Tokyo, Japan. Shaw, N. H. 1965. “Division of Tropical Pastures, Annual Report,” p. 42. CSIRO, Townsville, Queensland, Australia. Shaw, N. H. 1968. “Division of Tropical Pastures, Annual Report,” pp. 11-13. CSIRO, Townsville, Queensland, Australia. Siregar, M. E. 1983. “Proceedings 5th World Animal Production Conference, pp. 613-614. Tokyo, Japan. Skerman, P. J. 1977. In “F.A.O. Plant Production and Protection Series,” No. 2. U.N., Rome. Stevenson, F. J. 1986. “Cycles of Soil: Carbon, Nitrogen, Phosphorus, Sulfur, Micronutrents.” Wiley, New York. Sumberg, J. E. 1985. Trop. Agric. 62, 17-19. Takahashi, M., and Ripperton, J. C., 1949. Hawaii Agric. Exp. S t n . Bull. 100,56. Tangendjaja, B., Lowry, J. B., and Wills, R. B. H. 1986. Aust. J . Exp. Agric. 26(3), 315-318.

54

GRAEME BLAIR E T A L .

Tham, K. C., Wong, C. C., and Ajit, S. S. 1977. MARDI Res. Bull. 5 , 10-20. Tiedemann, A. R., and Klemmedson, J. 0. 1973. Soil Sci. SOC.Am. Proc. 37, 107-1 11. Topark-Ngarm, A. 1983. Leucaena Res. Rep. 4,77-78. Torres, F. 1983. Agrofor. Syst. 1,323-333. Vallis, I. 1978. Nitrogen relationships in grass/legume mixtures. In “Plant Relations in Pastures” (J. R. Wilson, ed.), pp. 190-202. CSIRO, Melbourne. Vallis, 1. 1979. I n “Nitrogen Relationships in Pasture Systems of Southern Queensland,” Proceedings of a workshop held at Toowoomba, June 5 , 1979, pp. 3-26. Dept. Primary Industries, Queensland. Van Den Beldt, R. J. 1983. In “Leucaena Research in the Asian-Pacific Region,” Proceedings of a workshop held in Singapore, November 1982, pp. 103-108. IDRC, Ottawa. Van Den Beldt, R., and Huxley, P. 1982. “Institutions Studying Nitrogen Fixing Trees,” Nitrogen Fixing Tree Association, miscellaneous publication 82- 106 NFTA, Waimanalo, Hawaii. Vercoe, T. K. 1989. I n “Trees for the Tropics” (D. J. Boland, ed.). ACIAR, Canberra. Virginia, R. A. 1986. For. Ecol. Manage. 16,69-79. Whiteman, P. C. 1980. “Tropical Pasture Science.” Oxford Univ. Press, Oxford. Wildin, J. H. 1980. Queensl. Agric. J . 106(3), 194-197. Wong, C. C., Izham, A., Chen, C. P., Hassan, W., Aminah, A., and Eng, P. K. 1983. Proc. 7th Annu. Con$ MSAP pp. 115-127. Yost, R. S., Evans, D. O., and Saidy, N. A. 1985. Trop. Agric. 62(1), 20-24.

ADVANCES IN AGRONOMY, VOL. 44

STATISTICAL ANALYSES OF MULTI LOCATION TRIALS Jose Crossa Biometrics and Statistics Unit International Maize and Wheat Improvement Center (CIMMYT) 06600 Mexico D. F., Mexico

I. Introduction 11. Conventional Analysis of Variance A. Limitations B. Components of Variance 111. Joint Linear Regression A. Statistical and Biological Limitations B. Other Measurements of Yield Stability C. Stability, Risk, and Economic Analysis IV. Crossover Interactions V. Multivariate Analyses of Multilocation Trials A. Principal Components Analysis B. Principal Coordinates Analysis C. Factor Analysis D. Cluster Analysis VI . AMMI Analysis A. AMMI Analysis with Predictive Success VII. Other Methods of Analysis VIII. General Considerations and Conclusions References

I. INTRODUCTION Multilocation trials play an important role in plant breeding and agronomic research. Data from such trials have three main agricultural objectives: ( a )to accurately estimate and predict yield based on limited experimental data; ( b )to determine yield stability and the pattern of response of genotypes or agronomic treatments across environments; and (c) to provide reliable guidance for selecting the best genotypes or agronomic treatments for planting in future years and at new sites. Agronomists in particular use multilocation trials to compare combinations of agronomic factors, such as fertilizer levels and plant density, and 55 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

56

JOSE CROSSA

on this basis make recommendations for farmers. Breeders compare different improved genotypes to identify the superior ones. Variation in yield responses among certain agricultural production alternatives (genotypes, agronomic treatments, and cropping systems), when evaluated in different environments, is known in the classical sense as interaction. This interaction is part of the behavior of the genotype or agronomic treatment and confounds its observed mean performance with its true value. Assessing any genotype or agronomic treatment without including its interaction with the environment is incomplete and thus limits the accuracy of yield estimates. Therefore, a significant portion of the resources of crop breeding and agronomy programs is devoted to determining this interaction through replicated multilocation trials. Data collected in multilocation trials are intrinsically complex, having three fundamental aspects: (a) structural patterns; (b) nonstructural noise; and (c) relationships among genotypes, environments, and genotypes and environments considered jointly. Pattern implies that a number of genotypes respond to certain environments in a systematic, significant, and interpretable manner, whereas noise suggests that the responses are unpredictable and uninterpretable. The function of the experimental design and statistical analyses of multilocation trials is to eliminate as much as possible of the unexplainable and extraneous variability (noise) contained in the data. Gauch (1988) mentions that statistical analysis of multilocation trials may have two different objectives. With the first, postdictiue criteria, a statistical model is constructed for a data set, and success is measured in terms of the model’s ability to fit this data set, with consideration of parsimony (reduced model with minimal degrees of freedom). With the second, predictive criteria, the data from a yield trial are partitioned into (a) data used to construct a model (modeling data) and (b) data used to validate the model (validation data). Success is measured in terms of the model’s ability to fit the validation observations. Gauch points out that predictive assessment of a multilocation trial is expected to provide more accurate yield estimates than a model chosen by postdictive criteria. Although many countries conduct extensive trials, little attention has been devoted to the most effective analyses of the data generated. This chapter reviews some of the conventional statistical analyses and stability methods for yield trials, Statistical and biological limitations are discussed. New methodologies for analyzing multilocation trials as well as multivariate analyses for assessing yield stability are presented.

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

57

II. CONVENTIONAL ANALYSIS OF VARIANCE Consider a trial in which the yield of G genotypes is measured in E environments each with R replicates. The classic model for analyzing the total yield variation contained in GER observations is the analysis of variance (Fisher, 1918, 1925). The within-environment residual mean square measures the error in estimating the genotype means due to differences in soil fertility and other factors, such as shading and competition from one plot to another. After removing the replicate effect when combining the data, the GE observations are partitioned into two sources: (a) additive main effects for genotypes and environments and (b)nonadditive effects due to genotypeenvironment interaction. The analysis of variance of the combined data expresses the observed (YG)mean yield of the ithgenotype at thejZhenvironment as

Yu = p

+ Gi + Ej + GEu + eu

(1)

where p is the general mean; Gi, Ej, and GEu represent the effect of the genotype, environment, and genotype-environment interaction, respectively; and eii is the average of the random errors associated with the rth plot that receives the ithgenotype in thejZhenvironment. The nonadditivity interaction as defined in (1) implies that the expected value of the ithgenotype in thePhenvironment (Y,>depends not only on the levels of G and E separately but also on the particular combination of levels of G and E. The presence of a significant genotype-environment interaction complicates interpretation of the results. Freeman (1985) mentioned two intuitive ways of overcoming this problem. The first is to examine the data and see if the interaction is due to one observed outlier. A high residual might be due to a recording error. In that case the outlier may be replaced by its expected value, using a model such as (1) without considering the interaction term. One can then determine whether the interaction in model (1) disappears. The second approach is to find a different scale on which the results can be expressed. Frequently, the logarithm transformation of the data removes interaction without rank changes and achieves both equality of variance within treatments and normality of residuals. However, crossover or rank change interactions cannot be removed by log transformation. When genotype-environment interaction exists and is not due to one outlier and cannot be removed by a suitable transformation, it should be assumed that an underlying model such as (1) is representing the data.

58

JOSE CROSSA

A. LIMITATIONS

Hill (1975) outlined the advantages of analysis of variance in obtaining unbiased estimates of genetic and genotype-environment interaction variance components. He failed, however, to recognize its limitations in describing further structures in the nonadditive component. In practice, if there is little variation in residual mean squares from one environment to another and the experiments are of equal size, the pooled error variance is found by averaging the residual mean squares of all environments. This combined experimental error is used to test the null hypothesis that the genotype differences are the same in all environments. This analysis is open to criticism, however, if error variances are heterogeneous. The F-test of the genotype-environment interaction mean squares against the pooled error variance is biased toward significant results. Cochran and Cox (1957 p. 554) point out that in agricultural experimentation, loss of sensitivity of the F-test is equivalent to discarding 10% to 20% of the data. A correct test of significance, by weighting each genotype mean by the inverse of its estimated variance, has been used by Yates and Cochran (1938) and Cochran and Cox (1957). The weighted analysis gives less weight to environments that have a high residual mean square. The sum of squares for genotype-environment interaction is inflated by errors in the weights; however, it can be reduced to a quantity that is distributed approximately as chi-square. The disadvantage of weighted analysis is that weights may be correlated to environment yield responses (with high-yielding environments showing higher error variance and low-yielding sites presenting lower error variances). This would mask the true performance of some genotypes in certain environments. It is recommended that less weight be assigned to agricultural environments of less importance (Patterson and Silvey, 1980). The genotype mean square is influenced by the pooled error variance, the variance of genotype-environment interaction, and the variance among true genotype means. The ratio of the genotype mean square to the genotype-environment interaction provides a test for the null hypothesis that there are no differences among the true genotype means. A criticism of this F-test is that, if the interaction variance is not the same for all of its components (some components of the interaction are much higher than others), too many significant results are obtained. In a trial of genotypes, this may occur when some genotypes are relatively unresponsive to a change in environment whereas others have a marked response. It is

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

59

recommended that the genotypes be further partitioned into a set of orthogonal components and that all of these components be tested for their interaction with the environment (Cochran and Cox, 1957). Often, the analysis of variance test of the significance of the genotypeenvironment interaction declares it not significant when in fact it is agronomically or genetically important and its sum of squares accounts for a large proportion of the total variation (Zobel et al., 1988). This may occur because the interaction contains a large number of degrees of freedom. One of the main deficiencies of the combined analysis of variance of multilocation yield trials is that it does not explore any underlying structure within the observed nonadditivity (genotype-environment interaction). Analysis of variance fails to determine the pattern of response of genotypes and environments. The valuable information contained in (G- I) (E-1) degrees of freedom is practically wasted if no further analysis is done. Since the nonadditive structure of a data matrix has a nonrandom (pattern) and random (noise) component, the advantages of the additive model are lost if the pattern component of the nonadditive structure is not further partitioned into functions of one variable each. Williams (1952), Mandel (1961, 1969, 1971), and Gollob (1968) have delineated methods for analyzing and interpreting two-way tables with interaction. They show that the sum of squares for interaction can be further partitioned in multiplicative components related to eigenvalues. The interaction part of Eq (1) can be expressed in the form GEU = klul;slj

+ k 2 ~ 2 ; ~ +2 j k 3 ~ 3 ; ~ +3 j . . .

(2)

h

then GEU=

knunisN, and n- 1

YU = p

+ Gi + Ej + (ck,vnisnj)+ eU h

(3)

n-l

where k, is the singular value of the nthaxis (kn2is the eigenvalue), u,; is the eigenvector of the Shgenotype for the nthaxis, si is the eigenvector of the

c h

fh

environment for the nthaxis, and

n- I

h

uni =

C si

=

I . This result links

n- I

the analysis of variance with the principal components analysis. This analysis is called Additive Main effect and Multiplicative Interaction (AMMI) and is considered in this chapter in the discussion of nonconventional analysis of variance.

60

JOSE CROSSA

B. COMPONENTS OF VARIANCE Analysis of variance of multilocation trials is useful for estimating variance components related to different sources of variation, including genotypes and genotype-environment interaction. Variance components have been widely used in genetics and plant breeding (Comstock and Moll, 1964; Cockerham, 1964; Gardner, 1964). In general, variance component methodology is important in multilocation trials, since errors in measuring the yield performance of a genotype arise largely from genotype-environment interaction. Therefore, knowledge of the size of this interaction is required to ( a ) obtain efficient estimates of genotype effects and (b) determine optimum resource allocations, that is, the number of plots and locations to be included in future trials. In a breeding program, variance component methodology is used to measure genetic variability and to estimate the heritability and predicted gain of a trait under selection. For balanced multilocation trials, that is, those with the same number of experimental units (genotypes or agronomic treatments) observed per site, estimation of the variance component is accomplished using the analysis of variance method. Each of the mean squares is known to estimate a linear function of the variance components defined in the model. These linear functions are called expected mean squares. By solving simultaneous equations, linear functions of the mean squares can be obtained that estimate each variance component. This method is limited to balanced data, and its main advantage is that it produces the best unbiased point estimators of the variance components (Graybill and Hultiquist, 1961). However, there is nothing intrinsic in the method to prevent negative estimates. The interpretation of a negative estimate of a nonnegative parameter creates controversy. In practice, the negative estimate can be accepted and used or a value of zero can be used instead. Thompson (1961, 1962) gives some rules for ignoring the negative component and reestimating the others. Genetic and genetic-environment variance components can be estimated by the maximum likelihood method. The disadvantage of these estimators, in the case of balanced data, is that they are biased downward (Patterson and Thomson, 1975). This problem can be overcome by using the restricted maximum likelihood (REML) method (Robinson, 1987). This method is analogous to the analysis of variance, and both produce identical estimators for balanced data. For unbalanced experiments, including incomplete block designs, estimating the expected mean squares can be difficult, and the analysis of variance method for variance component estimation is not necessarily a

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

61

desirable approach. Unbalancedness in multilocation trials can have many different causes, including shortage of seed, testing of some genotypes only at some locations (or in some years), and the addition of new genotypes to the trial system and discarding of others. General methods for calculating variance components in nonorthogonal data by means of REML analysis have been developed by Patterson and Thomson (1971, 1975).

Ill. JOINT LINEAR REGRESSION Another important model for analyzing and interpreting the nonadditive structure (interaction) of two-way classification data is the joint linear regression method. This approach has been extensively used in genetics, plant breeding, and agronomy for determining yield stability of different genotypes or agronomic treatments. The genotype-environment interaction is partitioned into a component due to linear regression (b;)of the ithgenotype on environmental mean and a deviation (d$: (GE)jj = biEj

+ djj

(4)

and Y j j= p

+ G; + Ej + (b;Ej + d j j )+ e j j

(5)

This model uses the marginal means of the environments as independent variables in the regression analysis and restricts the interaction to a multiplicative form. It was first proposed by Yates and Cochran (1938) in their analysis of a barley yield trial. The method divides the (G-1) (E-1) df for interaction into G-1 df for heterogeneity among genotype regressions and the remainder (G-1) (E-2) for deviation. Further details about interaction are obtained by regressing the performance of each genotype on the environmental means. Eberhart and Russell (1966) proposed pooling the sum of squares for environments and genotype-environment interactions and subdividing it into a linear effect between environments (with 1 df), a linear effect for genotype-environment (with G- 1 df), and a deviation from regression for each genotype (with E-2 df). Thus, not until the 1960s was it possible to solve the intractable problem of genotype by environment interaction by means of a regression approach. Part of the genotype’s performance across environments or genotype stability is expressed in terms of three empirical parameters: the mean performance, the slope of the regression line, and the sum of squares

62

JOSE CROSSA

deviation from regression. Although joint regression has been principally used for assessing the yield stability of genotypes in a plant breeding program, it may also be used for agronomic treatments. It has also been used to estimate biometrical genetical parameters (Bucio Alanis et al., 1969). When attention is focused on environments, the converse analysis may be performed by regressing each environment’s yields on the genotype means (Fox and Rathjen, 1981). Freeman (1973), Hill (1975), and Westcott (1986) have provided comprehensive reviews of regression methods for studying genotypeenvironment interactions. Several statistical and biological limitations of the regression method should be noted.

A. STATISTICAL A N D BIOLOGICAL LIMITATIONS The first statistical criticism of regression analysis is that the genotype mean (x variable) is not independent of the marginal means of the environments ( y variable). Regressing one set of variables on another that is not independent violates one of the assumptions of regression analysis (Freeman and Perkins, 1971; Freeman, 1973). This interdependence may be a major problem for small numbers, but not when the number of genotypes is large (say 15 to 20). If the standard set for stable yield is based on very few genotypes (say 4), each estimated stability coefficient involves regressing one genotype on an average to which it contributes one-fourth. Biological and algebraic interdependency also exists between slopes and sums of squares due to deviations from regression. Hardwick and Wood (1972) concluded that this is a necessary adjunct of the line-fitting procedure. The second statistical limitation is that errors associated with the slopes of genotypes are not statistically independent, because the sum of squares for deviation, with (C-1) (E-2) df, cannot be subdivided orthogonally among the G genotypes. The third statistical problem with regression analysis is that it assumes a linear relationship between interaction and environmental means. When this assumption is violated,the effectiveness of the analysis is reduced, and results may be misleading (Mungomery et al., 1974). In fact, the analysis requires that a high proportion of the genotype by environment effects should be attributable to linear regression (Perkins, 1972; Freeman, 1973). A nonlinear relationship between interaction and environmental effects has been proposed by Pooni and Jinks (1980), and Hill and Baylor (1983) have used an orthogonal contrast analysis of variance that subdivides the

63

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

variation over environments (years and sites) for each entry into sources due to environment linear and quadratic effects. Freeman and Perkins (1971) have criticized Eberhart and Russell’s partitioning of the pooled sum of squares for environments and genotypeenvironment interaction, noting that the 1 df sum of squares for the linear component between environments is the same as the total sum of squares for environments with E-1 df. A major biological problem with regressing genotype means on environmental means arises when only a few very low or very high yielding sites are included in the analysis. The fit of a genotype may be largely determined by its performance in those few extreme environments, with possibly misleading results (Hill and Baylor, 1983; Westcott, 1986). An example is presented by Westcott (1986) from the barley yield trial data of Yates and Cochran (1938), in which regression coefficients for the yield of the genotypes were calculated for all of the trials and for all except the highestand lowest-yielding site (Table I). The exclusion of one extreme point had a strong influence on the slope of genotypes 2 , 4 , and 5 , even though the lowest-yielding site was only 41.1 units apart from the grand mean. Crossa (1988) found that excluding 1 very low yielding site out of 20 or 1 high-yielding site out of 17 influenced the estimates of slopes and deviations from regression for some genotypes. The performance of some genotypes at only one site overshadowed their general response at most of the other sites. The author concluded that regression analysis should be used with caution when the data set includes results from a few extremely low or high yielding locations. Another biological criticism of the regression method is that the relative stability of any two genotypes depends not only on the particular set of locations included in the analysis but also on the other genotypes that are included in the regression calculation. It has been shown that the stability of a genotype depends on the mean performance of the group with which Table I Regression Coefficients of Five Barley Genotypes” Genotype

All sites included

Excluding highestyielding site

Excluding lowestyielding site

1 2 3 4 5

0.84 0.99 0.95 1.61 0.61

0.87 0.79

0.93 0.96 0.99 1.46 0.65

* From Wescott (1986).

1

.oo

1.92 0.43

64

JOSE CROSSA

that entry is being compared (Knight, 1970; Witcombe and Wittington, 1971; Mead et al., 1986; Crossa, 1988). Furthermore, it is possible that the ranking of two genotypes' stability coefficients may be reversed when they are compared with two other sets of genotypes. The stability of a particular genotype is unsatisfactory if its response is different from the mean response of the group with which it is being compared (Easton and Clements, 1973). This can be seen in Table 11, which gives the deviations from regression for 6 entries, considered ( a ) as members of the original set of 25 entries and (b) as an isolated group. It can be seen that the entry Raven x 65 RN 85 was originally a stable line (604) but appears unstable when considered as a member of the subset of 6 entries (6078). Crossa (1988) estimated Eberhart and Russell's stability parameters for genotypes considered, along with others, as a subset of the original group of 27 entries. When 7 genotypes were considered in isolation, deviation from regression of some genotypes changed drastically. This result confirmed that the yield stability of one entry, as determined by regression, varied according to the average response of the rest of the group. The author also pointed out that, in trying to determine which genotype is superior, plant breeders have difficulty reaching a compromise between the yield mean, slope, and deviation from regression, because the genotype's response to environments is intrinsically multivariate and regression tries to transform it into a univariate problem (Lin et al., 1986). An alternative approach to overcoming the dependency present in the regression analysis-one especially suitable for agronomic treatments-is to consider the joint distribution of a pair of treatments, say A and B, and to regress the yield differences (A-B)on the mean yield (A-B/2) (R. Mead,

Table I1 Deviation from Regression of 6 Genotypes when Considered as Members of the Original Group of 25 Genotypes and as an Isolated Group" Genotype

Member of 25 genotypes

Isolated group

Aotea Arawa 1020,Ol Nadadores 63 Hi-61 X Aotea Raven x 65 RN 85

8,942 5,267 9,645 8,235 12,402 604

1,347 4,839 3,486 6,425 9,543 6,078

From Easton and Clements (1973).

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

65

personal communication). Assuming an approximately linear relationship between both treatments, a positive slope would indicate that B is more stable than A. If a large percentage of the genotype-environment interaction sum of squares can be explained by the heterogeneity of regressions, then the joint regression method can efficiently describe the pattern of adaptation in the response of genotypes. However, Baker (1969), Byth et al. (1976), Eagles and Frey (1977), and Shorter (1981) reported that a very small portion (9- 16%) of the genotype-environment sum of squares is attributable to linear regression in various situations. Shorter (1981) concluded that, if this is the most common situation in field crops, the joint regression method of analysis is of little value. Moll ef al. (1978) studied the interaction of several populations of maize with environments, using the Eberhart and Russell procedure with the modification of Mather and Caligari (1974). The interaction sum of squares was divided into two parts: differences among genotypes in their variability among environments and differences in correlations among pairs of entries. Moll et al. found that heterogeneity of regression coefficients among genotypes may be due to heterogeneity of variance. Using results from Bruckner and Frohberg (1987) on kernel weight of 20 spring wheats tested in 15 environments, Baker (1988a) pointed out that the high correlation between regression coefficients and estimated variances over environments suggests that heterogeneity of slopes is explained by heterogeneity of variance. B. OTHERMEASUREMENTS OF YIELDSTABILITY Other methods of determining genotype stability are based on genotype-environment interaction effects and are briefly examined next. Plaisted and Peterson (1959) computed combined analysis of variance for each pair of genotypes included in a trial. The variance component of the genotype-environment interaction is estimated for each pair and each genotype. The genotype with the smallest mean variance component contributes less to the total interaction and is considered the most stable. Wricke (1962, 1964) defined the concept of ecovalence as the contribution of each genotype to the genotype-environment interaction sum of squares. The ecovalence (W;)or stability of the ithgenotype is its interaction with environments, squared and summed across environments, and expressed as w.= [_Yu. .- -y.1 . -- y J. - -y..32 (6) where

r, is the mean performance of genotype i in thejZhenvironment and

66

JOSE CROSSA

-

Yi. and y.j_are the genotype and environment mean deviations, respectively, and Y.. is the overall mean. Accordingly, genotypes with low ecovalance have smaller fluctuations from the mean across different environments and are therefore more stable. Shukla (1972) defined the stability variance of genotype i as its variance across environments after the main effects of environmental means have been removed. Since the genotype main effect is constant, the stability variance is based on the residual (GEu + eu) matrix. Lin and Binns (1988) defined the superiority measure (Pi)of the ithgenotype as the mean square of distance between the ith genotype and the genotype with maximum response as Pi = [n(Y;. - M . . ) 2 + (Yo - Y;.+ Mj. + M..)2]/2n

(7)

where Yu is the average response of the ithgenotype in thefh environment, Yi. is the mean deviation of genotype i, Mjis the genotype with maximum response among all genotypes in thejth location, and n is the number of locations. The first term of the equation represents the genotype sum of squares, and the second term is the genotype-environment sum of squares. The smaller the value of Pi, the less its distance to the genotype with maximum yield and the better the genotype. A pairwise genotypeenvironment interaction mean square between the maximum and each genotype is also calculated. This method is similar to that of Plaisted and Peterson (1959), except that ( a )the stability statistics are based on both the average genotypic effects and genotype-environment interaction effects and (b) each genotype is compared only with the one maximum response at each environment. Lin et al. (1986) reviewed nine stability measurements frequently used in biological research and grouped them into four categories, depending on whether they are based on the deviations from the average genotype effect or on the genotype-environment interaction effects. The authors defined three different parametric concepts of stability statistics. A genotype is stable if ( a ) its among-environment variance is small; ( 6 ) its response to environment is parallel to the mean response of all genotypes included in the trial; and (c) the residual mean square from the regression model on the environmental index is small. Stability methods based on the genotypeenvironment interaction sum of squares correspond to type b, whereas the Eberhart and Russell method is type c. As the authors point out, these parametric concepts of stability are relatively simple and address only some aspects of stability without giving an overall picture of the genotype’s response. A genotype may be considered to have type b stability and simultaneously type c instability. Since a genotype’s response to environment is multivariate, Lin et al. (1986) proposed using cluster analysis to classify genotypes.

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

67

C. STABILITY, RISK,A N D ECONOMIC ANALYSIS One of the main aims of breeders and agronomists is to recommend to farmers new agriculture production alternatives (genotypes, agronomic treatments, and cropping systems) that are stable under different environmental conditions and minimize the risk of falling below a certain yield level. Subsistence farmers using low levels of inputs in unfavorable environments tend to be reluctant to adopt new technology. Given the uncertainty of their circumstances, these farmers’ main concern is not so much to increase production as to avert catastrophe. Conventional regression analysis considers only three components of stability: (a) response to changing environment (regression coefficient); (b)yield variability; and (c) mean yield level. However, this assessment of stability is incomplete and inappropriate unless it is related to risk probability (Barah et a / . , 1981; Mead et al., 1986). The concept of risk efficiency of a particular genotype involves a tradeoff between its average yield and variance. A genotype is risk efficient if no other genotype has the same yield with lower variance or the same variance with higher mean yield. The mean-standard deviation analysis provides a method in which the benefits of reduced yield variability are measured against loss in yield (Binswanger and Barah, 1980). This analysis requires that the breeder or farmer specify how mean and standard deviation are “trade off.” Mean-standard deviation analysis translates the stability parameters of a genotype (slope and deviation from regression) into economic benefit for the farmer. Mean-standard deviation analysis and regression analysis were compared on yield data of pearl millet genotypes tested for 5 years in India and Pakistan (Witcombe, 1988). The results of both analyses were similar in all the environments and the standard deviation predicted well the values of deviation from regression. In comparing the risk stability of two cropping systems (two crops versus one crop), Mead et al. (1986) define risk as the probability of yield falling below certain prespecified levels. The authors describe a general method of expressing stability related to risk probability by adjusting a bivariate distribution to the data and then estimating a theoretical continuous risk curve. The method can be used for assessing the risk stability of any two genotypes or agronomic treatment. The stochastic dominance procedure (Anderson, 1974; Menz, 1980) ranks different agricultural alternatives according to farmers’ risk aversion and selects those with high risk efficiency. It is assumed that each alternative has a probability distribution of yield,Ai), and therefore a cumulative ~ firstdistribution function, F(i). Then, the Ai) is said to dominate A J by

68

JOSE CROSSA

degree stochastic dominance if all the values of the yield distribution of alternative i are greater than those of alternative j. Second- and thirddegree stochastic dominance appear when the distributions of yields are not easily separated. The importance of stochastic dominance is, unlike mean-standard deviation analysis, that the breeder or farmer does not have to specify the trade-off between average yield and variance. Under yield uncertainty a major problem is how to make trade-offs among conventional stability statistics, for example, mean yield, slope, and deviation from regression. The central concept in safety-first decision strategies is the assumption that breeders and farmers prefer genotypes with a small chance of producing small yields. Eskridge (1990) addressed this issue by developing safety-first selection indices based on four different stability approaches: (a) the variance of a genotype across environments (EV); (b) the regression coefficient used by Finlay and Wilkinson (1963) (FW); (c) the stability variance of Shukla (1972) (SH); and (4the regression coefficient and deviation from regression defined by Eberhart and Russell (1966)(ER). The rank correlations between the mean genotype rankings and the four selection indices show that FW, SH, and ER produce similar rankings (>0.65). The mean ranking, on the other hand, is poorly correlated with EV (0.152) and only moderately correlated with FW, SH, and ER (0.45 < rank correlations < 0.7). Only one genotype was ranked near the bottom for all indices. A safety-first index is useful for selecting genotypes in the presence of genotype-environment interaction (Eskridge, 1990)because: (a) it weights the importance of stability relative to yield; (b) it can be used with different types of univariate stability statistics for any trait; and (c) it is more likely to identify superior varieties when high costs are associated with low yields.

IV. CROSSOVER INTERACTIONS Interaction in the classic sense exists because the responses of genotypes are not parallel over all environments. In agricultural production, changes in a genotype’s rank from one environment to another are important. These are called crossovers or qualitative interactions, in contrast to noncrossovers or quantitative interactions (Baker, 1988b,c; Gail and Simon, 1985). With a qualitative interaction, genotype differences vary in direction among environments, whereas with quantitative interactions, genotypic differences change in magnitude but not in direction among environments. If significant qualitative interactions occur, subsets of genotypes are to be recommended only for certain environments, whereas

69

STATISTICAL ANALYSES O F MULTILOCATION TRIALS

with quantitative interactions the genotypes with superior means can be used in all environments. Therefore, it is important to test for crossover interactions (Baker, 1988b). Consider, for example, genotypes A and B tested in environments 1 and 2 Let and 7 A 2 be the means of genotype A in environments 1 and 2, respectively, and and be the means of genotype B in environments 1 and 2, respectively. The genotype effects in each environment are defined as

rA,

rBl rB2

dl = F A 1 d2 = Y.42 -

FBI F,2

No interaction and all types of interactions can be illustrated in the space of genotype effects by plotting d2 and dl (Fig. 1). Then, the line where dl = d2 represents the situation in which there are differences in genotype means but not in genotype-environment interac-

No genotype effect at orlgln A*

over Interaction 7 d7<0

B4

FIG.1. The space of genotype effects d , and d2. From Gail and Simon (1985).

70

JOSE CROSSA

tion; when d l = d2 = 0 (at the origin), there are no differences in genotype means and no interaction. Qualitative or crossover interaction will occur in the second and fourth quadrants when d1<0 and d2>0 or d1>0 and dz
V. MULTIVARIATE ANALYSES OF MULTILOCATION TRIALS Multivariate analysis has three main purposes: ( a ) to eliminate noise from the data pattern (i.e., to distinguish systematic from nonsystematic variation); ( b )to summarize the data; and (c) to reveal a structure in the data. In contrast with classic statistical methods, the function of multivariate analysis is to elucidate the internal structure of the data from which hypotheses can be generated and later tested by statistical methods (Williams and Gillard, 1971, cited by Gauch, 1982b). Multivariate analyses are appropriate for analyzing two-way matrices of G genotypes and E environments. The response of any genotype in E environments may be conceived as a pattern in E-dimensional space, with the coordinate of an individual axis being the yield or other metric of the genotype in one environment. Two groups of multivariate techniques have been used to elucidate the internal structure of genotype-environment interaction:

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

71

1. Ordination techniques, such as principal components analysis, principal coordinates analysis, and factor analysis, assume that data is continuous. These techniques attempt to represent genotype and environment relationships as faithfully as possible in a low-dimensional space. A graphical output displays similar genotypes or environments near each other and dissimilar items are farther apart. Ordination is effective for showing relationships and reducing noise (Gauch, 1982a, 1982b). 2. Classification techniques, such as cluster analysis and discriminant analysis, seek discontinuities in the data. These methods involve grouping similar entities in clusters and are effective for summarizing redundancy in the data.

A. PRINCIPAL COMPONENTS ANALYSIS

Principal components analysis is one of the most frequently used multivariate methods (Pearson, 1901; Hotelling, 1933; Gower, 1966). Its aim is to transform the data from one set of coordinate axes to another, which preserves, as much as possible, the original configuration of the set of points and concentrates most of the data structure in the first principal components axes. In this process of data reduction, some original information is inevitably lost. Principal components analysis assumes that the original variables define a Euclidean space in which similarity between items is measured as Euclidean distance. This analysis can effectively reduce the structure of a two-way genotype-environment data matrix of G (genotypes) points in E (environments) dimensions in a subspace of fewer dimensions. The matrix can also be conceptualized as E points in G dimensions. The model is written as

where the terms are defined as in (2). Under certain conditions, principal components analysis is a generalization of the linear regression analysis (Williams, 1952; Mandel, 1969; Johnson, 1977; Digby, 1979). Mandel (1971) analyzed a two-way data matrix by applying the AMMI analysis. The first step in his solution was to conduct an analysis of variance with terms for two main effects and the interaction between rows and columns. The residual table (i.e., the row-column interaction) was partitioned into multiplicative terms where eigenvalues and eigenvectors are obtained. Finally, the relationships between the first two eigenvectors, which accounted for most of the variation, were examined.

72

JOSE CROSSA

Freeman and Dowker (1973) used principal components analysis to interpret the causes of genotype-environment interaction in carrot trials. Hirosaki et af. (1975) found that principal components analysis was more efficient than the linear regression method in describing genotypic performance. On the other hand, Perkins (1972) reported that principal components analysis was not useful for studying the adaptation of a group of inbred lines of Nicotiana rustica. Principal components analysis combined with cluster analysis was effective in forming subgroups among 29 populations of faba bean (Vicia faba L.), which differed in mean performance and response across environments (Polignano et af., 1989). Principal components have also been used by Suzuki (1968), Goodchild and Boyd (1979, and Hill and Goodchild (1981). Zobel et al. (1988) presented analysis of variance and principal components analysis for seven soybean genotypes yield-tested in 35 environments (Table 111). The genotype by environment interaction sum of squares of the analysis of variance was large but not significant. The principal components analysis with the first three principal axes accounting for 76% of the variation is found to be statistically efficient but undesirable for describing the additive main effects. Kempton (1984) used AMMI analysis for summarizing the pattern of genotype responses across environments with different levels of nitrogen. The first principal component is the axis that maximizes the variation among genotypes. The second principal component is perpendicular to the first and maximizes the remaining variation. The display of the genotypes and environments along the first two principal component axes for the interaction table of residuals is called the biplot (Gabriel, 1971, 1981). Table I11 Analysis of Variance and Principal Components Analysis of a Soybean Trial' Principal components analysis

Analysis of variance Source

df

Mean square

Treat Geno Env GE Error

244 6 34 204 667

574*** 1499***

a

3105***

125 111

From Zobel et al. (1988).

*** Significant at 0.001 probability level.

Source Treat PCA 1 PCA 2 PCA 3 Res Error

df 244 41 39 37 I27 667

Mean square 574*** 2599***

471***

264*** 44 111

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

73

Figure 2 represents the biplot of 12 genotypes and 14 environments (7 sites each with low and high nitrogen levels). The component of the total interaction due to nitrogen level is small, since the biplot shows that highand low-nitrogen trials are closely associated. Environments represented by vectors of similar orientation but different length usually give similar genotype rankings. Zobel et af. (1988) and Crossa et af. (1990) used the same model to analyze a series of soybean and maize trials, respectively. Additive main effects for genotype and environments are first fitted by the analysis of variance. Then, multiplicative effects for genotype by environment interaction are calculated by principal components analysis. The biplot of the model helps to visualize the overall pattern of response as well as specific interactions between genotypes and environments. Ordination techniques such as principal components analysis may have limitations. First, in reducing dimensionality of multivariate data, distortions may sometimes occur. If the percentage of variance accounted for by the first principal components axes is small, individuals that are really far apart may be represented by points that are close together (Gower, 1967).

*i4

I 3L

Axla 1

FIG.2. First two principal component axes for genotypes ( 0 )and environments based on residual yields. (Sites are coded from I to 5 ; L and H are trials with low and high nitrogen levels, respectively.) From Kernpton (1984).

74

JOSE CROSSA

In this case, higher axes may be inspected to identify points with large displacement not revealed in lower dimensions. Second, a lack of correlation between variables prevents few dimensions from accounting for most of the variation (Williams, 1976). Third, sometimes the components do not have any obvious relationship to environmental factors. Fourth, contrary to analysis of variance, which assumes a complete additive model and treats the interaction as a residual, principal components analysis assumes a complete multiplicative model without any description of the main effects of genotypes and environments (Zobel et al., 1988). This is important in the context of multilocation trials, in which genotype means are of primary interest. Principal components analysis confounds the additive (main effects of genotypes and environments) structure of the data with the nonadditivity (genotype-environment interaction). The fifth limitation is that nonlinear association in the data prevents principal components analysis from efficiently describing the real relationships between entities (Williams, 1976). The linear regression method uses only one statistic, the regression coefficient, to describe the pattern of response of a genotype across environments, and, as mentioned previously, most of the information is wasted in accounting for deviation. Principal components analysis, on the other hand, is a generalization of linear regression that overcomes this difficulty by giving more than one statistic, the scores on the principal component axes, to describe the response pattern of a genotype (Eisemann, 1981).

B. PRINCIPAL COORDINATES ANALYSIS Principal coordinates analysis (Gower, 1966)is a generalization of principal components analysis in which any measure of similarity between individuals can be used. Its objective and limitations are similar to those of principal components analysis. Principal coordinates analysis was used in combination with cluster analysis (“pattern” analysis) to study the adaptation of soybean lines evaluated across environments in Australia (Mungomery et al., 1974; Shorter et al., 1977). The authors found these analyses to be useful for helping breeders choose among test sites for early screening of breeding lines. Principal coordinates analysis was employed to examine the use of a reference set of genotypes to monitor genotype-environment interaction (Fox and Rosielle, 1982a) and also to assess methods for removing environmental main effects to provide a description of environments (Fox and Rosielle, 1982b).

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

75

A spatial method for assessing yield stability, in which principal coordinates analysis is based on a suitable measure of similarity between genotypes, has been proposed by Westcott (1987). As pointed out by Crossa (1988), the method has several advantages: ( a ) it is trustworthy when used for data that include extremely low or high yielding sites; ( b ) it does not depend on the set of genotypes included in the analysis; and (c) it is simple to identify stable varieties from the sequence of graphic displays. The spatial method has been extensively used by Crossa et al. (1988a,b, 1989) to assess the yield stability of CIMMYT’s maize genotypes evaluated across international environments.

C. FACTOR ANALYSIS Factor analysis is an ordination procedure related to principal components analysis, the “factors” of the former being similar to the principal components of the later. A large number of correlated variables is reduced to a small number of main factors (Cattell, 1965),and variation is explained in terms of general factors common to all variables and in terms of factors unique to each variable. The axes of the general factors may be rotated to oblique positions to conform to hypothetical ideas. Factor analysis has been used to understand relationships among yield components and morphological characteristics of crops (Walton, 1972; Seiler and Stafford, 1985). Jardine et al. (1963) used an oblique rotation to indicate four relatively independent factors related to bread wheat baking quality. Peterson and Pfeiffer (1989) applied principal factor analysis to study the underlying structures and relationships of test sites, based on winter wheat performance. The authors grouped the original 56 locations into seven regions, which can be considered megaenvironments for winter wheat adaptation. The association between secondary factors was used to identify transitional environments between the seven major regions.

D. CLUSTER ANALYSIS Cluster analysis is a numerical classification technique that defines groups or clusters of individuals. Two types of classification can be distinguished. The first is nonhierarchical classification, which assigns each item to a class. Relationships among classes are not characterized, so this type is useful in the early stages of data analysis. The second type is hierarchical

76

JOSE CROSSA

classification, which groups individuals into clusters and arranges these into a hierarchy for the purpose of studying relationships in the data. Cluster analysis requires a measure of similarity between the individuals to be classified, and it imposes a discontinuity in the data. The method has been used to study genotype adaptation by simplifying the pattern of responses and to subdivide genotypes and environments into more homogeneous groups. Comprehensive reviews of the application of cluster analysis to the study of genotype-environment interactions can be found in Lin et af. (1986) and Westcott (1987). Some of the disadvantages of cluster analysis are: (a) numerous hierarchical grouping algorithms exist, and each of them may produce different cluster groups; (b)the truncation level of the classificatory hierarchies may be decided arbitrarily; (c) many different similarity measures can be used (Lin et af., 1986, listed nine), yielding different results; and (d) cluster analysis may produce misleading results by showing structures and patterns in the data when they do not exist (Gordon, 1981, cited by Westcott, 1987).

VI. AMMI ANALYSIS The additive main effect and multiplicative interaction (AMMI) method integrates analysis of variance and principal components analysis into a unified approach (Bradu and Gabriel, 1978; Gauch, 1988). It can be used to analyze multilocation trials (Gauch and Zobel, 1988; Zobel et al., 1988; Crossa et al., 1990). AMMI analysis first fits the additive main effects of genotypes and environments by the usual analysis of variance and then describes the nonadditive part, genotype-environment interaction, by principal components analysis. The AMMI model is given by Eq. (3). The AMMI method is used for three main purposes. The first is model diagnosis. AMMI is more appropriate in the initial statistical analysis of yield trials, because it provides an analytical tool for diagnosing other models as subcases when these are better for a particular data set (Bradu and Gabriel, 1978; Gauch, 1985). The second use of AMMI is to clarify genotype-environment interactions. AMMI summarizes patterns and relationships of genotypes and environments (Kempton, 1984; Zobel et al., 1988; Crossa et af., 1990). The third use is to improve the accuracy of yield estimates. Gains have been obtained in the accuracy of yield estimates that are equivalent to increasing the number of replicates by a factor of two to five (Zobel et al., 1988; Crossa et af., 1990). Such gains may be used to

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

77

reduce costs by reducing the number of replications, to include more treatments in the experiment, or to improve efficiency in selecting the best genotypes. This last benefit has obvious implications for breeding programs and particularly for maize hybrid testing systems, in which designs with fewer replicates per location are used (Bradley et al., 1988). A. AMMI ANALYSIS WITH PREDICTIVE SUCCESS Traditional analysis of variance of multilocation trials is intended to forecast agricultural performance, but it focuses only on postdictive assessment of genotype yield responses without evaluating the model’s predictive accuracy with validation data not used in constructing the model. Gauch (1985, 1988) emphasized the model’s success in predicting validation data (prediction criteria), in contrast to its success in fitting its own data (postdiction criteria). Because multilocation trials are used for selecting genotype or agronomic treatments for farmers’ fields in new environments, model evaluation should measure predictive success. Gauch proposed that AMMI analysis be used with prediction criteria. Prediction assessment consists of splitting data into two subgroups, modeling data and validation data, and comparing the success of several models by computing their sum of squared difference (SSD) between model predictions and validation data. A small value of SSD indicates good predictive accuracy. Several models are then constructed and compared empirically in terms of their ability to predict the validation data: AMMIO, which estimates only the additive main effects of genotypes and environments and retains none of the principal components axes (PCA); AMMI1, which combines the additive main effects from AMMIO with the genotype-environment interaction effects estimated from the first principal component axis (PCA 1);AMMI2 and so on, up to the full model with all PCA axes. The predictive values of the full model are equal to the average of the replicates selected at random for modeling. Results of postdictive AMMI analysis of a trial consisting of 15 soybean genotypes evaluated in 15 environments are given in Table IV (Gauch, 1988). The postdictive evaluation using F-test at 5% showed that three PCAs of the interaction are significant; therefore, the model, including the two main effects, has 103 df. However, this information includes pattern and noise (systematic and nonsystematic variation). Prediction assessment, on the other hand, does discriminate between pattern and noise and indicates AMMI with one interaction PCA as the best predictive model (Fig. 3). This model has 55 df-14 for genotypes, 14 for environ-

78

JOSE CROSSA Table IV AMMI Analysis for a Soybean Trial"

df 14 14 196 27 25 23 21

Environment Genotype G x E PCA I PCA 2 PCA 3 PCA 4 Residual Error

100

210

ss

MS

38,798 2,552 6,880 2,348 1,250 1,010 736 1,536 4,649

2,77 I *** 182*** 35*** 87*** 50*** M***

35 15

22

From Gauch (1988).

*** Significant at 0.001 probability level.

7000

J

72M) -

0

m m

7400-

E

Ti

1

7600-

7m-

Bwo I

1

28

55

80

103

124

143

160 175 188

208 22013/4

FIG.3. Sum of squared difference (SSD) between model prediction and validation data for IS models (AMMIO with 28 df to the full model with 224 df). From Gauch (1988).

79

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

ments, and 27 for the interaction PCA 1. Further interaction PCAs will capture mostly noise and therefore do not help to predict validation observations. The interaction of 15 soybean genotypes with 15 environments is best predicted by the first principal component of genotypes and environments. Thus, the model is

Yo = p

+ Gi + Ej + k l v , i ~ +l j eii

(9)

From (9) it can be seen that, when a genotype or an environment has an interaction PCA score of nearly 0, it has a small interaction. When both have PCA scores of the same sign, their interaction is positive; if different, their interaction is negative. For data in which AMMI1 is found to be the best predicted model, a graphical display of the genotype and environment interaction PCA I and their mean effects should be useful for revealing favorable patterns in genotype response across environments. Figure 4 gives the mean on the x axis and the AMMI interaction PCA 1 scores on the y axis of 17 maize genotypes tested in 36 environments (Crossa et al., 1990). Three groups of genotypes with different genetic composition can be seen: ( a )group 1 includes genotypes 13, 14,15, and 17, which contain temperate germplasm from the U.S. Corn Belt and southern Europe; (b) group 2 comprises genotypes 1, 2,3,4, and 5, which are from subtropical regions and have intermediate maturity; and (c) group 3 contains genotypes 6 to 12 and 16, which are derived from lowland tropical maize types from Mexico and the Caribbean islands. Interaction PCA 1 scores arrange the environments in a sequence from tropical environments 3020-

00

'0:

00

0

04

00

5 -10:

a

0 0

0

.

-20-

CO

-301

017

-40

013

015 014

0

-60

0

1880

2880

3880

4880

5880

6880

I

7880

Mean (kg ha-')

FIG. 4. Plot of the means (kg ha-') and PCA 1 scores of 17 maize genotypes ( 0 ) and 36 environments (0). From Crossa er a / . (1990).

80

JOSE CROSSA

(positive PCA 1) to temperate environments (negative PCA 1). The two temperate environments with the greatest negative PCA 1 scores favor temperate germplasm (group 1). At the other extreme of the diagram, tropical environments tend to favor genotypes from group 2 and 3.

VII. OTHER METHODS OF ANALYSIS Many other approaches might be employed for studying genotypeenvironment interactions. Several of them have not been examined systematically or extensively used for different crops. In most yield trials, environments are measured by the average yield of the genotypes or agronomic treatments. However, it is important to collect, analyze, and interpret physiological and environmental variables for ( a ) studying their relationships with genotype performance and (b)understanding the causes of the observed genotype-environment interaction (Westcott, 1986; Eisemann and Mungomery, 1981). The differential physiological responses of genotypes to edaphic and climatic factors, especially those related to nutrient efficiency and stress tolerance, are relevant to genotype-environment interaction (Baker, 1988a,c). The multilinear regression method, in which environmental data are used as independent variables, can be employed for predictive purposes (Knight, 1970; Feyerherm and Paulsen, 1981; Haun, 1982). Hardwick (1972) and Hardwick and Wood (1972) used physiological and environmental variables to develop a predictive multiple linear regression model. Principal components analysis, combined with multiple regression, may be useful for reducing the number of environmental variables to be included in the final analysis (Perkins, 1972). Principal components analysis was used by Holland (1 969) to summarize and interpret environmental data. However, it is of limited use, because the importance of a certain variable in the analysis may not be related to the extent of genotype response (Eisemann and Mungomery, 1981). Most of the exploratory or geometrical methods can be applied to the analysis of multilocation trials, although their use for this purpose has not been investigated. Ordination techniques, such as weighted average (Rowe 1956), polar ordination (Bray and Curtis, 1957), reciprocal average (Fisher, 1940), and detrended correspondence analysis (Hill and Gauch, 1980) have been used in community ecology to discover structures in data matrices (Gauch, 1982b). Their use in examining the pattern of genotype (or environment) responses needs investigation.

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

81

Canonical discriminant analysis has been used to allocate environments according to their interaction with genotypes (Seif ef al., 1979). The stratified ranking method was used by Fox et al. (1990)for analyzing general adaptation of a large international triticale data base. The technique scores the number of locations for which each line occurred in the top, second, and bottom one-third of the entries in each trial. A line that occurred in the top one-third of the entries across locations was considered well adapted. Unbalanced data often occur in multilocation trials as a result of ( a ) missing plots or ( b ) combining results of different experiments that do not have the same set of treatments. For incomplete data, missing plot values can be fitted, and the genotype-environment interaction sum of squares can be further partitioned into principal components (Freeman, 1975). An algorithm for inputting missing values and then fitting the additive main effect and multiplicative interaction (AMMI) model has recently been developed (Gauch and Zobel, 1990).

VIII. GENERAL CONSIDERATIONS AND CONCLUSIONS Data from multilocation trials help researchers estimate yields more accurately, select better production alternatives, and understand the interaction of these technologies with environments. Several methodologies have been presented for efficient statistical analysis of such data. For geneticists, plant breeders, and agronomists, parametric stability statistics, obtained by linear regression analysis, are mathematically simple and biologically interpretable. However, this method has major disadvantages: ( a ) it is uninformative when linearity fails; (6)it is highly dependent on the set of genotypes and environments included in the analysis; and (c) it tends to oversimplify the different response patterns by explaining the interaction variation in one dimension (regression coefficient), when in reality it may be highly complex. There is a danger in sacrificing relevant information for easy biological and statistical interpretation. A broad range of multivariate methods can be used to analyze multilocation yield trial data and assess yield stability. Although some of them overcome the limitations of linear regression, the results are often difficult to interpret in relation to genotype-environment interaction (as is the case with principal components analysis and cluster analysis). Certain multi-

82

JOSE CROSSA

variate techniques or a combination of them offer very relevant biological information and are statistically simple. The integration of certain ordination and classification multivariate methods into so-called “pattern” analysis and factor analysis and the biplot method are valuable tools for grouping environments or genotypes showing similar response patterns. The combination of analysis of variance and principal components analysis in the AMMI model, along with prediction assessment, is a valuable approach for understanding genotype-environment interaction and obtaining better yield estimates. Agronomic predictive assessment with AMMI can be used to analyze the results of on-farm trials. More research is needed to quantify the probability of successful selection of a genotype or agronomic treatment when using AMMI predictive value, compared with the probability of its selection based on the predictive value of other models. Only qualitative or crossover interactions are relevant in agriculture. Therefore, appropriate statistical analysis for quantifying and testing changes in rank from one environment to another is required. More attention has to be devoted to the collection, analysis, and interpretation of environmental and physiological variables. This will help to characterize particular genotypes and geographical regions and therefore better explain certain aspects of the interaction.

ACKNOWLEDGMENTS The author thanks Drs. Kent Eskridge, Paul Fox, and Hugh Gauch for their helpful comments on the manuscript.

REFERENCES Anderson, J. R. 1974. R e v . Marketing Agric. Econ. 42, 131-184. Azzalini, A., and Cox, D. R. 1984. J. R . Stat. Soc. 46,335-343. Baker, R. J. 1969. C a n . J . Plant Sci. 49,743-751. Baker, R. J. 1988a. I n “Proceedings of the Second International Conference on Quantitative Genetics.” Sinauer, Sunderland, Massachusetts. Baker, R. J. 1988b. C a n . J . Plant Sci. 68,405-410. Baker, R. J. 1988c. “IS1 Atlas of Science: Animal and Plant Sciences.” Barah, B. C., Binswanger, H. P., Rana, B. S., Rao, N. P. 1981. Euphytica 30,451-458. Binswanger, H. P., and Barah, B. C. 1980. Research Bulletin No. 3. International Crops Research Institute for the Semi-Arid Tropics (ICRISAT). Palanchero, India. Bradley, J. P., Knittle, K. N., and Troyer, A. F. 1988. J . Prod. Agric. 1, 34-38. Bradu, D., and Gabriel, K. R. 1978. Technometries 20,47-68. Bray, J. R., and Curtis, J. T. 1957. Ecol. Monogr. 27,325-349.

STATISTICAL ANALYSES O F MULTILOCATION TRIALS

83

Bruckner, P. L., and Frohberg, R. C. 1987. Crop Sci. 27,31-36. Bucio Alanis, L., Perkins, J. M., and Jinks, J. L. 1969. Heredity 24, 115-127. Byth, D. E., Eisemann, R. L., and De Lacy, I. H. 1976. Heredity 37,215-230. Cattell, 1965. Biometrics 21, 190-215. Cochran, W. G., and Cox, G. M. 1957. “Experimental Designs,” 2nd Ed. Wiley, New York. Cockerham, C. C. 1964. “Statistical Genetics and Plant Breeding,” Vol. 982, pp. 53-93. National Academy of Science-National Research Council, Washington, D.C. Comstock, R. E., and Moll, R. H. 1964. “Statistical Genetics and Plant Breeding,” Vol. 982, pp. 164-196. National Academy of Science-National Research Council, Washington, D.C. Crossa, J. 1988. Theor. Appl. Genet. 75,460-467. Crossa, J., Westcott, B., and Gonzalez, C. 1988a. Theor. Appl. Genet. 75,863-868. Crossa, J . , Westcoot, B., and Gonzalez, C. 1988b. Exp. Agric. 24,253-263. Crossa, J., Westcott, B., and Gonzalez, C. 1989. Euphytica 40,245-251. Crossa, J . , Gauch, H. G., and Zobel, R. W. (1990). Crop Sci. 30,493-500. Digby, P. G. N. 1979. J. Agric. Sci. 93, 81-86. Eagles, H. A., and Frey, K. J. 1977. Crop Sci. 17,253-256. Easton, H. S., and Clements, R. J. 1973. J. Agric. Sci. 80,43-52. Eberhart, S. A., and Russell, W. A. 1966. Crop Sci. 6,36-40. Eisemann, R. L. 1981. I n “Interpretation of Plant Respnse and Adaptation to Agricultural Environments.” Univ. of Queensland, St. Lucia, Brisbane. Eisemann, R. L., and Mungomery, V. E . 1981. In “Interpretation of Plant Response and Adaptation to Agricultural Environments.” Univ. of Queensland, St. Lucia, Brisbane. Eskridge, K. M. 1990. Crop Sci. 30,369-374. Feyerherm, A. M., and Paulsen, G. M. 1981. Agron. J. 73,277-282. Finlay, K. W., and Wilkinson, G. N . 1963. Aust. J. Agric. Res. 14,742-754. Fisher, R. A. 1918. Trans. R. SOC.Edinburgh 52,399-433. Fisher, R. A. 1925. ‘Statistical Methods for Research Workers.” Oliver & Boyd, London. Fisher, R. A. 1940. Ann. Eugenics 10,422-429. Fox, P. N., and Rathjen, A. J. 1981. Ausf. J . Agric. Res. 32,691-702. Fox, P. N., and Rosielle, A. A. 1982a. Crop Sci. 22, 1171-1175. Fox, P. N., and Rosielle, A. A. 1982b. Euphytica 31,645-656. Fox, P. N., Skovmand, B., Thompson, B. K., Braun, H.-J., and Cormier, R. 1990.Euphyfica 47,5744. Freeman, G. H. 1973. Heredity 31,339-354. Freeman, G . H. 1975. Appl. Star. 24, 1-46. Freeman, G. H., 1985. J . Appl. Star. l2,3-10. Freeman, G. H., and Dowker, B. D. 1973. Heredity 30,97-109. Freeman, G . H., and Perkins, J. M. 1971. Heredity 27, 15-23. Gabriel, K. R. 1971. Biometriku 58,453-467. Gabriel, K. R. 1981. “Interpreting Multivariate Data.” Wiley, Chichester. Gail, M., and Simon, R. 1985. Biometrics 41, 361-372. Gardner, C. 0. 1964. “Statistical Genetics and Plant Breeding,” Vol. 982. pp. 225-252. National Academy of Science-National Research Council, Washington, D.C. Gauch, H. G. 1982a. Ecology 63, 1643-1649. Gauch, H. G. 1982b. “Multivariate Analysis in Community Ecology,” 1st Ed. Cambridge Univ. Press, London and New York. Gauch, H. G. 1985. Mimeo 85-7. Cornell Univ. Ithaca, New York. Gauch, H. G. 1988. Biometrics 44,705-715. Gauch, H. G., and Zobel, R. W. 1990. Theor. Appl. Genet. (in press).

84

JOSE CROSSA

Gauch, H. G., and Zobel, R. W. 1988. Theor. Appl. Genet. 76, 1-10. Gollob, H. F. 1968. Psychometrika 33,73-116. Goodchild, N. A., and Boyd, W. J. R. 1975. Aust. J . Agric. Res. 26,209-217. Gordon, A. D. 1981. “Classification.” Chapman & Hall, London. Gower, J. C. 1966. Biometrika 53,325-338. Gower, J. C. 1967. Statistician 17, 13-28. Graybill, F. A., and Hultquist, R. A. 1961. Ann. Math. Stat. 32,261-269. Hardwick, R. C. 1972. Nature (London) New Biol. 236, 191-192. Hardwick, R. C., and Wood, J. T. 1972. Heredity 28,209-222. Haun, J. R. 1982. Agric. Meteor. 27, 191-207. Hill, J. 1975. J. Agric. Sci. 85,477-493. Hill, J., and Goodchild, N. A. 1981. Theor. Appl. Genet. 59,317-325. Hill, M. O., and Gauch, H. G. 1980. Vegeratio 42,47-58. Hill, R. R., and Baylor, J. E. 1983. Crop Sci. 23,811-815. Hirosaki, S., Okuno, T., and Shiyomi, M. 1975. “JIBP Synthesis,” Vol. 6. Univ. of Tokyo Press, Tokyo. Holland, D. A. 1969. Exp. Agric. 5, 151-164. Hotelling, H. 1933. J. Educ. Psycho/. 24,417-441. Jardine, R., Moss, H. J., and Mullaly, J. V. 1963. Aust. J . Agric. Res. 14,603-621. Johnson, G. R. 1977. Crop Sci. 17,837-842. Kempton, R. A. 1984. J. Agric. Sci. 103, 123-135. Knight, R. 1970. Euphytica 19,225-235. Lin, C. S., and Binns, M. R. 1988. Can. J. Planr Sci. 68, 193-198. Lin, C. S., Binns, M. R., and Lefkovitch, L. P. 1986. Crop Sci. 26,894-900. Mandel, J. 1961. J. Am. Stat. Assoc. 56,878-888. Mandel, J. 1969. J. Res. Natl. Bur. Stand. (U.S.) 11,411-429. Mandel, J. 1971. Technometrics 13, 1-18. Mather, K., and Caligari, P. D. S. 1974. Heredity 33,43-59. Moll, R. H., Cockerham, C. C., Stuber, C. W., and Williams, W. P. 1978. Crop Sci. 18, 641-645.

Mead, R., Riley, J., Dear, K., and Singh, S. P. 1986. Biometrics 42,253-266. Menz, K. M. 1980. Field Crop Res. 3,33-41. Mungomery, V. E., Shorter, R., and Byth, D. E. 1974. Aust. J . Agric. Res. 25,59-72. Patterson, H. D., and Silvey, V. 1980. J. R . Star. SOC.143,219-252. Patterson, H. D., and Thompson, R. 1971. Biometrika 58,545-554. Patterson, H. D., and Thompson, R. 1975. Proc. 8th I n t . Biornetric Conf. pp. 197-207. Pearson, K. 1901. Philos. Mag. 2,559-572. Perkins, J. M. 1972. Heredity 29, 51-70. Peterson, C. J., and F‘feiffer, W. H. 1989. Crop Sci.29,276-282. Plaisted, R. L., and Peterson, L. C. 1959. A m . Potato J . 36,381-385. Polignano, G. B., Ugenti, P., and Pemno, P. 1989. Euphytica 40,31-41. Pooni, H. S., and Jinks, J. L. 1980. Heredity. 45,389-400. Robinson, D. 1987. Statistician 36,3-14. Rowe, J. S. 1956. Ecology 37,461-473. Seif, E., Evans, J. C., and Balaam, L. N. 1979. Aust. J . Agric. Res. 30, 1021-1026. Seiler, G. J., and Stafford, R. E. 1985. Crop Sci.25,905-908. Shorter, R. 1981. I n “Interpretation of Plant Response and Adaptation to Agricultural Environments.” Univ. of Queensland, St. Lucia, Brisbane. Shorter, R., Byth, D. E., and Mungomery, V. E. 1977. Aust. J. Agric. Res. 28,223-235. Shukla, G. K. 1972. Heredity 29,237-245.

STATISTICAL ANALYSES OF MULTILOCATION TRIALS

85

Suzuki, S. 1968. Proc. Int. Congr. Genet. 12th. 1968 1. Thompson, W. A. 1961. Bull. Int. Inst. Stat. 34, 1-4. Thompson, W. A. 1962. Ann. Math. Stat. 33, 273-289. Walton, P. D. 1972. Crop. Sci. U,731-733. Westcott, B. 1986. Heredity 56,243-253. Westcott, B. 1987. J. Agric. Sci. 108,267-274. Williams, E. J. 1952. Biometrika 39,65-81. Williams, W. T. 1976. “Pattern Analysis in Agriculture Science.” CSIRO, Melbourne and Elsevier, Amsterdam. Williams, W. T., and Gillard, P. 1971. Aust. J . Agric. Res. 22, 245-260. Witcombe, J. R. 1988. Euphytica 39, 11-18. Witcombe, J. R., and Wittington, W. J. 1971. Heredity 26, 397-411. Wricke, G. 1962. Z. Pjlanzenzuecht. 47,92-96. Wricke, G. 1964. Z. Pjlanzenzuecht. 52, 127-138. Yates, F., and Cochran, W. G. 1938. J. Agric. Sci. 28,556-580. Zobel, R. W., Wright, M. J., and Gauch, H. G. 1988. Agron. J. 80,388-393.

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY, VOL. 44

EVALUATION AND DOCUMENTATION OF GENETIC RESOURCES IN CEREALS A. B. Damania Genetic Resources Unit International Center for Agricultural Research in the Dry Areas (ICARDA) Aleppo, Syria

I. 11. 111. IV . V. VI. VII.

Introduction Evaluation of Cultivated Wheat Evaluation of Cultivated Barley Genetic Resources from Ethiopia Evaluation of Wild and Primitive Forms of Wheat and Barley Documentation of Genetic Resources Summary and Conclusions References

I. INTRODUCTION Agriculture is a relatively recent historical phenomenon having begun just over 10,000 years ago in the near East and later in Central America. Through the increase of food after the beginning of the agricultural (neolithic = food producing) revolution, the human species has incredibly multiplied its own population at the expense of the rest of the world’s biota (Reed, 1969). Early farmers initiated a series of partly conscious selections that have resulted in the landraces we see today. Plant breeding activity did not begin until the mid-1800s. This activity gathered pace after the turn of the century and breeders such as Strampelli were already using wild and primitive forms in breeding programs following the rediscovery of Mendel’s work (Maliani and Bianchi, 1979). Vavilov (1926) was the first to realize the need for a broad genetic base for crop plant improvement. But after the Second World War, massive aid projects led to the development of high-yielding cultivars that began steadily replacing the local varieties (landraces), thus narrowing the genetic base of several vital crops such as wheat, barley, and rice. By the 1960s, an urgent need was expressed in two symposia (Frankel and Ben87 Copyright Q 1990 by Academic Press, Inc. All rights of reproduclion in any form reserved.

88

A. B. DAMANIA

nett, 1970; Frankel and Hawkes, 1975a) to preserve the older more variable germplasm and wild relatives, which began to be referred to as genetic resources. The term “genetic resources” per se excludes breeding lines and recently released varieties (Frankel and Hawkes, 1975b), which are composed of gene combinations rather than the genes themselves. The two symposia also thoroughly reviewed the need for immediate and systematic exploration and collection on a worldwide scale of genetic resources of food and other commercial crops. An International Board for Plant Genetic Resources (IBPGR) was formed in 1972 by the Technical Advisory Committee of the Consultative Group on International Agricultural Research to undertake the plant collection and conservation recommended by the symposia. With the establishment of the IBPGR, collection and conservation of representative samples of genetic variability in landrace populations were accelerated and a large number of samples began to accumulate in the cold stores of the major genebanks. Genetic resources merely stored safely are of little value to plant breeders unless they are evaluated and the resulting data made widely available. Evaluation is, therefore, an essential link between conservation and use. In fact, Frankel (1987) categorically stated that genetic resources have been utilized without elaborate characterization, but never without evaluation. The next step was to evaluate the collected samples to identify sources of useful traits in order that the material be better utilized than in the past. The evaluation process for large collections follows several distinct stages: (a) seed multiplication and preliminary evaluation; ( 6 ) systematic evaluation of the entire collection; and (c) further evaluation of selected accessions (Chang, 1985). The utilization aspect of genetic resources was recently reviewed by Brown et al. (1989), wherein factors that are likely to limit or facilitate this process are discussed. Genetic resources workers discriminate between “characterization” and “evaluation” (Erskine and Williams, 1980; Hawkes, 1985), although this fact is not widely known. Characterization is defined as recording information only once on those traits that are highly heritable, easily visible, and expressed in all environments, for example, grain patterns and isozyme profiles. Characterization provides a standardized record of readily observable morphological characters that, together with passport (origin) data, identify an accession in the genebank. Evaluation, on the other hand, is the assessment of more variable traits for potential use in breeding, such as plant height, time to maturity, disease resistance, and protein content. This is done in several ways: growing the material in different environments, exposing it to various abiotic and biotic stresses, assessing grain quality, and selecting the best lines for the desirable attri-

EVALUATION OF GENETIC RESOURCES IN CEREALS

89

butes. “Preliminary evaluation” refers to screening that is frequently carried out during multiplication, at a single location and without the use of replicates prior to the incorporation of the samples into the collection (Damania et al., 1983; Erskine and Williams, 1980). The information generated by characterization and evaluation not only improves utilization of the germplasm but also rationalizes storage space by identifying duplicates and eliminating redundant germplasm. In spite of the wide recognition given to the importance of conserving and evaluating genetic resources (Frankel and Bennett, 1970; Hawkes, 1971, 1983; Frankel and Hawkes, 1975a; Harlan, 1975; Plucknett et al., 1983; Holden and Williams, 1984; Rogalewicz, 1985), little was known about the variability in the primitive forms, old landraces, and wild relatives of cereals and much work still remains to be done. Detailed evaluation of stored populations of different origins allows an understanding of the patterns of variability. The population structure of a species is defined as the totality of ecological and genetical relationships among individual members that may coevolve as a result of gene exchange, but may also diverge under localized forces of evolutionary change (Jain, 1975). Landraces and primitive cultivars are products of many years of crop evolution, and it is vital to preserve their genetic composition during and after evaluation. Cases have been reported where polymorphic cereal populations have undergone radical changes in their genetic composition in one growing cycle (Shevchuk, 1973). However, in the case of samples collected from village markets or those that are subjected to biased sampling methods, it is sufficient to safeguard and maintain their genes and not necessarily their gene frequencies within populations. It is now widely recognized that extensive surveys of geographic areas for genetic variability and computerized documentation of the evaluation results are needed for the utilization of large collections of cereal genetic resources. The problems of describing geographic variability data and development of statistical methods for categorizing sets of population samples from diverse localities have been discussed by Gabriel and Sokal (1969), and the analysis of variability between and within genotypes and environments has been discussed by Freeman and Dowker (1973). Computers with appropriate statistical packages have greatly facilitated this task of documentation. This chapter reviews the status of evaluation and documentation of genetic resources of wheat and barley. The work of a multidisciplinary team of evaluators that include germplasm scientists, taxonomists, cytogeneticists, pathologists, biochemists, and physiologists is presented, followed by conclusions and suggestions for future avenues of research.

90

A. B. DAMANIA

II. EVALUATION OF CULTIVATED WHEAT The evaluation of cereal genetic resources collections goes back to the time when their value to crop improvement began to be appreciated by the breeders. Cultivated wheat has been the most extensively evaluated crop among cereals, which is in keeping with the prominent position it enjoys in terms of production tonnage and importance as a food crop. Qualset and Puri (1973) evaluated heading time in a world collection of durum wheat (Triticumdurum Desf.) and presented results for each geographic area from which samples were analyzed. They found a wide range of heading time in about 3700 samples studies and identified those that were highly photosensitive. Puri and Qualset (1978) also researched effect of seed and seeding rate on yield and other characteristics of durum wheat and found a positive correlation between large seed size and yield. In another study, geographic patterns of phenotypic diversity for qualitative traits of more than 3000 samples of durum wheats were evaluated in the United States Department of Agriculture (USDA) world collection at Tulelake, California, by Jain et al. (1975). Observations were recorded on leaf sheath glossiness, glume pubescence and color, awn color, kernel color, and basal spike node fertility. Variability for each character was found usually within each geographical region. Although Jain et al. (1975) admit that the collection was small and not representative, they found centers of diversity among material from Ethiopia, India, and the Mediterranean countries. Some of the areas known to be important sources of genetic resources were poorly represented in the study, a fact emphasizing the need for intensification of efforts in exploration and conservation. Spagnoletti Zeuli and Qualset (1987) have reported an evaluation study on the same durum wheat entries for spike characters. Five clusters were delineated among 26 country origins and an east to west clinal pattern was detected that represented a gradient in unimproved to improved types. In a previous study, Porceddu (1976) evaluated 2400 wheat landraces from the world collection of durum wheat and recorded information on growth habit, beginning of shooting, heading time, culm length, and number of elongated internodes. Distinct differences were found to exist among samples from different countries. Multivariate analysis showed that samples from Mediterranean countries were very similar, probably due to old trade links. Similarity in other samples from Cyprus, Egypt, Jordan, and Palestine was also noted. Another discovery was the significant similarity between accessions from Turkey and the United States. This was attributed to the history of transport of germplasm from Turkey and subsequent inclusion in U.S. varieties. Konzak et al. (1973) also evaluated wheaL germplasm from Turkey for reaction to mildew at

EVALUATION OF GENETIC RESOURCES IN CEREALS

91

Pullman, Washington, U.S.A., and estimated yellow berry content of the grain produced. Porceddu and Scarascia-Mugnozza (1983) carried out similar studies for ascertaining variability in landraces of durum wheat from Algeria, Ethiopia, and Italy and found that there was clear separation between Ethiopian and Italian material, but Italian landraces were more variable among themselves than Ethiopian ones and material from Algeria was more variable than both the other two. Hence, differences among landrace populations from the same country were highly significant for all 1 1 characters studied, except kernel weight and spike density. Using multivariate analysis also, Kosina ( 1980) evaluated structure and caryopsis quality of some hybrids of spring wheat. Ehdaie and Waines (1989) described genetic variability in T. aestiuum from Iran. They concluded that local landraces, such as those found in Iran, could be improved by selection for shorter genotypes with fewer tillers per plant, but with larger and heavier grains. Morphological and physiological variability in T. aestiuum collected from Afghanistan was also reported comprehensively by Tani and Sakamoto (1987). There are traits, such as resistance to diseases and tolerances to certain types of soils, for which variability can only be observed at particular sites. Such traits are economically important and every effort must be made to record and document them by carrying out evaluation at sites where the incidence of that particular stress is the greatest, such as the so-called disease “hot spots.” For example, for screening against resistance to Septoria tritici (leaf blotch), ICARDA uses a humid and high-rainfall site located on the Mediterranean coast in Syria in addition to artificial inoculation. For experiments on tolerance to salinity, a drought-affected site on the shores of salt lake Jabboul in northern Syria is used. Jana et al. (1983) evaluated 3000 durum wheat accessions from various countries at this site, and 10 lines were found to be highly tolerant to combined stresses of salinity and drought. However, it is known that salinity is highly variable in the field and if experiments do not comprise several replicates, laboratory confirmation with tests such as chlorophyll influorescence (Smillie and Nott, 1982) may be used to help identify salt-tolerant lines. Selected bread wheat and barley plants from landraces grown in Nepal and Pakistan were examined and variability for certain qualitative traits described by Witcombe (1975). Barley from Nepal was found to be more variable than barley from Pakistan, whereas in the case of wheat the reverse was true. Murphy and Witcombe (198 I ) further analyzed landraces of wheat from northern India by growing single plants under glasshouse conditions in Wales, U.K. Multivariate analysis of data on quantitative traits was used to distinguish between introduced material and indigenous germplasm on the basis of means recorded on single plants. This data

92

A. B. DAMANIA

analysis method was used to detect modern varieties so they could be excluded from genebanks as genetic resources. However, variability studies based on observations of quantitative traits on single plants in a controlled environment are inconclusive and should be verified by field studies before an inference is drawn. Damania et al. (1985) evaluated populations of wheat and barley landraces from Nepal (Triticum aestivum and Hordeum vulgare L.) and the Yemen Arab Republic (T. turgidum L. and H . distichum L.) for morphological variability and days-to-heading under field conditions using sufficiently large samples. Observations were recorded for 18 characters on 50 randomly selected plants from each landrace. There were significant differences in variability among regions in the Yemen and river valleys in Nepal, as well as among landraces in the same regions or river valleys. It was concluded that landrace variability within primary or secondary centers of diversity could not be fully evaluated without growing the plant material in the field under conditions similar to the original habitat, although an initial impression of the extent of variability can be obtained through the application of polyacrylamide gel electrophoresis (PAGE) of seed proteins (Damania et a / ., 1983). Variability in high-molecular-weight glutenin subunits in landraces of hexaploid wheats from Afghanistan was also evaluated by Lagudah et al. (1987) using PAGE. The variability seemed to be independent of the altitude and geographical location of the collection site. A world collection of bread wheat (T. aestivum L.) maintained by the USDA was systematically analyzed for protein and lysine content by Vogel et al. (1973). Lewontin (1974) has expounded the advantages of electrophoretic surveys of proteins as measures of genetic diversity and reviewed the technical limitations and conceptual opportunities offered to plant biologists using this technique. Wheats can also be evaluated for quality and study of genomes and genotypes with the use of PAGE (Konarev et al., 1979). Kobrehel and Gautier (1973) studied peroxidase patterns of primitive and modern wheats by PAGE and found that brownness in macaronis can be associated with the compositions of the peroxidases. Damania (1985) demonstrated the usefulness of this technique in evaluating landraces of hexaploid wheats and their possible utilization in improving the bread-making quality of modern varieties. However, predictions on good or poor gluten quality based solely on presence or absence of certain bands in an electrophoretic gel may not be conclusive. It has been pointed out that the band with relative mobility (Rm) 45 and the band of Rm 42, which were indicators for good pasta and bread-making qualities in durum and bread wheat, respectively, were merely genetic markers, whereas other proteins (low-molecular-weight glutenins) were responsible for gluten viscoelasticity (Pogna et al., 1988).

EVALUATION OF GENETIC RESOURCES IN CEREALS

93

Information on the genetic variability of a sample is extremely useful and the eventual objective of every evaluator should be to describe the variation on the basis of a list of differences between and within samples in the sequence of nucleotides in the deoxyribonucleic acid (Erskine and Williams, 1980). At present, the study of storage protein (prolamins) variants by electrophoresis is the most convenient and rapid method available for detecting genetic differences at the DNA level in a cereal collection (Damania et al., 1983). Documentation of data derived from electrophoretic studies of storage proteins in cereal genetic resources has been reviewed by Konarev et al. (1979). However, it may be argued that heterogeneity in storage proteins alone is of little value to the breeders or genetic conservationists because its correlation with any single agronomic character is obscure. Nevertheless, these markers can perhaps monitor the relative genetic diversity with a greater degree of accuracy (Brown, 1978; Damania, 1983) than field studies with an inadequate number of traits. A prescreening procedure for identification of the ploidy levels and chromosomal aberrations, such as deletions with the use of electrophoresis, has been described by Damania (1985). This technique can also be used as a tool for elimination of duplicate germplasm stocks (Damania and Somaroo, 1988).

Ill. EVALUATION OF CULTIVATED BARLEY Barley is one of the most dependable cereal crops in harsh environments. It is grown in semiarid areas as well as in cold, short-season areas. Local varieties and landraces of barleys occupy nearly 80% of the cultivated areas in West Asia and North Africa and these should be collected before they are lost. Ward (1962), in one of the early efforts to characterize a large number of germplasm samples, evaluated 6200 lines from the USDA world collection of barleys and recorded observations on seven qualitative characters. He concluded that only a very small portion of the potential genetic diversity in barley was represented in the collection. Farmers still rely on barley landraces that have a stable performance in the dry areas and rarely outyield modern homogeneous varieties (Ceccarelli et al., 1987). The variability within these landraces is large and compares well with that within populations of its wild relative, Hordeum spontaneum, growing in the same region (Jana and Pietrzak, 1988). Variability for agronomic traits in landraces of barley has been evaluated by several workers. De Pace et al. (1978) crossed a set of six old Italian wheat varieties (female parents) with five breeding lines from the International Wheat and Maize Improve-

94

A. B. DAMANIA

ment Center (CIMMYT). The F3 progenies were grown in the field and variability for flag-leaf size, heading, and maturity time was recorded on single plants. It was concluded that utilization of genetic resources distant from the present varieties is advantageous as it permits breakage of linkage groups. The Cereal Improvement Program at ICARDA collected barley landraces from 33 locations in the drier regions of Syria and Jordan (Weltzien, 1982a). On subsequent evaluation it was found that some samples from Syria had a definite cold requirement to induce flowering (Weltzien, 1982b), indicating that under local conditions this trait may contribute to yield stability by preventing winterkill of early-planted material. Information of this kind is valuable for a crop improvement program, not only by showing which characteristics contribute to adaptation but also by indicating the degree of plasticity present in the indigenous landraces for which varietal improvement may be undertaken. Tolbert et al. (1979) carried out a diversity analysis on the USDA world collection of barleys and Ceccarelli et al. (1987)evaluated genetic diversity of barley landraces from Syria and Jordan for agronomic, morphological, and quality traits. Considerable diversity was observed between as well as within collection sites. Single-plant progenies were identified with larger yields and more desirable expression of agronomic characters than the original landraces. On the other hand, Murphy and Witcombe (1986) performed discriminant and reciprocal averaging analysis on single plants of covered and naked barleys from the Himalaya and confirmed their difference in a multivariate way. Sixteen populations of two-rowed barleys from the Yeman Arab Republic were screened for loose smut (Ustilago nuda) disease and found to be highly resistant when compared to the six-rowed barley landraces from Nepal (Damania and Porceddu, 1981). Incidence of diseases varies depending on climatic factors and inferences made of natural infestation in a single season should be avoided as they can be misleading. In another screening of landrace material, van Leur et al. (1989) tested 280 barley lines collected from different sites in Syria and Jordan. Large variability in response to four disease-yellow rust, scald, powdery mildew, and covered smut-was observed. No consistent association between the environmental conditions of the collection site and the level of resistance of the landrace lines could be found. Variation of flavonoids between barley lines was studied and found to be greatest in the Near East (Afghanistan and Iran) and Ethiopia (Frost et al., 1975). Allard et al. (1970) reviewed some studies on enzyme variability in a world collection of barleys and reported on the geographical distribution of different enzyme variants and factors responsible for the development and maintenance of the observed variability in the samples.

EVALUATION OF GENETIC RESOURCES IN CEREALS

95

In any survey of the distribution of genetic variability within a crop species of economic value, or its wild relatives, the most obvious pattern that emerges is the variability associated with broad geographical regions (Allard, 1970). It is important to sample not only the broadest geographical variability but also to have samples from the extremity of distribution of the species, that is, the marginal areas. According to the hypothesis of “peripheral diversity” put forward by Yamashita (1980), it has been suggested that there is considerable accumulation of diversity where a species has reached its geographical limits as a result of physical or climatic barriers that it cannot traverse. This variability needs to be collected, followed by evaluation for economically useful traits.

IV. GENETIC RESOURCES FROM ETHIOPIA A mention must be made of the very substantial amount of variability in genetic resources still to be found in Ethiopia (Porceddu and Perrino, 1973; Mengesha, 1975) and the considerable evaluation studies carried out on this material. This region of Africa has been often quoted as a secondary center of diversity of tetraploid wheats, and Vavilov (1932)was able to find a high amount of variability among and within samples. Porceddu et al. (1973) described 145 morphologically different phenotypes in a collection of tetraploid wheats from Ethiopia. The influence of altitude on yield and quality in cereals in Ethiopia was reported by Alkamper (1974). Bekele (1984) carried out an extensive evaluation study on populations of 153 tetraploid and 72 hexaploid wheats, respectively, collected from several regions in Ethiopia; 633 1 and 2700 individual plants were studied for 11 morphological characters. Jain et al. (1975) emphasized the importance of both regional and local patterns of genetic variability to genetic resources conservationists as well as to plant breeders. Analyses of regional patterns of variability for characters to determine the relative contribution of various regions to genetic resources showed differences in the level of importance of useful germplasm material. Polymorphism in characters was higher for tetraploid than hexaploid wheats, indicating that Ethiopia is almost certainly a center for diversity for tetraploid wheats. The contribution to total variability varied among characters and in both forms the contribution to the total phenotypic diversity was highest at the lowest level (within localities), followed by the differences among populations in a region, and among regions themselves. Qualset (1975) discussed optimal sampling strategy for germplasm possessing a single desirable trait in a center of diversity, using disease resistance in Ethiopian barleys as an example. The identity of genes in barley

96

A . B. DAMANIA

accessions from Ethiopia when compared with genes in primitive barley samples from the area of origin in the fertile crescent of the Middle East would provide evidence for the migration of this crop from the primary center of origin to Ethiopia. Negassa (1985a) evaluated the Ethiopian barleys for resistance to powdery mildew, studied the patterns of phenotypic diversity in a collection, and suggested that the Arussi-Bale highlands in Ethiopia was the secondary center of origin for H . vulgare (Negassa, 1985b). In another significant evaluation work, Negassa (1986) studied a collection of 293 entries of tetraploid and hexaploid wheats from Ethiopia. Some characters showed clinal variability, whereas others had localized concentrations. More significantly, a gene center for grain quality was identified in southern Ethiopia and this was confirmed by Dominici e f a / . (1988), who found lines with gliadin (storage protein of wheat) banding patterns not reported before from Ethiopian wheat landraces through an electrophoretic study. During 1967-1968, Kyoto University organized a collecting mission that covered Ethiopia (Sakamoto and Fukui, 1972).The principal objective was to investigate the variability pattern of cultivated plants and collect samples of their local strains. Seven hundred and seven accessions of wheat and barley were collected. Preliminary observations and analysis of variance on morphology and spike coloration have been published. Report of such information after evaluation of collected material is very useful for planning future collecting activities and every expedition must follow this example. The Ethiopian gene center certainly appears to be harboring interesting material when germplasm with such a high level of variability has already disappeared from other centers of diversity. Hence, further collections could be made to conserve genetic resources before the expanding human population and deforestation lead to genetic erosion.

V. EVALUATION OF WILD AND PRIMITIVE FORMS OF WHEAT AND BARLEY Wild relatives and primitive forms are those from which the present crop plants were domesticated and that continue to survive in conjunction with the cultivated species or by themselves as weeds. Several of these forms have become tolerant to biotic and abiotic stresses in their natural habitats, thus enhancing their value to agriculture as additional sources of genetic variability. However, their usefulness to plant breeders depends on their cytogenetic affinity and barriers to hybridization, the amount of

EVALUATION OF GENETIC RESOURCES IN CEREALS

97

germplasm available for evaluation, and the desirable traits revealed, and whether repeated backcrossing is needed to eliminate undesirable traits introduced from the wild or primitive parent. The evaluation of wild relatives of crop plants, including wheat and barley, has also been reviewed by Harlan (1984). Many regions in the world have succeeded in improving crop production through the introduction of high-yielding varieties. However, these varieties have not met with great success in West Asia and North Africa because of their intolerance to an environment where moisture is limited and inputs are very low (Srivastava and Damania, 1989). To improve and stabilize cereal production in the drylands, characters such as earliness plus tolerance to drought, temperature extremes, low plant nutrients, and diseases need to be incorporated into varieties. The genes for these characters can probably be found in wild relatives and primitive forms that are well adapted to such environments through independent survival over a very long period of time. Wheat and barley plant breeders in developed countries consider landraces and primitive forms as unadapted germplasm and Hallauer and Miranda (1981) termed as exotics all germplasm that does not have immediate use without selection for adaptation to a given area. The exploitation of wild and primitive forms in wheat breeding has been insufficient for four reasons. First, collections of wild progenitors in the past have been fragmentary as well as scanty and material available in collections is not representative (Croston and Williams, 1981). Second, work on wild forms has primarily concentrated on evolutionary (Feldman and Sears, 1981) and taxonomic studies (Bowden, 1959; Chennaveeraiah, 1960; Morris and Sears, 1967; Kimber and Feldman 1987). Third, wild relatives are not well adapted to Europe and North America and hence are used mainly as single-gene donors. Finally, variability between and within populations of wild species has not been looked at in adequate detail and utilization has hardly commenced (Srivastava et al., 1988). The principal reason for utilizing wild species such as Aegilops, Agropyron, and Triticum dicoccoides Korn. in wheat breeding has been the transfer of genes for disease resistance, salinity tolerance, and highprotein content, respectively, when these desirable traits are not found in the available genetic stocks. However, confirmation of the sources of resistance and cataloging of genes for resistance should precede any attempt at transfer from the wild to the cultivated forms. This would involve collecting and evaluating in-depth relevant germplasm collections for genetic variability. In-depth evaluation would consist of replicated tests at different locations, resulting in genetic analyses and observations on traits such as disease resistance.

98

A. B. DAMANIA

It is also useful to gain knowledge of genetic relationships between the wild donor and cultivated recipient species because without ready crossability, desirable traits cannot be easily transferred. Thus, wild progenitors of the cultivated species should be considered an important source of variability for broadening the genetic bases of cultivated crops (Harlan, 1976; Hawkes, 1977; Lange and Balkema-Boomstra, 1988). Brown (1978) states that the genetic resources of wild relatives of crop plants should be systematically evaluated, for these sources of genes will supplement, and even rival, the primitive landraces in their effectiveness in crop improvement programs. There is an opinion among certain workers, especially those involved in breeding for favorable environments, that the variability in landraces of wheat and barley has been fully exploited and for further progress in introducing useful genes one should turn to wild relatives and other alien germplasm of the secondary and tertiary gene pools. Strampelli, working in Italy, was among the first plant breeders in Europe to utilize wild and primitive genetic resources, especially Triticum villosum Schur. and T . spelta L., to improve wheat in 1906 (Maliani and Bianchi, 1979). Elliot (1957) transferred stem rust resistance to common wheat from Agropyron elogatum and Riley et al. (1968) transferred yellow rust resistance from Aegilops comosa Sibth. to cultivated wheat by genetically induced homologous recombination. One of the most detailed evaluation studies of morphological, physiological, genetic, and cytological characteristics of Aegilops and Triticum species was carried out at the University of Kyoto (Kihara et al., 1965). A geographical survey of species of wheat and its wild relative Aegilops was conducted following the University of Kyoto expedition to Afghanistan, Iran, and Pakistan in 1955. Emphasis was placed on collecting Aegilops squarrosa L., the probable donor of the D genome to cultivated wheat; other species were also sampled whenever possible. There was evidence that bread wheat arose as a hybrid between cultivated emmer ( T . dicoccum Schub.) and A . squurrosu. This study was not only illustrated with photographs, but also contained valuable information on utilization of the germplasm. Cox et a/. (1989) evaluated 212 accessions of A. squarrosa from the University of Kyoto collection using polyacrylarnide gel electrophoresis and found gliadin diversity to be higher than in the D genome of cultivated bread wheat. Genetic variability in Aegilops species and primitive forms of wheat has been studied by several researchers (Hillel et al., 1973; Sharma et a/., 1981; Dhaliwal et al., 1986; Waines et al., 1987), whereas in the past more emphasis was placed on Triticum dicoccoides, the wild progenitor of wheat (Lawrence et al., 1958; Avivi, 1979) as reviewed by Poyarkova (1988). Also, Gerechter-Amitai and Stubbs (1970) found the source of resistance to yellow rust in T. dicoccoides from Palestine.

EVALUATION OF GENETIC RESOURCES IN CEREALS

99

In another evaluation of primitive and wild types of wheat comprising 75 accessions of T . boeoticum Boiss., T . araraticum Jakubz., and T . dicoccoides; 42 of T . urartu Tuman.; and 34 of A . squarrosa seeking resistance to leaf rust, Hessian fly, and greenbug, Gill et al. (1989) found that all species except T . dicoccoides had resistance to both leaf rust and Hessian fly. Only A . squarrosa contained some resistance to greenbug, T . boeoticum had an intermediate response, but all other species were susceptible. Multiple resistance lines to leaf rust and Hessian fly were identified, but only A . squarrosa had two lines resistant to all three pests. Hillel et al. (1973) evaluated A . longissima Schweinf. and A . speltoides Tausch, two loosely related species that differ in their mating systems, to assess the direct effect of the mating system on the amount of genetic variability. They found that for most of the 36 quantitative characters examined, the differences between populations, the total variances of the populations, and the mean within-species variances were greater in the selfer (longissima) than the outbreeder (speltoides). These differences were attributed to the low probability of a successful gene flow in A . longissima, with each isolated population adapted to a specific microenvironment. Ninety-three accessions of cultivated emmer wheat ( T . dicoccum), five each of two wild tetraploid wheats ( T . dicoccoides)and T . araraticum, and the cultivated varieties ‘Modoc’ ( T . durum) and ‘Anza’ ( T . aestiuum) were evaluated for plant height, seed weight, flour protein content, and flour lysine content, as well as several morphological and grain quality characters by Sharma et al. (1981). Variability among lines for each trait in different species was significant except plant height in dicoccoides and araraticum. Primitive and wild wheats were higher in protein and lysine content, but lower in spike weight and 1000-kernel weight than the two modern cultivars. It seems that selection for larger kernels has resulted in a drop in the protein content of seeds from the wild species to the modern wheats. Lawrence et al. (1958) also found higher protein and lysine contents in wild wheats than in cultivated bread and durum wheats, but only one accession of T . dicoccoides and T . monococcum L. were analyzed. Avivi (1978) evaluated 47 samples of T . dicoccoides for protein content in kernels and found it to be highly variable, ranging from 19% to 28%. Waines (1983) did a comparative study of primitive diploid wheats such as T . monococcum with modern polyploid varieties and advocated the former’s direct usage as commercial varieties. Among the several positive characteristics of diploid wheats mentioned were less extraneous genetic material, only seven linkage groups and hence easier to manipulate than polyploid wheats, greater ecogeographic distribution and hence wider adaptability, and resistances to pests and diseases. However, Waines (1983) did not fully evaluate grain quality aspects such as pasta and dough

100

A. B. DAMANIA

products, which are vital to the success of any wheat variety, although one may tend to agree with him that exploratory research should be carried out to see if diploid wheats have a future as commercial varieties. Hordeum spontaneum C. Koch. is now recognized as the only progenitor of cultivated barley (Harlan, 1979). Wild barleys, H. spontaneum and H . murinum, and cultivated varieties from Afghanistan were collected by the Kyoto University Scientific Expedition. The morphological, physiological, and genetic characterization are described by Takahashi et al. (1965). Resistance to barley powdery mildew was also found in H. spontaneum, but the reaction was classified as being different from that of the cultivated forms. The overall diversity in this collection was reported to be low and the wild forms were less variable than the cultivated. Jana et al. (1987) studied genetic diversity in morphological characters in H. spontaneum and cultivated barleys from the Near East and also found that the latter were more variable than the former. However, it must be considered that this wild form in the near East has a long history of survival, has undergone millenia of natural selection pressures and therefore is better adapted to harsh environments (if not as variable as the cultivated landraces), and hence is a very valuable genetic resource for abiotic stress tolerance genes (Grando et al., 1985). An evaluation of H. spontaneum accessions was also carried out by Ceccarelli and Grando (1987) to assess the amount of useful genetic variability within this species. The results indicated that the progenitor is a useful source of genes for a number of economically important characters for breeding barley in the dry areas. Bakhteyev (1979) evaluated a collection of 77 samples of the same wild species originating from Iran, Iraq, and Turkey at an experimental farm in the Soviet Union. Considerable variability was observed in this study and the species merits breeders’ interest especially for adaptation to dry areas. Landraces of primitive wheats have been evaluated for several economically important traits. Blum et al. (1987) evaluated 13 accessions of T. compactum Host., a form of wheat cultivated in Syria and Palestine in the prepottery neolithic period (Renfrew, 1969), for grain quality. It was found, while comparing these with some modern cultivars of bread and durum wheat, that the latter were superior in all characters tested except protein content and quality. In a comparative study of wild and primitive forms of Triticum, Asins and Carbonell (1986) provided information for future use on intraspecific variability differences among species. Robertson et al. (1979) evaluated the genetic variability in primitive wheats for seedling root numbers, as it is well known that this character is probably responsible for drought resistance in wheats. Among the 16 species studied, the highest root numbers

EVALUATION OF GENETIC RESOURCES IN CEREALS

101

were found in T. durum and the lowest in T . araraticum; seed weight was positively correlated with root number. A test of progenies showed that this character was stable from one generation to the next over two environments. O’Toole and Bland (1987) have reviewed genotypic variation in root systems of cultivated wheat and Aegilops. Electrophoretic techniques also have been applied to the study of variability of storage proteins in wild and primitive forms of wheat (Cole er al., 1981; Damania er al., 1988) and barley (Jana et al., 1987). Using this technique, Nevo et al. (1979) showed that there was greater variability in wild and weedy barleys ( H . sponraneum) collected from Palestine than in a composite cross of cultivated barley that included more than 6000 varieties in its parentage. This result was surprising since cultivated barley is conspicuously more variable than its wild and weedy relatives from the Near East (Harlan, 1984). However, Jana and Pietrzak (1988) reported almost identical levels of variability between the two in material collected from Greece, Jordan, Syria, and Turkey. Nevo et al. (1982, 1988) studied genetic diversity within and between Turkish populations of wild emmer, T . dicoccoides, utilizing electrophoretic and statistical analysis, and reported that climatic selection played an important role in genetic differentiation of populations of this species and that the wild gene pool represents a significant genetic resource for utilization in wheat improvement. In another study of the same material, Nevo and Payne (1987) reported variability in seed storage protein and the use of certain high-molecular-weight glutenins for improvement of bread-making qualities of wheat. Variability for kernel proteins in 841 accessions of T . dicoccoides has been reported also by Mansur-Vergara et al. (1986) using gel electrophoresis. The protein content measured ranged from 15% to 25% with some accessions having high protein content as well as large kernel size. A high percentage of protein content was also reported by Srivastava and Damania (1989) in dicoccoides accessions from Syria and by Avivi (1978) in those from Palestine. In a study of 167 accessions of T . dicoccum from 23 different countries of origin using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), Vallega and Waines (1987) identified a total of 20 alleles out of which 9 were different from those reported by earlier work. The newly discovered alleles enhance the genetic variability available to improve the industrial quality of wheats, and some of them may facilitate basic research on the relationship of industrial quality with highmolecular-weight glutenin subunit number. Success in future crop improvement depends largely on the ability to exploit existing genetic resources in wild relative and primitive forms,

102

A. B . DAMANIA

especially for less favorable environments. A better utilization and exploitation of these resources requires greater understanding of critical issues related to evolutionary pathways, geographical distribution, genetic diversity, and multicharacter associations. Variability studies, such as those described earlier, have provided this vital information. However, much more work needs to be done before evaluation can be instrumental in greater use of genes from wild relatives and primitive forms.

VI. DOCUMENTATION OF GENETIC RESOURCES An efficient system for dissemination of evaluation data on genetic resources material held in genebanks is essential if it is to be of use to breeders. Databases of genetic resources information on world collections have been established and contain a formidable amount of evaluation information that needs analysis and documentation (Ford-Lloyd, 1978). For example, 12,129 accessions of barley from a world collection were evaluated by ICARDA at one location (Somaroo et al. 1986, 1988); there was significant variability among the landraces for such characters as days to heading, plant height, 1000-grain weight, protein/lysine ratio, and resistance to diseases. In recent years much effort has been devoted to making these databases as comprehensive and mutually compatible as possible. However, less effort is being channeled toward considering how the results of these studies might be put to use. This is because most users are interested primarily in a very restricted aspect of the data (Williams et af., 1980). Plant breeders may be interested in an immediate problem such as resistance to a particular pathogen or a specific agronomic trait. Alternatively, their interest may reflect locally prevailing environmental conditions. For example, breeders at ICARDA are not interested in plants adapted for favorable conditions, as such environments are not common in West Asia and North Africa, where the Center operates (Damania and Srivastava, 1989). However, an international genetic resources center might be expected to take a wider view and Williams et al. (1980) contend that genetic resources evaluation data presently stored in computer memory banks contain more information of practical value than is immediately apparent without proper analysis. A detailed study of network analysis of genetic resources data has been attempted (Williams et af.,1980; Robinson et af., 1980; Burt et al., 1980). Internationally agreed descriptor states are not used when evaluation data are exchanged between genetic resources workers. Considerable time

EVALUATION OF GENETIC RESOURCES IN CEREALS

103

and effort would be saved if a lengthy description of the method and intervals used for recording each character is avoided. The IBPGR has been convening small working groups of scientists for the purpose of arriving at an internationally agreed upon list of descriptors for describing information for wheat (IBPGR, 1985) and barley (IBPGR, 1982). However, experience has shown that it is too time-consuming to record observations on all descriptors mentioned in the descriptor lists, as the method of recording and the selection of characters are very much dictated by the region and the needs of local breeders at each institution concerned. It is also necessary to set out data in a standard format using a generally agreed upon series of descriptors and descriptor states for the crop. In this way the data can be entered into computers, retrieved, and exchanged among institutions with the least possible confusion and optimum efficiency (Hawkes 1985). For the purpose of utilization, systematic analysis and description of samples is useful in both distinguishing between populations and identifying duplicates, as well as in providing information on the extent of variation within a given germplasm collection. It is axiomatic that the more documentation on a collection, the greater the chance of its rational utilization. Information from the site where a particular sample was collected may be extremely important. For instance, at ICARDA, germplasm that is described as having a short maturity period receives immediate attention of the breeders as this trait is very useful to escape drought and high temperatures during grain filling in the dry areas. Therefore, information recorded by germplasm collectors at site would be very valuable later when the samples are evaluated. Peeters (1988) studied statistically a large barley germplasm collection at Cambridge and reported that despite extensive collecting activity in recent years and subsequent exchange between countries, combinations of characters have remained substantially different in germplasm by country gene pool. Material from the United States now contains more variability in toto than material from any other country. Subsequently, Peeters and Martinelli (1989) used hierarchical cluster analysis to classify entries from this collection according to their degree of similarity and concluded that this statistical analysis procedure could be used as a tool to classify entries to their respective gene pools even when no passport data are available. Often those responsible for entering data recorded at a collecting site into a computer data base believe that lengthy descriptive notes made at the collection site, for example, notes on disease observations or peculiarities of the habitat, are not relevant and hence should be omitted. Nothing could be more erroneous. Although it is recommended that passport data

104

A. B. DAMANlA

must not be encumbered with vast amounts of morphologicaldescriptions, it should certainly contain disease and habitat data and, more importantly, comments from the farmers as to the useful features that distinguish their material from the rest. Inadequate passport data very often inhibit effective utilization of collected germplasm. It has been repeatedly pointed out to collectors and genebank managers that passport data divulge extremely valuable and in many cases the only available information on the ecological adaptation of an accession and hence no effort should be spared to fill this important gap in documentation of germplasm (Frankel, 1987). Systematic description of samples for discrete traits has been limited to cataloging the phenotypic variation because of constraints in relating genotype with phenotype. Quantitative morphoagronomic traits are also currently used in characterization. These traits are controlled by several genes, each contributing a small effect that is quite often blurred by the environment. Consequently, the correlation between genotype and phenotype is obscured. Certain evaluation studies have used ranking as a method of describing results of economically important traits such as yield. This ranking may change from one site to another for some quantitative characters such as plant height and days to heading (Damania, 1983). Such unstable characters cannot be adequately described when studied at a single location. Thus, the concept of multilocation testing becomes imperative. We cannot commit ourselves to hard and fast rules regarding the selection of a representative sample, but it must be stressed that an evaluation that partially covers the total variability can only be of limited value at best. That is, if raw data are misinterpreted or incorrectly fed into the main data base, self-consistency is lost and the entire task becomes futile. Unfortunately, not all the samples assembled in our genebanks were collected with the aim of preserving genetic variability of populations in danger of extinction. On the contrary, several genetic resources collecting expeditions were targeted to filling certain gaps, such as finding resistant lines to specific stresses or studying relationships between wild and cultivated species. Therefore, genetic material from such expeditions represents only a fraction of the existing variability present in a particular area. This being so, it can provide useful genes for current breeding goals, but may be inadequate for tomorrow’s needs (Porceddu, 1976). The major collections of wheat and Barley now contain several thousand accessions. Such numbers may be too large for detailed evaluation. In response to this, there is a recent trend toward developing “core collections,” which are subsets upon which detailed evaluation work may

EVALUATION OF GENETIC RESOURCES IN CEREALS

105

be concentrated (Brown, 1989). The remainder of the large collection constitutes a reserve still maintained in storage and available when a desired trait cannot be found in the core (Chapman, 1989). The ability of genetic resources managers to respond to requests from breeders for material depends very much on the adequate description of the accessions and the ability to query the information in a computerized data base. Hitherto, insufficient emphasis has been placed on recording passport data, and their absence is a major constraint to curators in assessing the range of variability in their collections and in identifying gaps (Williams, 1989).

VII. SUMMARY AND CONCLUSIONS Genetically uniform cultivars are employed by cereal farmers in the major cereal-producing countries of the world. Because plant breeding is essentially a process for exploitation of genetic variability, breeders could also examine means of conserving already existing genetically variable germplasm as well as creating new varieties. Mak and Harvey (1982) have described the composite cross technique that creates, as well as exploits, genetic variability, using the USDA barley world collection as a model. This may be one of the ways to proceed for other cereal crops. When precise objectives of evaluation are known at the outset, the task becomes relatively simple. In the case of most wild and primitive forms, evaluation aims to reveal potentially useful variability for direct use in the breeding programs. This may necessitate initial characterization in nurseries and cataloging of passport information, followed by a more detailed field study in collaboration with the end-users of the germplasm. Inferences regarding geographical variability, even on the basis of evaluation of a world collection, cannot be considered truly representative as they represent findings based on the composition of a collection that may be comprehensive for some regions and deficient for others. Furthermore, variability studies based on collections made several years ago may not accurately reflect the variability to be found at present in the same area. It is presumed that, after observation of dramatic degradation of the environment and genetic erosion, there would be considerable decline in variability if not extinction of indigenous germplasm in several previously generich regions in the world (Hawkes, 1981). The utilization of germplasm collections in crop improvement for the major cereals has revealed the following:

106

A. B. DAMANIA

1. The use of exotic germplasm. The successful use of landraces and wild species in cereal improvement has been more extensive in the developed countries, which lack indigenous germplasm, than in the less developed countries. In recent years the International Agriculture Research Centers (IARC) have contributed substantially to generation and distribution of improved adapted germplasm with genes from landraces and wild forms to the national programs and other institutions, as the germplasm developed for more favorable environments has not succeeded in the dry areas. 2. Constraints to the use of exotic germplasm. Many plant breeders were reluctant to devote a greater part of their resources for the exploitation of landraces and wild species in the past. This was because the potential value of these germplasms for the stressed environments was not fully appreciated. However, in recent years, varieties targeted for lowinput, rain-fed agricultural systems possess genes from adapted landraces and even direct usage of selections of the best lines isolated from landraces have been recommended for release. 3. Support for plant genetic resources programs. Extensive use of landraces, primitive forms, and wild species will be more tenable for harsh environments when the process of conservation, evaluation, documentation, and exchange of germplasm is strengthened and adequately funded. Donor countries and international agencies could increase support for utilization of indigenous landraces and primitive forms in the recently established breeding programs of the developing world. 4. Use of computers and statistical program packages. Computer programs designed for analyzing a large quantity of evaluation data have greatly reduced the time and effort needed for arriving at tangible results. This in turn has led to the publication ofgermplasm catalogs, which have facilitated dissemination of information on genetic resources collections to actual users, allowing for greater utilization of the services rendered by genebanks. However, breeders prefer to receive a short list of accessions with specific traits to choose from rather than large genetic resources catalogs.

Electrophoretic techniques that permit rapid mass screening of samples are increasingly recognized as powerful research tools for the study of genetic variations in populations. A wider application of gel electrophoresis in the evaluation of plant genetic resources is expected. The use of restriction fragment length polymorphisms (RFLP) for evaluating genetic diversity has been described (Bernatsky and Tanksley, 1989), and is the best available means for detecting differences at the DNA level on samples of reasonable size. It would be useful if such techniques were utilized to a greater extent than at present on genetic resources collections.

EVALUATION OF GENETIC RESOURCES IN CEREALS

107

The breeding objectives of the developed countries, mostly located in favorable environments, are different from those of the West Asia and North Africa region. The advanced breeding programs in the former countries began utilizing exotic landraces about 100 years ago and have fully exploited them, so they no longer seek variability but only single genes from the wild relatives and grasses of the tertiary gene pool. The breeding objectives for West Asia and North Africa, on the other hand, are to develop varieties adapted to withstand harsh environments and low inputs. Hence, selections from landraces and crosses with wild and pnmitive forms are undertaken to produce well-adapted germplasm for targeted agroecological zones. In recent years, world collections of cereals have been evaluated by many scientists working in different countries who were searching for economically useful genes or gene combinations. Confidence has been expressed that such materials are a usable source of breeding stocks, although they still require thorough assessment. Large-scale evaluation, if carried out thoroughly, is an expensive, arduous, and time-consuming process. Therefore, it is imperative to carefully select the traits that one wishes to evaluate in consultation with the breeders. Further, not all material in a collection may be of immediate interest. Priorities need to be discussed and selection of the traits made on the basis of their importance to the actual user. Such a procedure will assure optimal utilization of physical facilities, manpower, and financial resources. Finally, breeding objectives change, sometimes rapidly, and hence evaluation needs to be adaptive to a certain extent to succeed. ACKNOWLEDGMENTS The author thanks Drs. W. Erskine, S. Ceccarelli, and J . Valkoun for their comments and suggestions on a draft of this review and the Government of ltaly for financial support.

REFERENCES Alkamper, von J. 1974. Z. Acker. Pfianzenbau 140, 184-198. Allard, R. W. 1970. I n “Genetic Resources in Plants-Their Exploration and Conservation” (0.H. Frankel and E. Bennett, eds.), pp. 97-108. IBP/Blackwell, London. Allard, R. W., Kahler, A. L., and Weir, B. S. 1970. In “Proceedings of Second International Barley Genetic Symposium” (R. A. Nilan, ed.), pp. 1-13. Washington State Univ. Press, Pullman. Asins, M. J., and Carbonell, E. A. 1986. Theor. Appl. Gener. 72,551-558. Avivi, L. 1978. In “Proceedings of Fifth International Wheat Genetics Symposium” (S. Ramunajam, ed.), pp. 372-380. ICAR, New Delhi. Avivi, L. 1979. Genet. Agric. 4, 27-38. Bakhteyev, F. Kh. 1979. Z . Pjianzenzuechr. 83,211-221.

108

A. B. DAMANIA

Bekele, E. 1984. Hereditas 100, 131-154. Bernatsky, R., and Tanksley, S. D. 1989. In “The Use of Plant Genetic Resources” (A. H. D. Brown, 0. H. Frankel, D. R. Marshall, and J. T. Williams, eds.), pp. 353-362. Cambridge Univ. Press, Cambridge. Blum, A., Sinmena, B., Golan, G., and Mayer, J . 1987. Plant Breed. 99,226-233. Bowden, W. M. 1959. Can. J . Bot. 37,657-684. Brown, A. H. D. 1978. Theor. Appl. Genet. 52, 145-157. Brown, A. H. D. 1989. In “The Use of Plant Genetic Resources” (A. H. D. Brown, 0. H. Frankel, D. R. Marshall, and J. T. Williams, eds.), pp. 136-156. Cambridge Univ. Press, Cambridge. Brown, A. H. D., Frankel, 0. H., Marshall, D. R., and Williams, J. T. (eds.). 1989. “The Use of Plant Genetic Resources”. Cambridge Univ. Press, Cambridge. Burt, R. L., Pengelly, B. C., and Williams, W. T. 1980. Agro-Ecosystems 6, 119-127. Ceccarelli, S., and Grando, S. 1987. Genet. Agric. 41, 131-142. Ceccarelli, S., Grando, S., and van Leur, J. G. 1987. Euphytica 36,389-405. Chang, T. T. 1985. Iowa State J . Res. 59,379-397. Chapman, C. 1989. In “IBPGRTraining Courses: Lecture Series 2. Scientific Management of Germplasm: Characterization, Evaluation and Enhancement” (H. T. Stalker and C. Chapman, eds.), pp. 55-63. IBPGR, Rome. Chennaveeraiah, M. S. 1960. Acta Horti. Gotob. 23, 85-178. Cole, E. W., Fullington, J . G., and Kasarda, D. D. 1981. Theor. Appl. Genet. 60, 17-30. Cox, T. S., Harrel, L. G., Lubbers, E. L., and Amri, A. 1989. Annu. Wheat Newsln. 35, 140- 142. Croston, R. P., and Williams, J. T. 1981. “A World Survey of Wheat Genetic Resources.” IBPGR, Rome. Damania, A. B. 1983. Ph.D. Thesis, Univ. of Birmingham, Birmingham, U.K. Damania, A. B. 1985. Indian J . Genet. 45, 213-220. Damania, A. B., and Porceddu, E . 1981. Plant Genet. Resour. Newsl. 48,2-3. Damania, A. B., and Somaroo, B. H. 1988. Rachis, 7,54-56. Damania, A. B., and Srivastava, J. P. 1989. I n “Proceedings of the Third International Symposium on Genetic Aspects of Plant Mineral Nutrition” (N. El-Bassam and B. C. Loughman, eds.), pp. 425-430. Kluwer, Dordrecht. Damania, A. B., Porceddu, E., and Jackson, M. T. 1983. Euphyrica 32,877-883. Damania, A. B., Jackson, M. T., and Porceddu, E. 1985.2. Pj7anzenzuecht. 94, 13-24. Damania, A. B., Tahir, M., and Somaroo, B. H. 1988. In “Proceedings of the Seventh International Wheat Genetics Symposium” (T. E. Miller and R. M. D. Koebner, eds.), pp. 937-941. IPSR, Cambridge. De Pace, C., Porceddu, E., and Blanco, A. 1978. In “Proceedings of the Fifth International Wheat Genetics Symposium” (S. Ramunajam, ed.), pp. 680-694. ICAR, New Delhi. Dhaliwal, H. S., Gill, K. S., Singh, P. J., Multani, D. S . , and Singh, B. 1986. Crop Improu. 16, 145-449. Dominici, L., Grottanelli, A , , Tomassini, C., Lafiandra, D., Porceddu, E., and Damania, A. B. 1988. Rachis 7,34-36. Ehdaie, B., and Waines, J. G. 1989. Euphyrica 41, 183-190. Elliot F. C., 1957. J. Hered. 48,77-81. Erskine, W., and Williams, J. T. 1980. Plant Genet. Resour. Newsl. 41, 19-33. Feldman, M., and Sears, E. R. 1981. Sci. Am. 244,98-109. Ford-Lloyd, B. V. 1978. In “Conservation of Plant Genetic Resources” ( J . G. Hughes, ed.), pp. 42-45. Univ. of Aston, Birmingham, U.K. Frankel, 0. H. 1987. IBPGR Southeast Asia Newsl. 11, 13.

EVALUATION O F GENETIC RESOURCES IN CEREALS

109

Frankel, 0. H., and Bennett, E. (eds.), 1970. “Genetic Resources in Plants-Their Exploration and Conservation.” IBPIBIackwell, London. Frankel, 0. H., and Hawkes, J. G. (eds.), 1975a. “Crop Genetic Resources for Today and Tomorrow.” Cambridge Univ. Press, Cambridge. Frankel, 0. H., and Hawkes, J. G. 1975b. I n “Crop Genetic Resource for Today and Tomorrow” (0.H . Frankel and J. G. Hawkes, eds.), pp. 1-1 I. Cambridge Univ. Press, Cambridge. Freeman, G. H., and Dowker, B. D. 1973. Heredity 30,97-109. Frost, S . , Holm, G., and Asker, S. 1975. Heriditas 79, 133-142. Gabriel, K. R., and Sokal, R. R. 1969. Sysr. Zool. 18,259-278. Gerechter-Amitai, Z. K., and Stubbs, R. W. 1970. Euphyrica 19, 12-21. Gill, B. S., Raupp, W. J., Wilson, D. L., Gill, K. S., and Howell, K. D. 1989. Annu. Wheat Newsl. 35, 138-139. Grando, S.. Falistocco, E., and Ceccarelli, S. 1985. Genet. Agric. 39,65-76. Hallauer, A. R., and Miranda, J. B. 1981. “Quantitative Genetics in Maize Breeding.” Iowa State Univ. Press, Ames, Iowa. Harlan, J. R. 1975. Science 188,618-621. Harlan, J. R. 1976. Crop Sci. 16,329-333. Harlan, J. R. 1979. In “Evolution of Crop Plants” (N. W. Simmonds, ed.), pp. 93-98. Longman, London. Harlan, J. R. 1984. In “Crop Genetic Resources: Conservation and Evaluation” ( J . W. H. Holden and J. T. Williams, eds.), pp. 212-222. Allen and Unwin, London. Hawkes, J. G. 1971. Outlook Agric. 6,248-253. Hawkes, J. G. 1977. Euphytica 26,615-621. Hawkes, J. G. 1981. I n “Plant Breeding 11” (K. J. Frey, ed.), pp. 57-83. Iowa State Univ. Press, Ames, Iowa. Hawkes, J. G. 1983. “The Diversity of Crop Plants.” Harvard Univ. Press, Camridge, Massachusetts. Hawkes, J. G. 1985. I n “Proceedings of Eucarpia Genetic Resources Section. International Symposium on Evaluation for the Better Use of Genetic Resources Materials” (V. Rogalewicz, ed.), pp. 69-77. Research Institute of Plant Production, Prague. Hillel, J., Feldman, M. W., and Simchen, G. 1973. Heredity 30, 141-167. Holden, J. H. W., and Williams, J. T. (eds.). 1984. “Crop Genetic Resources, Conservation and Evaluation.” Allen and Unwin, London. IBPGR 1982. “Descriptors of Barley.” IBPGR/FAO, Rome. IBPGR 1985. “Revised Descriptors List for Wheat (Triticum spp.).” IBPGR/CEC, Rome. Jain, S. K. 1975. In “Crop Genetic Resources for Today and Tomorrow” (0.H. Frankel and J. G. Hawkes, eds.), pp. 15-36. Cambridge Univ. Press, Cambridge. Jain, S. K., Qualset, C. O., Bhatt, G. M., and Wu, K. K. 1975. Crop Sci. 15,700-704. Jana, S . , and Pietrzak, L. N . 1988. Genetics 119,981-990. Jana, S . , Srivastava, J. P., and Gautam, P. L . 1983. I n “Proceedings of the Sixth International Wheat Genetics Symposium” (S. Sakamoto, ed.), pp. 137-141. Kyoto Univ., Kyoto. Jana, S., Pietrzak, L. N., Morris, M. I., Srivastava, J. P., Holwerda, B. C., and Thai, K. M. 1987. I n “Proceedings of the V International Barley Genetics Symposium,” (S. Yasuda and T. Konishi, eds.), pp. 63-73. Institute for Agricultural and Biological Sciences, Okayama. Kihara, H., Yamashita, K., and Tanaka, M. 1965. In “Cultivated Plants and Their Relatives” (K. Yamashita, ed.), pp. 1-118, Kyoto Univ., Kyoto. Kimber, G., and Feldman, M. 1987. “Wild Wheat-An Introduction.” Univ. of Missouri, Columbia, Missouri.

110

A. B. DAMANIA

Kobrehel, K., and Gautier, M. F. 1973. In “Genetics and Breeding of Durum Wheat” ( G . T. Scarascia-Mugnozza, ed.), pp. 527-536. Univ. of Bari, Ban, Italy. Konarev, V. G., Gavnlyuk, I. P., Gubareva, N. K., and Peneva, T. I. 1979. Cereal Chem. 56, 272-278. Konzak, C. F., Behre, T., Davis, M. A., and Sadam, M. 1973. In “Genetics and Breeding of Durum Wheat” (G. T. Scarascia-Mugnozza, ed.), pp. 179. Univ. of Ban, Ban, Italy. Kosina, R., 1980. Z. Pfanzenzuecht. 85,294-307. Lagudah, E. S., Flood, R. G., and Halloran, G. M. 1987. Euphytica 36,3-9. Lange, W., and Balkema-Boomstra, A. G. 1988. In “Cereal Breeding Related to Integrated Cereal Production” (M. L. Jornaand L. A. J . Slootmaker, eds.), pp. 157-478. Eucarpia, Wageningen, Netherlands. Lawrence, J. M., Day, K. M., Huey, E., and Lee, B. 1958. Cereal Chem. 35, 169-179. Lewontin, R. C. 1974. “The Genetic Basis of Evolutinary Change.” Columbia Univ. Press, New York. Mak, C., and Harvey, B. L. 1982. Euphytica 31,85-92. Maliani, C., and Bianchi, A. 1979. Genet. Agric. 33, 1-14. Mansur-Vergara, L., Konzak, C. F., Gerechter-Amitai, Z. K., Grama, A., and Blum, A. 1986. Theor. Appl. Genet. 72,296-301. Mengesha, M. 1975. In “Crop Genetic Resources for Today and Tomorrow” (0.H. Frankel and J. G. Hawkes, eds.), pp. 449-453. Cambridge Univ. Press, Cambridge. Moms, R., and Sears, E. R. 1967. In “Wheat and Wheat Improvement” (K. S. Quinesberry and L. P. Reitz, eds.), pp. 19-87. American Society of Agronomy, Madison, Wisconsin. Murphy, P. J., and Witcombe, J. R. 1981. In “Proceedings of FAO/UNEP/IBPGRTechnical Conference on Crop Genetic Resources.” IBPGR, Rome. Murphy, P. J., and Witcombe, J. R. 1986. Theor. Appl. Genet. 71,730-735. Negassa, M. 1985a. Hereditas 102, 123-138. Negassa, M. 1985b. Hereditas 102, 139-150. Negassa, M. 1986. Plant Breed. 97, 147-162. Nevo, E., and Payne, P. I. 1987. Theor. Appl. Genet. 74,827-836. Nevo, E., Zohary, Brown, A. H. D., and Haber, M. 1979. Euolution 33,815-833. Nevo, E., Golenberg, E., and Beiles, A. 1982. Theor. Appl. Genet. 62,241-254. Nevo, E., Beiles, A., and Kaplan, D. 1988. Heredity 61,31-45. O’Toole, J . C., and Bland, W. L. 1987. Adu. Agron. 41,91-145. Peeters, J. P. 1988. Theor. Appl. Genet. 76, 17-24. Peeters, J. P., and Martinelli, J. A. 1989. Theor. Appl. Genet. 78, 42-48. Plucknett, D. L., Smith, N. J. H., Williams, J. T., and Anishetty, N. M. 1983. Science 220, 163-169. Pogna, N., Lafiandra, D., Feillet, P., and Autran, J. C. 1988. J . Cereal Sci. 7,211-214. Porceddu, E. 1976. Z. Pfanzenzeucht. 77,314-329. Porceddu, E., and Penino, P. 1973. Plant Genet. Resour. Newsl. 30,33-36. Porceddu, E., and Scarascia-Mugnopzza, G. T. 1983. In “Proceedings of the Sixth International Wheat Genetics Symposium” (S. Sakamoto, ed.), pp. 241-252. Kyoto Univ., Kyoto. Porceddu, E., Perrino, P., and Olita, G. 1973. In “Genetics and Breeding of Durum Wheat” (G. T. Scarascia-Mugnozza, ed.), pp. 181-200. Univ. of Bari, Bari, Italy. Poyarkova, H. 1988. Euphytica 38, 11-23. Pun, Y. P., and Qualset, C. 0. 1978. Phyton 36,41-51. Qualset, C. 0. 1975. In “Crop Genetic Resources for Today and Tomorrow” (0.H. Frankel and J. G. Hawkes, eds.), pp. 81-96. Cambridge Univ. Press, Cambridge. Qualset, C. O., and Pun, Y. P. 1973. In “Genetics and Breeding of Durum Wheat” ( G . T. Scarascia-Mugnozza, ed.), pp. 165-178. Univ. of Bari, Bari, Italy.

EVALUATION O F GENETIC RESOURCES IN CEREALS

111

Reed, C. A. 1969. In “The Domestication and Exploitation of Plants and Animals” (P. J. Ucko and G. W. Dimbleby, eds.), pp. 361-380. Duckworth, London. Renfrew, J. M. 1969. I n “The Domestication and Exploitation of Plants and Animals” (P. J. Ucko and G. W. Dimbleby, eds.), pp. 149-172. Duckworth, London. Riley, R., Chapman, V., and Johnson, R. 1968. Nature (London) 217,383-384. Robertson, B. M., Waines, J. G., and Gill, B. S. 1979. Crop Sci. 19,843-847. Robinson, P. J., Burt, R. L., and Williams, W. T. 1980. Agro-Ecosystems 6, I 1 1-1 18. Rogalewicz, V. (ed.). 1985. “Proceedings of the Eucarpia Genetic Resources Section. International Symposium on Evaluation for the Better Use of Genetic Resources Materials.” Research Institute of Plant Production, Prague. Sakamoto, S., and Fukui, K. 1972. Kyoto Univ.Afr. Stud. 7, 181-225. Sharma, H. C., Waines. J. G., and Foster, K. W. 1981. Crop Sci. 21,555-559. Shevchuk, T. 1973. Plant Genet. Resour. Newsl. 29,2-6. Smillie, R. M., and Nott, R. 1982. Plant Physiol. 70, 1049-1054. Somaroo, B. H., Adham, Y. J., and Mekni, M. S. 1986. “Barley Germplasm Catalog I , 1986.” ICARDA, Aleppo, Syria. Somaroo, B. H., Abiad. G. F.. Humeid, B. 0.. and Srivastava, J. P. 1988. “Barley Germplasm Catalog 11, 1988.” ICARDA, Aleppo, Syria. Spagnoletti Zeuli, P. L., and Qualset, C. 0. 1987. Crop Sci. 27,235-241. Srivastava, J. P., and Damania, A. B. 1989. In “The Use of Plant Genetic Resources”, (A. H. D. Brown, 0. H. Frankel, D. R. Marshall, and J. T. Williams, eds.), pp. 88-104. Cambridge Univ. Press, Cambridge. Srivastava, J. P., Damania, A. B., and Percetti. L. 1988. I n “Proceedings of the Seventh International Wheat Genetics Symposium” (T. E. Miller and R. M. D. Koebner, eds.), pp. 153-158. IPSR, Cambridge. Takahashi, R., Hayashi. J., Yasuda, S., and Hiura, U. 1965. In “Cultivated Plants andTheir Relatives” (K. Yamashita. ed.), pp. 141-168. Kyoto Univ., Kyoto. Tani, Y., and Sakamoto, S. 1987. In “Domesticated Plants and Animals of the Southwest Eurasian Agro-Pastoral Culture Complex” (S. Sakamoto, ed.). pp. 47-69. Kyoto Univ., Kyoto. Tolbert, D. M., Qualset, C. O., Jain, S. K., and Craddock, J. C. 1979. Crop Sci. 19,789-794. Vallega, V., and Waines, J . G. 1987. Theor. Appl. Genet. 74,706-710. van Leur, J. A. G., Ceccarelli, S., and Grando, S. 1989. Plant Breed. 103,324-335. Vavilov, N. I . 1926. Bull. Appl. Bot. Plant Breed. 16, 1-245. Vavilov, N. I. 1932. Bull. Appl. Bot. Plant Breed. Leningrad, Suppl. N o . 55. Vogel, K. P., Johnson, V. A., and Mattern, P. J. 1973. Agric. Exp. Sin., Univ. of Nebraska, Coll. Agric. Res. Bull. 258. Waines, J. G., 1983. In “Proceedings of the Sixth International Wheat Genetics Symposium” ( S . Sakamoto, ed.), pp. 115-122. Kyoto Univ., Kyoto. Waines, J. G., Ehdaie, B., and Barnhart, D. 1987. Genome 29,41-46. Ward, D. J. 1962. USDA Tech. Bull. 1276, United States Government Printing Office, Washington, D.C. Weltzien, E. 1982a. Plant Genet. Resour. Newsl. 52, 5-6. Weltzien, E. 1982b. Rachis 1,6-7. Williams, J. T. 1989. I n “The Use of Plant Genetic Resources” (A. H. D. Brown, 0. H. Frankel, D. R. Marshall, and J. T. Williams, eds.), pp. 235-244. Cambridge Univ. Press, Cambridge. Williams, W. T., Burt, R. L., Pengelly, B. C., and Robinson, P. J. 1980. Agro-Ecosystems 6, 99-109. Witcombe, J. R. 1975. Euphytica 24,431-434. Yamashita, K. 1980. Z. Pjanzenzuecht. 84, 122-132.

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY, VOL. 44

MODELING CROP ROOT GROWTH AND FUNCTION Betty Klepper and R. W. Rickman United States Department of Agriculture Agricultural Research Service Columbia Plateau Conservation Research Center Pendleton, Oregon 97801

I . Introduction 11. Early Models 111. Desirable Model Features 1v. Model Components A. Root Classification B. Root Growth Parameters for Modeling C. Root-Soil Relationships in Growth D. Root-Soil Relationships in Uptake Functions V. Some Existing Root Growth and Function Models VI. Limitations to Development of Root Growth Models References

I. INTRODUCTION Many crop growth models that quantify plant uptake of water and solutes from soils require a quantitative description of the root system and its location in the soil profile. They require information like root length density distribution with depth and changes in that distribution over time. To sample the soil-root system and measure its properties over time in the growing season is labor-intensive. Therefore it is useful to have models that relate the generation of new root material and the decay of old roots to plant properties and to soil conditions at various profile depths. Furthermore, root growth and death models are needed for use in calculating fluxes of carbohydrates into below-ground organs for the detailed, physiologically based crop growth models presently being built. Such information as numbers and locations of root meristems, elongating zones, and aging zones, with their associated rates of water and nutrient uptake and their suites of exudates, are needed for interfacing to models that will eventually be written to describe rhizosphere microbial dynamics and

114

BETTY KLEPPER A N D R. W. RICKMAN

activities of specific root pathogens. This chapter describes some of the ideas and approaches that are being taken in developing root growth models for incorporation into ecosystem-level crop models.

II. EARLY MODELS Most early root growth models were developed to provide a “root sink” to absorb the soil water or nutrients lost from the soil profile. They contained little specific information on the root system itself. If one assumes that root length density profiles can be used to express the root sink, then one appropriate model has been a simple exponential distribution of roots with depth (Gerwitz and Page, 1974). This empirical model is sufficient for many crops and gives a simple mathematical formula to calculate the root sink activity at any soil profile depth. However, under management with intermittent rainfall, such models do not predict observed root profiles. For example, Fig. 1 shows the root length density profile for cotton roots in soil profiles that were either well watered or allowed to dry (Klepper er al., 1973). An exponential decrease of roots with depth would describe fairly adequately the roots in the wet profile, but not in the dry one. Unfortunately, root sink activity is not always proportional to the root length density profile (Reicosky et al., 1972), making it difficult to provide root data appropriate for use in water uptake models. The lack of agreement of observed rooting with a macroscopic (root sink) description of root function and the additional lack of fit with empirical exponential root distributions have led to the need for describing root systems at the microscopic level. The first attempts at the microscopic approach presented the root as a cylinder surrounded by a concentric cylinder of soil. Root length density was obtained from root observations or empirical distributions with depth and various estimates of water or nutrient uptake were computed using appropriate diffusion coefficients. The level of our knowledge of root systems is such that we no longer have to resort to direct measurement or to these empirical models to provide estimates of root length density changes to local soil properties. Researchers have proceeded to generate biologically based models that account for root physiology. Some of these models allow computations of root growth and decay to respond dynamically to shoot physiological activity and to changing soil conditions at different depths in the soil profile (e.g., Huck and Hillel, 1983).

MODELING CROP ROOT GROWTH AND FUNCTION

115

ROOT DENSITY ( cm/cm3) 0.0

n

0.6

W

x !3 a w n 1.2 July

1.8 1

I

I

We1 I - w a t e r e d

I

I

I

Profile

I

Drying Profile

FIG. 1. Root length density profiles for cotton plants in well-watered and drying soils. After Klepper et d.(1973).

Ill. DESIRABLE MODEL FEATURES Figure 2 shows the desired output from a root growth model for use in a model of root function. The top panel of Fig. 2 shows the time-averaged root length densities for each depth for a month during the growing season. These data come from a drying experiment on cotton (Gossypiurn hirsuturn) at the Auburn Rhizotron (see Table 5 of Browning et al., 1975). In the rhizotron, nondestructive root counts on glass viewing surfaces were made every other day. From the slopes of these graphs in Fig. 2 we can calculate that the average daily change rate over all depths was 0.0081 m.m-3.sec-' (0.07 ~ m . c m - ~ . d a y - ' ) ; it ranged from 0 to 0.028 m.m-3.sec-' (0.24 ~ m . c m - ~ - d a y - ' )For . a 0.1-m depth increment under 1 m2 of land surface, this result means that an average change in length of 70 m of root per day was detectable by the nondestructive

116

BETTY KLEPPER AND R. W. RICKMAN

!iiV DAY 6

DAY 16

DAY 26

RLD (crn/crn') RLO (crn/crn') RLD (crn/crn') 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5

0 120 W a 150

180

FIG.2. Root length density at five depths over a 30-day period in the growing season. Data are from Browning et a / . (1975).

techniques used in this rhizotron experiment. The line graphs (Fig. 2, bottom) shown for three dates in the growing season represent the usual sampling frequency of the most intensive field trials (every 10 days). These data illustrate the inability of such infrequent sampling to capture root growth rates for model validation. The figure demonstrates the necessity of frequent sampling to calibrate and validate root growth models. Designation of the spatial arrangement of roots with respect to the crop row and the soil profile should be relatively flexible in root growth models to allow for studies involving plow pans, straw layers, high-fertility zones from band placement, and other special treatments. For most purposes, however, horizontal layers 0.1 to 0.3 m thick have been used in a onedimensional geometry. The detailed assortment of root systems into twodimensional arrays of cubes 1 cm on a side (Lambert et al., 1976; BarYosef et d , 1982) is useful if differentiation is made of the roots near the crop row and of their interaction with banded fertilizers or other amendments.

GENETIC MANIPULATION OF THE COWPEA

145

(2n=4x=32) revealed enhanced resistance to the wilt fungus Verticillum alboatrum in protocloned and callicloned populations (Latunde-Dada, 1983; Latunde-Dada and Lucas, 1983, 1988). Some of the genotypes in these populations exhibited differences in gross morphology and ploidy. It was also possible, through the incorporation of toxic metabolites of the fungus V . albo-atrum into plant growth media, to select tissues and plantlets that were incrementally more resistant to this wilt fungus than the unselected populations of the parental variety Europe (Latunde-Dada and Lucas, 1988). Similar results on cellular selection have been reported in the interaction between maize and Helminthosporium maydis (Gengenbach and Green, 1975; Gengenbach et al., 1977), oilseed rape and Phoma lingam (Sacristan, 1982), potato and Phytophthora infestans (Behnke, 1979, 1980), and sugarcane and Helminthosporium sacchari (Nickell, 1977). The underlying causes of genetic somaclonal variation include, among others, polyploidy , chromosome rearrangements, point mutations, somatic crossovers, transposable elements, and rearrangements and recombinations among cytoplasmic organelle genomes (Orton, 1984). These events may be visualized as mitotic accidents that constitute a means of generating biological diversity. It remains possible to harness this powerful mode of evolution in extending further the gene pool of the cowpea and enhancing its adaptability to the challenges of the Rain Forest Belt of Nigeria, in particular, as well as elsewhere.

CULTURE B. EMBRYO Mention has been made of the problem of sexual incompatibility barriers among species of the genus Vigna. A determined assessment of this constraint suggests the need for techniques that in breaking and circumventing these barriers, will aid the exploitation of gene transfer from dissimilar species into Vigna unguiculata. Three techniques of plant tissue culture, namely, embryo culture, somatic hybridization, and transformation, are suggested for this task. The methods of embryo culture have made possible the in uitro crosspollination and fertilization of flowers, and the rescuing as well as nurturing of the resulting hybrid proembryos and embryos, which rarely survive in natural crosses. Examples abound of successful interspecific hybrids obtained via embryo culture, including the crosses Brassica oleracea X B . campestris (Nishi et al., 1959; Snell, 1978) to produce Brassica napus in higher frequencies, and Nicotiana tabacum X N . rependa, N . stocktonii, and N . nesophila (Reed and Collins, 1978). In these cases the possibility has been expressed (Brettell and Ingram, 1979) for the transfer

118

BETTY KLEPPER AND R. W. RICKMAN

with statistical distributions is perfected, the location of roots in the soil will no longer be routinely assumed to be horizontally uniform. It should become possible to quantify the degree of exploration of variously structured soil profiles by tap, modified tap, and fibrous root systems (Logsdon et al., 1988; Logsdon and Allmaras, 1989). The validation of detailed physiologically based root growth and function models is very labor-intensive. With minirhizotron techniques (Taylor, 1987), root length density evaluations can be done more frequently than with coring techniques, but validation of most root growth models is still feasible only at intervals of every few days. Consequently, relatively long time steps of a day or two are probably adequate for root growth models. However, some aspects of root function models (water or nutrient uptake) may require shorter time steps because the diurnal transpiration pattern imposes a diurnal pattern on water uptake rates.

IV. MODEL COMPONENTS A. ROOT CLASSIFICATION Early models merely generated root length density profiles and made no differentiation as to age or root type. However, we need models that distinguish between long-lived roots such as taproots, which function primarily as transporting roots during most of their life span, and the shorter-lived roots that serve primarily absorptive functions. In many models, it is possible to keep track of age classes of roots because discrete lengths of root at each depth increment are generated each day (Diggle, 1988). The mix of root age classes is sensitive to antecedent soil conditions. For example, the root system shown in Fig. 1 for the drying profile on July 29 would have a high proportion of old roots in the top half of the profile and a high proportion of young roots in the bottom half, but the well-watered profile would have a more even distribution of young and old roots because roots were being generated throughout the profile at approximately the same rate at which they were decaying. Thus, age distribution patterns are dynamic and models should be sufficiently sensitive to capture this aspect of rooting patterns. Other root classifications may be of use in modeling. The distinction of the seminal and crown roots of cereals is important because of the difference between these two root types in time of appearance and in potential depth of rooting (Elkins, 1987). Distinctions among axes and first- and second-order branch roots may have some use if they are explicit in a

MODELING CROP ROOT GROWTH AND FUNCTION

119

model because the higher-order branches are often ephemeral absorbing roots and main axes and sometimes first-order laterals are the central transporting roots. Distinctions among roots functioning in transport and absorption have proven necessary in water extraction models that specify axial resistances to transport through roots (Klepper et al., 1983). One useful approach would be to organize roots into the three classes shown in Fig. 3. One class (A) is the downward-growing main taproots, axes, or early branches that take over downward-growing dominance from a taproot and serve primarily as transporting roots during most of their existence. Although these vertical roots certainly go through an absorbing phase early in their life span and may absorb some materials throughout their lives, their most significant function in plant physiology is in the transport of materials absorbed by younger, smaller branch roots. The second class (B) is first-order lateral roots, which may be horizontal or vertical in their growth direction and often grow at an angle for a distance of several centimeters before turning downward. Their function may be either transport or absorption depending on their age and the species being studied. Finally, the third class ( C ) of roots is second-order and higherorder laterals that are primarily absorptive and ephemeral. Unfortunately, available data sets for classifying roots into these categories are rare.

A - Vertical axis

B

-

First-order laterals

C

-

Higher-order laterals

MODIFIED TAPROOT SYSTEM

STRICT TAPROOT SYSTEM

y

F I BROUS SYSTEM

*B

A A

A

A

FIG.3. The three classes of roots as applied to a strict taproot system, a modified taproot system, and a fibrous system.

120

BETTY KLEPPER AND R. W. RICKMAN

This classification system works well for monocots, which have root systems composed of a series of downward-growingaxes. At germination the radical usually grows vertically downward. The radical is followed by other axes from successive nodes on the plant. These axes grow for a short distance at successively increasing angles from the vertical and then turn downward to vertical. Below the 0.3 m depth most cereal axes are vertically oriented if soil conditions permit. The net result of this pattern of axis production and growth is that successive axes explore wider and wider cylinders of soil as the season progresses. Figure 4 shows photographs of wheat roots excavated from the field using a 0.17 x 0.17-m-square sampling tube (Belford et al., 1986). A side was removed from the tube and 2-cm sections of soil were washed out, leaving 2-cm horizontal layers of soil in place to hold the roots in their original relationship to one another. The A roots are the main vertical axes that connect to the culms. The B and C roots are thinner and shorter and presumably are primarily absorptive. There is a tendency for roots of grasses to be less ephemeral than those of dicots (Huck and Hillel, 1983). Dicotyledonous root systems of annuals are generally of two types: taprooted and modified taprooted (Fig. 3). The taprooted plant has one central root that grows vertically and explores ever-deeper layers of soil as the season progresses. This taproot undergoes secondary growth and may become woody. In some taprooted species at least, the B roots grow horizontally or at an angle to the vertical for a time and then turn downward (Stone et al., 1983; Stone and Taylor, 1983). They may undergo secondary growth to become transporting roots.

B. ROOTGROWTHPARAMETERS

FOR MODELING

Figure 5 shows a hypothetical example of the information available from a root growth model with the three classes of roots distinguished and with the age distributions for roots in each of those classes at each depth. This output would provide more definition of root condition, age, and origin than that currently available in models describing uptake of materials from soils. Furthermore, the root length density profile breakdown into classes could be verified in the field because the root types have morphological criteria as their primary defining property. Some modelers would like to use root diameters to calculate root surface areas for each class. That calculation could be incorporated into the preceding output by assigningan average diameter for each class and/or by generating a root diameter in response to local soil conditions. Generally axes have larger diameters than first-orderlaterals, which in turn are wider

MODELING CROP ROOT GROWTH AND FUNCTION

121

FIG.4. The top meter of field-grown wheat root systems showing orthogonality of the axes and branches except for the upper 0.3 m, which has a conical arrangement.

122

BETTY KLEPPER AND R. W. RICKMAN

FIG. 5. Theoretical output from a root growth model with root length densities divided into the three classes of root (A, B, and C), with the age distribution of each root class shown in the bar graphs below.

than higher-order laterals, but the actual diameters encountered reflect plant species, plant conditions, and local soil conditions, thus making assignment of fixed diameters to each class of questionable validity. Also monocot axes often slough off their cortex, making their diameter smaller as they age. In dicots, secondary growth occurs and the diameters of A and B roots change over the season. Modelers of carbohydrate partitioning factors need information on the daily carbohydrate requirements for a growing root system. For each monocot root class, a root lineal density (weight per unit length) can be assigned and the carbohydrate needed to produce each length of root calculated. Respiratory losses are more difficult to assess since they vary

MODELING CROP ROOT GROWTH AND FUNCTION

123

both with root diameter class and with root age. Nevertheless, the ideal root growth model should account for respiratory losses within the root as well as for the losses of organic compounds exuded from the root and respired by rhizosphere organisms. For those models that use the number of rneristems to partition carbohydrates (Charles-Edwards, 1984), it would be useful to keep track of the number of meristems at each depth for each class of root. The number and location of meristems and zones of elongation would also be of interest to researchers concerned with root-pathogen interactions since young root tips are a point of attack for certain microorganisms such as Pythium (Cook and Haglund, 1982), whereas older surfaces are more subject to other organisms such as Fusarium (Skipp et al., 1982). Finally, some models require information on distance to the nearest neighbor to calculate the sphere of influence of the root. The formula (Barley, 1971) relating average distance between neighbors (4to the root length density (Lv),

d

=

4 (Lv)-‘I2

would probably suffice. However, this formula does not take into account the fact that roots tend to follow preferential paths (old root channels, worm holes, natural fractures in the soil) and tend to be clumped rather than random in distribution (Wang et a f . , 1986). It would also be desirable to retain some of the developmental relationships of root axes and their branch lengths so that the “plumbing” of the root system is explicit in the model. This is relatively easy to simulate but very difficult to validate. Laborious excavations and root-washing schemes are required (Belford et al., 1986). C. ROOT-SOILRELATIONSHIPSI N GROWTH Elongation rates of roots in soils depend on many factors, some of which have major impact and others of which are less important. Physical factors include soil strength, soil oxygen diffusion rate, soil temperature, and water potential. Chemical factors include soil pH, activities of certain ions such as aluminum, borate, and calcium, soil osmotic potential, and soil fertility. Basic concepts for use in modeling cell elongation were discussed by Lockhart (1965) and modified for roots in soil to include soil resistance to root penetration by Greacen and Oh (1972). The driving force for elongation is the turgor pressure in the expanding cells of the elongation zone. Resisting this are constraints in the cell wall and the constraint of soil. These concepts can be simplified to the following relationship:

124

BETTY KLEPPER AND R . W. RICKMAN

1dL =

L dt

(b(P-Y-M)

where

L is root length (m), (b is wall extensibility (S'aMPa-'), P is turgor pressure (MPa), Y is the minimum turgor required for expansion (MPa), and M is the resistance of soil to root penetration (MPa). Very little is known about the influence of environment, plant age, and other factors on (b, P, and Y, but most modelers of root growth in soil concentrate on the M term and assume that 4, P, and Yare constant. Since most crop models are intended for use in agricultural soils where toxic materials are not present at sufficient concentrations to influence root elongation and where tillage and other management history has provided adequate fertility and minimal soil compaction problems, it may be possible to simplify the modeling of root elongation rate with respect to soil properties at least for some cropping situations. For example, assume that the principal factor that changes during the growing season is soil water content. Because it is possible to relate soil strength to soil water content for the specific soil profile being modeled, it is possible to calculate root elongation rate as a function of soil strength, which could be determined separately for each depth increment from the soil water budget. It would be possible to use separate relationships for each depth increment, but in practice it is probably only necessary to specify the plow layer, any compacted zones, and the profile below the traffic-compaction zone. One property that could be specified for each layer might be the probability of a root encountering a low-strength spot in the soil. A high presence of old root and worm channels would raise this probability, for example. Although this profile simplification would not take care of acid subsoils or other common problems, it would be adequate for many agricultural soils. A more sophisticated approach has been taken by Jones and associates (Jones, 1990). They account for changes in soil resistance to root penetration based on soil texture, water content, bulk density, aeration, and temperature. The branching rate (number of branches per unit length of parent root) is also difficult to model because it varies both with local soil properties and with general plant vigor. Local pockets of nutrients such as P and N (Drew, 1975;Maizlich et al., 1980)can influence the branching rate considerably, as can local soil water content and aeration. Realistic ranges of

MODELING CROP ROOT GROWTH AND FUNCTION

125

values are available from the literature but sensitivity relationships to local soil properties such as water content, nutrient concentration, and strength are poorly understood. Even worse is our understanding of the relationship of root branching rate to plant vigor and photosynthetic history or the relationship of death of the main root apex to subsequent branching behavior. Nevertheless, we believe that fairly representative numbers for branching rates can be assumed from literature values for early modeling efforts. For wheat axes, the branching rate averaged about 3 per centimeter for one field experiment on winter wheat with a maximum of 8 per centimeter in a surface layer where ammonium nitrate had been applied in the early spring (Belford et al., 1987). The timing of branching for any particular length of root is a function of root maturity (age). Behind the zone of elongation is the zone of maturation, usually seen as the zone with young root hairs, and behind that is the zone of the root that produces branches. A certain amount of developmental time is required before a new segment of root can branch. For wheat (Triticum aestiuum L.), this period of branching can be measured in degree-days and is about 250 GDD (base temperature, OOC)(Klepper et al., 1984). For peas (Pisum satiuum L.), it is about 200 GDD (B. Klepper, 1984). By combining time and temperature into GDD, it is possible to simplify the modeling of root development (Gregory, 1987). In summary, the complex assortment of factors that influence root elongation rate, branching time, and branching rate can be simplified for early modeling efforts by using GDD for timing events and by using a relationship between soil water content and root elongation rate calibrated for the particular soil profile being worked with. Eventually, concepts such as those discussed by Gupta and Larson (1982) or Jones (Jones, 1990) may be used to generalize the relationship between strength and water content from soil textural information.

I N UPTAKE FUNCTIONS D. ROOT-SOILRELATIONSHIPS

Substances move into plant roots from the surrounding soil solution as a result of diffusion processes or “active” energy-requiring mechanisms or both. Formulas for calculating root uptake rates generally have an explicit expression for the activity (or concentration) of the material inside and outside the root. For example, in the case of water, the water potential difference provides the driving force for the diffusion of water from soil to root (Taylor and Klepper, 1978). For ion uptake, charge must also be

126

BETTY KLEPPER AND R . W. RICKMAN

accounted for (Nye, 1984). All of these models require some expression of the root length, root fresh weight, or root surface area over which the uptake occurs. This approach neglects activity of microorganisms in the rhizosphere, which most modelers ignore. These microorganisms can have significant impacts on nutrient uptake (Barber and Loughman, 1967; Cushman, 1984). Root uptake models can be written from either the macroscopic or the microscopic point of view. The macroscopic view looks at the whole soil profile and the root system as a uniformly “smeared” sink at any depth or a “root sink.” Uptake is assigned to the root by disappearance of the material from the soil. The microscopic view looks at the root as a cylinder surrounded by a concentric cylinder of soil that represents the average volume of soil that “belongs” to each root. In this approach, materials move radially toward the root across the rhizosphere and the total plant uptake is the sum of the uptakes from each length of root in each profile layer. Historically, macroscopic models preceded microscopic ones primarily because of the great difficulty in specifying the “effective” root length (surface area, etc.) for field conditions. Most present-day models are microscopic as a result of recent improvements in measurement of root systems (Bohm, 1979). Movement of nutrient ions from the soil or soil solution to the root surface, into the root, and to the shoot has been described by mechanistic models. For example, the nutrient uptake model of Claassen and Barber (1976) provides estimates of absorption of specific ion species based on their particular concentration in soil solution, diffusion characteristics, and a maximum intake rate by the root for that ion. The surface area of root present for absorption is provided as either an exponential or linear function of soil depth depending on the stage of crop growth. This model, as well as the modification provided by Cushman (1984) to account for competition between adjacent roots, focuses on uptake mechanisms as described by soil supply and transport properties and potential root intake rate. Two parameters (maximum intake rate and ion concentration in solution at one-half of the maximum intake rate) are generally used in the Michaelis-Menten relationship to describe the root absorption characteristics. When this model is applied to ions such as phosphate or potassium, Epstein (1976) suggested that two uptake mechanisms may apply, one at low ion concentration and another at high ion concentration. Work on interactions of root-inhabiting microorganisms with nutrient uptake (Barber and Loughman, 1967) has shown that the presence and kind of microorganism have a major effect on the uptake process. The plant root system in a field soil is effectively enclosed in a “glove” of living microbes.

MODELING CROP ROOT GROWTH AND FUNCTION

I27

These organisms apparently have first call on nutrients moving to the plant. At low soil solution nutrient concentrations, the amount of nutrient utilized by microbes at the root surface is a major fraction of the total amount removed from soil. The nutrient that actually moves into the shoot is that absorbed from soil in excess of consumption by root surface microbes. As summarized by Tinker (1984), root surface organisms can be either detrimental or beneficial to net nutrient supply to plant shoots depending on efficiency of absorption from soil and the fraction of nutrient passed on to the shoot. The two uptake mechanisms for low and high nutrient concentrations may be replaced by an uptake model that accounts for the activity of the surface-living organisms and the root itself or for the activity of the root cortex and the stele. Unfortunately, many experiments measure only the appearance of the element in the shoot and do not quantify the pools involved in the uptake pathway. Prediction of total nutrient uptake as measured by nutrient content of the shoot depends on both the ion absorption mechanism as discussed in the foregoing and the effective root length of the plant. Root length has been calculated as a mathematical (exponential or linear) distribution with depth, or as nonuniform and based on local soil environment and root senescence. The nature of the root length distribution predicted by a model will have a major impact on computed uptake of any nutrient. The amount of nutrient in the plant shoot will, in turn, influence the continued growth of both the shoot and the root. The interaction of root and shoot growth and function will be critical in any working whole-plant growth model. Since shoot growth will depend on the amount of limiting nutrient provided by the root system, the root growth will depend on the amount of carbohydrate provided by the shoot, the feedback and control of one on the other are vital for accurate estimation of total plant development and growth. Very few current simulations attempt to predict the shoot and root growth and function of a crop using linked mechanistic models for both. Root respiration estimates are sometimes made to help provide a realistic distribution of shoot-produced photosynthate, but the difficulty of obtaining measures of such quantities is accentuated by the consideration of the root surface microorganisms, which themselves are estimated to consume from 1% to 10% of total shoot photosynthate (Tinker, 1984). Progress in modeling nutrient uptake will occur when the root-shoot interaction is described to provide day-to-day estimates of root length increase. With an improved daily root length estimate, the uptake mechanism and the impact of root surface and rhizosphere organisms can be incorporated and daily uptake rates calculated.

128

BETTY KLEPPER AND R. W.RICKMAN

V. SOME EXISTING ROOT GROWTH AND FUNCTION MODELS In this section we propose to discuss some of the current concepts used to generate root growth models for crop growth simulations. The coverage is not intended to be complete. We merely wish to illustrate the different approaches that have been taken by modelers to date and to illustrate the usefulness of each approach. One of the earliest concepts of root modeling involved the partitioning of carbohydrates to create belowground biomass. For example, the partitioning coefficient has been assumed to be a constant fraction of the daily photosynthate produced or an amount left over from shoot requirements. One generalized perennial grass model requires roots to grow on a reserve carbohydrate pool (Coughenour et al., 1984). Johnson and Thornley (1983) allocated 10% of the carbohydrate to roots in an established grass sward model. A slightly more physiologically based concept is to partition photosynthate to roots as afunction of the ratio of root to shoot phytomass. For example, Fishman et a / . (1984) assumed that sink strength is proportional to sink size, as measured by current dry matter, in partitioning photosynthate among the leaves, stems, roots, and tubers of potato. This same concept is used by Skiles et al. (1982)in a model for grasses, based on ideas used in the ELM model, which was developed as a part of the U.S. Grassland Biome of the International Biome Project. Skiles et a / . (1982) use a simple proportionality until senescence begins, at which time they allocate a species-dependent constant fraction of photosynthate to roots. They also make root respiration rate and mortality depend on temperature and water stress, lending an environmental sensitivity to the simulation. Temperature has been used to regulate allocation of dry matter to wheat roots, with the minimum allocation occurring at the optimum shoot growth temperature and with increasing proportions going to roots as temperatures either rise or fall from 20°C (O’Leary et a / . , 1985). These authors use a degree-day (4°C base) model to simulate maximum rooting depth of wheat. Restrictions caused by dry soil are obtained by downscaling root growth rate from the maximum rate in proportion to how far the soil available water falls below 30% of the maximum in the zone into which roots are extending. Another approach has involved both carbon and nitrogen nutrition (Davidson, 1969; Reynolds and Thornley, 1982; Johnson, 1985).The partitioning algorithm used by Reynolds and Thornley (1982) adjusts the shoot to root partitioning of photosynthate depending primarily on the nitrogen to carbon ratio in the plant’s labile storage pool, although there is an addi-

MODELING CROP ROOT GROWTH AND FUNCTION

129

tional factor that determines the degree of control exerted by the plant on partitioning. Johnson and Thornley ( 1985) incorporated these concepts into a grass crop model that divides root structure into four components (growing, newly expanded, medium-aged, and senescing). This compartmentalization allows calculation of a “live root structural dry weight” and also permits separate uptake parameters and respiratory maintenance costs to be inserted for the different ages of root. The model also makes nitrogen uptake by roots dependent on root content of soluble carbohydrate. A different approach was taken by Huck and Hillel (1983), who partitioned carbohydrates to roots and shoots according to the relative values of plant water potential in the plant parts. This means that under optimal moist soil conditions, shoot growth is favored, but roots in moist soil become more and more favored as the profile dries. Their model also takes some account of the effect of local soil temperature. It generates root length by assuming constant relationships between root mass and root length and diameter. Root death depends on soil temperature and the concentration of reserve carbohydrate. These same concepts were incorporated into ROOTSIMU V 4.0 (Hoogenboom and Huck, 1986) and ROOTSIMU V 4.3 (Hoogenboom et al., 1987),which also take account of the local decreases in root extension and branching as soil dries. Of similar complexity is “RHIZOS,” a root simulation that generates daily root length changes in small cells (volumes of soil, generally I cm3) placed specifically with respect to the crop row (Lambert et al., 1976; Bar-Josef et al., 1982; Whisler et al., 1986). Soil mechanical impedance is calculated for each cubic centimeter of soil and the difference between root turgor pressure and the threshold turgor required for expansion is calculated. These parameters are used to calculate the potential root growth reduction caused by mechanical impedance and soil water potential. Further potential reductions are calculated daily for impacts of unfavorable soil temperature and oxygen partial pressures. Then the most limiting of these potential reductions is used to calculate the rate of change of root dry weight in each cell. Finally, the carbohydrate left over from making required shoot parts is distributed to the cells where root growth conditions are most favorable. Jones and associates ( C . A. Jones, Temple, Texas) have refined the impacts of local soil conditions on root growth even further. They address aluminum toxicity and calcium deficiency through user-entered soilspecific parameters; soil texture through a reduction in root growth proportional to the volume fraction of particles larger than 2 mm in diameter; soil strength through input values of soil water content, bulk density, and sand content; aeration through consideration of water-filled pores, critical

130

BETTY KLEPPER AND R . W. RICKMAN

values of water-filled porosity for the soil, and a plant-dependent factor for sensitivity to flooding; and temperature from the local temperature of each soil layer and a minimum and optimum temperature for growth characteristic of the species. They assume that 1% of the root dry weight in a soil layer normally dies each day, but allow low soil moisture or poor aeration to increase this rate up to a maximum of 2% per day. A completely different approach is taken in “WHTROOT” by Porter et al. (1986). They describe cereal root system growth and development as degree-day driven. Each nodal axis develops at a set point in plant developmental time, grows vertically at a constant rate (mm per GDD, TB = O ) , and branches after 250 GDD have elapsed. The model contains no restrictions to root growth from stresses, and it assumes different rates of elongation, lineal densities, and diameters for vertical axes, first-order laterals, and higher-order laterals. Several concepts from this model were incorporated into a root mapping model for fibrous root systems (Diggle, 1988).

VI. LIMITATIONS TO DEVELOPMENT OF ROOT GROWTH MODELS Development of well-validated root growth and function models awaits improvements in three areas of technology. First, improvements in numerical solutions to complex equations are needed to allow the routine use of computers in the solution of uptake equations. Second, more time- and cost-effective methods of measuring roots under field conditions will be needed before the massive quantities of data required for confident model validation can be obtained. Finally, we have insufficient knowledge about properties of roots and about root-soil relationships. Such “simple” questions as “What fraction of the root system is active in taking up materials?” are not thoroughly answered. We still know very little about concentrations of materials and other conditions at the root-soil interface. We still do not know the impact of diurnal hysteresis at the root surface on water uptake rates. We have very poor quantitative relationships between root growth and function and numerous environmental factors such as aeration, time of day, and root age. Our understanding of the integration of root system activity over the whole soil profile is fairly good for water uptake where continuity of the tensions in the xylem system causes intercommunication among roots, but for uptake of other materials, such as nitrogen, we know very little about how uptake rates are integrated over the entire root system. The morphological and anatomical changes in roots as related to profile depth, environ-

MODELING CROP ROOT GROWTH AND FUNCTION

131

mental and endogenous factors, and genetics are only beginning to be studied. Nevertheless, models serve the useful purpose of focusing clear attention on deficiencies in our knowledge and, along with field and laboratory investigations, are an important part of the progress being made in understanding root growth and function.

REFERENCES Barber, D. A., and Loughman, B. C . 1967J. Exp. Bot. 18, 170-176. Barley, K. P. 1971. “Encyclopedia of Science and Technology,” 3rd Ed. McGraw-Hill, New York. Bar-Yosef. B.. Lambert, J. R., and Baker, D. N. 1982. Truns. ASAE25,1268-1273 and 1281. Belford, R. K . , Rickman. R. W., Klepper. B., and Allmaras. R. R. 1986. Agron. J . 78, 757-760. Belford R. K . . Klepper, B., and Rickman. R. W. 1987. Agron. J. 79,310-319. Bohm. W. 1979. “Methods of Studying Root System, Ecological Studies,” Vol. 33. SpringerVerlag. New York. Browning. V. D., Taylor. H. M.. Huck. M. G.. and Klepper. B. 1975. Aitbirrn Uniu. Agric. Exp. Stn. Btrll. 467. Charles-Edwards, D. A. 1984. Ann. Bot. 53,699-704. Claassen. N.. and Barber, S. A. 1976. Agron. J . 68,961-964. Cook, R. J., and Haglund, W. A. 1982. Wushirigtori Strite Uniu. Res. Bid/. XB0913. Coughenour, M. B., McNaughton, S. J., and Wallace. L. L. 1984. Ecol. Modell. 23,101-134. Cushman. J. H. 1984. Soil Sci. 138, 164-171. Davidson. R. L. 1969. A n n . Bot. 33,561-569. Diggle. A. J . 1988. Plant Soil 105, 169-178. Drew, M. C. 1975. New Phytol. 75,479-490. Elkins, C. B. 1987. Agron. Ahstr.. p. 110. Epstein, E. 1976. I n “Encyclopediaof Plant Physiology” (U.Luttgeand M. G. Pitman. eds.). N Series, Vol. 2, Part B. pp. 70-94. Springer-Verlag, New York. Fishman, S., Talpaz, H., Dinar, M., Levy, M., Arazi, Y.. R o m a n . Y., and Varsharsky, S. 1984. Agric. Syst. 14, 159-169. Gerwitz, A., and Page, E. R. 1974. J. A p p l . Ecol. 11,773-782. Greacen. E. L., and Oh, J. S. 1972. Notirrr (London)N i w B i d . 235, 24. Gregory, P. J. 1987. In “Root Development and Function” (P. J. Gregory, J. V. Lake. and D. A. Rose, eds.), pp. 147-166. Cambridge. Gupta, S. C., and Larson, W. E. 1982. I n “Predicting Tillage Effects on Soil Physical Properties and Processes,” pp. 151-178. American Society of Agronomy, Madison, Wisconsin. Hoogenboom, G., and Huck, M. G. 1986. Alabamu Agric. Exp. Stn. Agron. Soils D e p . Ser. 109. Hoogenboom. G . , Huck. M. G., and Hillel. D. 1987. Adu. Irrig. 4,331-387. Huck, M. G., and Hillel, D. 1983. Adu. Irrig. 2, 273-333. Huck, M. G., Hoogenboom, G., and Peterson, C. M. 1987. In “Minirhizontron Observation Tubes: Methods and Applications for Measuring Rhizosphere Dynamics” (H. M. Taylor, ed.), pp. 109-121. American Society of Agronomy, Madison, Wisconsin. Johnson, I. R. 1985. Ann. Bot.55,421-431. Johnson, I . R., and Thornley, J. H. M. 1983. PIant Cell Enuiron. 6,721-729.

132

BETTY KLEPPER AND R. W. RICKMAN

Johnson, 1. R., and Thornley, J. H. M. 1985. Plant Cell Enuiron. 8,485-499. Jones, C. A., Bland, W. L., Ritchie, J. T., and Williams, J. R. 1990. Simulation of root growth. In “Modeling Plant and Soil Systems” (J. Hanks and J. T. Ritchie, eds.). Agronomy Society of America, Madison, Wisconsin. Klepper, B., Taylor, H. M., Huck, M. G., and Fiscus, E. L. 1973. Agron. J. 65,307-310. Klepper, B., Rickman, R. W., and Taylor, H. M. 1983. Agric. Water Manage. 7, 115-141. Klepper, B., Belford, R. K., and Rickman, R. W. 1984. Agron. J . 76, 117-122. Lambert, J. R., Baker, D. N., and Phene, C. J. 1976. Proc. U S . - U . S . S . R . Seminar, Moscow, Kishiner, Riga, pp. 1-32. Lockhart, J. A. 1965. J. Theor. Biol. 8,264-275. Logsdon, S . D., and Allmaras, R. R. 1989. Agron. Abstr.. p. 285. Logsdon, S. D., Allmaras, R. R., Wu, L.. and Swan, J. B. 1988. Agron. Absrr. p. 280. Maizlich. N . A., Fritton. D. D., and Kendall, W. A. 1980. Agron. J. 72,25-31. Nye, P. H. 1984. I n “Roots, Nutrient and Water Influx and Plant Growth” (S. A. Barber and D. R. Bouldin, eds.), Special Publication 49, pp. 89-100. American Society of Agronomy, Madison, Wisconsin. O’Leary, G. J., Connor, D. J., and White, D. H. 1985. Agric. Syst. 17, 1-26. Porter, J. R., Klepper, B., and Belford, R. K. 1986. Plant Soil92, 133-145. Reicosky, D. C., Millington, R. J., Klute, A., and Peters, D. B. 1972. Agron. J. 64,292-297. Reynolds, J. F.. and Thornley. J. H. M. 1982. Ann. Bor. 49,585-597. Skiles, J. W., Hanson, J. D., and Parton, W. J. 1982. I n “Analysis of Ecological Systems: State of the Art in Ecological Modelling” (W. K. Lavenroth, G. V. Skogerboe, and M. Flug, eds.), pp. 467-473. Elsevier, Amsterdam. Skipp, R. A., Christensen, M. J., and Caradus, J. R. 1982. N. Z. J . Agric. Res. 25,87-95. Stone, J. A.. and Taylor, H. M. 1983. Agron. J. 75,613-681. Stone, J. A., Kaspar, T. C., and Taylor, H. M. 1983. Agron. J. 75, 1050-1054. Taylor, H . M. 1987. ASA Spec. Pub/. 50. Taylor, H. M., and Klepper, B. 1978. Adu. Agron. 30,99-128. Tinker, P. B. 1984. Plant Soil 76, 77-91. Wang, J., Hesketh, J. D., and Wooley, J. T. 1986. Soil Sci. 141,432-437. Whisler, F. D.. Acock, B., Baker, D. N., Fye, R. E.. Hodges, H. F., Lambert, J. R., Lemmon, H. E., McKinion. J. M., and Reddy, V. R. 1986. Adu. Agron. 40, 141-208.

ADVANCES IN AGRONOMY. VOL. 44

GENETIC MANIPULATION OF THE COWPEA (Vigna unguiculata [L.] Walp.) FOR ENHANCED RESISTANCE TO FUNGAL PATHOGENS AND INSECT PESTS A. 0. Latunde-Dada Department of Crop Production College of Agricultural Sciences Ogun State University Ago-lwoye Ogun State, Nigeria

I . Introduction Insect Pests Ill. Fungal Pathogens IV. Tissue Culture Technology A. Somaclonal Variation and Cellular Selection B. Embryo Culture C. Somatic Hybridization D. Genetic Transformation V. Conclusions and Epilogue References 11.

I. INTRODUCTION The cowpea, Vigna unguiculata (L.) Walp., is perhaps the most important papilionaceous grain legume in the Third World, particularly Africa. Although conjectural claims to the contrary abound (Vavilov, 1951 ; Steele, 1976; Zhukovskii, 1962, Piper, 1913; Sauer, 1952), the dominant worldview is that the crop originated in West Africa, probably in the subhumid savanna grasslands of Nigeria, the area of its greatest diversity (Faris, 1965; Rawal, 1975; Lush and Evans, 1981; Ng and Marechal, 1985). It is, however, agreed that the cowpea was domesticated about 4000 years ago from the wild progenitors V . unguiculata ssp. dekindtiana vars. dekindriana (in sub-Sahelian West Africa) and mensensis (in the humid and

133 Copyright B 1990 by Academic Press. Inc. All rights of reproduction in any form reserved.

134

A. 0. LATUNDE-DADA

subhumid zones) and distributed thereafter throughout sub-Saharan Africa, reaching the Middle East by 2000 B.C. and India by 1500 b . c . (Lush and Evans, 1981; Steele, 1976). The Indian cultigroups of V. unguiculata ssp. unguiculata, namely, Biflora (the catjang bean) and Sesquipedalis (the yardlong or asparagus bean), arose probably by selection from the early cowpea domesticates (Faris, 1965). The crop is believed to have been introduced into Mediterranean Europe before 300 B.C. and reached the Americas only recently, during the Great Slave Trade of the sixteenth and seventeenth centuries A.D. (Steele, 1976). The cowpea is today cultivated throughout sub-Saharan Africa, Southeast Asia, Latin America, and the United States (Fig. 1). Nigeria is the world's largest cowpea producer, with production levels ranging between 40% (Rachie, 1985) and 75%(Cobley and Steele, 1977) of the world's output. The cowpea plant is a herbaceous annual with habits ranging from determinate to completely indeterminate types. The stem is nonpubescent and may be erect, semierect, prostrate, or twining. The inflorescence is a simple raceme borne on an elongated axillary peduncle on which the flower opening is acropetal. The pod, a legume, contains seeds that are arranged in a single row within the pod wall. The taproot system develops both secondary and tertiary roots along its length, all of which are capable of nodulation, hence symbiotic nitrogen fixation. The mutualistic interaction between the cowpea root and soil bacteria of the genus Rhizobium results annually in an estimated production level of 198 kg/ha of nitrogen, a level higher than that of Arachis hypogaea (124 kg/ha) or Glycine max

+ dispersal rouln

FIG. 1. Major cowpea-growing areas of the world, with West Africa as the center of origin.

GENETIC MANIPULATION OF THE COWPEA

135

(103 kg/ha) (Nutman, 1975). The cowpea plant by this feat not only provides for its own nitrogen requirements but also contributes to the replenishment of nitrogen in tropical soils, and assists in cutting the everincreasing capital outlay on nitrate fertilization when intercropped with tropical cereals. The cowpea plant is drought tolerant. Vigna unguiculata is diploid with a chromosome number of 2n = 22. It belongs to the tribe Phaseoleae of the subfamily Papilionoideae. It is classified further into the subtribe Phaseolinae and group Phaseolastrae. The genus Vigna is pantropic, comprising 7 subgenera and about 170 species (Cobley and Steele, 1977). The taxonomy of the cowpea at levels of taxon below the generic appears complex. Marechal’s approach to the classification (Marechal e t al., 1978; Marechal, 1982) is given in Table I . Sexual incompatibility barriers exist in the Phaseolinae and so far interspecific crosses are yet to be reported with Vigna unguiculata (Baudoin and Marechal, 1985). Nevertheless, these authors regard V . neruosa (section Catiang) and V . frutescens (section Liebrechtsia) to be close enough to introgress with the cowpea: no hybrid has been reported between these close relatives and the cultivated cowpea. All members of the cultigroups of V . unguiculata ssp. unguiculata are interfertile and also sexually compatible with the wild varieties of the subspecies dekindtiana (Lush and Evans, 1981; Ng and Marechal, 1985). In Nigeria, the cowpea is cultivated solely for its seed, which is harvested dry and eaten either boiled or processed into different dish forms. The seed contains about 25% (w/w) protein (Bressani, 1985) and constitutes an important and cheap source of protein for most people in West Africa, especially the poor. Moreover, the cowpea seed, compared with some other pulses, i s relatively low in toxic substances and antimetabolites. These qualities have for long promoted its acceptability among the peoples of Nigeria. The cowpea crop is well adapted to the savanna grasslands of Nigeria. About 80% of the country’s output comes from this region (Emechebe, 1981). In this region, it is cultivated either solely or in mixtures with cassava and such cereal crops as maize, sorghum, and millets. Sole crops of cowpea are more commonly cultivated in the southern humid Rain Forest Belt, where, in spite of an annual rainfall pattern and distribution that can theoretically support three crops of cowpea, only 20% of the Nigerian output is produced. Figure 2 shows the attempts made at the Ogun State University, Ago-Iwoye, Nigeria (about 50 km south of Ibadan, and in the Rain Forest Belt of Nigeria), to produce three crops of cowpea over two consecutive seasons (A. 0. Latunde-Dada, unpublished data). Cowpea production, in the Rain Forest Belt of Nigeria and elsewhere in Africa is limited by a number of constraints. The grain output from most

136

A. 0. LATUNDE-DADA Table 1 Taxonomy of the Cowpea, Vignu unguiculutu Walp." ~~

Taxon

Cowpea

Family: Subfamily:

Leguminoseae Papilionoideae

Tribe: Subtribe:

Phaseoleae Phaseolinae

Group: Genus: Subgenus:

Phaseolastrae Vigna Vigna

Section:

Catiang

Species:

V . unguiculata

Subspecies:

V. unguiculata

Cultigroups:

Unguiculata (cowpea)

Related taxonomic forms (a) Caesalpinioideae (b) Mimosoideae (a) Cajaninae (b) Glycinae Dolichastrae Phaseolus (a) Macrorhyncha ( b ) Plectotropis (c) Sigmoidotropis (d) Ceratotropis (e) Haydonia (0 Lasiocarpa (a) Vigna (b) Comosae (c) Reticulatae (d) Liebrechtsia (e) Macrodontae (a) V . neruosa (b) V . frutescens (in section Liebrechtsia) (a) V . unguiculata ssp. unguiculata spp. srenophylla (b) V . unguiculata spp. tenuis (c) V . unguiculata spp. dekindtianab (a) BIFLORA = CYLINDRICA (catjang bean) (b) SESQUIPEDALIS (asparagus or yardlong bean) (c) TEXTILIS

Sources: MarCchal et a / . (1978); Markchal, (1982). Vigna unguiculata ssp. dekindtiana comprises the following varieties: (a) dekindtiana, (b) pubescens, (c) mensensis, and (d) protracta.

local varieties is between 100 and 300 kg/ha (IITA Research Briefs, 1984). Apart from the obvious genetic inadequacies (such as extreme viny growth habit, compulsive photoperiodism, low flowering and pod-setting abilities, and low yields) of these local varieties, other problems abound. These include susceptibility to insect pests, lack of resistance to viral, bacterial, and fungal diseases, lack of tolerance of excessive moisture levels, weed infestation, and inadequate soil nutrient supply. These problems have been reviewed elsewhere (Summerfield et al., 1974; Thottappilly and Rossel, 1985; Muleba and Ezumah, 1985; Poku and Akobundu, 1985).

IIL

GENETIC MANIPULATION OF T H E COWPEA

350

300

1

J

A

S

O

N

I37

D

FIG.2. Annual rainfall distribution pattern for Ago-lwoye. Horizontal arrows represent periods during which cowpea (vars. TV x 3236, Ife Brown, IT82E-60, and IT81D-994) was cultivated. During period 1, plants were largely free from foliar pathogens and insect pests; pods required additional artificial drying, however. During period 11, the incidence of Ootheca mutabilis and Aphis cracciuoru was high and damage due to web blight, Pythium stem rot, and anthracnose was severe in the susceptible varieties. Pods were picked compulsorily when light yellow and were dried artificially. Web blight was the greatest challenge, especially in Ife Brown and IT83E-60, during period 111. Additional drying was unnecessary, and yields were highest during this period.

Advances have been made in the various ramifications of cowpea research during the past two decades to ameliorate the gloomy situation just presented. The achievements of two agricultural research institutes situated in the Rain Forest Belt of Nigeria are worth considering. The Institute of Agricultural Research and Training (IART) of the Obafemi Awolowo University of Ibadan was responsible for earlier research during the 1960s that led to the development and release of the varieties Westbred and Prima, and later Ife Brown and Ife Bimpe (Franckowiak et al., 1973; Ojomo, 1975). The varieties Ife Brown and Ife Bimpe are short, determinate, and early, flowering within 35 days of sowing. When given adequate care, Ife Brown yields about 1700 kg/ha (Singh and Ntare, 1985) but lacks demonstrable resistance to insect pest and disease problems. Since its inception in 1967, the International Institute of Tropical Agriculture (IITA), also in Ibadan, has virtually taken the lead in Nigeria and indeed international cowpea research. The VITA series of cowpea varieties released originally by IITA in the 1970s have since been supplanted in

138

A. 0. LATUNDE-DADA

Nigeria by such popular IITA varieties as TV X 3236, IT82E-60, IT82D716, IT84E-124, and lately IT84S-2246-4. IITA continues to draw on the rich genetic diversity of the cowpeas and has succeeded in a seemingly endless release of lines and varieties that combine high-yielding characteristics with multiple pest and disease resistance (Table 11), and without any apparent impairment of the seed protein levels. The modern IITA varieties of cowpea grown in the Rain Forest Belt are day-neutral, early, determinate, short, and erect and do not require staking. Moreover, they have peduncles that elongate rapidly during the early reproductive phase to display the maturing pods high up and above the leaf canopy. These pods mature uniformly, thereby facilitating synchronous harvesting. In general, the potential yield of these IITA varieties is 2 t/ha; nevertheless, there are plans and projections afoot to develop varieties with yield potentials in the Table I1 Some IITA Cowpea Varieties Bred for Multiple Resistance to Disease and Insect Pests”

Insect pest resistance‘

I. 2. 3. 4. 5.

6. 7. 8. 9. 10.

1I . 12.

Variety”

A

B

VITA3 ITSID-994 IT81D-1020 IT81D-1137 IT82D-812 TV x 3236 IT82D-716 IT83S-742-11 IT83S-742-I3 IT83S-728-5 IT84E-124 IT84S-2246-4

+

+ + + + +

+ +

C

+ + + +

+

+ +

D

+ +

+

+ + +

a Sources: “Varietal Improvement of Cowpea.” IITA Crop Production Training Series. Singh, S. R. 1986. Trop. Grain Legume Bull. 32, 10-24. Varieties I , 2,3,4,5,7,8,9. 10, and 12 possess multiple resistance to diseases. IT81D-I 137, for instance, is resistant to a total of 1 1 fungal, bacterial, and viral diseases. TV x 3236 and IT84E-124 are resistant to at least 7 such diseases. A, B, C, and D represent, respectively, leaf hoppers, aphids, thrips, and bruchids; + denotes resistance; a blank space denotes lack of resistance to the insect pest.

GENETIC MANIPULATION OF THE COWPEA

139

region of 6 tlha (IITA Research Briefs, 1984). However, it must be stated that this projected target may be rather overambitious for a field crop such as cowpea growing for 60-odd days. A yield potential of between 3 and 4 t dry seed/ha may be more feasible. The attainment of even this lower yield requires, on the one hand, a reevaluation of the nagging problems of insect pests and fungal diseases peculiar to the various ecological zones. Second, it also requires the adoption of a technology that in consonance with conventional plant breeding methods used so far at IITA seeks to both increase the genetic diversity of the cowpea and improve the selection of desirable characteristics. These two considerations are explored in this chapter.

II. INSECT PESTS Eight out of the 31 insect pest species identified in the cowpea by Singh and Jackai (1985) constitute the major preharvest entomological pest problem in the Rain Forest Belt of Nigeria. These insects are:

1. The cowpea leaf beetle, Ootheca mutabilis (Shalberg), which, though a sporadic pest, may cause total crop loss through foliage destruction. It is also an important virus vector. 2. The cowpea bud thrip, Megalurothrips sjostedti (Trybom) = Taeniothrips sjostedti (Trybom), which may cause total crop loss through interference with flowering. 3. The cowpea aphid, Aphis craccivora (Koch), which causes up to 35% damage when pods are infested and also transmits a number of virus pathogens. 4. The cowpea pod borer, Maruca testulalis (Geyer), the caterpillars of which may cause 60% or more damage when coincident with those of Lasperyresia ptychoru (Meyrick), a minor lepidopterous pest; and the following cowpea coreids or pod suckers: 5. Acanthomia tomentosicollis (Stalberg) = Clavigralla tomentosicollis (Stal) 6. Acanthomia horrida (Germar) = Clavigralla shadabi (Dolling) 7. Anocplocnemis curvipes (Fabricius) 8. Riptortus dentipes (Fabricius) These coreid bugs feed massively on young developing pods and foliage of cowpea and may cause up to 60% yield loss. Figure 3 illustrates the activities of these pests in relation to the phenology of the crop. Sources of resistance have been identified within the massive collection

140

A. 0. LATUNDE-DADA

FIG.3. Spans of activity of the major preharvest insect pests of cowpea, V. unguicduta, in relation to the crop’s phenology. Shaded areas of bars denote span of peak activity. Source: Singh, S. R. 1980. In “Biology of Breeding for Resistance to Arthropods and Pathogens in Agriculture Plants” (M. K. Harris, ed.), Texas A & M Univ. Bulletin, MP-1451, p. 398-42 1 .

of cowpea germplasm (about 12,000 accessions) at IITA for thrips ( M . sjostedti), aphids ( A . cracciuora),and the major harvest insect pest, Callosobruchus maculatus (Fabricius), a curculionid beetle. These resistance genes have been bred successfully in different combinations into such cowpea lines as TVx3236, IT81D-1020, IT82D-716, IT81D-994, and IT84E-124. The list of IITA varieties with insect pest resistance is given in Table 11. Sources of resistance to Ootheca mutabilis, Maruca testulalis, and the coreid pod suckers are as yet unidentified. For these pests, chemical control by insecticide applications remains the only pragmatic means of protection. Given the medium-sized farms replete in an underdeveloped country such as Nigeria, and the current depressingly low economic fortunes of the country, the rising costs of farm inputs continue to cut rather deeply into the farmer’s profit margin. Moreover, the prices of such insecticides as Cymbush Sherpa, Decis, and Thiodan, which are prominent items on the IITA’s production package for cowpeas, are high enough to edge these all-important chemicals out of the reach of the small-time farmer and thus depress his yield. In a virtually insect-pest-susceptible variety like Ife-Brown, this could mean total crop loss, and even in IITA advanced breeding lines a fourfold reduction in yield (IITA Research Briefs, 1984). For this class of farmer, efforts that are aimed at introducing

GENETIC MANIPULATION OF THE COWPEA

141

into existing cowpea varieties and lines effective and durable resistance to these destrutive insect pests remain the only panacea to the problem. Such research must aim at introducing, for example, genes for antibiosis, nonpreference mechanisms as well as leaf pubescence for the control of Ootheca mutabilis, genes for antibiosis for the control of Maruca testulalis, and genes for scabrous and pubescent pods for the control of the coreid pod suckers. These may be achieved, theoretically, by attempting crosses between the cowpea and other closely related Vigna species (such as V . uexillata) as well as wide crosses with the more distantly related genera of the Phaseoleae. The existence of incompatibility barriers has been highlighted earlier.

Ill. FUNGAL PATHOGENS Fungal diseases account for losses in the cowpea surpassed only by the damage due to insect pests (Emechebe and Shoyinka, 1985). Out of these, anthracnose and brown blotch (caused by Colletotrichum lindemuthianum [Sacc. & Magn.] Bri. & Cav., and C . capsici [Syd.], respectively), leaf spots (incited by Cercospora canescens Ell. & Mart., and Pseudocercospora cruenta [Deighton] Sacc.), web blight (caused by Rhizoctonia s o l d [hill. & Delacr.] Bourd & Galz), and wilts (caused by Fusarium oxysporum f.sp. tracheiphilum Schlect, ex Fr. and Sclerotium rolfsii Sacch.) have been identified as the most damaging on cowpea in the Rain Forest Belt of Nigeria (Emechebe and Shoyinka, 1985). Sources of resistance to some of these diseases have been identified in the IITA germplasm accessions and varieties and lines such as TVx3236, IT82E-716, IT82D699, IT83D-326-Z, IT8 1 D- 1 137, IT835-1 1, and IT8 1D- 1020 have been bred and selected from multiple resistance to various fungal diseases. More significantly, these lines combine this characteristic with degrees of insect pest resistance and high grain yield potentials also. Resistance to web blight, a destructive foliar disease, remains largely unresolved. This disease is caused by the imperfect fungus Rhizoctonia solani (with Thanatephorus cucumeris [Frank] Donk. as its basidiomycete teleomorph) and is a major problem in cowpea production in the humid Rain Forest Belt. No cowpea variety is wholly resistant to this disease, to this author’s knowledge. The development and deployment of an effective and durable means of protection through breeding methods is therefore imperative. Although descriptive studies abound on the biology and control of a number of the fungal diseases mentioned (Williams 1975; Emechebe, 1981;

142

A. 0. LATUNDE-DADA

Onesirosan and Barker, 1971; Emechebe and Shoyinka, 1989, there are deficiencies in our understanding of the biochemistry, physiology, and molecular aspects of the host-pathogen interactions. Previous studies have shown that Vigna unguiculata produces an array of phytoalexins, including kievitone, phaseollin, phaseollidin, vignafuran, and medicarpin, in response to both fungal attack and abiotic elicitation (Martin and Dewick, 1979; Patridge and Keen, 1976; Preston et al., 1975; Dixon et al., 1983). It is also known that these fungi, particularly species of the genera Colletotrichum, Cercospora, and Rhizoctonia, elaborate in axenic culture metabolites with phytotoxic and elictor activities (Frantzen et al., 1982; Anderson, 1978, MarrC, 1980; Iacopellis and De Vay, 1987). The precise role of these various compounds in the interactions between cowpea and its Nigerian isolates of fungal pathogens is not clear at the present time. Furthermore, the possibility exists for using these metabolites of both fungal and host origins in screening and selecting populations of cowpea genotypes as well as hybrid lines for enhanced and novel resistance to diseases. This was found to be true in the interaction between the legume Medicago sativa and Verticillum albo-atrum, a wilt fungus (LatundeDada, 1983). The efficient screening and selection of lines with desirable characteristics can be greatly enhanced by the adoption of the in vitro methods of plant tissue culture. Furthermore, these methods can be harnessed in evolving a better understanding of the interactions between the cowpea plant and its major fungal pathogens-an area of research the IITA has yet to explore. Of greater importance, however, is the potential of the techniques and methods of plant tissue culture in assisting in the generation of both novel and enhanced resistance to insect pests and fungal pathogens of the cowpea.

IV. TISSUE CULTURE TECHNOLOGY The techniques and methods of plant tissue culture were developed initially for investigating aspects of the control of growth and morphogenesis in plants. In response to Haberlandt’s prescient article of 1902 in which he propounded the intrinsic ability of excised plant organs and isolated cells to grow in culture and exhibit totipotency, this area of research was pursued vigorously in subsequent years. As media techniques improved, success was recorded in the establishment of root cultures (White, 1934), callus cultures (Gautheret, 1939; Nobecourt, 1939; White, 1939), embryo cultures (van Overbeek et al., 1942), cell suspension cultures (Muir et al., 1954), protoplast cultures (Cocking, 1960), and

GENETIC MANIPULATION OF THE COWPEA

143

anther cultures (Bourgin and Nitsch 1967). Attention soon switched to the totipotency of cultured plant protoplasts. This was achieved in 1971 in the tobacco (Takebe et a f . ,1971), although the regenerability of isolated plant cells had been demonstrated earlier in tobacco and carrot callus cultures (Skoog and Miller, 1957; Steward, 1958). The list of plants regenerated from tissue and isolated protoplast cultures increases by the day. Nevertheless, a large number of genera in the Leguminoseae remain either unresearched or recalcitrant. Although the temperate forage legumes lucerne (alfalfa, Medicago satiua), the clovers (Trifofiumspp.), and fenugreek ( Trigonelfafoenum-graecum) have been cultured and regenerated from protoplasts since the early 1980s (Kao and Michayluk, 1980; Santos et a f . , 1980; Johnson et a f . , 1981; Bhojwani and White, 1982, Gresshoff, 1980; Xu et al., 1982), the protoplasts of the pulses are still recalcitrant to culturing. A number of recent reports point, however, to the imminent surmountability of the tissue cultural problems in these crops. For example, it has been possible to regenerate whole plants from callus cultures of Glycine spp. (Christianson et a f . , 1983; Gamborg et al., 1983; Lippmann and Lippmann, 1984; Li et a f . , 1985; Barwale et al., 1986), Pisum satiuum (Jacobsen and Kysely, 1984), Phaseolus uufgaris (Martin and Sondahl, 1984), and Viciafaba (Griga et a f . , 1987).The probability is therefore high that the problems of culturability and regenerability will be overcome in due course even in the cowpea. Efforts are currently being made in this author’s laboratory to achieve this purpose. Plant improvement employs genetic solutions to maximize the various aspects of crop productivity. In the case of the cowpea, the yield has been raised 10-fold within two decades of research, thereby breaking the initially low and grossly primitive yield barrier of 200 kg/ha. This achievement was made solely through conventional mass selection, pedigree breeding, pureline breeding, sexual hybridization, and backcross breeding with a backup of appropriate insecticide application packages. Sustained crop productivity, however, requires an appreciably high annual rate of increase, about 2% (Anonymous, 1980), in the face of the ubiquitous problems of pests, diseases, and changing environmental factors. To effectively achieve such levels and further increase crop productivity the breeder requires both reliable sources of genetic variability for traits that need to be improved and an efficient selection regime for genotypes that are improvements on the preexisting cultivars. The techniques of plant tissue culture are, on these two grounds, already proving to be an asset. These techniques include: (a) somaclonal variation and cellular selection; (b) embryo culture; (c) somatic hybridization and cybridization; and (4transformation or transgenosis. The pressing need to employ the techniques of somaclonal variation and

144

A. 0. LATUNDE-DADA

embryo culture will be advocated in this chapter. These techniques may represent easy but nevertheless cost-effective means of extending the genetic variability of V. unguiculata and selecting desirable traits, and may ensure the survival of zygotes and embryos resulting from difficult and wide interspecific and intergeneric crosses among members of the Phaseolinae and Phaseolastrae. In addition to these techniques, the desirability of both somatic hybridization (and cybridization) and transformation will be highlighted. A. SOMACLONAL VARIATION AND CELLULAR SELECTION Unmutagenized cultured cells and protoplasts are now known to exhibit variation and the ability to regenerate variable plant populations. The variations have been shown to be either genetic or epigenetic in nature (Green, 1977; Skirvin, 1978; Thomas et al., 1979; Krikorian, 1982). It is therefore possible to induce, inadvertently through normal tissue culture cycles, variability arising from mitotic events and so increase the store of heritable genetic variability for the plant in question. Termed somaclonal variation (Larkin and Scowcroft, 1981), this phenomenon has been demonstrated in a large number of crop plant speces, including potato (Solanum tuberosum; Shepard et al., 1980; Thomas et al., 1982; Karp et al., 1982; Sree Ramulu et al., 1983), rice (Oryza sativa; Henke et al., 1978; Oono, 1981; Sun et al., 1983), wheat (Triticum aestivum; Ahloowahlia, 1982; Larkin et al., 1984), maize (Zea mays; Larkin and Scowcroft, 1983; Brettell et al., 1980; Kemble et al., 1982), sorghum (Sorghum bicolor; Gamborg et al., 1977), sugarcane (Saccharum of$cinarum; Heinz et al., 1977, Liu 1981), and lucerne (=alfalfa, Medicago sativa; Reisch and Bingham 1981; Latunde-Dada and Lucas, 1983). In addition, the heritability of various traits has been reported in the somaclonal variants of these species. Of greater importance and significance, however, has been the demonstration of enhanced resistance to disease in somacloned populations of some of these crop species. Enhanced resistance to eyespot (caused by Helminthosporium sacchari), Fiji disease (caused by Ustilago scitaminea), and downy mildew (caused by Sclerospora sacchari) was identified among somaclonal variants in sugarcane (Heinz et al., 1977; Nickell, 1977); some protocloned potato plants were found to be more resistant to both early blight (caused by Alternaria solani) and late blight (caused by Phytophthora infestans) than the parental variety Russet Burbank (Sheperd, 1981; Sheperd et al., 1980, 1983). Previous work on somaclonal variation in the temperate autotetraploid forage legume Medicago sativa

GENETIC MANIPULATION OF THE COWPEA

145

(2n=4x=32) revealed enhanced resistance to the wilt fungus Verticillum alboatrum in protocloned and callicloned populations (Latunde-Dada, 1983; Latunde-Dada and Lucas, 1983, 1988). Some of the genotypes in these populations exhibited differences in gross morphology and ploidy. It was also possible, through the incorporation of toxic metabolites of the fungus V . albo-atrum into plant growth media, to select tissues and plantlets that were incrementally more resistant to this wilt fungus than the unselected populations of the parental variety Europe (Latunde-Dada and Lucas, 1988). Similar results on cellular selection have been reported in the interaction between maize and Helminthosporium maydis (Gengenbach and Green, 1975; Gengenbach et al., 1977), oilseed rape and Phoma lingam (Sacristan, 1982), potato and Phytophthora infestans (Behnke, 1979, 1980), and sugarcane and Helminthosporium sacchari (Nickell, 1977). The underlying causes of genetic somaclonal variation include, among others, polyploidy , chromosome rearrangements, point mutations, somatic crossovers, transposable elements, and rearrangements and recombinations among cytoplasmic organelle genomes (Orton, 1984). These events may be visualized as mitotic accidents that constitute a means of generating biological diversity. It remains possible to harness this powerful mode of evolution in extending further the gene pool of the cowpea and enhancing its adaptability to the challenges of the Rain Forest Belt of Nigeria, in particular, as well as elsewhere.

CULTURE B. EMBRYO Mention has been made of the problem of sexual incompatibility barriers among species of the genus Vigna. A determined assessment of this constraint suggests the need for techniques that in breaking and circumventing these barriers, will aid the exploitation of gene transfer from dissimilar species into Vigna unguiculata. Three techniques of plant tissue culture, namely, embryo culture, somatic hybridization, and transformation, are suggested for this task. The methods of embryo culture have made possible the in uitro crosspollination and fertilization of flowers, and the rescuing as well as nurturing of the resulting hybrid proembryos and embryos, which rarely survive in natural crosses. Examples abound of successful interspecific hybrids obtained via embryo culture, including the crosses Brassica oleracea X B . campestris (Nishi et al., 1959; Snell, 1978) to produce Brassica napus in higher frequencies, and Nicotiana tabacum X N . rependa, N . stocktonii, and N . nesophila (Reed and Collins, 1978). In these cases the possibility has been expressed (Brettell and Ingram, 1979) for the transfer

146

A. 0. LATUNDE-DADA

of genes for novel resistance to Plasmodiophora brassicae (club root incitant) into B. napus, and to Phytophthora parasitica var. nicotianae (causes black shank disease) into commercial tobacco. In the Leguminoseae, Honma (1956)reported the production by embryo culture of a hybrid plant from the cross Phaseolus uulgaris X P . acutifolius. This hybrid retained the floral and vegetative characteristics of P . vulgaris, the female parent, and was more resistant to halo blight caused by Pseudomonas phaseolicola. Successful crosses have also been reported in Lotus pedunculatus x L . tenuis (Lautour et al., 1978) and Trifolium ambiguum x T. repens (Williams, 1978). These examples suggest the likelihood of overcoming the barriers in natural Vigna x Vigna and Vigna X Phaseolus crosses. With necessary improvements in media and techniques it might be possible to overcome, through in uitro fertilization of excised ovules and subsequent culture, stigmatic and style incompatibility mechanisms in these species. Even in cases commonly encountered in the Phaseolinae, where the incompatibility barrier is expressed in the early stages after fecundation (MarCchal, 1982), the hybrid embryo might be rescued from aborting as demonstrated by Braak and Kooistra (1975) in the cross Phaseolus ritensis x P . uulgaris. The possibility therefore exists for the introduction of pod and leaf pubescence genes from Vigna pubescens, for example, into V . unguiculata. C. SOMATIC HYBRIDIZATION Somatic hybridization seeks in practice to overcome sexual incompatibility barriers between distinctly different plant species by the fusion of their protoplasts, which are basically plant cells minus their walls. The methods of isolation, culture, and fusion of plant protoplasts, the basic components of this technology, have been comprehensively described (Evans and Cocking, 1977; Gamborg and Bottino, 1981).Protoplast fusion is commonly brought about through the mediacy of such fusogens as high calcium levels plus alkalinity in the neighborhood of pH 10, polyethylene glycol (PEG), and sodium nitrate (Power et al., 1970; Keller and Melchers 1973; Kao and Michayluk, 1974). The initial fusion products, called heterokaryons, result from a plasmogamic mixing of cytoplasms containing dissimilar nuclei, mitochondria, and chloroplasts, within which exist opportunities for plasmogenic recombinations. These heterokaryons soon develop cell walls and, given karyogamy, produce a somatic hybrid tissue from which hybrid whole plants may be regenerated (Gamborg and Bottino, 1981; Cocking et al., 1981). In some circumstances, cybrids (i.e., tissues, organs, or plants in which plasmogamy but not nuclear fusion

GENETIC MANIPULATION OF THE COWPEA

147

occurred, followed by the abortion of one of the incompatible nuclei) and chimaeras (plasmogamy , but no nuclear fusion occurring; subsequent tissue, organ, or plant possesses different nuclei in different cells) may result (Schieder and Vasil, 1980; Cocking et al., 1981). Although interspecific and intraspecific sexual hybrids are usually genetically uniform, populations of regenerated plants derived via the somatic hybridization of the same species in question contain more variability (Evans, 1983). This author attributed the variability in somatic hybrids (and cybrids, as well as chimaeras) to nuclear instability or incompatibility, mitotic recombination, somaclonal variation, and organelle segregation. Furthermore, Cocking et al. (1981) observed that protoplast fusions result in a novel “cytoplasmic mix” within which exist great opportunities for mitochondrial and chloroplastic recombinations. Evidence has been given (Belliard et al., 1979) for recombinant mitochondrial genomes in Nicotiana tabacum following somatic hybridization. Thus with organelle segregations and recombinations, a wider spectrum of nuclearcytoplasmic genetic combinations is feasible through protoplast fusion. Techniques exist for ascertaining the unique differences between the various products of plant protoplast fusion and the parental types (i.e., their hybridity), and for their isolation from mixed cultures of parental material. These have been reviewed elsewhere (Gamborg and Bottino, 1981; Schieder and Vasil, 1980). The majority of hybrid plants that have been produced via protoplast fusion and somatic hybridization are also obtainable via sexual crossing (Carlson et al., 1972;Power et al., 1976).Nevertheless, somatic hybridization and hybrid plant regeneration have been achieved in the sexually incompatible Nicotiana species (e.g., N . tabacum X N . nesophila, N . tabacum x N . rependa, and N . tabacum X N . stocktonii; Evans et al., 1981), Lycopersicon esculentum x Solanum tuberosum (Melchers et al., 1978; Sheperd et al., 1983), and Daucus carota X Petroselinium hortense (Dudits et al., 1980). In the Nicotiana examples mentioned, Evans et al. (1981) reported the transfer of tobacco mosaic virus resistance genes from wild tobacco species of the section Rependae (genus Nicotiana) into Nicotiana tabacum via somatic hybridization. Although interspecific heterokaryons and hybrid tissues have been obtained on fusing the protoplasts of various legume species (Kao et al., 1974; Constabel et al., 1975, 1976, 1977; Kao and Michayluk, 1974), none has resulted in hybrid plant regeneration. With recent advances in legume tissue and protoplast culture, however, it is envisaged that the constraints in the culturability and regenerability of the protoplasts of the pulses will be overcome shortly. The removal of this bottleneck would provide the basis for successful somatic hybridization among species of the genera

148

A. 0. LATUNDE-DADA

Vigna, Glycine, and Phaseolus, including the attendant spin-offs in the transfer of genes for pest and disease resistance into the cultivated cowpea. The families Umbelliferae and Cruciferae have recently broken the “monopoly” of success in the throughput of intergeneric and interspecific somatic hybrid plants hitherto seemingly restricted to the family Solanaceae (Schenck and Robbelen, 1982; Kameya et ul., 1981).

D. GENETICTRANSFORMATION Finally, genetic transformation through the use of naked DNA, cell organelles, liposomes, plant viruses, plasmids, and cosmids may become applicable in the near future in cowpea improvement. These recombinant DNA and organelle transfer techniques have been adequately reviewed elsewhere (Davey et a / . , 1980; Kleinhofs, 1977; Hess, 1977; Giles, 1977, Barton and Brill 1983; Fraley and Horsch, 1983;Gardner, 1983). Efforts in the genetic transformation of plants have, however, been more greatly focused around the use of gene vectors to facilitate both the uptake and stabilization of foreign DNA inside the plants cells. These vectors include the Ti- and Ri-plasmids of the oncogenic bacteria Agrobacterium tumefaciens and A . rhizogenes, phytopathogenic DNA viruses such as the caulimo- and gemini-viruses, and more recently novel plasmid constructs with hybrid and chimaeric genes (Herrerra-Estrella et a / . , 1983b; Hain et al., 1985; Potrykus et al., 1985a,b; Lorz et a / . , 1985). These hybrid and chimaeric gene constructs are “fitted” with promoter sequences and expression signals, such as gene IV of the cauliflower mosaic virus, the nopaline synthase promoter (pNOS), and the polyadenylation signal of the octopine synthase gene of Agrobacterium tumefaciens, which are recognized in most plant cells (Herrerra-Estrella et al., 1983b; Lorz et al., 1985). It has been possible thus far to transform successfully some dicotyledonous and monocotyledonous whole-plant and tissue systems with foreign genes for resistance to kanamycin, methotrexate, and chloramphenicol and for ribulose 1,5-bisphosphate carboxylase (Paszkowski et al., 1984; Herrerra-Estrella et a / . , 1983a, 1984). On the basis of these findings, it can be expected that recombinant DNA techniques may be used as successfully in plant cells as is already the case in mammalian and microbial systems (DeBlock et al., 1984). The applicability of these techniques to the cowpea hinges rather heavily both on the establishment of protocol for the efficient regeneration of whole plants from isolated protoplasts of the crop and on the recognition and isolation of useful foreign genes required for its genetic engineering. One such group of foreign genes codes for the polypeptide insecticide

GENETIC MANIPULATION OF THE COWPEA

149

endotoxins (delta-endotoxin and beta-endotoxin) in the bacterium Bacillus thuringiensis, a biological control agent. The delta-endotoxin, in particular, has an insecticidal spectrum confined to insect species within the order Lepidoptera (Luthy, 1980), the taxon to which both Maruca testulalis and Laspeyresia prychora belong. The delta-endotoxin gene has recently been cloned in Escherichia coli and the chances are high from cloning this gene in such plant transformation vectors as the Ti-plasmid and other related plasmid constructs discussed earlier. Therefore, it is pertinent here to envisage populations of transformed protocloned cowpea plants fitted with, and expressing, the gene for a nontoxic (to humans) and highly specific insecticide, and resistant to the pod borer Maruca testulalis!

V. CONCLUSIONS AND EPILOGUE The dramatic gains in cowpea productivity during the last two decades are largely attributable to advances made in the IITA Grain Legume Improvement Program. This period witnessed the development and release of improved cowpea varieties fortified with genes for multiple resistance to diseases and pests and that possess significantly higher yield capacities. Moreover these varieties enjoy considerably high levels of consumer acceptance. The variety TVx3236, for example, has easy “cookability” and IT84E-124, which was released much later, is as resistant to thrips as TVx3236 but more acceptable to farmers in the Nigerian Rain Forest Belt because of its earliness (60 days) and brown rough seed coat. The IITA has recently outshone itself by releasing the variety IT84S2246-4, a vast improvement over ITS4E-124. This new variety has brown rough seeds, is resistant to thrips, aphids, and bruchids, demonstrates multiple resistance to 10 diseases, and is fast replacing IT84E-124. Such is the pace of innovations at the IITA! The yield histories of such crops as wheat, potatoes, field beans, and groundnuts suggest, however, that dramatic gains in crop yield may become ephemeral without research on the proper maintenance to promote their sustenance and improvement (Wittwer, 1979). The challenges of the tropical environment, the habitat of most of the world’s cowpeas, necessitate a continual release of cowpea varieties in the face of newly evolving races and strains of fungal pathogens and insect pests as well as other environmental vagaries. The currently successful resistant cowpea varieties may, as they become more widely cultivated, be in danger of succumbing to inevitably more virulent races of pathogens and pests. Herein

150

A. 0.LATUNDE-DADA

lies the predicament of the boom-and-bust cycle, which is particularly applicable to situations where resistance is either monogenic or oligogenic. A call is made here for cowpea varieties with polygenic resistance to pests and diseases, as this ensures durability. Although the lead time for crop varietal development and release through conventional breeding methods has been put at between 10 and 20 years (Pluckett and Smith, 1986), the adoption of cellular methodologies may considerably shorten this period (Scowcroft and Larkin, 1985). This chapter has sought to explore the potentialities of some areas of tissue culture technology in the improvement of cowpea. it has been necessary to bring into focus the successes of cellular methodologies in other crops, particularly sugarcane, potato, and alfalfa, in which yields have been increased and sources of novel and enhanced pest and disease resistance identified. Successful interspecific hybridization between a line of V. unguiculata ssp. dekindtiana var. pubescens and the cultivated cowpea (var. IT82E-124) through embryo rescue and culture has been reported (Fatokun and Singh, 1987). The progeny of this cross exhibited characters intermediate between those of both parents but resembled the wild parent in having hairy shoots and leaves. i t remains to be seen if such crosses between the cultivated cowpea and its hairy close and wide relatives would, aided by in vitro methods, effect the introduction of pod hairiness into the crop, thereby protecting against the destructive coreid bugs. It is certain that these cellular methodologies will, in the cowpea, increase our understanding of the mechanisms of disease development, permit early detection of infection, and enable a more rapid thoroughput, as well as promote the engineering of resistant varieties. REFERENCES Ahloowahlia, B. S. 1982. Crop Sci.22, 405-410. Anderson, A J. 1978. Phytoparhology 68, 189-194. Anonymous 1980. “Entering the Twenty First Century” Vol. 2, The Global 2000 Report to the President. Council on Environmental Quality and the Department of State, Government Printing Office, Washington, D.C. Barton, K. A . , and Brill, W. J. 1983. Science 219,671-676. Barwale, U. B., Kerns, H. R., and Wildholm, J. M. 1986. PIanta 167,473-481. Baudoin, J. P., and Marechal, R. 1985. In “Cowpea Research, Production and Utilization” (S. R. Singh and K . 0. Rachie, eds.), pp. 3-9. Wiley, Chichester. Behnke, M. 1979. Theor. Appl. Genet. 55,69-71. Behnke, M. 1980. Theor Appl Genet 56, 151-152. Belliard, G., Cede], F., and Pelletier, G . 1979. Nature (London)281,401-403. Bhojwani, S . , and White, D. R. 1982. Plant Sci. Lett. 25,255-262. Bourgin, J. P., and Nitsch, J. P. 1967. Ann. Physiol. Veg. 9,377-382.

GENETIC MANIPULATION O F T H E COWPEA

151

Braak, J. P., and Kooistra, E. 1975.Euphytica 24,669-679. Bressani, R. 1985.In “Cowpea Research, Production and Utilization” (S. R. Singh and K. 0. Rachie, eds.), pp. 353-359. Wiley, Chichester. Brettell, R. I. S., and Ingram, D. S. 1979.Biol. Rev. 54, 329-345. Brettell, R. I. S., Thomas E., and Ingram, D. S. 1980.Theor. Appl. Genet. 58,55-58. Carlson, P. S.,Smith, H. H., and Dearing, R. D. 1972.Proc. Narl. Acad. Sci. U.S.A. 69, 2292-2294. Christianson, M. L., Warnick, D. A , , and Carlson, P. S. 1983.Science 22,632-634. Cobley, L. S., and Steele, W.M. 1977.“An Introduction to the Botany of Tropical Crops,” 2nd Ed. Longmans, London. Cocking, E . C. 1960.Nature (London) 187,927-929. Cocking, E. C., Davey, M. R., Pental, D. and Power, J. B. 1981. Nature (London) 293, 265-270. Constabel, F., Dudits, D., Gamborg, 0. L., and Kao, K. N. 1975.Can. J Bot. 53,2092-2095. Constabel, F., Webber, G . , and Kirkpatrick, J. W. 1977.C. R . Seances Acad. Sci. Ser. D: 285,319-321. Constabel, F., Webber, G., Kirkpatrick, J. W. and Pahl, K. 1976.Z. Pjlazenphysiol. 79, 1-7. Davey, M. R., Freeman. J. D., Draper, J., and Cocking, E. C. 1980.In “Tissue Culture Methods for Plant Pathologists” (D. S. Ingram and J. P. Helgeson, eds.), pp. 209-219. Blackwell, Oxford. DeBlock, M., Herrera-Estrella. L., Van Montagu, M., Schell, J., and Zambryski, P. 1984. EMBOJ3,1681-1689. Dixon, R. A., Dey, P. M., and Lamb, C. J. 1983.Adu. Enzymol. R e / . Areas Mol. Biol. 55, 1-136. Dudits, D.,Hadlaczky, G., Lazar, G., and Hydm, Z. 1980.In “Plant Cell Cultures: Results and Perspectives” (F. Sala, B. Parisi, R. Cella, and 0. Ciferri, eds.), pp. 207-214. Elsevier/North-Holland, Amsterdam. Emechebe, A. M. 1981. Samaru J. Agric. Res. 1,20-26. Emechebe, A. M., and Shoyinka, S. A. 1985.In “Cowpea Research, Production and Utilization” (S. R. Singh and K. 0. Rachie, eds.), pp. 174-192. Wiley, Chichester. Evans, D. A. 1983.Biotechnology May, 253-261. Evans, D. A., Flick, C. E., and Jenson, R. A. 1981. Science 213,907-909. Evans, P. K.,and Cocking, E. C. 1977.In “Plant Tissue and Cell Culture” (H. E. Street, ed.), 2nd ed. Blackwell, Oxford. pp. 103-135. Faris, D. G. 1965. Can. J Genet. Cyrol. 7,433-452. Fatokun, C. A., and Singh, B. B. 1987.Plant Cell Tissue Organ Cult. 9,229-233. Fraley, R. T., and Horsch, R. B. 1983.In “Genetic Engineering of Plants: An Agricultural Perspective” (T. Kosuge, C. P. Meredith, and A. Holaender, eds.), pp. 177-194. Plenum,New York. Franckowiak, J. D., Ojomo, 0. A., and Barker, L. N. 1973. “Ife Brown, a New Cowpea Cultivar for Western Nigeria.” Paper prepared for the 9th annual meeting of the Agricultural Society of Nigeria, Ilorin, Nigeria. Frantzen, K. A,, Johnson, L. B., and Stuteville, D. L. 1982.Phyoparhology 72,568-573. Gamborg, 0 . L., and Bottino, P. J. 1981.In “Advances in Biochemical Engineering: Reactors and Reactions” (A. Fiechter, ed.), pp. 239-263. Springer-Verlag, New York. Gamborg, 0. L., Shyluck, J. P., Brar, D. S., and Constab, F. 1977.Plant Sci. Lett. 10,67-74. Gamborg, 0. L., Davis, B. P., and Stahlhut, R. W. 1983.Plant Cell Rep. 2, 209-212. Gardner, R. C. 1983. In “Genetic Engineering of Plants: An Agricultural Perspective” (T. Kosuge, C. P. Meredith and A. Hollaender, eds.), pp. 121-142.Plenum, New York.

152

A. 0. LATUNDE-DADA

Gautheret, R. J. 1939. C . R. Seance Acad. Sci. Paris 208, 118-121. Gengenbach, B. G., and Green, C. E. 1975. Crop Sci. 15,645-649. Gengenbach, B. G., Green, C. E., and Dononvan, C. M. 1977. Proc. Natl. Acad. Sci. U . S . A . 74,5113-5117. Giles, K. L. 1977. In “Applied and Fundamental Aspects of Plant Cell, Tissue and Organ Culture” ( J . Reinert and Y. P. S. Bajaj, eds.), pp. 536-550. Springer-Varlag, Berlin. Green, C. E. 1977. HortScience l2, 131-134. Gresshoff, P. M. 1980. Bor. Gaz. (Chicago) 141, 157-160. Griga, M., Kubalakov, M., and Tejklova, E. 1987. Plant Cell Tissue Organ Cult. 9, 167-177. Haberlandt, G. 1902. Sber. Akad. Wiss. Wien. 111,69-92. Hain, R., Stabel, R., Czernilofsky, A. P., Steinbib, H. H., Herrera-Estrella, L., and Schell, J. 1985. Mol. Gen. Genet. 199, 161-168. Heinz, D. J., Krishnamurthi, M., Nickell, L. G., and Maretzki, A. 1977. In “Applied and Fundamental Aspects of Plant Cell, Tissue and Organ Culture” ( J . Reinert and Y. P. S. Bajaj eds.), pp. 3-17. Springer-Verlag, Berlin. Henke, R. R., Mansur, M. A., and Constantin, M. J., 1978. Physiol. Plant. 44, 11-14. Herrera-Estrella, L., De Block, M., Messens, E., Hernalalsteens, J. P., Van Montagu, M., and Schell, J. 1983a. EMBO J . 2,987-995. Herrera-Estrella, L., Depicker, A., Van Montagu, M., and Schell, J. 1983b. Nature (London) 303,209-213. Herrera-Estrella, L., Van den Broeck, G., Maenhaut, R., Van Montagu, M., Schell, J., Timko, M., and Cashmore, A. 1984. Nature (London)310, 115-120. Hess, D. 1977. I n “Applied and Fundamental Aspects of Plant Cell, Tissue and Organ Culture” ( J . Reinert and Y. P. S. Bajaj, eds.), pp. 506-535. Springer-Verlag, Berlin. Honma, S. 1956. J. Hered. 47,217-220. Iacobellis, N. S., and De Vay, J. E. 1987. Physiol. Mol. Plant Pathol. 30,421-432. IITA. 1984. Res. Briefs 5(4), 2, 5 , 10. Jacobsen, H. J., and W. Kysely, 1984. Plant Cell Tissue Organ Cult. 3,319-324. Johnson, L. B., Stuteville, D. L., Higgins, R. K., and Skinner, D. Z. 1981. Plant Sci. Lett. 20, 297-304. Kameya, T., Horn, M. E., and Wildholm, J. M. 1981. Z. Pfanzenphysiol. 104,459-466. Kao, K. N., and Michayluk, M. R. 1974. Planta 115,353-367. Kao, K. N., and Michayluk, M. R. 1980. Z. Pfanzenphysiol. 96, 135-142. Kao, K. N., Constabel, F., Michayluk, M. R., and Gamborg, 0. L. 1974. Planta 210, 2 15-227. Karp, A., Nelson, R. S., Thomas, E., and Bright, S. W. J. 1982. Theor. Appl. Genet. 63, 265-272. Keller, W. A., and Melchers, G. 1973. Z. Naturforsch. 28,737-741. Kemble, R. J., Flavell, F. B., and Brettell, R. I. S. 1982. Theor. Appl. Genet. 62,213-217. Kleinhofs, A. 1977. Ann. Rev. Gent. 11,79-101. Krikorian, A. D. 1982. Biol. Rev. 57, 151-218. Larkin, P. J., and Scowcroft, W. R. 1981. Theor. Appl. Genet. 60,197-214. Larkin, P. J., and Scowcroft, W. R. 1983. I n “Genetic Engineering of Plants” (A. Hollaender, C. P. Meredith, and T. Kosuge, eds.), pp. 289-314. Plenum, New York. Larkin, P. J., Ryan, S. A., Brettell, R. I. S., and Scowcroft, W. R. 1984. Theor. Appl. Genet. 67,443-455. Latunde-Dada, A. 0. 1983. Ph.D. Thesis, Univ. of Nottingham, Nottingham, U.K. Latunde-Dada, A. O., and Lucas, J. A. 1983. Plant Sci. Lett. 32,205-211. Latunde-Dada, A. O., and Lucas, J. A. 1988. Plant Sci. 58, I 1 1-1 19. Lautour, G. De., Jones, W. T., and Ross, M. D. 1978. N. Z. J . Bor. 16,61-66.

GENETIC MANIPULATION OF T H E COWPEA

153

Li, B. J., Langridge, W. H. R., and Szalay, A. A. 1985. Plant Cell Rep. 4,344-347. Lippmann, B., and Lippmann, G. 1984. Plant Cell Rep. 3,215-218. Liu, M. C. 1981. In “Plant Tissue Culture” (T. A. Thorpe, ed.), pp. 299-323. Academic Press, New York. Lorz, H., Baker, B., and Schell, J. 1985. Mol. Gen. Genet. 199, 178-182. Lush, W. M., and Evans, L. T. 1981. Euphytica 30,579-587. Luthy, P. 1980. FEMS Microbiol. Lett. 8, 1-7. MarCchal, R. 1982. Taxon. 31,280-283. Martchal, R., Manscherpa, J. M., and Stainier, F. 1978. Boissiere 28, 1-273. MarrC, E. 1980. Prog. Phytochemisiry 6,253-284. Martin, I. S . , and Sondahl, M. R. 1984. Plani Physiol. 117,97-103. Martin, M., and Dewick, P. M. 1979. Phytochemistry 18, 1309-1317. Melchers, G., Sacristan, M. D., and Holder, S. A. 1978. Carlsberg Res. Commun. 43, 203-218. Muir, W. H., Hilderbrandt, A. C., and Ricker, A. J. 1954. Science 119,877-887. Muleba, N., and Ezumah, H. C. 1985. In “Cowpea Research Production and Utilization” (S. R. Singh and K. 0. Rachie, eds.), pp. 289-295. Wiley, Chichester. Ng, N. Q.. and R. MarCchal, 1985. In “Cowpea Research, Production and Utilization” (S. R. Singh and K. 0. Rachie, eds.), pp. 11-21, Wiley, Chichester. Nickell, L. G. 1977. Crop Sci. 17,717-719. Nishi, S. J., Kawata, J., and Toda, M. 1959. Jpn. J. Breed. 8, 215-222. Nobecourt, P. 1939. C. R . Seances SOC.Biol. 130, 1270-1271. Nutman, P. S. 1975. In “Symbiotic Nitrogen Fixation in Plants, International Biological Program” (P. S. Nutman, ed.), Vol. 7. Cambridge Univ. Press, Cambridge. Ojomo, 0. A. 1975. I n “Proceedings of IITA Collaborators’ Meeting on Grain Legume, 9-13 June 1975” (R. A. Luse and K. 0. Rachie, eds.). IITA, Ibadan, Nigeria. Onesirosan, P. T., and Barker, L. N. 1971. Planf Dis.Rep. 55,820-822. Oono, K. 1981. In “Plant Tissue Culture” (T. A. Thorpe, ed.), pp. 273-298. Academic Press, New York. Orton, J. T. J. 1984. Adu. Plant Paihol. 2, 153-189. Paszkowski, J., Shillito, R. D., Saul, M., Mandak, V., Horn, T., Horn, B., and Potrykus, I. 1984. EBMO J . 3,2717-2722. Patridge, J. E., and Keen, N . T. 1976. Phytopathology 66,426-429. Piper, C. V. 1913. “The Wild Prototype of the Cowpea.” United States Department of Agriculture, Bureau of Plant Industries, Circular 124. Plucknett, D. L., and N. J. H. Smith, 1986. Bioscience 36,40-45. Poku, J. A., and Akobundu, I. 0. 1985. I n “Cowpea Research, Production and Utilization” (S. R. Singh and K. 0. Rachie, eds.), pp. 327-334. Wiley, Chichester. Potrykus, I . , Paszkowski, J., Saul, M. W., Petruska, J., and Shillito, R. D. 1985a. Mol. Gen. Genet. 199, 169-177. Portrykus, I., Saul, M. W., Petruska, J., Paszkowski, J., and Shillito, R. D. 1985b. Mol. Gen. Genet. 199, 183-188. Power, J. B., Cumming, S. E., and Cocking, E. C. 1970. Nature (London)225, 1016-1018. Power J. B., Frearson, E. M., Hayward, C., George, D., Evans, P. K., Berry, S. F., and Cocking, E. C . 1976. Nature (London) 263,500-502. Preston, N . W., Chamberlain, K., and Skipp, R. A. 1975. Phytochemistry 14, 1843-1844. Rachie, K. 0. 1985. In “Cowpea Research, Production and Utilization” (S. R. Singh and K. 0. Rachie, eds.), pp. 21-28. Wiley, Chichester. Rawal, K. M. 1975. Euphytica 24,699-707. Reed, S. M., and Collins, G. B. 1978. J . Hered. 69,311-315.

154

A. 0. LATUNDE-DADA

Reisch, B., and Bingham, E. T. 1981. Crop Sci. 21,783-788. Sacristan, M. D., 1982. Theor. Appl. Genet. 61, 193-200. Santos, A. V. P., Dos, Outka, D. E., Cocking, E. C., and Davey, M. R. 1980. PJanzenphysiol. 99,261-270. Sauer, C. 0. 1952. “Agricultural Origins and Dispersals.” MIT Press, Cambridge, Massachusetts. Schenck, H. R., and Robbelen, G. 1982. PJanzenzuecht. 89,278-288. Schieder, 0.. and Vasil, I. K. 1980. In “International Review of Cytology Supplement IIB: Perspectives in Plant Cell and Tissue Culture” (I. K. Vasil, ed.), pp. 21-46. Academic Press, London. Scowcroft, W. R., and Larkin, P. J. 1985. In “Comprehensive Biotechnology: The Principles, Application and Regulations of Biotechnology in Industry, Agriculture and Medicine” (M. Moo-Young, ed.), Vol. 4, pp. 153-168. Pergamon, Oxford. Shepard, J. F. 1981. Annu. Reu. Phytoparhol. 19, 145-166. Shepard, J. F., Bidney, D., and Shahin, E. 1980. Science 28, 17-24. Shepard, J. F., Bidney, D., Barsby, T., and Kemble, R. 1983. Science 219,683-688. Singh, S . R., and Jackai, L. E. N. 1985. In “Cowpea Research, Production and Utilization” (S. R. Singh and K. 0. Rachie, eds. pp. 217-231. Wiley, Chichester. Singh, B. B., and Ntare, B. R. 1985. In “Cowpea Research, Production and Utilization” (S. R. Singh and K. 0. Rachie, eds.), pp. 105-1 15. Wiley, Chichester. Skirvin, R. M. 1978. Euphytica 27,241-266. Skoog, F., and Miller, C. 0. 1957. Symp. Soc. Exp. Biol. 11, 118-130. Snell, C. L. 1978. In “Interspecific Hybridization in Plants” (E. Sanchez-Monge and F. Garcie-Oimedo, eds.), pp. 339-343. Univ. Politecnica de Madrid, Madrid. Sree Ramulu, K., Dijkhuis, F. and Roest, S. 1983. Theor. Appl. Genet. 65,329-338. Steele, W. M. 1976. In “Evolution of Crop Plants” (N. W. Simmonds, ed.), pp. 339-343. Longman, London. Steward, F. C. 1958. A m . J. Bot. 45,709-713. Summerfield, R. J., Huxley, P. A., and Steel, W. 1974. Field Crop Abstr. 27, 301-312. Sun, Z.-X., Cheng-Zhang, Zheng, K.-L., Qi, X.-F., and Fu, Y.-P. 1983. Theor. Appl. Gener. 67967-73. Takebe, I., Labib, G., and Melchers, G. 1971. Naturwissenshafren 58,318-320. Thomas, E., King, P. J., and Potrykus, I. 1979. Z. PJlanzenzuecht. 82,l-30. Thomas, E., Bright, S. W. J., Franklin, J . , Lancaster, V. A., and Miflin, B. J. 1982. Theor. Appl. Genet. 62,65-68. Thottappilly, G . , and Rossel, H. W. 1985. In “Cowpea Research, Production and Utilization” (S. R. Singh and K. 0. Rachie, eds.), pp. 155-171. Wiley, Chichester. Van Overbeek, J., Conklin, M. E., and Blankeslee, A. F. 1942. A m . J . Bot. 29,472-477; cited by Yenug, E. C., Thorpe, T. A., and Jensen, C. J. 1981. In “Plant Tissue Culture: Methods and Applications in Agriculture” (T. A. Thorpe, ed.), pp. 253-271. Academic Press, New York. Vavilov, N. I. 1951. Chron. Bor. 13, 1-364. White, P. R. 1934. Plant Physiol. 9,585-600. White, P. R. 1939. A m . J. Eor. 26,59-64. Williams, E. 1978. N . Z . J . Eot. 16,499-506. Williams, R. J. 1975. PANS 21,253-267. Wittwer, S. H. 1979. Bioscience 29,603-610. Xu, X-H, Davey, M. R., and Cocking, E. C. 1982. Z. PJlanzenphysiol. 107,231-235. Zhukovskii, P. M. 1962. “Cultivated Plants and Their Wild Relatives.” Commonwealth Bureau of Plant Breeding (transl. by P. S. Hudson), Cambridge.

ADVANCES IN AGRONOMY, VOL. 44

NITROGEN FIXATION BY LEGUMES IN TROPICAL AND SUBTROPICAL AGRICULTURE Mark B. Peoples’ and David F. Herridge2

’ Australian Centre for International Agricultural Research (Project 8800) CSIRO Division of Plant Industry Canberra, A.C.T. 2601, Australia Australian Centre for International Agricultural Research (Project 8800) New South Wales Agriculture and Fisheries Tamworth, New South Wales 2340, Australia

I. Introduction 11. Methods of Assessing N2 Fixation A. ISN-Isotopic Techniques B. N-Difference Method C. Ureide Method D. N Balance E. Acetylene Reduction Assay F. N Fertilizer Equivalence G. Nodule Evaluation 111. N2 Fixation in Legume Production Systems A. Food Legumes B. Forages, Multipurpose Tree Legumes, Cover Crops, and Green Manures IV. Contribution of Legume N to Plant and Animal Production A. Direct Transfer B. Contribution of Legume Residues to Crop Production C. Contribution of Legumes to Livestock Systems V. Strategies to Enhance N2 Fixation A. Plant Breeding and Selection B. Rhizobia, Inoculation, and Plant Nodulation C. Crop and Soil Management V1. Concluding Remarks References

Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

156

MARK B. PEOPLES AND DAVID F. HERRIDGE

I. INTRODUCTION Soils in the humid tropics and subtropics are commonly leached, highly weathered, and low in both total and plant-available nitrogen (N). Agricultural systems in these regions are subjected to high annual losses of N (see Barthomomew, 1977; Greenland, 1977; Wetselaar and Ganry, 1982) and to rapid land degradation because of the intense cultivation and the evershortening of fallow periods in traditional farming systems in response to increasing human populations (Lal, 1989). In drier regions, overgrazing and harvesting of trees for fuelwood have reduced productivity and accelerated desertification (N. Sanginga, personal communication). This decline in soil quality is a serious global problem. The current annual rate of loss of arable land to soil degradation is estimated at 5-7 million ha (Lal, 1989). Crop yields in the tropics and subtropics are often limited by N supply, and animal productivity and fecundity are restricted by poorquality forage (Gibson et al., 1982; Bayer and Waters-Bayer, 1989). Apart from high-return cash crops, N fertilizers are used only to a limited extent to improve plant productivity because of their high costs, the low per capita incomes and limited credit facilities of most farmers, and lack of effective infrastructures for their production and distribution (Wood and Myers, 1987). It is in these situations that the opportunities are greatest for the exploitation of the legumes’ capacity to symbiotically fix atmospheric Nz. Commonly consumed legume seeds contain 17-34% protein (Wijeratne and Nelson, 1987) and most tropical legume forages contain 2-4% N (Harricharan et al., 1988; Little et al., 1989). Potentially, these protein-rich legume products can be produced in N-impoverished soils without additions of fertilizer N , and in any soil without depleting soil N reserves. The residues can contribute N to the soil organic pool when incorporated or fed to animals. However, these desirable characteristics of legumes can be realized only when large amounts of atmospheric N2 are fixed. The successful formation of a functional symbiosis is dependent on many physical, environmental, nutritional, and biological factors and cannot be assumed to occur as a matter of course (see Gibson et al., 1982; Haque and Jutzi, 1985; O’Hara et al., 1988). Failure to fix NZin soils low in mineral N will result in depressed legume production, and high rates of fertilizer N may be required to achieve seed and plant yields similar to those of a wellnodulated crop (compare nonnodulating and nodulating soybean data in Table I). In a soil with higher concentrations of mineral N, the legume can compensate for poor N2 fixation by scavenging N from the soil. Although

157

NITROGEN FIXATION BY TROPICAL LEGUMES Table I

Effects of N Fertilizer on the Seed Yield and Productivity of a Nonnodulating Soybean (Glycine mar CV. Lee) and the Nonlegume Ragi (Eleusine coracana) Compared to a Nodulating, N2-Fixing Soybean Cultivar (cv. Shilajeet) Grown Without N Fertilizera Nodulated soybean N fertilizer treatment (kg N/ha) Oh

20" 40 80

160

Nonnodulating soybean

Ragi

Seed yield (t/ha)

Total crop N (kg N/ha)

Seed yield (t/ha)

Total crop N (kg N/ha)

Seed yield (t/ha)

Total crop N (kg N/ha)

2.39 -

232 -

1.51 1.55

I03 1 I6 152

I .52

81 90

LSD(p = .05)

1.87

2.11 2.43 0.18

1.78

106 127

233

2.06 2.50 2.72

26

0.50

17

185

158

" Data derived from Chandel er a / . (1989). Soil in unamended field plots contained 0.081% total N N fertilizer applied as urea.

production in this situation may not be impaired, the net result of cropping with a legume with deficient nodulation is an exploitation of N reserves. Soil N fertility is lost and the potential benefit of the legume in a cropping sequence will not be realized. Because the relationship between N2 fixation and legume growth or production is not always direct and obvious, considerable effort has been directed into the development of methods for measuring N2 fixation (e.g., Weber, 1966; Hardy et al., 1973; Fried and Broeshart, 1975;Mariotti et al., 1983;McClure et al., 1980; Pate et al., 1980; Herridge, 1982b; Bergersen et a / . , 1989). Only field measurements of N2 fixation will allow realistic evaluation of the contribution made by legumes to a system's N balance and provide a basis for developing strategies to better manage the soilcrop system. Such measurements allow more precise identification of limitations on legume growth and N2-fixing capacity. In this chapter we discuss the merits of the different techniques that have been used to quantify N2 fixation and review the role of symbiotic N2 fixation in the N economies of the many legume-based production systems of the tropics and subtropics. Finally, we review a number of strategies that may lead to improved N2 fixation in these systems.

158

MARK B. PEOPLES AND DAVID F. HERRIDGE

II. METHODS OF ASSESSING N2 FIXATION There is no single “correct” way of measuring symbiotic N2 fixation. No one technique can be relied upon to provide an accurate measure of N2 fixation for every legume species grown under all soil or environmental variables. Each technique has unique advantages and limitations, however, some methods are more likely to provide reliable and quantitative estimates of N2 fixation than others. Methods employed in legume field studies will be summarized in the following sections. Additional details and comparisons of the various methodologies may be found in other more comprehensive reviews (La Rue and Patterson, 1981; Herridge, 1982a; Danso, 1985; Bergersen, 1988a; Ledgard and Peoples, 1988; Peoples et al., 1988; 1989a).

A. ‘ S N - ITECHNIQUES ~ ~ ~ ~ ~ ~ ~ There are two stable isotopes of N, I4N and lsN. The heavy isotope, ”N, occurs in atmospheric N2 at a constant abundance of 0.3663 atom% (Mariotti et al., 1983). If the isotopic concentrations are different in two sources of N (soil N and atmospheric N2), the proportion P of plant N arising from one of them (in this case N2) can be calculated as

Where x , y , and b are, respectively, the isotopic compositions of the plantavailable soil N, of a N2-fixing plant growing in that soil, and of a legume fully dependent on N2 fixation for growth. In practice, x is obtained from the isotopic composition of a non-N2-fixing reference plant (a nonlegume, uninoculated legume, or nonnodulating legume genotype) that is totally dependent on soil N (Fig. 1). As the reliance of the legume on N2 fixation for growth is reduced, the value of y approaches x; conversely, y approaches b as P increases. When the ”N concentration of plant-available soil N is higher than atmospheric N2, N2 fixation will result in a gradual decline in the ”N composition of the legume from the level measured in the reference plant as N is progressively assimilated from the air (Fig. 1). This is sometimes described as “”N dilution” (Chalk, 1985). Although this is a particular application of “isotope dilution,” the term is also used for different purposes in experiments with soil-plant systems (e.g., Nishio and Fujimoto, 1989). Assumptions inherent in ”N methods are that:

159

NITROGEN FIXATION BY TROPICAL LEGUMES N

A

.n

FIG. 1. Diagrammatic representation of the principles involved in "N techniques for measuring N2 fixation. The utilization of atmospheric N2 results in a lower "N composition in the legume than measured in the nonfixing reference plant and is represented as a reduced area in the "N sector of the leguminous plant. After Peoples er a / . (1988).

1. The reference plant lacks the ability to fix N2 and the l5N/I4N ratio measured in its products of growth is the same as plant-available soil mineral N. 2. The legume and non-N2-fixing reference plant explore a soil N pool of identical l5N/I4Nabundance.

If a suitably precise mass spectrometer is available, the very small differences in "N abundance between atmospheric N2 and soil (e.g., 0.368-0.373 atom% "N in soil mineral N), which occurs naturally in many agricultural soils, can be utilized for N2 fixation measurements (Shearer and Kohl, 1986). More usually, the differences between N2 and soil N are extended by incorporation of "N-enriched compounds into the soil (e.g., of between 5 and 95 atom% "N; Chalk, 1985). Alternatively, although less frequently, differences between I5N composition of air and soil may also be widened by applying "N-depleted (i.e., depleted in "N content relative to atmospheric N2) materials to soil (e.g., 0.0016 atom% I5N ammonium sulfate; van Kessel and Roskoski, 1988).

160

MARK B. PEOPLES AND DAVID F. HERRIDGE

1 . "N Enrichment

The use of methods involving the artificial adjustment of soil "N concentrations to measure N2 fixation has been extensively reviewed (e.g., Chalk, 1985; Hauck and Weaver, 1986; Danso, 1988; Ledgard and Peoples, 1988; Witty et al., 1988; Peoples e t a l . , 1989a). A major assumption is that the legume and reference plant or crop absorb the same relative amounts of N from the added "N and from the soil. The high cost of instrumentation to measure 15N and expense of "N materials (U.S.$lO/g of ["Nlnitrate of 10% enrichment) are limitations to the widespread use of this potentially very accurate method. The main advantage of the technique is that it provides a "time-averaged'' estimate of P, which is the integral of any changes in plant reliance on N2 fixation that may have occurred during the measurement period. The calculation of P is independent of yield, although it is necessary to measure dry matter and N yield to determine the amount of N2 fixed. Estimates of P are calculated from the following: P = l -

(atom% 15N excess legume N) (utom% "N excess soil-derived N)

where atom% "N excess = (atom% "N sample) - (atom% "N air N2), and atom% "N of air N2 = 0.3663. It can be seen that Eq. (2) is related to (1). The atom% "N of soil-derived N is generally estimated from the "N enrichment of non-N2-fixing reference plants grown in the same soil over the same period as the legume. It may be possible under certain conditions to determine it directly from measurements on the soil mineral N (Chalk et al., 1983).The choice of an appropriate non-N2-fixing reference plant is the single most important factor affecting the accuracy of estimates of P. Figures 2a-c illustrate three possible legume-reference plant relationships that are likely to occur in field experiments using "N enrichment. When "N-labeled materials are applied to soil (most commonly in the form of soluble inorganic fertilizer), a highly enriched layer is usually created near the surface. Enrichment of soil mineral N generally declines rapidly with depth. In Fig. 2a the legume roots have explored deep into the region not artificially enriched with 15N,whereas roots of the reference plant have been restricted to the uppermost enriched zone. The uptake of any soil N by the legume from the unenriched region will, when the lsN/I4N compositions of the legume and reference are compared, lead to an overestimation of NZfixation. In the reverse situation (Fig. 2b), where roots of the reference plant have grown below the "N-enriched zone, uptake of N from the unenriched soil

NITROGEN FIXATION BY TROPICAL LEGUMES

161

zone of I5N enrichment

\~............. .............. ................ ................. ..........................

*dd

..

decreasing w i t h depth

~~~~

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

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

I5N enriched

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

-

n -. atu a l 15N .. -.r -. ..

abundance

(:::::I

:fl.y&@: :::::::::::::::::::

.... .................... .... .'qy.V ..................... ..............................

FIG.2. Diagrammatic representation of four situations that can occur during "N experimentation in the field. Figures (a) to (c) illustrate differences in the root growth of a nodulated legume and a non-N2-fixingreference plant in hypothetical I5N enrichment studies. Application of "N-enriched material to the soil surface has created a zone where soil N is artificially enriched with 15N. However, the "N enrichment decreases down the soil profile until natural abundance levels are reached at depth. The fourth case study (d) illustrates a "N natural abundance study where the 15N composition of plant-available soil N is relatively uniform down the soil profile. After Peoples et a / . (1989a).

will result in a lowered "N content of the reference material and will lead to an underestimation of fixation by the legume. It is in the ideal situation (Fig. 2c), where roots from both test and reference plants explore a similar volume of soil of similar "N enrichment, that accurate determinations of symbiotic N2 fixation can be expected. However, even under these conditions, further errors can arise if there are substantial changes over time in isotopic composition in the soil N pool combined with differences in the patterns of N uptake by the legume and reference plants. A decline in soil

162

MARK B. PEOPLES AND DAVID F. HERRIDGE

lsN/14Nratio with time results from a loss of plant-available "N due to uptake, leaching, or immobilization and the continuing release of I4N compounds via mineralization of soil organic matter. Together, these factors can lead to estimates of fixation that vary with choice of reference species. Immobilized forms of "N have been used to ensure a more gradual release of labeled N and to provide a more stable enrichment of soil. These include the use of "N-labeled plant material, slow-release "N formulations (e.g., "N incorporated into gypsum pellets), or the application of soluble "N salts with a readily available carbon source (e.g., sucrose) so that the "N is bound in the soil biomass (Boddey et al., 1984; Giller and Witty, 1987). Errors associated with mismatching of the reference and N2-fixing legume are most important when P is small. 2. Natural "N Abundance

Almost all N transformations in soil result in isotropic fractionation. The net effect is often a small increase in the I5Nabundance of soil N compared with atmospheric N2 (Shearer and Kohl, 1986). In looking at such small differences in I5N concentration, data are commonly expressed in terms of parts per thousand (6"N or 0/00): 6"N

=

1000 x

(atom% "N sample) - (atom% "N standard) (atom% "N standard)

(3)

where the standard is usually atmospheric N2 (0.3663 atom%). By definition, the 6"N of air Nz is zero. The natural abundance method gives an integrated estimate of P over time as with "N enrichment studies, but it can be applied to established experiments (provided nonfixing reference material is available) because no pretreatment, that is, "N application, is necessary. An estimate of P is obtained by using the equation (analogous to (1) and (2)) P =

(6"N soil N) - (6"N legume N) (6I5Nsoil N) - B

(4)

The 6"N of soil N is commonly obtained from a non-N2-fixing reference plant. Similar estimates of P may also be obtained using direct soil measurements (Bergersen et a / . , 1989). The value B is a measure of isotopic fractionation during N2 fixation and is determined by analysis of the 6"N of total plant N accumulated by a nodulated legume grown in N-free media. Isotopic fractionation during N2 fixation is minimal but not zero and should be taken into account. The value of B should ideally be prepared for

NITROGEN FIXATION BY TROPICAL LEGUMES

163

each new legume species studied (Peoples et al., 1989a). Some species characteristically exhibit slight enrichment or depletion of 615N in various plant parts when grown with N2 as the sole source of N (see Yoneyama et al., 1984; Bergersen et al., 1988; Kohl et al., 1989). Critics of the natural abundance method frequently cite the enrichment of I5N in nodules as a potential source of difficulty when quantifying N2 fixation. The use of the appropriate B value (e.g., in soybean it is - 1.30 O/OO when analyzing only shoots, or a value of -0.79 if whole plants [shoot plus roots] are harvested; Bergersen et al., 1988)when calculating P minimizes possible errors. Even so, field results have indicated that well-nodulated roots make little contribution to total crop N and therefore have little effect on estimates of P (Bergersen et al., 1989). There appears to be no evidence for significant rhizobial strain-induced changes in 615N enrichment or depletion of different plant organs or whole-plant B values in tropical legumes (M. B. Peoples, unpublished data). Nonetheless, there can be dynamic changes in 615N of plant parts during organ development, so estimates of P should be based on 615N of whole plants or total shoot N and not of single leaves or individual plant parts (Bergersen et al., 1988). Although the principles of the natural abundance technique are the same as those underlying I5N enrichment, the main limitations are quite different (reviewed by Mariotti et al., 1983; Shearer and Kohl, 1986; Bergersen, 1988b; Ledgard and Peoples, 1988; Peoples et al., 1989a).An isotope ratio mass spectrometer capable of accurately measuring differences of 0.1 O/OO (about 0.00004 atom% "N) is required and sample collection and preparation require great care to: 1. Avoid contamination with "N-enriched materials. 2. Prepare uniform dry matter samples to avoid variation due to tissue differences in "N abundance. 3. Avoid losses of minute quantities of nitrogen during Kjeldahl digestion and distillation, or during concentration of distillates before analysis on the mass spectrometer. Such losses can change the "N abundance and lead to poor replication of samples.

The accuracy of the technique will ultimately depend on the levels of natural "N abundance of the soil. Low and/or variable soil 6"N values will be unsuitable for assessing N2 fixation. A range of levels of 615N has been measured in the mineral N component of tropical and subtropical soils (Table 11; see also Ruschel et al., 1982). Fortunately, the natural "N abundance of plant-available N is often sufficiently high and uniform for reliable N2 fixation measurements in arable agricultural systems where soils are regularly cultivated (Table 11). Although a gradual change in the 6"N of mineral N has been reported with soil depth for some rice-growing

164

MARK B. PEOPLES AND DAVID F. HERRIDGE

Table I1 Range of Levels of Natural Abundance of "N in Plant Available Soil N from Tropical and Subtropical Soils Location Australia New South Wales Queensland Western Australia India Indonesia Java Sumatra Malaysia Sungei Buloh Taiping Kuala Terengganu Philippines Luzon Leyte Thailand Chiang Mai Mae Taeng

Site use

I5N natural abundance (O/OO)

SE"

Reference"

0.4 0.2

I 2 3 4 5

Cropping Cropping Pasture/cropping Cropping Cropping

8.7- 13.9" 6.6 2.6-14.8" 6.3 4.0-5.2"

Tree legume stand Cropping Tree legume stand

2.0 6.0 2.5-4.2'

Rubber plantation Cropping Rubber plantation

3.2-3.6 10.1-16.6 1.2-4.3'

Cropping Tree legume stand

5.1 - 13. 5 c 1.0-3.8"

0.7

Cropping Cropping Cropping

8.0 3.9 4.5-8.4'

0.2 0.5

0.7 0.2

6

7 0.8 8 9 10

11 12 12

Standard error of the estimate mean at one site. I . Herridge et al. (1990); 2. M. B. Peoples and M. Bell (unpublished data); 3. H. V. A. Bushby (unpublished data); 4. Ofori et ul. (1987); 5. Yoneyama et a/. (1990); 6. D. P. Nurhayati, T. Ibrahim, and M. B. Peoples (unpublished data); 7. Norhayati er al. (1988); 8. A. W. Faizah and M. B. Peoples (unpublished data); 9. Watanabe er al. (1987); 10. A. S. Almendras, P. J. Dart, and M. B. Peoples (unpublished data); I I . Rerkasem e t a / . (1988); 12. A. Bhromsiri, C. Sampet, and M. B. Peoples (unpublished data). Range of values detected at different field sites. a

soils (Watanabe et al., 1987) (possibly a result of repeated, heavy applications of N fertilizers), analysis of many other soils suggests that this may not be a general phenomenon. More often the level of "N natural abundance of plant-available N is relatively constant with soil depth (as depicted in Fig. 2d; Watanabe er al., 1987; Bergersen, 1988b; G. L. Turner, M. B. Peoples, and F. J. Bergersen, unpublished data) and does not appear to change rapidly with time (see reference plant and soil data in Ofori er al., 1987; Norhayati et al., 1988; Bergersen et al., 1989; Herridge et al., 1990). Therefore, the major limitation of 'N enrichment techniques (i.e., choice

NITROGEN FIXATION BY TROPICAL LEGUMES

165

of appropriate reference plant) is likely to be less critical with the natural abundance method. Certainly, field studies have shown that the use of different reference plants does not greatly influence the calculation of P (Rerkasem et al., 1988; Bergersen et al., 1989; Yoneyama et al., 1990). Where natural abundance and "N enrichment techniques have been compared, field estimates of N2 fixation were similar, with similar precision (see Ofori et al., 1987; Ledgard and Peoples, 1988; Norhayati etal., 1988).

B. N-DIFFERENCE METHOD The simplest estimates of N2 fixation reported in the literature were obtained by measuring the total amount of N accumulated by legume crops. Such determinations are based on the arbitrary assumption that the legumes derive all of their N from N2 fixation. Values obtained will almost certainly overestimate N2 fixation. An adaptation of this N yield technique has been used to assess N2 fixation by tree species, where increments of soil N under trees and N contained in the plant biomass have been measured and summed (Dommergues, 1982). Results obtained using such a method should also be treated with caution since the observed increments cannot be attributed solely to symbiotic N2 fixation. Other processes, such as extraction of N by roots from deep soil horizons or from the water table, could contribute to the accumulation of N. A true measure of N2 fixation based on legume N yield can only be obtained when the contribution of soil N to total plant N is determined. This is often estimated by growing a non-N2-fixing crop in the same soil under identical conditions as the legume, usually in adjacent plots. The difference in total N accumulated by the legume (NI) and nonfixing control (Nnf) crops is regarded as the contribution of N2 fixation to legume growth. Thus, N2 fixation is calculated as N2 fixed = NI

-

Nnf

(5)

Using the crop N data for the unfertilized plots from Table I as an example, N2 fixed by the nodulated soybean is calculated to be 128 Kg N/ha using the nonnodulating soybean as the nonfixing control and 150 kg N/ha using the nonlegume Ragi. These calculations will be dependent on the accuracy of measurements of crop dry matter and N content, which in turn will largely be determined by errors associated with sampling and subsampling (Hunt et al., 1987). There are two basic assumptions inherent in the use of the N-difference procedure:

166

MARK B. PEOPLES AND DAVID F. HERRIDGE

1 . The N contained in the non-N2-fixingcontrol plants is derived only from soil N. 2. The legume and control crops assimilate the same amount of soil N.

The N-difference method is a relatively simple procedure and can be used when facilities for only total N analysis are available. However, because of the underlying requirement that the legume and control plants utilize equivalent amounts of soil N, the choice of control crop is of utmost importance. Ideally the two plant types should explore the same rooting volume, have the same ability to extract and utilize soil mineral N, and accumulate soil N over the same period of time. A non-N2-fixing control plant may be ( a ) a non-legume; (6) an uninoculated legume (requires the soil to be free of effective Rhizobium spp.); or (c) a nonnodulating legume, preferably an isoline of the test legume. Unfortunately, there are often substantial differences between fixing and nonfixing plants in their capacities to use soil N. Even when a supposed “ideal” control plant is used (e.g., a nonnodulating isoline of the test legume), errors in calculating N2 fixation may result because of differences in root morphologies (Boddey et al., 1984). Residual levels of soil mineral N may also be higher following a legume crop than a nonfixing crop (Herridge and Bergersen, 1988). Such observations led Evans and Taylor (1987) to propose a modification of the N-difference procedure to improve accuracy when legume and control are not well matched. In this method the amount of mineral N in the soil under the legume (Soil (1)) and nonfixing (Soil (nf)) crops is measured at the completion of crop growth and the difference between the two added to the difference in crop N yields. Thus Eq. ( 5 ) becomes N2 fixed = [NI - Nnf]

+ [Soil (I)

-

Soil (nf)]

(6)

The importance of choosing a particular control plant depends ultimately on the level of plant-available soil N and the amount of N2 fixed. If the control plants accumulate very much less N than the legume, differences in uptake of soil N between the two plant types will not greatly influence the determination of N2 fixation. Comparisons of N2 fixation calculated by N-difference with estimates by ”N dilution have often shown good agreement in soils low in N or where recovery of ”N label is equal in the legume and control (Chalk, 1985). Under such conditions, however, agreement of estimates is mathematically inevitable and in itself does not constitute an independent confirmation that either technique is correct (Urquiaga and Boddey, 1987).

NITROGEN FIXATION BY TROPICAL LEGUMES

167

C. UREIDEMETHOD Xylem sap carries N-containing compounds from the roots to the shoots of field-grown legumes originating from ( a ) nodules as assimilation products of N2 fixation and (b) soil mineral N taken up by the roots. The first stable product of N2 fixation in the legume nodule is ammonia. The ammonia is released from the bacteroids to be assimilated into glutamine and glutamate in the infected host cell via the coupled activity of the enzymes glutamine synthetase (GS) and gluatmate synthase (GOGAT). Table I11 Nodulated Legumes That Transport Ureides as a Major Nitrogenous Component of Xylem Sap" Genus Albizia Cajanus Calopogonium Cenirosema Codariocalyx Cyamopsis Desmodium Glycine Hardenbergia Kersiingiella Lablab Macroptilium Macrotyloma Phaseolus Psophocarpus Pueraria Tedehegi Vigna

Voandzeia

Speciesh lophaniha cajan caeruleum pubescens gyroides reiragonoloba discolorlrensoniiluncinaium

max geocarpa purpureus atropurpureum ungorum lunatusluulgaris ietragonolobus jauanicalphaseoloides triqueirum angularislmungolradiatal

trilobalunguiculatal umbellaia subierranea

Forty percent or more of total xylem sap N is estimated to be in the form of allantoin and allantoic acid. Adapted from Pate and Atkins (1983);Schubert (1986); Peoples ei al. (1989a), and includes unpublished information from D. F. Herridge, B. Palmer, M. B. Peoples, F. D. Dakora, C. A. Atkins, and J. S. Pate.

168

MARK B. PEOPLES AND DAVID F. HERRIDGE

Despite the production of glutamine as the initial product of ammonia assimilation, it is rarely the major N solute transported in the xylem of nodulated plants (Pate and Atkins, 1983; Peoples et al., 1987). Secondary reactions involving the transfer of the amide- or amino-N of glutamine to other products comprise important metabolic processes within the functional nodule (Pate and Atkins, 1983; Schubert, 1986). In many nodulated legumes of tropical origin, the ureides, allantoin and allantoic acid, are the predominant N compounds exported from the nodule and transported in the xylem (Table 111; e.g., soybean, see Fig. 3). Nitrate and ammonium ions are the two forms of mineral N taken up by plant roots. In most agricultural soils, where nitrification takes place rapidly, nitrate is considered to be the dominant N source for plant growth (Stevenson, 1982). Solutes derived from soil mineral N under these conditions will be transported in the xylem as free nitrate or as organic products of nitrate reduction. Characteristically, nitrate reduction assumes a minor role in assimilating nitrate in the roots of most tropical and subtropical legume species (Andrews, 1986; Wallace, 1986), and, as a consequence, much of the incoming nitrate is transported in an unreduced form (Fig. 4).

R

0001

-1

- N

+ NO3

FIG.3. Composition of N solutes of xylem sap collected from fully symbiotic (nodulated plants maintained on N-free complete nutrients) and nonnodulated (fed zero or 10 mM nitrate) soybean plants. Saps were collected as bleeding sap from decapitated roots or from detached nodules. Asn = asparagine; Gln = glutamine. After Peoples and Gibson (1989).

169

NITROGEN FIXATION BY TROPICAL LEGUMES

The portion of the nitrate that is metabolized will be reduced to ammonia by the combined action of root nitrate reductase and nitrite reductase (Pate and Atkins, 1983). The ammonia will then be assimilated by GS and GOGAT as in the nodule, though, unlike the nodule, ureides do not play a major role in the subsequent metabolism or transport of this N from the roots. Rather it is the amide, asparagine, that is the exported end product of nitrate reduction (Fig. 3). The same appears to be true if ammonium rather than nitrate is taken up by roots. The incoming ammonium ions are incorporated into amides and amino acids rather than ureides (Peoples et al., 1989b). Therefore, in the absence of nodules, ureides represent only a very minor component of the total N of xylem sap (Fig. 3). Because there are substantial differences in the principal forms of N transported in the xylem between highly symbiotic and unnodulated or poorly fixing legume plants, it is possible to use the abundance of ureides relative to the other N components in xylem sap as an indirect 'measure of the proportion (P)of plant N derived from N2 fixation (Fig. 4). Glasshouse calibration experiments (McClure et al., 1980; Pate et al., 1980; Rerkasem et al., 1988; Peoples et al., 1989b; Herridge and Peoples, 1990) indicate that there is a predictable progressive decrease in xylem ureides and a compensatory increase in amino acid and nitrate contents as N2 fixation PARTIALLY SYMBIOTIC

FULLY SYMBIOTIC

N,

N,

SHOOT AXIS

1 ly

+ NO;

NON- SYMBIOTIC NO;

A

NoDwM 1

4' 1 a

RooT

N2T -w NO;

b

XYLEM TRANSPORT OF:

a

C

UREIDE I AMINO COMPOUNDS NITRATE 6 AMINO COMPOUNDS

FIG.4. Pathways of N transport from the root systems of ureide-exportinglegume species when growing in (a) nitrate-free soil, (b) moderate levels of soil nitrate, and (c) high levels of soil nitrate.

170

MARK B. PEOPLES AND DAVID F. HERRIDGE

declines and uptake of combined N by roots increases (Fig. 5). The list of ureide-exporting species so far identified (Table 111) suggests that the ureide method is potentially applicable to the most important food legumes (Table V; but not groundnut, chick-pea, or lentil; Peoples et al., 1986; 1987) and to many forage and shrub legumes (Tables IX and X). To date, the ureide method has been used in field studies in the tropics more to detect the presence of fixation activity or to differentiate treatment effects (Kumar Rao et al., 1981; Neves et al., 1985; Kucey et al., 1988a,b; Norhayati et al., 1988) than to quantify NZfixation (Rerkasem et al., 1988; Hughes and Herridge, 1989; Herridge et al., 1990). Sampling of legumes for N solutes may be achieved by (Herridge, 1982b; Peoples et al., 1989a; Herridge and Peoples, 1990): ( a ) collecting xylem sap exuding spontaneously from the root stumps of decapitated plants; (b)recovering xylem sap from freshly detached stems or branch segments using a mild vacuum; or (c) preparing aqueous extracts of the soluble N pool of stem tissue. Concentrations of N solutes can vary markedly from plant to plant and may be dependent on individual plant water status, tissue metabolism, and transpiration rate. However, the effects of such concentration changes can be minimized by presenting N solute data in the form of a ratio. The abundance of ureides, therefore, is usually expressed as a proportion of total sap N. Thus 100 X (4 x ureides) Relative ureide - N (%) = (4 X ureides) + amino + nitrate (7) where ureides, amino acids, and nitrate are given in molar concentrations.

FIG.5. Changes in the composition of N solutes of xylem sap collected as root-bleeding sap or vacuum-extracted sap of nodulated soybean fed a range of constantly maintained levels of [''Nlnitrate. Derived from Herridge and Peoples (1990).

NITROGEN FIXATION BY TROPICAL LEGUMES

171

Technically, the field sampling of xylem contents is quick and simple, and analysis of the individual N solute components (i.e., ureides, amino compounds, and nitrate) can be done by colorimetric analyses in test tubes (Peoples et al., 1989a). There is no need for expensive or sophisticated equipment, and many samples can be collected and analyzed on a single day. Sampling need not be totally destructive (Herridge et al., 1988) and since it is confined to the accessible aerial parts of the legume, the method may overcome problems associated with measuring N2 fixation in twining ground cover and forage legumes and in woody perennial legumes (Peoples et al., 1988). There appears to be little problem associated with applying the ureide method to woody legumes. Large volumes of sap can be recovered from short branch segments and there is no evidence of branch-to-branch variation in the composition of sap from field-grown trees (D. F. Herridge and D. P. Nurhayati, unpublished data). However, despite some preliminary studies on shrubs (Hansen and Pate, 1987; van Kessel et al., 1988), little is known about the forms of N transported in the xylem of woody legumes, and the use of the procedure is restricted currently to those few species that have been already identified as ureide exporters (Table 111). Although preliminary investigations comparing a range of food legume species (soybean, green gram, black gram, cowpea, pigeon pea and common bean) indicate that differences in the relationship between relative ureide-N and N2 fixation are likely to be minor (Fig. 6), the ureide method

loo]

0

s!

5

E a,

a

0

A A

60

0

soybean BlackGrarn GreenGrarn

a @

cowpea DryBean Pigeon pea

20

40

1 0

A

0

60

80

100

P (“10) FIG. 6. Relationships between the relative abundance of ureide-N in xylem sap and P (determined by a I5N technique) for a range of tropical food legumes. From D. F. Herridge and M. B. Peoples (unpublished data).

172

MARK B. PEOPLES AND DAVID F. HERRIDGE

should not be used for interspecific comparisons without first calibrating the species under study. Apart from possible species differences in N transport, the effect of sampling procedures and physiological, environmental, and nutritional variables should also be identified before glasshouse-derivedrelationships can be applied legitimately to field-grown plants. Various factors that have been considered are summarized in Table IV. Of the potential sources of error listed, time delays between harvesting Table IV Potential Limitations in the Use of the Ureide Method to Evaluate Nz Fixation by Legumes in the Field

Variable/Comments Plant species: Principal N compounds transported from nodules are characteristic of a species. Method only valid for those legumes that export ureides. Method is an indirect measure-must establish relationship between N solute composition and N2-fixing status. Cultivar: Appears to be insignificant. Rhizobium strain: Conflicting reports on the effect of strain on relative ureide-N. Significance in estimating N2 fixation yet to be evaluated, but probably minor. Plant age: May require more than one calibration curve to cover all stages of growth. N stress and senescence: N solute relationships may be invalid under severe N stress or senescence (total xylem N solute concentrations of less than 1-2 pmol/ml) since ureides may also be synthesized from degradation products of nucleic acids. Source of soil N: Apparently no difference to relative ureide-N if legume takes up nitrate or ammonium. Sampling of sap: Volumes and solute concentrations vary on a diurnal basis, but relative ureide-N is constant between A.M. and P.M. If using vacuum to collect sap from stem, N solute composition unaffected by source or strength of vacuum. Branch-to-branch variation in relative xylem composition and effect of stem segment sampled appears small in tree legumes and field crops after flowering. However, there is a progressive change in relative ureide-N values if time delay between branch or stem detachment and vacuum extraction is much longer than 5 min. Storage of xylem samples: Sap composition stable for only 4 hr at 25-30°C. Stability improved if diluted 1 : 1 in absolute ethanol. Freeze for long-term storage. Prepared from Neves er al. (1985); Herridge era/. (1988); Peoples er a/. (1989b); Herridge and Peoples (1990), and includes unpublished data of D. F. Herridge, D. P. Nurhayati, and M. B. Peoples.

NITROGEN FIXATION BY TROPICAL LEGUMES

173

the shoot and vacuum-extracting xylem sap from the stem appear to be the most serious; however, such time-related changes in N solute composition can be avoided by collecting sap within 5 min of stem detachment (Herridge et al., 1988; Peoples et al., 1989b). The only other situation in which ureide data could be misleading is during severe N stress or senescence. Ureides under these conditions might be derived from purines released as degradation products of nucleic acids, or be mobilized from storage pools. Therefore, high levels of relative ureides in xylem would not necessarily reflect a high reliance on N2 fixation. Periods of N redistribution that might induce errors can be identified by monitoring plant N accumulation and nodulation and will be characterized by very low concentrations of total N (i.e., 1-2 mM) in the xylem stream. The major disadvantage of the ureide method is that it provides only a short-term measure of symbiotic dependence. If estimates of seasonal N2 fixation are required rather than a comparative measure of treatment effects on fixation activity, repeated measurements of xylem N solutes must be combined with sequential sampling for crop dry matter and N. Only then can total inputs of N be determined and partitioned between symbiotic and soil-derived N according to the estimates of P (Herridge et al., 1990). Such protocols can result in very close agreement between the ureide method and long-term, time-integrated field estimates using 'N techniques (Fig. 7) (Rerkasem e f al., 1988).

70 . o Site 1 60Site 2 -0 0

5

-

B

50-

0

Site 3 Site 4 Site 5

s

400

0

0

0 0

II X

0

1 0 20 30 40 50 6 0 70

P p"),natural15Nabundance

FIG.7. P estimated for field-grown soybean using I5N natural abundance and ureide methods. After Herridge er al. (1990).

174

MARK B. PEOPLES AND DAVID F. HERRIDGE

D. NBALANCE The N balance technique requires the measurement of changes in soil N and the estimation of all possible external N inputs (e.g., in rainwater, dust, animal droppings, by ammonia absorption, or through weathering) and outputs (denitrification, volatilization of nitrous oxides, ammonia, etc., leaching, erosion, and removal of crop or animal products within a given soil-plant system; see Greenland, 1977; Knowles, 1980; Wetselaar and Ganry, 1982). Any net increase in N under a legume is attributed to N2 fixation. Because the method relies on many independent and unrelated measurements, each made with differing accuracy, the confidence with which N2 fixation is calculated may be low. In particular, it is difficult to determine changes in soil N with precision, since any incremental change is likely to be small relative to the total amount of N present in the soil. Large inputs of fixed N2 therefore are necessary before any significant increase can be detected. For this reason, experimentation over several years is required (e.g., Firth et al., 1973; Wetselaar et al., 1973). E. ACETYLENE REDUCTION ASSAY

The acetylene reduction assay (ARA) arose from observations that the N2-fixing enzyme complex, nitrogenase, catalyzes the reduction of acetylene (C2H2) to ethylene (C2H4). The standard ARA method involves enclosing detached nodules or nodulated root systems in airtight containers and exposing them to an atmosphere containing C2H2. After an incubation period, gas samples are collected and analyzed for ethylene by using gas chromatography (Turner and Gibson, 1980; Upchurch, 1987). There has been widespread application of the ARA in many areas of N2 fixation research because of its rapidity, simplicity, high sensitivity, and low equipment and resource costs. A major difficulty in calculating absolute rates of N2 fixation from ARA data is in the precise relationship between moles of C2H2 reduced and moles of N2 fixed. The theoretical C2H2/N2conversion factor of 3 : 1 has commonly been used (six electrons are required for the reduction of N2 to NH3; only two electrons are used to reduce C2H2 to C2H4). However, actual measurements of the conversion ratio have rangedfrom 1.5 to 8.4 : 1 for pasture and food legumes and 1.6 to 4.8 : 1 for shrub legumes (Hardy et al., 1973; van Kessel et af., 1983; Hansen et af., 1987). The ARA does not relate directly to N2 fixation, but instead measures electron flux through nitrogenase. Since C2H2competitively inhibits H2 evolution, electron allocation for both N2 reduction and H2 production will therefore be measured

NITROGEN FIXATION BY TROPICAL LEGUMES

I75

by ARA (Upchurch, 1987). As the ratio of H2 produced to N2 fixed can be greater than 1 : 1 in different symbioses and may change on a diurnal basis or vary with plant age even the simple ranking of Rhizobiurn strains based on ARA can be inappropriate (Witty and Minchin, 1988; Sellstedt et af., 1989). A fundamental assumption of the ARA is that the rate of nitrogenase activity is not affected by the substitution of C2H2for N2. This assumption may not always be valid. In many legume species there can be substantial declines in nitrogenase activity after exposure to C2H2. In these cases, considerable errors can result if standard ARA procedures are used (Minchin et af.,1983; Davey and Simpson, 1988). Furthermore, the extent of the CzH2-induced decline can be influenced by plant and nodule age, plant stress, nitrate treatments, and changes in O2partial pressure, making quantitative comparisons of N2 fixation based on long-term C2H2 accumulation misleading (Carroll et af.,1987; Witty and Minchin, 1988; Davey and Simpson, 1989). The ARA provides only an instantaneous measure of nitrogenase activity under the prevailing assay conditions; its accuracy, therefore, has always been restricted by the requirement for repeated assays to adjust for marked diurnal and seasonal fluctuations. Further major errors can arise from incomplete recovery of total nodule population, nodule detachment or damage prior to assay, or declines in nitrogenase activity induced by removal of the shoot and washing or shaking of root systems to remove soil (Minchin et al., 1986; Witty and Minchin, 1988). Although in situ procedures with flow-through gas systems have been devised to overcome some of these technical problems (e.g., Denison et af., 1983; McNeill et af., 1989), the inherent limitations of the ARA just described must continue to make quantitative interpretation of field data difficult. When ARA has been compared with other measurement techniques under field conditions, ARA greatly underestimated N2-fixingactivity (e.g., Boddey et af.,1984; Kumar Rao and Dart, 1987).

F. N FERTILIZER EQUIVALENCE The N fertilizer equivalence technique assesses the amount of N2 fixed by a legume by: 1. Growing N-fertilized non-N2-fixingplants in plots alongside the unfertilized N2-fixing test legume. The N fertilizer level at which the yields of the nonfixing plants match those of the legume is equivalent to the amount of N2 fixed. The value obtained is usually expressed as fertilizer N equiva-

176

MARK B. PEOPLES AND DAVID F. HERRIDGE

lence. In the example presented in Table I, N fertilizer had to be applied at a rate of 160 kg N/ha before yields of either of the nonfixing crops approached that of nodulated soybean. 2. Comparing production of a nonfixing crop, usually a cereal, in land sown previously to either a legume or nonfixing species. The benefit may be expressed as a yield increase attributed to the legume in the absence of fertilizer N, or as the amount of fertilizer N required to boost yields in plots where no legumes had been grown to match the yields attained following a legume. This procedure has most commonly been used to estimate the value of a leguminous green manure to a following crop (Ladha et al., 1988). Best estimates will be given when all legume or nonlegume material is incorporated prior to planting the following nonfixing crop. Neither of the two alternative approaches provides a measure of N2 fixation. The result achieved with the second procedure will depend on the rate of decomposition of incorporated residues and release of plantavailable N and is liable to include any “nitrate-sparing” by the previous legume crop in addition to N contributed by N2 fixation (Herridge and Bergersen, 1988). Furthermore, differences in N fertilizer use efficiency and in losses of fertilizer N through volatilization, denitrification, and leaching that occur in the tropics (e.g., Sisworo et al., 1990) make direct comparisons of experimental data between different studies difficult. Nonetheless, the technique can provide a site-related estimate of the potential economic value of a legume in a rotation in terms of increased production and/or savings of fertilizer N. G. NODULEEVALUATION Since N2 fixation is dependent on the formation and maintenance of nodules, the degree of nodulation (as determined by nodule number, weight, or size, or by subjective rating for color and distributionon the root system; see Gueye and Bordeleau, 1988; Peoples et al., 1989a) has been used as a measure of symbiotic activity. Nodule evaluation is quick, convenient, and inexpensive. The extent of nodulation is correlated with NZ fixation in some field experiments (e.g., Bergersen et al., 1989), however, other studies indicate that even varietal comparisons of symbiotic capacity based on nodule weight can be unreliable (Duque et al., 1985). Nodule assessments can at best provide an indirect indication of a legume’s potential to fix N2 and cannot be used to quantify the amount of N2 fixed. Nonetheless, it is often a valuable adjunct to other measurements and can assist in data evaluation and interpretation since poor nodulation is often the basis of poor N2 fixation.

NITROGEN FIXATION BY TROPICAL LEGUMES

177

In the following sections we review the levels of N2 fixation achieved by different legumes in the tropics and subtropics and discuss constraints restricting fixation in the various production systems. The tabulated lists of estimates of NZ fixation are not exhaustive, but examples have been chosen to cover the range of quantities and proportions of N2 fixed. We have restricted citations to reports using the ureide method, N-difference, or ‘’N techniques and have only used ARA determinations when other estimates are not available. We have not considered reports where growth and N2 fixation in pots were extrapolated to an area basis, or where soil was amended with high C : N material to enhance fixation.

Ill. N2 FIXATION IN LEGUME PRODUCTION SYSTEMS Legumes are widely used for food, fodder, shade, fuel, timber, and green manure and as cover crops. The different agricultural systems in which they feature include (Yaacob et al., 1980): 1. Plantation systems: where legume cover crops, food crops, or shade trees are grown in the interrow space of plantation tree crops such as cocoa, tea, coffee, rubber, and oil palm. 2 . Tillage systems: where legumes are grown in rotation or intercropped with cereals. 3. Alternate tillage systems: principally represented by shifting cultivation and natural bush fallow. 4.Grazing systems: includes extensive grazing of natural vegetation in semiarid regions, intensive pastoral type of agriculture, and “cut-andcarry” systems.

Legumes will be discussed in the following sections on the basis of whether they are used as human or animal food or as a green manure. Both annual herbaceous and woody perennial legumes are included in the latter, although tree and shrub legumes are dealt with in detail elsewhere (see the contribution by Blair et al. in this volume).

A. FOODLEGUMES The two major constituents of human nutrition are calories and protein. Globally, plant sources contribute about 70% and animals 30% of human protein needs. In developing countries in the tropics and subtropics, where protein shortages are usually more prominent, plant sources provide ap-

178

MARK B. PEOPLES AND DAVID F. HERRIDGE

proximately 90% of the calories and up to 90% of human dietary protein. Of plant proteins, cereals contribute about 68%, legume seeds 19%, and roots, tubers, nuts, fruits, and vegetables (including many legumes) the remaining 13%. There is, however, wide variation in the relative consumption of the various legume sources by humans according to the spectrum of crops grown and the importance of animals in the rural economy (Rachie and Roberts, 1974; Wijeratne and Nelson, 1987). Many different food legumes are utilized as important components of lowland and upland cropping systems (Table V). Some species are grown

Table V Important Food Legumes Produced in the Tropics and Subtropics Genus Arachis Cajanus Canavalia Cicer Cyamopsis Glycine Kerstingiella

Species hypogaea cajan ensiformis gladiata arietinum tetragonoloba max geocarpa

Lablab Lens Macrotyloma Mucuna Pachyrhizus Phaseolus

purpureus culinaris uniforum pruriens erosus acutifolius lunatus vulgaris

Psophocarpus Vigna

tetragonolobus aconitifolia angularis mungo radiata urnbellata unguiculata subterranea

Voandzeia

Common name Groundnut Pigeon pea Jack bean Sword bean Chickpea Cluster bean Soybean Kersting’s groundnut Lablab bean Lentil Horse gram Velvet bean Yam bean Tepary bean Lima bean Common bean, kidney bean Winged bean Moth bean Adzuki bean Black gram Green gram Rice bean Cowpea Bambara groundnut

Collated from Rachie and Roberts (1974); Wood and Myers (1987).

Table VI Estimates of the Proportion (P)and Amount of Plant N Derived from Symbiotic N2Fixation for Commonly Grown Food Legumes in Tropical and Subtropical Systems Nz fixed

-

Species

Location

Groundnut

Australia Brazil Ghana India India Brazil Indonesia Nigeria Thailand

Pigeon pea Soybean

-4

W

Common bean Cowpea Bambarra groundnut

Brazil Kenya Brazil Indonesia Kenya Ghana

Treatment variable

Total crop N (kg N/ha)

Water supply Inoculation Cultivar Cultivar Season Site/season Rotation Inoculation Cultivar Tillage Inoculation Water supply Cultivar Phosphorus Site/season Rotation Phosphorus -

171-275 147- 163 126- 165 17-92 112-206 79-100 121-643 22 1-234 50-188 157-25 1 180 18-71 128-183 25-69 67-100 92-94

-

P 0.22-0.49 0.47-0.78

-

0.86-0.92 0.88 0.70-0.80 0.33 0.14-0.70 0.5 1-0.62 0.54-0.78 0-0.45 0.66 0.16-0.71 0.16-0.32 0.32-0.70 0.12-0.33 0.26-0.35 -

Amount (kg N/ha/crop) 37-133 68-116 32-134 109- 152 68-88 85-154 26-33 15-125 17-450 91-116 27-147 0-113 119 3-32 17-57 9-5 I 12-22 24-39 40-62

Method "N I5N N-diff. I5N "N I5N "N I5N 15N I5N I5N I5N "N I5N I5N "N I5N N-diff.

Reference" 1

2 3 4 5 2 6 7 8 9 10 11

12 13 2 6 14 15

L1 1. M. Bell and M. B. Peoples (unpublished data); 2. Boddey et al. (1990); 3. Dakora (1985b); 4. Giller er al. (1987); 5. Kumar Rao er al. (1987); 6. Sisworo e t a / . (1990); 7. Eaglesham et al. (1982); 8. Rennie et al. (1988); 9. Kucey er al. (1988b); 10. Kucey et al. (1988a); 11. Suwanarit e r a / . (1986); 12. Duque et al. (1985); 13. Ssali and Keya (1986); 14. Ssali and Keya (1984a); 15. Dakora (1985a).

Table MI Effect of Soil Mineral N and N Fertilizers on Crop N Productivity and the Proportion (P)and Amount of Crop Derived from N2 Fivation

Level

Species Groundnut

Location India

Soil mineral N (kg Nlha) -

Fertilizer N (kg Nlha) 0 100

Senegal

-

Chickpea

Australia

Soybean

Australia

10 (to 120 cm) 326 70 (to 120 cm) 260

India

-

200 15 60

06 100

-

0' 100

Nz fixed Total crop N (kg Nlha)

P

196 210 243 -

0.61 0.47 0.42 -

114 1 94 230 265 63 108 89 115

0.85 0.17 0.34 0.06 0.29 0.26 0.48 0.24

Amount (kg Nlhalcrop) 120 99 102 52 25 97 33 78 16 18 28 43 28

Method

Reference"

"N

1

"N

2

I5N

3

Ureide

4

I5N

1

-

Malaysia Day 45

Common bean Cowpea

z

+ further

Zambia

-

Kenya

-

Kenya

-

India

-

40 at sowing 20 as nitrate or 20 as urea 20 100 10 100 20 100 0 100 200

169 239 200 182-206 148-236 149 158 116 137 163 138 172

0.68 0.33 0.15 0.19-0.70 0.20-0.43 0.39

0.10 0.53 0.08 0.77 0.67 0.33

115 79 30 35-144d 3-101 58 16 62 11 125 92 57

I5N

5

"N

6

15N

7

15N

8

"N

1

1. Yoneyama et al. (1990); 2. Gibson er a/. (1982); 3. J. A. Doughton (unpublished data); 4. Herridge et a/. (1990); 5 . Norhayati et a / . (1988); 6. Munyinda et af. (1988); 7. Ssali and Keya (1986); 8. Ssali and Keya (1984b). Uninoculated. Inoculated. Comparison of cultivars.

182

MARK B. PEOPLES AND DAVID F. HERRIDGE

widely as oilseeds (groundnut and soybean) or pulses (pigeon pea, chickpea, cowpea, green and black gram, common bean), but often crops have only localized use (e.g., lablab bean, rice bean, bambarra groundnut, Kersting’s groundnut). The distribution of crop legumes is generally determined by their adaptation to particular climates and environments (Rachie and Roberts, 1974; Wood and Myers, 1987). Food legumes are grown commonly in areas where cereals such as paddy and upland rice, wheat, and sorghum are the major crops. They are grown primarily during the wet season, but can be grown in the dry season when residual soil moisture or irrigation water is available. Short-duration crops such as green gram may be grown opportunistically prior to or immediately after a cereal crop. Estimates of NZ fixation by the food legumes in various regions throughout the tropics and subtropics are presented in Tables VI, VII, and XIV. The range of estimates is large and appears to be unaffected by crop species or by country of location. For individual species, ranges of P were 0.22-0.92 for groundnut, 0.10-0.88 for pigeon pea, 0.17-0.85 for chickpea, 0-0.95 for soybean, 0.10-0.71 for common bean, and 0.08-0.89 for cowpea. These values resulted in amounts of N2 fixed ranging from 0 to 450 kg N/ha. The levels of fixation depend on water supply, inoculation, crop management practices, including applications of fertilizer N, and soil N fertility. There is a strong inverse relationship between the level of plant-available soil N (either present in soil or supplied as N fertilizer) and P (Tables VII and XIV). In almost all cases, P was reduced in the presence of higher N, resulting in generally lower amounts of N2 fixed. The exception occurred when an increase in N yield with higher soil N was relatively greater than the concomitant reduction in P (see data derived from Yoneyama et al., 1990). The negative relationship between plant-available soil N and P is arguably the single most important limitation to Nz fixation. Food legumes are not only grown in pure stands, but often they are interplanted with other species. Mixed cropping is practiced traditionally in many parts of Africa, Asia, and Latin America. Growing two or more species simultaneously in the same field intensifies crop production and more effectively exploits the environment. Combinations of crops are determined by the length of growing season and environmental adaptation, but usually early- and late-maturing crops are combined to ensure efficient utilization of the entire growing season. In tropical and subtropical regions, legumes such as cowpea, groundnut, soybean, chick-pea, common bean, pigeon pea, and rice bean are usually intercropped with maize (Zea mays), sorghum (Sorghum bicolor), millet (Pennisetum and Panicum spp.), or rice (Oryza satiua) (Ofori and Stem, 1987). The quantity of N2 fixed by the legume (Table VIII) in an intercrop depends on the species, morphology, legume density in the intercrop mixture, and crop management. Differences in the competitive abilities of the component crops for soil N

Table VIII

Effect of Intercropping Food Legumes with Cereals on the Proportion (P)and Amount of Crop N Derived from NtFmation Nz fixed Species

Location

Rice bean

Thailand

Intercrop ratio (legume :cereal) 1OO:O 75 : 25 50 :50

Level N fertilizer (kg N/ha) 0

8 Cowpea

25 :75 Australia Hawaii

100:o 71 : 29

P

Amount (kg N/ha/crop)

129 67 48 83 134 178 69 60 125 89 60-89

0.36 0.49 0.63 0.84 0.86 0.78 0.72 0.32 0.90 0.51 0.69 0.66 0.30-0.47

49b 68' 8Ib 56' 416 64' 97 57 62 32 87 59 18-42

28

0.34-0.44

10-12

119 83

0.54 0.58

150 136 122 103

0.79 0.81 0.59 0.64

135 144

25 :75

100:O

Total crop N (kg N/ha)

20 200 20 200 0

1oo:o

Method

Reference"

ISN/ureide

1

ISN

2

15N

3

ISN

4

64 48

"N

5

118 110 72 66

''N

6

(row spacing 40-120 cm) 75 :25 (row spacing 40-60 cm) India

100:O

66:33 Nigeria

1OO:O

62 :38

0 60 (to cereal rows) 25 100 25 100

1. Rerkasem er al. (1988); 2. Rerkasem and Rerkasem (1988); 3. Ofori ef a/. (1987); 4. van Kessel and Roskoski (1988); 5. Patra er a / . (1986); 6. Eaglesham et al. (1981). At time of harvest of companion maize 91 days after sowing as determined by "N. At final legume harvest at 147 days as derived from ureide sap analysis.

184

MARK B. PEOPLES AND DAVID F. HERRIDGE

can result in a stimulation of NZfixation (Table VIII) and ultimately lead to an increased N yield in the intercrop relative to the legume and cereal monocrops (e.g., Rerkasem et al., 1988; Rerkasem and Rerkasem, 1988). Legumes of indeterminate growth with a climbing habit are generally most successful (e.g., rice bean, see Table VIII). Graham and Rosas (1978) showed that NZfixation by climbing types of common bean was unaffected by intercropping with maize. However, with shorter stature legumes such as soybean and groundnut, shading by tall cereal crops can reduce both yield and N2 fixation (Wahua and Miller, 1978; Nambiar et al., 1983).

B. FORAGES,MULTIPURPOSE TREELEGUMES, COVER CROPS, A N D GREENMANURES The total productivity of the agricultural sectors of tropical and subtropical regions depends not only on food crops but also on livestock production. In Asia, for example, over 60% of food is produced by small landholders who keep livestock in mixed cropAivestock farming systems (Moong, 1986). Fodder for animals in such systems includes crop residues, weeds, tree leaves and planted forage crops (Moong, 1986; Little et al., 1989). Farmers prefer to feed their animals at minimum or no cost except labor, often using them to utilize weeds and crop by-products. The impact of inadequate nutrition on ruminant livestock under these conditions can be considerable (Bayer and Waters-Bayer, 1989). Deficiencies in protein contents of forages used in the tropics have been known to affect the rate of growth and maturation of ruminants (Harricharan et al., 1988). As a result, a diverse range of herbaceous forage (Table IX) and tree legumes (Table X) are being utilized increasingly in livestock production. Goodquality forage legumes can be either grazed or harvested and hand-fed as a high-protein supplement to grasses. Alternatively, legume leaves may be ground and rolled and the juices steam-coagulated into a protein concentrate suitable for monogastric animals (poultry and pigs); the fibrous residues are used as leaf cake for ruminants (Mustapha and Djafar, 1980). The woody perennial legumes (Table X) have a number of purposes. As well as providing animal fodder, they can be used to reclaim degraded wastelands, retard erosion, and provide shade, fuelwood, mulch, and green manure (Duhoux and Dommergues, 1985; Brewbaker, 1986; Brewbaker and Glover, 1988; Nair, 1988). The potential for the integrated use of these species is well illustrated in alley farming systems. Alley cropping involves growing arable crops in the interspaces (alleys) between rows of planted trees or woody shrubs that are periodically trimmed and maintained as hedgerows. The leaf and twig prunings are added to the soil

NITROGEN FIXATION BY TROPICAL LEGUMES Table IX Perennial and Annual Herbaceous Legumes Commonly Used for Forage in Tropical and Subtropical Agriculture" Genus Aeschynomene Alysicarpus Arachis Calopogonium Canavalia Centrosema Cicer Clitoria Crotalaria Desmodium Dolichos Dunbaria Eriosema Galactia Hegnera Indigofera Lablab Lotonis Macroptilium Macrotyloma Mucuna Neonotonia Psophocarpus Pueraria Rhynchosia Stilozobium Stylosanthes Tadehagi Teramnus Trifolium

Species falcata glumaceuslvaginalis glabratalhypogaealpintoi caeruleumlmucunoides ensiformis brasilianumlpascuoruml pubescenslvirginianum arietinum ternatea juncea heterocarpon I intortuml ovalifoliumluncinatum

striata obcordata hirsuta purpureus bainesii atropurpureum africanumlaxillarelunijlorum

cochinchinesislpruriens wightii palustris phaseoloides deeringianum capitatalguianensislhamatal

humilislscabra triquetrum labialisluncinatus africanumlrueppellianuml

semipilosumlsteudneril rembense Vigna

luteolalparkeriltrilobatalung-

Zornia

uiculatalvexillata dip hylla /laiifolia

,,Collated from Hutton (1970); Haque and Jutzi (1984); Blair er al. (1986); Gramshaw et al. (1989); Kretschmer (1989);Little et al: (1989).

185

186

MARK B. PEOPLES AND DAVID F. HERRIDGE Table X Important N2-Fiuing Shrub and Tree Legumes used in Tropical and Subtropical Agriculture" ~~~~

Genus Acacia

Aeschynomene Albizia Cajanus Calliandra Chamaecytisus Codariocalyx Desmanthus Desmodium Erythrina Flemingia Gliricidia Inga Leucaena Neptunia Prosopis Pterocarpus Robinia Sesbania Tephrosia

Species albidalauriculaeformisl holosericealmangiuml mearnsii afrasperalamericanal niloticalsca bra chinensislfalcatariallebbek cajan calothyrsus palmensis gyranslgyroides virgatus discolorldistortumlrensonii

macrophylla sepium jinicuil leucocephala oleracea albalcinerariaIjuliJloral pallida erinaceuslmarsupium pseudoacacia grandijloralpunctatal rostratalsesban candida

Collated from Duhoux and Dommergues (1985); Brewbaker (1986); Brewbaker and Glover (1988); Ladha et al. (1988); Nair (1988).

as green manure or mulch (Wilson et al., 1986; Nair, 1988). In the humid tropics, hedgerows are often grown along contours on slopes to decrease water runoff velocity and reduce sediment transport capacity. Sediments trapped by the contour hedges facilitate the formation of natural terraces (Lal, 1983). Cultivation of rubber (Hevea brasiliensis) and oil palm (Elaeis guineensis) represents important economic activities in many countries in the tropics. Trees are planted in rows 7 to 10 m apart (depending on density of planting and terrain), with 2 to 2.5 m between trees. During the first 5- to 6-year immaturity period after planting, the interrow space (70-80% of the land area), if left bare, may be subjected to invasion by undesirable weeds

NITROGEN FIXATION BY TROPICAL LEGUMES

187

that are costly to control and that harbor diseases and pests that can infest the developing saplings. Establishment of twining, perennial legume cover crops such as Calopogonium caeruleum, C. muconoides, Centrosema pubescens, Pueraria javanica, and P . phaseoloides in the interrow space is a common practice to restrict weed growth, control leaching and erosion in steep terrain, and return organic matter and N to the soil. Not only are the plantation crops healthier and the economic costs associated with weed control minimized, but the trees also mature quicker (Ismail et al., 1980; Yeoh and Phang, 1980; Agamuthu and Broughton, 1985). The use of green manure crops is not restricted to long-term alley and plantation systems. Green manures are also widely used in short-term annual rotation sequences with crops such as rice. Green manures in these instances may involve both annual food legumes and perennial species listed in Tables IX and X (Kulkarni and Pandey, 1988; Nair, 1988). Characteristics of ideal leguminous green manure crops for lowland rice farming systems have been summarized by Ladha et al. (1988). Important traits include: ( a )early establishment and high seedling vigor; ( b )early onset of N2 fixation and its efficient sustenance; ( c ) fast growth and ability to accumulate large biomass and N in 5-7 weeks; (d)ease of incorporation into soil; and ( e ) rapid decomposition and release of plant-available N. Because of the requirement for tolerance to flooding, there has been recent interest in Sesbania rostrata and species of Aeschynomene and Neptunia, all of which have the ability to produce nodules on both roots and stems and can grow and fix N2 under flooded and dry conditions (Ladha et al., 1988; Rinaudo et al., 1988). Despite the large number of genera and species used for forage and green manure in the tropics and subtropics (Tables IX and X; see also Kulkarni and Pandey, 1988), there are relatively few estimates of N2 fixation by the herbaceous legumes (Table XI) in these important production systems. Assessments of N2 fixation by woody legumes (Table XII) are even more restricted because of methodological problems associated with measurement. Nonetheless, the values obtained are similar to those presented for the food legumes (ranges of 2-183 and 3-231 kg Nlha for the herbaceous forage and shrub and tree legumes; see Tables XI and XII, respectively). Estimates of P tend to be higher for the herbaceous species than for food legumes. In the majority of studies reported, the P values indicated that between 75% and 100% of the legume N was derived from N2 fixation (cf. Table VI). It is virtually impossible to draw further conclusions from the data summarized in Tables XI and XI1 because of the few studies involved, the use of ARA in some instances, and the short periods of assessment, particularly with the woody species. We can speculate, however, that the levels of fixation are probably in the same order of magnitude as the food legumes and, like the food legumes, mediated through N yield and P.

Table XI Estimates of the Proportion (P)and Amount of Legume N Derived from NzFixation for Cover Crops and Herbaceous Forage Legumes in Mixed Grass Swards N2 fixed Species

Legume N yield (kg Nlha)

P

Amount (kg Nlha)

Period of measurement

Method

Reference"

Malaysia

299

0.50

150

Annual estimate

N-diff.

1

Ghana Australia

25-19Sb

0.94

24-183b

1 season Annual ave. over 2 years

N-diff. "N

2 3

Location ~ _ _ _ _ _ _

Centrosema pubescens Pueraria phaseoloides Crotalaria juncea Desmodium intortum Lotononis bainesii Macroptilium atropurpureum Stylosanthes guyanensis Trifolium repens

~

88

(0.8-1.0 range) 16-135b

0.92 (0.75-0.97 range)

15-246

Vigna luieola Lablab purpureus Macroptilium atropurpureum Pueraria phaseoloides

Sesbania cannabina Stylosanihes spp. (I0 accessions)

Ghana India Malaysia (% legume in sward) 0-40 41-60 61-80 81-100 Australia Australia

76-95 28-48

0.78-0.83 0.84-0.87

37-56 59-79 23-39

1 season year 1 year2

N-diff. "N

2 4

12

0.75 0.92 0.86 0.76 0.70-0.93

9

3 months

"N

5

24 44 30 136-202

22 38 23 126-141

Seasonal ave.

"N

6

2-86

0.83-0.92

2-75

ISN

7

2-34 3-102

0.60-0.88 0.73-0.90

2-20 3-84

year I (9-1 1 weeks regrowth) -year 4 year6

a 1. Agamuthu and Broughton (1985);2. Dakora (1985a); 3. Vallis et al. (1977);4. Shivaram et al. (1988);5 . Zaharah et al. (1986a); 6. Chapman and Myers (1987); 7. Vallis and Gardener (1985). Estimated from mean legume uptake of soil N and P value.

190

MARK B. PEOPLES AND DAVID F. HERRIDGE

Table XI1 Estimates of the Proportion (P)and Amount of Plant N Derived from N2 Fixation for Multipurpose Tree and Shrub Legumes Nz fixed Species Acacia holosericea Acacia pennatula Gliricidia sepium Inga jinicuil

Location

Legume N yield (kg N/ha)

P

Australia Senegal Mexico

Amount Period of (kg N/ha) measurement Method Reference" 12

10-20 -

0.30 -

Mexico Mexico

Leucaena Malaysia leucocephala Nigeria Tanzania

296-3 13 288-344

0.58-0.78 0.34-0.39

Sesbania

Senegal

214-230

0.36-0.51

Senegal

54-100

0.13-0.18

Annual estimate 3-6 6.5 months 34 Annual estimate 13 Annual estimate Annual 23-45 estimate 182-23 I 3 months 98- I34 6 months Annual I10 estimate 83-109 2 months

ARA

1

I5N ARA

2 3

ARA

3

ARA

4

ISN "N ARA

5 6 7

'SN

8

ISN

8

rostrata

Sesbania sesban

7-18

2 months

" I . Langkamp er a / . (1979); 2. Cornet et al. (1985); 3. Roskoski et a / . (1982); 4. Roskoski and var Kessel(1985); 5 . Zaharah e t a / . (1986b);6. Sanginga et a / . (1989); 7. Hogberg and Kvarnstrom (1982);8. Ndoye and Dreyfus (1988).

These factors are likely to be influenced in turn by plant genotype, its adaptation to the particular environmental conditions, and plant and soil management.

IV. CONTRIBUTION OF LEGUME N TO PLANT AND ANIMAL PRODUCTION A. DIRECTTRANSFER It is often assumed that a portion of the N2 fixed by an intercropped legume is made available to the associated nonlegume during the growing season, or that direct transfer of N from forage legumes to companion

NITROGEN FIXATION BY TROPICAL LEGUMES

191

grasses occurs in mixed pasture swards (Whitney, 1982; Ofori and Stern, 1987; Kang, 1988). Decaying roots and nodules are thought to be important in this transfer of N, although these organs generally contain only a small fraction of the total plant N, for example, 3-40 kg N/ha may be present in the nodulated roots of field-grown legumes (Chapman and Myers, 1987; Kumar Rao and Dart, 1987; Bergersen et al., 1989). The propotion of the total root system that might be decomposing, and thereby releasing mineral N to the soil during growth, has not been estimated. Only potential decomposition of nodule and root material and rate of release of mineral N have been inv'estigated in laboratory and glasshouse studies (Table XIII). Whether mineralization or immobilization processes dominate depends on the N content of the root system, which in turn will be influenced by the extent of nodulation (nodules contain 4-7% N ,whereas unnodulated roots generally contain 1-2.5% N over a wide range of tropical legumes; M. B. Peoples, unpublished). Values above 1.8-2.0% N apparently favor mineralization, and below 1.5% N immobilization is favored, but other factors such as root lignin content and tissue C : N ratio are also likely to be important (Table XIII) (Nnadi and Haque, 1988). Regardless of the controlling factors, the rate of release of plantavailable N from roots and nodules during decomposition may be too slow to directly benefit the cereal or grass component during a single growing season. However, the possibility of exudation of N from living roots should not be ignored (Poth et af., 1986). Evidence of N transfer from legume to cereal has been obtained in intercropping studies (Eaglesham et al., 1981; Bandyopadhyay and De, 1986; Patra et al., 1986), although not confirmed in other investigations (Kumar Rao et al., 1987; Ofori et af., 1987; Rerkasem and Rerkasem, 1988; van Kessel and Roskoski, 1988). Similarly, there is little evidence of N transfer in mixed pasture swards (Vallis et al., 1977). This suggests that direct transfer of N from legume to nonlegume may not be a rapid or common phenomenon. RESIDUES TO CROP PRODUCTION B. CONTRIBUTION OF LEGUME 1 . Food Legume Crops

The total amount of N in a legume crop (NI) comes either from N2 fixation (Nf) or from uptake of mineral N from the soil. In food legumes, total N is partitioned into either the harvested seed (Nls) (proportion partitioned into seed described by the harvest index for N, NHI) or the vegetative parts (leaves, stem, and nodulated roots), which generally remain as crop residues (Wood and Myers, 1987). The net potential contribu-

Table XI11 Decomposition of Tropical Legume Residues and Release of Plant-AvailableN

Tissue

Species

Nodules

Centrosema pubescens Centrosema pubescens Glycine max Vigna unguiculata Cajanus cajan Glycine max Lablab purpureus Phaseolus lunatus Trifolium steudneri Vigna unguiculata Voandzeia subterranea Stylosanthes guianesis Glycine max Vigna unguiculata Desmodium intortum Desmodium tortuosum Glycine max Macroptilium atropurpureum Stylosanthes guianesis Vigna unguiculata Desmodium intortum Macroptilium atropurpureum

Roots

Stem

Leaf

Litter

Tissue N content (% N)

Duration of decomposition (days)

4.7 (4.2Ib

Mineralization as a proportion of added legume N

(%I

Reference"

I12

+31

I

1.3 (16.5) 2.7 (7.0) 2.3 (18.1) 2.4 (14.1)

112 I12 I40 I40

+ I9 + 36 +23 +47

1.2 1.4 0.9-1.1 2.1

89 89 70 89

- 17' - 10 -50 + 19

3

1.9-2.0

70

-45

4

I .5 2.5 3.9

89 89 89

- 10

3

I .2

I68

-2

5

2.4 (18.8) 1.5 (30.8)

140 140

-5 - I7

2

4. I 5.0 4.4

I68 224 168

+7d +3d +30

5 6 5

5.9 (7.3) 4.4 4.5 2.6

140 168 224 168

+76 + 24 +46 +8

2 5 6 5

2.9 (15.9)

140

+36

2

2.8 1.9 3.4 2.1

I68 224 I68 224

+7d

5 6 5 6

+ 27

2

4 3

+49

- 10d + 16 +5

" 1. Chulan and Waid (1981); 2. Frankenberger and Abdelmagid (1985); 3. Nnadi and Balasubramanian (1978); 4. Nnadi and Haque (1988); 5 . Henzell and Vallis (1977); 6. Vallis and Jones (1973). Values in parentheses are tissue C : N ratios. ' Negative values indicate immobilization of N. Desmodium intortum contains high total polyphenols (4.5% versus 2.0-2.2% in Macroptilium atropurpureum) in both green leaves and abscised leaf litter.

'

NITROGEN FIXATION BY TROPICAL LEGUMES

193

tion of N2 fixation to the N balance of a soil following legume cropping may be considered as Net N balance = Nf-Nls

(8)

Nf = (P X NI)

(9)

where and

Nls = (NHI

X

N1)

Since leaf-fall during crop legume development and nodulated roots can each contain up to 40 kg N/ha (Sheldrake and Narayanan, 1979; Kumar Rao and Dart, 1987; Chapman and Myers, 1987; Bergersen et al., 1989), it is desirable that they are both included in NI to ensure an accurate estimate of the net N balance. When the quantities of N involved in plant growth, in N2 fixation, and in the seed are calculated for food legume crops, it is apparent that the net N balance is often low and in some instances negative (Table XIV). Levels of fixation achieved by many crops in the field may be high, but are not always sufficient to offset the N removed with the harvested seed. Clearly, if food legumes are to contribute substantial amounts of N to the soil, P must be considerably greater than NHI. High-productivity crops such as soybean would require P values greater than 80% to avoid a net loss of N from the system. The potential contribution, however, may be even further diminished if legume haulms and crop residues are used as fodder for animals. Despite the variable N balances summarized in Table XIV, reported benefits of tropical crop legumes to subsequent cereal crops are consistent and substantial and may persist for several seasons (MacColl, 1989), regardless of whether the legume was grown in monoculture or intercropped (Table XV). Improvements in cereal yield following monocropped legumes lie mainly in the 0.5 to 3 t/ha range, representing around 30% to 350% increases over yields in cereal-cereal cropping sequences. Responses to previously intercropped legumes are more modest and yield improvements of only 0.34 to 0.77 t/ha have been measured. When the contribution of the legume was quantified as fertilizer N equivalence, as much as 68 kg N/ha was required in the cereal-cereal sequence to achieve similar yield improvements (Table XV). If the benefits of crop legumes in rotations cannot be explained solely in terms of residual fixed N, then what are the sources of the benefits demonstrated in Table XV? Most certainly a number of factors can operate, the relative importance of each dictated by site, season, and crop sequence. Improvements in soil structure following legumes (Hearne, 19861,

Table XIV Net N Balances for Several Legume Crops Following Seed Harvest

Species

Seed yield" (kg N/ha)

Location

~~

Groundnut Pigeon pea Soybean

Ghana Thailand India Australia Cultivated fallow No-tillage fallow India Nigeria

Early Medium Late maturity

+25Nx + IOON

Total crop Nb (kg N/ha) ~~

54 116 39 49 28 262 I50 152 202 204 195

Nz fixed NHI'

P

0.42 0.47 0.50 0.41 0.21 0.80 0.61 0.58 0.87 0.80 0.80

0.79 0.61 0.10 0.46 0.51 0.95 0.77 0.88 0.55 0.89 0.65

Amountd (kg N/ha/crop)

Net N balance' (kg N/ha)

Method

Referencef

N-diff. "N N-diff.

1 2 3

I5N Ureide

4 5

N-diff. ISN

6 7

~

128 245 72 120 134 329 245 264 232 255 244

101

150 7 55 69 312 I80 232 128 227 159

+47

+34 -32 +6 +41

+50 +30 +80 -74 + 23 -36

Common bean Green gram Cowpea

Brazil Australia Thailand Australia Ghana Nigeria

+25Nw

+ lOON

12 89 32 80 65 48h 49 54h 49h

74 177 61 125 226 82 134 94 118

0.16 0.50 0.52 0.64 0.29 0.58 0.37 0.57 0.41

0.62 0.63 0.60 0.69 0.89 0.61 0.75 0.30 0.41

46 112 37 87 20 1 50 101 28 49

+ 34 +23 +5 +7 +I36 +2 +52 -26 0

"N I5N 'SN "N N-diff. 1SN

8 4 2 9 1

7

N removed in seed, Nls. Total crop N at seed harvest, NI. Nitrogen harvest index = Nls/NI. Amount of NZfixed, Nf = NI x P. Net contribution of legume residue N to soil = Nf - Nls. f 1. Dakora et a / . (1987); 2. Suwanarit et a / . (1986); 3. Kumar Rao and Dart (1987); 4. Chapman and Myers (1987); 5. Hughes and Herridge (1989); 6. Chandel e t a / . (1989); 7. Eaglesham e t a / . (1982);8. Ruschel et a / . (1982);9. Ofori et a / . (1987). Rate of N fertilizer application (kg N/ha). Determinatelindeterminate cultivar comparison. a

196

MARK B. PEOPLES AND DAVID F. HERRIDGE

Table XV Grain Yield Responses of Cereals to Previous Legume Crops Relative to a Cereal-Cereal Cropping Sequence

Crop sequence Legume-maize Pigeon pea Groundnut Soybean Green gram Groundnut Cowpea Groundnut Soybean Lablab bean Pigeon pea Legume-rice Soybean Green gram Legume-sorghum Black gram Green gram Legume-wheat Pigeon pea Black gram Green gram Cowpea Groundnut Green gram Cowpea

monocrop intercropb

monocrop intercropb monocrop intercropb monocrop intercropb

Increase in cereal yield (t/ha)

Relative increase

1.06 0.77 0.35 0.48 0.48 0.85 0.91 0.72 0.49 I .96 2.78

353 157 25 34 34 89 96 24 16 65 89

0.80 0.20

66 17

3.68 2.82

79 61

68 68

0.27 1.26 0.65 0.74 0.75 0.34 I .60 0.48 1 .M) 0.38

21 98 51 58 23 10 49 15 30 I1

28 12 68 16 38 13

~~

a

~

(%)

~~~

N fertilizer equivalence (kg N/ha) 40 20+

Reference”

I 2

60 60 9‘ 7” 33‘ 67”

3 4

8

~~

1. Kumar Rao er al. (1983); 2. Suwanarit et al. (1986); 3. Dakora er al. (1987); 4. MacColl

(1989); 5. Chapman and Myers (1987); 6. Doughton and MacKenzie (1984); 7. Singh and Verma (1985); 8. Bandyopadhyay and De (1986). Intercropped with sorghum or maize in previous growing season. Mean response over a 5-year period. Pigeon pea grown for 2 years before incorporation.

NITROGEN FIXATION BY TROPICAL LEGUMES

I97

breaking of cycles of cereal pests and diseases, and phytotoxic and allelopathic effects of different crop residues (Sanford and Hairstron, 1984) can result in extra yield from a rotation. But much of the unexpected benefit probably results from the use of a cereal-cereal rotation as the “bench mark” comparison for the performance of a legume-cereal sequence. Even when most of the N is removed in legume seed and there is little or no net gain of N by the soil (Table XIV), a much higher loss of N would be expected from a nonleguminous/nonfixing crop when its seed is harvested (Nnfs). This can be illustrated by inserting the data of Ofori et al. (1987) for monocrop cowpea and maize into Eq. (8): Cowpea: Nf = 87 kg N/ha; Nls = 80 kg N/ha Net N balance = 87 - 80 = + 7 kg N/ha Maize: Nf = 0; Nnfs = 57 kg N/ha Net N balance = 0 - 57 = -57 kg N/ha Whereas the real benefit of the cowpea is only 7 kg N/ha, the benefit to the following cereal, assessed by comparing yields of maize-cereal versus cowpea-cereal sequences would combine both the small contribution by the legume (+7) and the N loss through removal of maize cobs and seed (-57). Therefore the apparent benefit is 64 kg N/ha. At the time of seed harvest and incorporation of legume residues in the soil, the total N of the plant is unevenly distributed between the different vegetative organs. These organs also differ from one another in the extent to which they release mineral N to the soil and, ultimately, to following crops. Although N transformations occurring in decomposing crop residues are influenced by temperature, moisture, residue management, and soil characteristics (Bartholomew, 1965; Smith and Peterson, 1982) and may be modified by lignin and polyphenol content (Vallis and Jones, 1973), the absolute N content, tissue N concentration, and C : N ratio of the legume material are the primary factors influencing the rate of mineralization (Table XIII). The potential for legume leaves to contribute N to a subsequent crop can be considerable since they represent the single largest source of vegetative N remaining in the residue trash, and because their high N content and low C : N ratio favor mineralization (Table XIII). Recent field experiments with ”N-labeled soybean residues (F. J. Bergersen, G. L. Turner, R. R. Gault, M. B. Peoples, and J. Brockwell, unpublished data) indicated that some 34% of the shoot N of a cereal crop planted immediately following incorporation of soybean trash can be derived from mineralization of soybean leaf material (3.2% N). Soybean stems (1.6% N) contributed only 4% of the N in the cereal shoots, despite the fact that the amount of stem N in the trash was one-third of the N added as leaves. Roots of soybean contributed less than 1%. A notable feature of mineralization of N in plant residues is that after a short period of time the

198

MARK B. PEOPLES AND DAVID F. HERRIDGE

rate of mineralization is quite slow, regardless of the initial percentage of N composition of the residues (Henzell and Vallis, 1977). Therefore, N in crop residues that is not mineralized during the first season becomes available only very slowly thereafter (usually at less than 5-10% per year) for successive crops. Combining the factors that influence the contribution of legume N to cropping systems, it is possible to derive a conceptual model to relate the transfer of fixed N2 from a leguminous crop to a following crop (Myers and Wood, 1987): Contribution to a subsequent crop = N1 x P (1-NHI) X Fm

X

E (1 1)

where N1, P, and NHI are as defined previously, Fm is the proportion of legume N mineralized, and E is the efficiency of utilization of this mineral N. To maximize the contribution of legume N to a following crop, it is necessary to maximize N1, P, Fm, and E and minimize NHI. There is little control over the efficiency of use of the mineral N by a subsequent crop, but the other factors in the equation can be subjected to manipulation through different management strategies. Unfortunately it is not always possible to optimize these factors for food legumes grown in tropical and subtropical farming systems (Myers and Wood, 1987). To maximize N yield and, therefore, N2 fixation, the legume crop should be grown in the most favorable growing period, that is, the wet season. In most systems the wet season is reserved for cereal crops, particularly rice. As a result, the food legume is commonly grown under the less favorable conditions of the late wet or dry season, and maximum yields are not achieved. If the legume is grown for seed, it is desirable to maximize the harvest index, which in turn maximizes NHI. The higher the value of NHI, the lower the amount of residual N to be returned to the soil. Often the Fm will also be low because the quality of the residues is usually reduced by the mobilization and reallocation of N to reproductive parts during seed filling and senescence (Peoples and Dalling, 1988). When a green manure crop is grown, however, the entire crop is returned to the soil. The quantity of N and the N concentration in the legume material returned are likely to be higher than when seed is removed from a food legume. As a consequence, decomposition might also be expected to be rapid and Fm high (Table XIII). 2. Green Manures and Cover Crops

Leguminous green manures differ widely in N concentration and N yield. In a survey of 86 different species used in India as green manure for

199

NITROGEN FIXATION BY TROPICAL LEGUMES

rice, the N contents ranged from 2.0% to 4.9% (summarized by Ladha et a / . , 1988). Legume leaf material used in alley cropping in Africa ranged from 2.0% to 4.0% N for Tephrosia africana, Leucaena leucocephala, and Gliricidia sepium, 4.8% to 5.1% N for Sesbania rostrata and Psophocarpus palustris (Wilson et al., 1986). The input of N by these green manures can be considerable (equivalent to up to 532 kg N/ha; Table XVI), and rates of N accumulation can be exceptionally rapid (from 0.4 to 10.8 kg N/ha/day; Table XVI) (Ladha et al., 1988). Studies of residue decomposition in alley cropping systems suggest that 50% of the added legume N may be released within 1-9 weeks depending on initial N content and prevailing environmental conditions (Wilson et al., 1986). It is perhaps not surprising, therefore, to find substantial increases in yield of cereal crops following green manures (on the order of 0.6-4 t/ha; Table XVI), Table XVI Increase in Cereal Grain Yield by Green Manure Crops Relative to UnfertilizedFallow Duration of green manure Legume Legume used crop N input Cereal as green manure (days) (kg N/ha)

Increase in cereal yield (t/ha)

Relative increase (%)

Reference"

Rice

82-85

I

Maize

Aeschynomene SPP. Crotalaria juncea Cyamopsis tetragonoloba Glycine max Sesbania aculeata Sesbania cannabina Sesbania rostra Sesbania sesban Vigna unguiculata Leucaena leucocephala Gliricidia sepium

49

423-532

4.0-4.1

60

I10

2.4

75

2

60

87

2.I

66

2

60

196 108

0.6 2.3

46 72

3 2

94

1.o

77

3

52 84 60

267 92 I I3

3.8 2.0 2.5

181 I02 78

4 5 2

I68 -

300 42' 42' 40 80

I .9

I46 97 105 60

6 7 7 8 8

-

1.6 1.7 0.6 1.6

160

I . Alazard and Becker (1987); 2. Beri et al. (1989); 3. Chapman and Myers (1987); 4. Rinauda et a / . (1983); 5. Palm et al. (1988); 6. Sanginga et al. (1988); 7. Kang, 1988; 8. Haque and Jutzi (1984). N yield from hedgerow prunings during one season of alley-cropped maize.

'

200

MARK B. PEOPLES AND DAVID F. HERRIDGE

equivalent to 50-120 kg fertilizer N/ha (Ladha et al., 1988; Sanginga et al., 1988; Beri et al., 1989). Clearly green manures can provide a substantial contribution of N to the soil and greatly benefit subsequent crops. Nonetheless, it must be remembered that unless the green manure was derived from an alley cropping system, this N will generally have to be obtained at the expense of an alternative cash or food crop, and there will be a cost in terms of labor and materials required for the sowing, maintenance, harvesting, and incorporation of the leguminuous material. In the case of cover crops in plantation estates, however, the benefits more than offset the initial cost of legume establishment and maintenance. Nitrogen is one of the most important major nutrients for growth and yield of rubber and oil palm. The return of N and organic matter from legume covers to soil and reduction of leaching losses of N have inevitably led to better growth by the trees, earlier commencement of tapping, increased (up to 20%) rubber yield, and a stimulation of the growth and yield of fresh fruit bunches by oil palm. The contribution of organic N in leaf litter from perennial legume covers during the 5-year immaturity period of rubber trees has been measured to be 250-400 kg N/ha (Yeoh and Phang, 1980), and 120 kg N/ha/year in oil palm (Agamuthu and Broughton, 1985). Residual effects of cover crops, however, appear to be very much larger than these estimates based on aboveground sampling would suggest. Long-term trials have shown that trees would require at least 900 kg of fertilizer N/ha over 13 years to achieve similar rubber yields to trees under legumes (Ismail et al., 1980). Though the legume covers normally commence to be shaded out after about 4 years, the net residual effect of the covers on rubber extends well into the fifteenth year of tapping, or up to the twentieth year from planting. The long-term beneficial effect of legumes seems to result from improvements in soil physical properties and the increased rooting and exploitation of the soil by trees, coupled with the slow decline of the cover crops as the trees mature and the progressive release of N and other nutrients from legume leaf litter (Ismail et al., 1980; Agamuthu and Broughton, 1985). C. CONTRIBUTION OF LEGUMES TO LIVESTOCK SYSTEMS The role of legumes in animal production is likely to be very different under the contrasting management protocols used in the extensive grazing systems common in semiarid Africa, the pastoral systems of Australia and Latin America, and the opportunistic browsing or cut-and-carry practices employed in Asia. Therefore, only the main quantitative aspects concerning the contribution of forage legumes to livestock systems will be dis-

NITROGEN FIXATION BY TROPICAL LEGUMES

20 1

cussed. Details on particular management systems are provided elsewhere (Hutton, 1970; Mustapha and Djafar, 1980; Yaacob et al., 1980; Blair et al., 1986; Bayer and Waters-Bayer, 1989; see the contribution by Blair et al. in this volume). Tropical legumes can improve ruminant productivity by increasing both potential stocking rates and annual liveweight gains. Linear relationships have commonly been found between animal performance and dietary intake of leguminous material or legume content in a pasture sward (Mustapha and Djafar, 1980). Such responses may be attributed to increased forage production, increased forage intake, and/or improved forage quality. 1 . Forage Production

Decomposition of N-rich legume material can be rapid (Table XIII), and a large proportion of the N from ungrazed pasture legume residues is likely to be incorporated into the soil organic fraction or directly taken up by associated grasses (Vallis, 1983). As a consequence of this, appreciable amounts of organic N accumulate in soils under forage legumes (Henzell et al., 1966; Vallis, 1972; Yaacob et al., 1980). The long-term improvement in soil N status often leads to substantial increases in grass and total herbage yields (see Jones et al., 1967; Bryan and Velasquez, 1982; Haque and Jutzi, 1985). The inclusion of legumes can help slow the progressive decline in animal carrying capacity and plant productivity, which inevitably occurs as pastures age. In Kenya, for example, productivity of a mixed grass/ legume sward (3 11 kg N/ha/year; 12.7 t dry matter/ha/year) was similar to that of a pure grass pasture (250 kg N/ha/year; 12.7 t/ha/year) in the first year of establishment. Despite an increasing contribution by the legume component (Desmodium uncinatum) over a 4-year period (from 61 to 159 kg N/ha/year and 2.4 to 4.1 t/ha/year), total N yield and dry matter production by the mixed sward declined by around 30% and 50%, respectively. Productivity of the pure grass pasture decreased by 75% and 70% over the same period. The ability of the sown legume to reduce total herbage decline was comparable to fertilizer applications of 200 kg N/ha/ year to grass (Thairu, 1972).

2 . Forage Quality and Intake Animal production is dependent on dry matter digestibility and intake. Intake is in turn controlled by the extent and rate of digestion in the rumen.

202

MARK B. PEOPLES AND DAVID F. HERRIDGE

A minimum level of protein (and energy) is required if microbial growth in the rumen is not to be limited. Under practical feeding conditions, a dietary intake of fodder of 1.15% N (7.8% crude protein) is necessary to provide a maintenance ration and 1.74% N is considered adequate to meet the protein requirements of beef cattle (Mustapha and Djafar, 1980; Harricharan et al., 1988). Tropical grass species are often low in N content (e.g., 65 of a total of 85 grass samples collected from 25 different species fed to ruminants in Indonesia had N concentrations of 0.6-1.6% N; Little et al., 1989). Although it is possible to maintain the level of grass N above the critical minimum value with N fertilizer, the rates of application required (e.g., more than 400 kg N/ha/year; Mustapha and Djafar, 1980) make this an unrealistic option. Rather, legumes that generally contain 2-4% N (Harricharan et a / . , 1988; Little et al., 1989) are used as a high-protein supplement. Increasing the proportion of legume in the diet will not only improve both protein content and digestibility, but can also increase the voluntary intake of the entire diet (Mustapha and Djafar, 1980; Rekib et al., 1987). Furthermore, inclusion of legumes in a pasture can raise the N content of the associated grasses (Bryan and Velasquez, 1982).

V. STRATEGIES TO ENHANCE N2 FIXATION Herridge and Brockwell (1988) reported a study in which inoculation and N fertilizer treatments were imposed on an imgated crop of soybean (cv. Bragg) grown on a deep, alkaline black earth (vertisol) initially free of Bradyrhizobium japonicurn. The 16 treatments comprised four levels of inoculation (nil, normal [nl, loon, and 1000n)and four rates of fertilizer N (0, 100,200, and 300 kg N/ha). These treatments succeeded in generating different-sized populations of B. japonicum in the soil (0, 2.3, 5.0, and 6.5 loglo rhizobia/g soil, measured 2 hr after sowing) and different soil N fertilities (39, 94, 144, and 280 kg nitrate-N/ha to 0.9 m depth at sowing). Results clearly showed that soil nitrate repressed nodulation (Fig. 8a), the relative abundance of ureide-N in xylem sap (Fig. 8b), and P (Fig. 8c). The effect was magnified as soil nitrate concentrations increased, but the inhibition of N2 fixation was substantially ameliorated by increased numbers of rhizobia. Effects on crop dry matter yield (Fig. 8d) and N accumulation were less pronounced. The largest effects were observed at the lowest level of inoculation in the absence of N fertilizer and at the lowest level of fertilizer application in the absence of inoculation. Multiple regression analysis of the relationships between P and rhizobial

NITROGEN FIXATION BY TROPICAL LEGUMES

a

b

C

d

203

FIG. 8. Effects of fertilizer N and rhizobial inoculation on (a) nodulation, (b) relative abundance of ureide-N in xylem sap, (c) P,and (d) shoot dry matter of field-grown plants of Bragg soybean. Data derived from Herridge and Brockwell(l988).

204

MARK B. PEOPLES AND DAVID F. HERRIDGE

numbers and soil nitrate indicated that they were highly correlated. Thus,

P could be expressed as P(%) = 40 + 4.4(z) - 0.4(Soil) + 0.001(Soi1)2

(r2 = 0.80)

(12)

where z = loglo number of B. japonicum in the seed zone (3 to 15 cm depth) at sowing and Soil = amount of soil nitrate to 0.9 depth (kg N/ha) at sowing. Although this function described a specific crop and environment and was not intended as a universal equation describing the relationship between soybean N2 fixation and soil N fertility and rhizobial status, it does serve to illustrate the predictable and quantitative nature of N2 fixation. In this instance, fixation was regulated essentially by only two variables, with soil nitrate and rhizobial numbers accounting for 80% of the variation. Plant (crop) yield is another determinant of N2 fixation, particularly when the levels of soil N are moderate and there are sufficient numbers of effective rhizobia present. Yields in these instances are related to cultivar, mediated through growth rates or crop duration, or to nutritiodwater availability. Thus, P can be described by P(%) = (b + a/N1) x 100

(13)

The function is derived from N2 fixed = a

+ b(N1)

(14)

(e.g., see legend to Fig. 9). Examples of the strong relationship between N yield and N2 fixed, and from that between N yield and P, are presented in Fig. 9. Crops involved are soybean (Hardarson et al., 1984), pigeon pea (Kumar Rao and Dart, 1987), and faba bean (Viciafaba; Duc et al., 1988). The three experiments involved assessment of N yield and N2 fixation of a range of cultivars. Generally N yield varied with crop duration. Each data set is site specific, with the values for N yield for unnodulated plants (i.e., values for x when y = 0; Fig. 9a) being determined by the levels of soil nitrate. Modifying soil nitrate will shift the line of best fit either to the left or to the right. The reaction of the individual cultivars to nitrate affects the slope of the line. Increasing the proportions and amounts of N2 fixed by legumes therefore will be achieved by: ( a ) maximizing legume yield within the constraints imposed by agronomic and environmental considerations; ( b ) reducing the legume’s sensitivity to the suppressive effects of nitrate on nodulation and N2 fixation (nitrate tolerance); (c) optimizing the numbers and effectiveness of rhizobia in the rooting zone, through strain selection and inoculation techniques, and through plant breeding for promiscuous or selective nodulation; and (d)reducing the amount of nitrate in the rooting zone.

205

NITROGEN FIXATION BY TROPICAL LEGUMES

I loo

Faba bean Pigeon pea Soybean

1. S

80 60

40

20 0

0

100

200

300

0

100

200

300

Crop N (kg/ha)

FIG.9. Relationships between N yield and (a) N2 fixed and (b) P for field-grown crops of faba bean, pigeon pea, and soybean. Relationships in (a) are described by: (0)y = -32 ( ? = 0 0 . 9 8 ) ; ( 0 ) y = - 6 5 + ].OX(?= 1.00);(.)~= -127+ 1 . 0 ~ ( 3 = 0 . 8 8 ) .

+0.9~

In the following sections, examples of programs to enhance N2 fixation are examined. In each case, progress is achieved through one or more of the preceding factors. A. PLANTBREEDING A N D SELECTION The challenge to improve the N2 fixation capacity of the legumes through selection and breeding is complex because there are two components to consider: the host plant and the rhizobia. Although host genotype x strain interactions have been shown to occur with a number of agriculturally important legumes (Graham and Temple, 1984; Mytton, 1984), many of the programs selecting for enhanced N2-fixing potential have chosen to ignore this complication. Instead, the programs have either relied on nodulation by the indigenous rhizobia (Kueneman et al., 1984), utilized commercial inoculants (Betts and Herridge, 1987; Herridge and Betts, 1988), or prepared inoculants containing either single strains or small numbers of strains (Kumar Rao and Dart, 1987; Neuhausen et al., 1988). The focus, therefore, is entirely on the identification of variation in N2 fixation attributable to the host. 1 . Legume Yield

Although it may be a simple matter to identify cultivars of legumes with increased plant yields and therefore increased N2fixation (e.g., Hardarson

-& a

206

MARK B. PEOPLES AND DAVID F. HERRIDGE

et al., 1984; Kumar Rao and Dart, 1987), the real challenge is to select cultivars that maintain higher levels of N2 fixation and P at the same level of yield. As we argued earlier, agronomic and environmental considerations may limit the size of individual plants and the duration of the crop must be taken into account. With species such as common bean, low N yield remains a major constraint to N2 fixation. In fact, of the commercial crop legumes, the common bean is regarded as the weakest at fixing N2 and is supplied with fertilizer N in most cases. There is some evidence that the low N2-fixing potential of common bean may be associated with low specific nodule activity (Felix et al., 1981; Pereira and Bliss, 19891, which is associated with high levels of H2 production, that is, low relative efficiency (RE) of N2 fixation (Pacovsky el al., 1984; Piha and Munns, 1987a). Results of Hungria and Neves (1987) endorsed these findings by showing a strong inverse relationship between nodule H2 production, nodule specific activity, and plant N yield. Hydrogen evolution and therefore nodule RE were affected by both host cultivar and Rhizobium strain. In other programs, high levels of fixation were associated with late maturity and climbing habit (Graham and Rosas, 1977; Rennie and Kemp, 1983; Piha and Munns, 1987b). This implied a simple relationship between leaf area duration and N2 fixation (see also Wynne et al., 1982), or patterns of carbohydrate partitioning within the plant (Graham and Halliday, 1977). A breeding program by Bliss and coworkers at the University of Wisconsin has produced new genotypes of common bean with increased plant vigor, increased N yields, and higher levels of N2 fixation. Selected hybrid lines have displayed between three- and sevenfold increases in total N fixed and two- to fourfold increases in P relative to the commercial parent cultivar Sanilac (Table XVII, Experiment 1). For all the hybrids, however, fixation capacity was substantially less than the capacity of Puebla 152, the high-fixing donor. In a second experiment, the superiority of hybrid line 24-21 was obvious. The line displayed higher rates of growth and N accumulation (+36%) and more total growth (+79%) than the commercial parent, while retaining the short season and determinate characteristics (Table XVII, Experiment 2). 2 . Nitrate Tolerance

Nitrate is one of the most potent inhibitors of N2 fixation (Table VII) (Streeter, 1988). Development of symbioses in which P is maintained at near maximum levels in the presence of high soil nitrate could provide the biggest single advance in the improvement of N2 fixation by legumes. Plant mutagenesis has been used to generate phenotypes exhibiting

207

NITROGEN FIXATION BY TROPICAL LEGUMES Table XVII Summary of Data from Two Experiments from a Breeding Program to Increase Nz Fixation by Common Bean“

Experiment 1 N 2 fixation Parent or line

P

(mg N/plant)

Seed yield (g/plant)

Maturity (days)

Determinate

(mg N/plant)

Sanilac 24-17 24-21 24-55 Puebla

0.12 0.48 0.25 0.22 0.57

76 583 216 192 852

18 31 19 23 38

85 I10 91 94 120

Yes No Yes Yes No

591 1068 I045 668 1429

a

Amount

Experiment 2 Total N

Derived from Attewell and Bliss (1985).

greater nodule mass under field conditions (e.g., Rosaiah et al., 1987) or forming high numbers of nodules in the presence of nitrate (e.g., “supernodulating” soybean; Carroll et af., 1985). Grafting experiments indicate that the supernodulating trait identified in soybean is mediated through the shoots, probably by the nonproduction of factors regulating the number of successful rhizobial infections that develop into nodules (Delves et af., 1986). Extreme supernodulating mutants can form up to 10-fold more nodules in both the absence and presence of nitrate, but this trait results in substantial reductions in root and shoot growth (Day et al., 1986), and despite higher P values, the mutants are only capable of fixing more N than the original wild-type parent under very high nitrate conditions (Hansen et af., 1989). Another approach that has been considered is the selection of mutants with a reduced ability to utilize nitrate, that is, lowered nitrate reductase activity (Nelson et al., 1983; Carroll and Gresshoff, 1986). From biochemical characterizations of such mutants it would appear that legumes can contain several constitutive and nitrate-inducible nitrate reductase isoenzymes (Nelson et af., 1984; Streit and Harper, 1986), and a genotype completely lacking nitrate reductase has yet to be reported. Nodulation and N2 fixation are still apparently sensitive to nitrate in those mutants already identified as having lowered nitrate reductase activities. Species differ considerably in their symbiotic tolerance to mineral N and sufficient natural variation may already exist among legume lines and cultivars so that it might not be necessary to resort to mutagenesis procedures to induce further variation for breeding purposes (e.g., Hardarson el

208

MARK B. PEOPLES AND DAVID F. HERRIDGE

al., 1984; Harper and Gibson, 1984; Gibson and Harper, 1985). A program was commenced in 1980 under this premise to screen 489 diverse genotypes of soybean for tolerance to nitrate (Betts and Herridge, 1987). In the first two cycles of screening, all genotypes were grown in sand-filled pots in the glasshouse and supplied with either nitrate-free nutrients or nutrients containing 2.5 mM nitrate. Plants were sampled at late flowering for plant growth, nodulation, and N2 fixation (relative ureide-N in xylem sap and plant parts). Results indicated that genotypes of Korean origin displayed higher-than-average levels of symbiotic activity in the presence of nitrate. Of the original 19 Korean lines, 80% were included in the second screening, and 47% were selected for subsequent field screening. Only 5% of the remaining 470 genotypes were selected as high-fixing after the second glasshouse screening. In the third year, 40 genotypes (including the 32 identified high-fixing lines) were sown into a high-nitrate soil (260 kg N/ha to a depth of 1.2 m) in the field (Herridge and Betts, 1988). The genotypes showing highest levels of nodulation and N2fixation under these conditions were all Korean types (Table XVIII). They had shoot biomass similar to the commercial cultivars Bragg and Davis, but had reduced uptake of soil N as indicated by higher recoveries of soil nitrate from the Korean plots immediately after seed harvest (Table XVIII). Seed yield of the Korean lines was around 30% less than that of the commercial varieties. Correlation matrices among the indices of nodulation, N2 fixation, plant growth, and seed yield revealed independence between the symbiotic- and yield-related characters. Therefore, the nitrate-tolerant Korean Table XVlll Measurements of Nodulation and N1 Fixation by, and Growth and Yield of, NitrateTolerant and Commercial Genotypes of Soybean in a High-Nitrate Field Soil. * ~~

Genotype Nitrate-tolerant lines Korean 466 Korean 468 Korean 469 Commercial cultivars Bragg Davis

Nodulation weight (mg/plant)

Nodulation number

Shoot Seed Residual dry matter yield soil nitrate' @/plant) (tlha) (kg N/ha)

Pb

316 254 I76

34.5 16.8 19.5

0.31 0.18 0.22

45.9 43.3 41.6

I .6 1.7 1.4

64 19 16

24 40

2.0 1.3

0 0

39.7 48.5

2.2 2.2

45 n.a.

Data from Herridge and Betts (1985, 1988). Calculated using the ureide method according to Herridge and Peoples (1990). To a depth of I .2 m.

209

NITROGEN FIXATION BY TROPICAL LEGUMES

lines were used as high-fixing donor parents in a breeding program with selection for both seed yield and N2-fixingcapacity. The four outstanding Korean lines (K464, K466, K468, and K469) were crossed with the commercial cultivars Forrest, Bossier, Reynolds, and Valder. Approximately 1500 seeds from F2 combinations were sown into a high-nitrate soil in the field in 1986. Seedlings were individually tagged and assessed during growth for plant and seed characteristics, growth habit, and agronomic type. Symbiotic N2 fixation of individual plants was evaluated during early pod fill using a nondestructive xylem ureide procedure (Herridge ef al., 1988). The mean level of N2-fixingactivity of the Fz’s was surprisingly constant for the 1 1 combinations and was between that of the commercial parents (P = 0-0.20) and the high levels displayed by 3 of the 4 Korean parents (P = 0.40) (Fig. 10). Although the average levels of fixation of the F2’s were below those of the best Korean lines, 35 individual F2’s displayed equally high levels, that is, P > 0.40. Culling of the F2 population for progression to the F3generation was made on the basis of N2 fixation (xylem relative ureides > 32%, see Fig. lo), plant type, and seed color. To establish the heritability of N2 fixation, the seed of 100 F2 lines, selected to cover all 1 1 families and range of xylem ureide levels, was sown in a replicated experiment in 1987. Levels of N2 fixation were again assessed for F3 plants by the ureide method and compared with data for the

selected for F3 generation

Mean

634 F2‘s rejected as low-fixing

s FIG.10. Ranges of relative abundance of ureide-N in xylem sap for the 1 1 F2 families and for the commercial and Korean parents. Plants were grown in high-nitrate soil in the field. The horizontal line indicates the cutoff value (32%) for selection of material for the F3generation.

210

MARK B. PEOPLES AND DAVID F. HERRIDGE

same lines the previous season. When data were divided into groups designated by a single common parent, correlations (broad-sense heritabilities) between F2 and F3 xylem ureide data were higher (range r = 0.24-0.72) relative to the pooled (Le., all parents) data ( r = 0.27). This implied that the separate populations behaved differently in the genetic control of N2 fixation. The generally significant correlations between F2 and F3 relative ureides indicated that nitrate tolerance was under quantitative genetic control with a broad-sense heritability of between 0.24 and 0.72 (Rose et al., 1989). B. RHIZOBIA, INOCULATION, AND PLANTNODULATION Legume inoculation is a long-established and successful practice. Vincent (1965) and others (e.g., Allen and Allen, 1958) have argued that it is a desirable practice in most agricultural soils throughout the world. Date (1977), however, cautioned that the need to inoculate was not universal and should be carefully determined for each individual situation before investing in inoculant production and use. There are three major groups of legumes that can be distinguished on the basis of compatibility with a range of strains of Rhizobium (Table XIX). At one extreme is a group of legumes that can form an effective symbiosis with a wide range of strains. Members of this group were nodulated by “cowpea-type” rhizobia and these Rhizobium spp. are so widespread in tropical soils that such legumes seldom respond to inoculation. Yet even within this supposedly “promiscuous” group that can be some host-strain specificity in terms of the symbiotic effectiveness of the associations formed (Gibson et al., 1982). At the other extreme are legumes with very specific rhizobial requirements. These specificities are most relevant when the legume is introduced to new areas. Response to inoculation of these legumes is usually successful provided that adequate numbers of rhizobia are applied at sowing. The third and intermediate group of legumes nodulate with many strains of Rhizobium, but effectively fix N2 with only a limited number of them. Thus inoculation and nodulation failures are more frequent because the inoculum strain is unable to compete with the ineffective but established soil populations of rhizobia. There are a number of conditions under which soils may be devoid of Rhizobium to form an effective symbiosis with a legume and which may warrant inoculation: (a) the absence of the same or a symbiotically related legume in the immediate past history; ( 6 ) poor nodulation when the same crop was grown previously; (c) when the legume follows a nonleguminous

NITROGEN FIXATION BY TROPICAL LEGUMES

21 1

Table XIX Legumes Grouped on the Basis of Nodulation and N2 Fmation with a Range of Rhizobium Speciesa

Nodulate effectively with a wide range of strains. Genera listed forming one loose group. Albizia Alysicarpus Arachis Calliandra Calopogonium Cajanus Canavalia Clitoria Croialaria Dolichos Eryihrina

Galaciia Gliricidia Indigofera Lablab Lespedeza Macropiilium Macroiyloma Mimosa Pachyrhizus Pongamia Neonotonia

Psophocarpus Pueraria Rhynchosia Siylosanihes (several subgroups) Tephrosia Teramnus Vigna Voandzeia Zornia

Nodulate with a range of strains but often ineffectively. Genera listed forming individual groups with some crossing between groups. Subgroups distinguishable. Acacia Adesmia Aeschynomene

Astragalus Centrosema ( 2 subgroups) Desmanihus Desmodium (2 subgroups)

Psoralea Sesbania(2 subgroups)

Nodulate effectively with specific strains only. Genera listed forming specific groups. Cicer Coronilla Glycine max Hedysarum Lathyrus Lens Leucaena a

Lotononis-Listia ( 3 subgroups) Lotus ( 3 subgroups) Lupinus (2 subgroups) Medicago Melilotus Onobrychis Ornithopus

Phase o1us Pisum Trifolium (many subgroups) Trigonella Vicia

After Peoples et al. (1989a).

crop in a rotation; (d)in land reclamation; and, (e) when environmental conditions are unfavorable for Rhizobium survival (e.g., very acidic or alkaline soils, under prolonged flooding, or very hot, dry conditions prior to planting). However, as a farming practice, inoculation generally remains the exception rather than the rule (Vincent, 1982). Exacting technology is essential for the production and distribution of inoculants (Date and Roughley, 1977),and the success of inoculation in the field depends on the procedure used and operator competence (Brockwell, 1980; Brockwell et al., 1988). In Australia and the United States, legume inoculation has played a funda-

212

MARK B. PEOPLES AND DAVID F. HERRIDGE

mental role in the establishment of legume-based pasture and cropping systems, but less use has been made of inoculants elsewhere. In Latin America, only two countries use inoculants to any extent and even in Brazil, the largest producer of the seed legumes, common beans are fertilized with N rather than inoculated (Freire, 1982). Inoculation responses in tropical soils appear to be confined to crops such as soybean which have specific Rhizobium requirements (Ayanaba, 1977; Halliday, 1985). Typically responses are substantial when indigenous, infective rhizobia are absent and soil nitrate is low (Table XX). In a high-nitrate soil, nodulation of, and NZfixation by, inoculated plants can be suppressed. However, nodulation and NZfixation (but not necessarily yields) could be increased by very high rates of inoculation ( 1 0 0 0 ~normal; see Table XX; Bergersen et al., 1989). When populations of infective rhizobia exist in high numbers in soils, they present a formidable barrier to the successful exploitation of superior Table XX Effect of Inoculation on Nodulation, N2Fixation, and Productivity of Bragg Soybean Grown in Various Backgroundsof Soil Nitrate and Bradyrhizobium japonicum" ~

Treatment B. japonicum-free soil Low nitrate Uninoculated Normal inoculation High inoculation High nitrate Uninoculated Normal inoculation High inoculation

High B. japonicum soil Low nitrate Uninoculated Normal inoculation

Nodule massb (mglplant)

P'

Crop dry matte@ (tlha)

Crop Nd (kg N/ha)

Seed yield (t/ha)

66 I68 I95

I .7 3.3 3.0

0 72 334

0.11 0.61

4.9 6.8 7.0

0 4

0 0 0.17

8.9 8.0 8.5

205 I96 213

3.2 2.9 3.3

9.8 9.3

258 24 I

2.2 2.0

50

I29 146

0.44

Derived from Herridge et a!. (1987); Herridge and Brockwell(1988). Sampled either 62 or 70 days after planting. Calculated by the ureide method according to Herridge and Peoples (1990). Values represent the mean of four sap samplings between days 70 and 109. Sampled when shoot dry matter and N were at a maxima. a

NITROGEN FIXATION BY TROPICAL LEGUMES

213

strains of Rhizobium used as inoculants (see Table XX) (Devine, 1984). In the United States, large populations of soybean Rhizobium have become established in soil with cropping so that now less than 10% of nodules are formed on soybean by the inoculant and yield responses are rare (Berg et al., 1988; Halliday, 1985). Research programs in several laboratories (e.g., Devine, 1984; Cregan and Keyser, 1986) currently aim to produce soybean cultivars that bypass the resident rhizobia in the soil to be nodulated by better, selected inoculant strains (this assumes that fixation and N supply are limited by the effectiveness of indigenous rhizobia). A similar strategy was employed for groundnut by Nambiar et al. (1984), who exploited host X strain specificity with cultivar Robut 33-1 and strain NC 92 to obtain consistent yield responses (mean of 16% over nine experiments) in soils containing moderate to high numbers of infective rhizobia and where all uninoculated plants were well nodulated. A contrasting strategy is to develop varieties that can be effectively nodulated by the resident soil rhizobia. Nangju (1980) observed that soybean genotypes from Southeast Asia nodulated successfully with the indigenous rhizobia in Nigeria, but U .S.-bred cultivars nodulated poorly without inoculation. Hybridization of the Asian and U.S. types has resulted in high-yielding lines capable of fixing large amounts of N without inoculation (Kueneman et al., 1984). C. CROPA N D SOILMANAGEMENT

Both yield and P can be influenced by crop and soil management; we examine two important practices. 1 . Tillage

Cultivation accelerates the oxidation of organic matter in soils (Doran, 1980) and generally results in higher nitrate-N in the profile (e.g., Thomas et al., 1973; Dowdell et al., 1983). Cultivation may also decrease the rates of denitrification (Doran, 1980; Rice and Smith, 1982), immobilization (Rice and Smith, 1982), and leaching (Thomas et al., 1973) of nitrate compared to in untilled soils. Additional fertilizer N may be required by cereals under reduced tillage treatments, but for legumes, the lowered soil nitrate levels should result in enhanced N2 fixation. No-till systems can also modify and improve soil structure to create more favorable soil moisture and temperature regimes for plant growth (Lal, 1989). In cropping systems research involving soybean in a moderate rainfall

214

MARK B. PEOPLES AND DAVID F. HERRIDGE

environment (1000 mm annually) in coastal, subtropical Australia, nodulation and N2 fixation were substantially improved under no-tillage, compared with the cultivated system (Table XXI). The increased N2 fixed resulted primarily from increased P (0.88 for no-till versus 0.73 for cultivated), since yields were essentially unaffected by tillage practice. N balances were positive for both systems, although substantially higher for no-tillage. In a drier region of subtropical Australia (650 mm annual average rainfall), the largest effect of no-tillage on Nz fixation was through yield. In 7 of the 11 crops grown over a 3-year period, plant growth and N accumulation were increased under no-tillage (Fig. 11). Soil nitrate levels were reduced under no-tillage relative to the cultivated plots. At the high-fertility (Site A, Fig. 11) and the low-fertility (Site C, Fig. 11) sites, N2 fixation by soybean was effectively suppressed in all treatments as a result of a combination of severe moisture stress restricting nodulation and crop growth (both sites) and high levels of soil nitrate (A site only). At the moderate-fertility site (Site B, Fig. 1l), soybean showed increased nodulation, crop yields, and seed production in the no-tillage treatment (Table XXII). Although P (assessed using the ureide method, see Table XXII) was also higher for the no-till plants in these experiments, the greatest effect on total N2 fixed was through increased yield (355 versus 236 kg N/ha; equivalent to a 50% increase). 2 . Cropping Intensity and Sequence

The quantity of soil nitrate available to a legume can be influenced by the recent cropping. Irrigated soybeans grown immediately after an oat crop Table XXI N Budgets for Soybean Grown with Cultivation or No-Tillage"**

Tillage

Nodule mass' (mg/plant)

Crop N (kg N/ha)

Seed N (kg N/ha)

Fixed N" (kg N/ha)

N balance'

Cultivated No-tillage

86 139

245 264

150 I52

I80 232

+ 30 +80

Values shown are means of four crops grown over 3 years of experimentation. Taken from Hughes and Herridge (1989). Sampled between 40 and 46 days after sowing. Assessed using the ureide method. Fixed N - seed N.

215

NITROGEN FIXATION BY TROPICAL LEGUMES

I

-20 ! 1983

1984 1985 Year FIG. 11. Relative changes in total N uptake by four crop legumes under no-tillage compared with cultivated cropping. Soybean data for 1983, 1984, and 1985 are the means of one, two, and three varieties, respectively. From D. F. Herridge and J. F. Holland (unpublished data).

fixed 244 kg Nlha compared with 143 kg N/ha fixed by soybeans growing in previously fallowed soil; the proportions of fixed N in seed were 68% and 33%, and net N balances were +39 and -44 kg N/ha, respectively (Bergersen et al., 1985). At another site, winter cereal cropping removed 130 kg N/ha and left 19 mg/kg soil (0-30 cm) of extractable mineral N at soybean planting, compared with 38 mg/kg in winter-fallowed soil. Both treatments produced similar soybean yields (3.5 and 3.3 t/ha, respectively), but P was 0.57 in the winter-cropped and only 0.09 in the winter-

Table XXll Effect of Tillage on Soil Nitrate, Nodulation, and N2Fixation and Crop Growth by FieldGrown Soybean (cv. Forrest)"

Tillage

Soil nitrateb (kg N/ha)

Nodule mass (mg/plant)

Crop N (kg N/ha)

Seed N (kg N/ha)

Relative ureide-N' (%)

Cultivated No-tillage

214 185

69 219

236 355

161 182

55 64

a

D. F. Herridge and J. F. Holland, unpublished data. At sowing, 0 to 1.2 m depth. Mean values of eight samplings.

216

MARK B. PEOPLES AND DAVID F. HERRIDGE

fallowed soils (Bergersen et al., 1989). Thus with proper choice of rotation, cropping systems can be managed for improved N2 fixation. In a different approach, intercropping maize and rice bean was found to increase P above the levels attained by the monocropped legume due to competition with maize for soil N (Table VIII). (Rerkasem et al., 1988; Rerkasem and Rerkasem, 1988). The net result was a higher total N yield of the intercrop (maize plus legume) relative to combined weighted N yield of maize and ricebean monocrops.

VI. CONCLUDING REMARKS We have considered the amounts of N fixed by various legumes in tropical and subtropical systems, examined the benefits of the legumes and of the fixed N to productivity of those systems, and argued that reduced N2-fixing activity is associated with low plant yield and reduced P . Efficient management of legumes to maximize the benefits depends on accurate assessment of N2 fixation in the field. This knowledge not only provides an insight into the N economy of the legume, but adds to our understanding of the general N cycle. Using the information gained, strategies can be developed to solve problems involving N in agricultural and natural systems. Solutions will come in the form of cropping and tillage systems to enhance N2 fixation, improved legume genotypes, improvements in methods of inoculation, and adoption of management practices to stimulate the establishment of large populations of desirable microbes in the soil. Data from individual experiments will be used not only to solve problems and provide a base for management decisions, but also for developing principles and, eventually, functional models. Fixation could then be predicted or estimated, even by farmers, given relevant information on cropping history, climatic data, plant species, and precrop soil nutrient tests.

REFERENCES Agamuthu, P., and Broughton, W. J . 1985. Agric. Ecosys. Environ. 13, 111-123. Alazard, D., and Becker, M. 1987. Plant Soil 101, 141-143. Allen, E. K., and Allen 0. N . 1958. In “Encyclopedia of Plant Physiology” (W. Ruhland, ed.), pp. 48-1 18. Springer-Verlag, Berlin. Andrews, M. 1986. Plant Cell Environ. 9,511-519. Attewell, J., and Bliss, F. A. 1985. In “Nitrogen Fixation Research Progress’’ (H. J. Evans, P. J. Bottomley, and W. E. Newton, eds.), pp. 3-9. Nijhoff, Boston.

NITROGEN FIXATION BY TROPICAL LEGUMES

217

Ayanaba, A. 1977. In “Biological Nitrogen Fixation in Farming Systems of the Tropics” (A. Ayanaba, and P. J. Dart, eds.), pp. 181-187. Wiley, Chichester. Bandyopadhyay, S. K., and De, R. 1986. J . Agric. Sci. 107,629-632. Bartholomew, W. V. 1965. In “Soil Nitrogen Agronomy No. 10” (w. V. Bartholomew, and F. E. Clark, eds.), pp. 285-306. American Society of Agronomy, Madison, Wisconsin. Bartholomew, W. V. 1977. In “Biological Nitrogen Fixation in Farming Systems of the Tropics” (A. Ayanaba and P. J. Dart, eds.), pp. 27-42. Wiley, Chichester. Bayer, W., and Waters-Bayer, A. 1989. Expl. Agric. 25,277-289. Berg, R . K . , Jr., Loynachan, T. E., Zablotowicz, R. M., and Lieberman, M. T . 1988. Agron. J . 80,876-88 I . Bergersen, F. J. 1988a. In “Biotechnology of Nitrogen Fixation in the Tropics” (Z. H. Shamsuddin, W. M. W. Othman, M. Marziah, and J. Sundram,eds.). pp. 105-115. Univ. Pertanian Malaysia, Serdang, Malaysia. Bergersen, F. J. 1988b. In “Biotechnology of Nitrogen Fixation in the Tropics” (Z. H. Shamsuddin, W. M. W. Othman, M. Marziah, and J. Sundram, eds.), pp. 137-144. Univ. Pertanian Malaysia, Serdang, Malaysia. Bergersen, F. J., Turner, G. L., Gault, R. R., Chase, D. L., and Brockwell, J. 1985. Aust. J . Agric. Res. 36,411-423. Bergersen, F. J . , Peoples, M. B., and Turner, G. L. 1988. Aust. J. Plant Physiol. 15,407-420. Bergersen, F. J . , Brockwell, J., Gault, R. R., Morthorpe, L., Peoples, M. B.. and Turner, G. L. 1989. Aidst. J . Agric. Res. 40, 763-780. Beri, V., Meelu, 0. P., and Khind, C. S. 1989. Trop. Agric. (Trinidad) 66, 11-16. Betts, J. H., and Herridge, D. F. 1987. Crop Sci. 27, 1156-1 161. Blair, G. J., Ivory, D. A., and Evans, T. R. (eds). 1986. “Forages in Southeast Asian and South Pacific Agriculture,” Proceeding No. 12, p. 202. ACIAR, Canberra. Boddey, R. M., Chalk, P. M., Victoria, R. L., and Matsui, E. 1984. SoilBiol. Biochem. 16, 583-588. Boddey, R. M., Urquiaga, S., Suhet, A. R., Peres, J. R., and Neves, M. C. P. 1990. SoilBiol. Biochem. (in press). Brewbaker, J. L. 1986. In “Forages in Southeast Asian and South Pacific Agriculature” ( G . J. Blair, D. A. Ivory, and T. R. Evans, eds.), pp. 43-50. ACIAR, Canberra. Brewbaker, J. L., and Glover, N. 1988. I n “Sustainable Agriculture: Green Manure in Rice Farming,” pp. 29-43. IRRI, Manila. Brockwell, J. 1980. In “Methods for Evaluating Biological Nitrogen Fixation” (F. J. Bergersen, ed.), pp. 417-488. Wiley, Chichester. Brockwell, J., Gault, R. R., Herridge, D. F., Morthorpe, L. J., and Roughley, R. J. 1988. Aust. J . Agric. Res. 39, 965-972. Bryan, W. B., and Velasquez, E. R. 1982. In “Biological Nitrogen Fixation Technology for Tropical Agriculture” (P. H. Graham and S. C. Harris, eds.), pp. 355-360. CIAT, Cali, Columbia. Carroll, B. J., and Gresshoff, P. M. 1986. Plant Physiol. 81,572-576. Carroll, B. J . , McNeil, D. L., and Gresshoff, P. M. 1985. Plant Physiol. 78,34-40. Carroll, B. J., Hansen, A. P., McNeil, D. L., and Gresshoff, P. M. 1987. Aust. J . Plant Physiol. 14,679-687. Chalk, P. M. 1985. SoilBiol. Biochem. 17, 389-410. Chalk, P. M., Douglas, L. A., and Buchanan, S. A. 1983. Can. J . Microbiol. 29,1046-1052. Chandel, A. S., Pandey, K. N., and Saxena, S. C. 1989. Trop. Agric. (Trinidad)66,73-71. Chapman, A. L., and Myers, R. J. K. 1987. Aust. J. Exp. Agric. 27, 155-163. Chulan, A., and Waid, J. S. 1981. In “Nitrogen Cycling in South-East Asian Wet Monsoonal Ecosystems” (R. Weselaar, J. R. Simpson, andT. Rosswall, eds.), pp. 150-153. Australian Academy of Science, Canberra.

218

MARK B. PEOPLES AND DAVID F. HERRIDGE

Comet, F., Otto, C., Rinaudo, G., Diem, H. G., and Dommergues, Y. 1985. Acta Oecol. Oecol. Plant. 6,211-218. Cregan, P. B., and Keyser, H. H. 1986. Crop Sci. 26,911-916. Dakora, F. D. 1985a. I n “Biological Nitrogen Fixation in Africa” (H. Ssali and S. 0. Keya, eds.), pp. 59-71. The Nairobi Rhizobium MIRCEN, Kenya. Dakora, F. D. 1985b. I n “Biological Nitrogen Fixation in Africa” (H. Ssali and S. 0. Keya, eds.), pp. 324-339. The Nairobi Rhizobium MIRCEN, Kenya. Dakora, F. D., Aboyinga, R. A., Mahama, Y., and Apaseku, J. 1987. MIRCEN J . 3,389-399. Danso, S. K. A. 1985. I n “Biological Nitrogen Fixation in Africa” (H. Ssali and S. 0. Keya, eds.), pp. 224-244. The Nairobi Rhizobium MIRCEN, Kenya. Danso, S. K. A. 1988. I n “Nitrogen Fixation by Legumes in Mediterranean Agriculture” (D. P. Beck and L. A. Materon, eds.), pp. 345-358. Nijhoff, Dordrecht. Date, R. A. 1977. I n “Biological Nitrogen Fixation in Farming Systems of the Tropics” (A. Ayanaba and P. J. Dart, eds.), pp. 169-180. Wiley, Chichester. Date, R. A., and Roughley, R. J. 1977. I n “A Treatise on Dinitrogen Fixation. IV. Agronomy and Ecology” (R. W. F. Hardy and A. H. Gibson, eds.), pp. 243-275. Wiley, Chichester. Davey, A. G., and Simpson, R. J. 1988. Aust. J . Plant Physiol. 15,657-667. Davey, A. G., and Simpson, R. J. 1989. J . Exp. Eot. 40, 149-158. Day, D. A., Lambers, H., Bateman, J., Carroll, B. J., and Gresshoff, P. M. 1986. Physiol. Plant. 68,375-382. Delves, A. C., Mathews, A., Day, D. A., Carter, A. S., Carroll, B. J., and Gresshoff, P. M. 1986. Plant Physiol. 82,588-590. Denison, R. F., Sinclair, T. R., Zobel, R. W., Johnson, M. N., and Drake, G. M. 1983. Plant Soil 70, 173-182. Devine, T. E. 1984. I n “Biological Nitrogen Fixation” (M. Alexander, ed.), pp. 127-154. Plenum, New York. Dommergues, Y. 1982. I n “Biological Nitrogen Fixation for Tropical Agriculture” (P. H. Graham and S. C. Harris, eds.), pp. 395-425. CIAT, Cali, Columbia. Doran, J. W. 1980. Soil Sci. SOC.A m . J . 44,765-771. Doughton, J. A., and MacKenzie, J. 1984. Aust. J . Exp. Agric. Anim. Husb. 24,244-249. Dowdell, R. J., Crees, R., and Cannell, R. Q. 1983. J . Soil Sci. 34, 367-379. Duc, G., Mariotti, A., and Amarger, N. 1988. Plant Soil 106,269-276. Duhoux, E., and Dommergues, Y. 1985. I n “Biological Nitrogen Fixation in Africa” (H. Ssali and S. 0. Keya, eds.), pp. 384-400. The Nairobi Rhizobium MIRCEN, Kenya. Duque, F. F., Neves, M. C. P., Franco, A. A., Victoria, R. L., and Boddey, R. M. 1985. Plant Soil 88, 333-343. Eaglesham, A. R. J., Ayanaba, A., Ranga Rao, V., and Eskew, D. L. 1981. Soil B i d . Biochem. 13, 169-171. Eaglesham, A. R. J., Ayanaba, A., Rango Rao, V., and Eskew, D. L. 1982. Plant Soil 68, 183-192. Evans, J., and Taylor, A. C. 1987. J . Aust. h i . Agric. Sci. 53, 78-82. Felix, J. F., Obaton, M., Messiaen, C. M., and Salsac, L. 1981. Plant Soil 63,427-438. Firth, P., Thitipoca, H., Suthipradit, S., Wetselaar, R., and Beech, D. F. 1973. Soil Biol. Biochem. 5,41-46. Frankenberger, W. T., and Abdelmagid, H. M. 1985. Plant Soil 87,257-271. Freire, J . R. 1982. Plant Soil 67, 227-239. Fried, M., and Broeshart, H. 1975. Plant Soil 43,707-71 1. Gibson, A. H., and Harper, J. E. 1985. Crop Sci. 25,497-501. Gibson, A. H., Dreyfus, B. L., and Dommergues, Y. R. 1982. I n “Microbiology of Tropical Soils and Plant Productivity” (Y. R. Domrnergues, and H. G. Diem, eds.), pp. 37-73. Nijhoff, The Hague.

NITROGEN FIXATION BY TROPICAL LEGUMES

219

Giller, K. E., and Witty, J. F. 1987. Soil Biol. Biochem. 19,459-463. Giller, K. E., Nambiar, P. T. C., Srinivasa Rao, B., Dart, P. J., and Day, J. M. 1987. Biol. Fertil. Soils 5 , 23-25. Graham, P. H., and Halliday, J. 1977. In “Exploiting the Legume-Rhizobium Symbiosis in Tropical Agriculture” ( J . M. Vincent, A. S. Whitney, and J. Bose, eds.), Univ. Hawaii College Tropical Agriculture Miscellaneous Publication 145. pp. 3 13-334. Univ. Hawaii, Honolulu, Hawaii. Graham, P. H., and Rosas, J. C. 1977. J . Agric. Sci. 88,503-508. Graham, P. H., and Rosas, J. C. 1978. J . Agric. Sci. 90,311-317. Graham, P. H., and Temple, S. R. 1984. Plant Soil 82,315-328. Gramshaw, D., Read, J. W., Collins, W. J., and Carter, E. D. 1989. In “Persistence ofForage Legumes” (G. C. Marten, A. G. Matches, R. F. Barnes, R. W. Brougham, R. J. Clernents, and G. W. Sheath, eds.), pp. 1-21, American Society of Agronomy, Madison, Wisconsin. Greenland, D. J. 1977. In “Biological Nitrogen Fixation in Farming Systems of the Tropics” (A. Ayanaba and P. J. Dart, eds.), pp. 13-25. Wiley, Chichester. Gueye, M., and Bordeleau, M. 1988. MlRCEN J. 4,365-375. Halliday, J. 1985. In “Nitrogen Fixation Progress” (H. J. Evans, P. J. Bottomley, and W.E. Newton, eds.), pp. 675-681. Nijhoff, Boston. Hansen, A. P., and Pate, J. S. 1987. J. Exp. Bot. 38, 1446-1458. Hansen, A. P., Pate, J. S., and Atkins, C. A. 1987. J. Exp. Bot. 38, 1-12. Hansen, A. P., Peoples, M. B., Gresshoff, P. M., Atkins, C. A., Pate, J. S., and Carroll, B. J. 1989. J . EXP.Bot. 40,715-724. Haque, I., and Jutzi, S. 1985. In “Biological Nitrogen Fixation in Africa” (H. Ssali and S. 0. Keya, eds.), pp. 340-371. The Nairobi Rhizobium MIRCEN, Kenya. Hardarson, G., Zapata, F., and Danso, S . K. A. 1984. Plant Soil 82,397-405. Hardy, R. W. F., Burns, R. C., and Holsten, R. D. 1973. SoilBiol. Biochem. 5,47-81. Harper, J. E., and Gibson, A. H. 1984. Crop Sci. 24,797-801. Hamcharan, H., Moms, J., and Devers, C. 1988. Trop. Agric. (Trinidad)65, 132-136. Hauck, R. D., and Weaver, R. W. (eds.). 1986. “Field Measurement of Dinitrogen Fixation and Denitrification,” Special Publication No. 18, p. 115. Soil Science Society of America, Madison, Wisconsin. Hearne, A. B. 1986. Field Crops Res. 13, 159-175. Henzell, E. F., and Vallis, 1. 1977. In “Biological Nitrogen Fixation in Farming Systems of the Tropics” (A. Ayanaba and P. J. Dart, eds.), pp. 73-88. Wiley, Chichester. Henzell, E. F., Fergus, I. F., and Martin, A. E. 1966. Aust. J . Exp. Agric. Anim. Husb. 6 , 157- 160. Herridge. D. F. 1982a. I n “Biological Nitrogen Fixation Technology for Tropical Agriculture” (P. H. Graham and S. C. H a m s , eds.), pp. 593-608. CIAT, Cali, Columbia. Herridge, D. F. 1982b. Plant Physiol. 70, 1-6. Herridge, D. F., and Bergersen, F. J. 1988. In “Advances in Nitrogen Cycling in Agricultural Ecosystems” ( J . R. Wilson, ed.), pp. 46-65. C.A.B. International, Wallingford, Connecticut. Herridge, D. F., and Betts, J. H. 1985. In “Nitrogen Fixation Progress” (H. J. Evans, P. J. Bottomly, and W. E. Newton, eds.), pp. 32. Nijhoff, Boston. Herridge, D. F., and Betts, J. H. 1988. Plant Soil 110, 129-135. Herridge, D. F., and Brockwell, J. 1988. Soil Biol. Biochem. 20,711-717. Herridge, D. F., and Peoples, M. B. 1990. Plant Physiol. 93,495-503. Hemdge, D. F., Roughley, R. J., and Brockwell, J. 1987. Aust. J. Agric. Res. 38,75-82. Herridge, D. F., O’Connell, P., and Donnelly, K. 1988. J. Exp. Bot. 39, 12-22. Herridge, D. F., Bergersen, F. J., and Peoples, M. B. 1990. Plant Physiol. 93,708-716.

220

MARK B. PEOPLES AND DAVID F. HERRIDGE

Hogberg, P., and Kvarnstrom, M. 1982. Planr Soil 66,21-28. Hughes, M., and Herridge, D. F. 1989. Ausr. J . Exp. Agric. 29,671-677. Hungria, M., and Neves, M. C. P. 1987. Planr Soil 103, I1 1-121. Hunt, P. G., Burnham, K. P., and Matheny, T. A. 1987. Agron. J . 79,425-428. Hutton, E. M. 1970. Adv. Agron. 22,2-73. Ismail, T., Lock, C. S. , and Pushparajah, E. 1980. I n “Proceeding of Legumes in the Tropics,” pp. 403-414. Univ. Pertanian Malaysia, Serdang, Malaysia. Jones, R. J., Griffiths Davies, J., and White, R. B. 1967. Ausr. J . Exp. Agric. Anim. Husb. 7, 57-65. Kang, B. T. 1988. I n “Advances in Nitrogen Cycling in Agricultural Ecosystems” ( J . R. Wilson, ed.), pp. 333-348. C.A.B. International, Wallingford, Connecticut. Knowles, R. 1980. In “Methods for Evaluating Biological Nitrogen Fixation” (F. J. Bergersen, ed.), pp. 557-582. Wiley, Chichester. Kohl, D. H., Reynolds, P. H. S., and Shearer, G. 1989. Planr Physiol. 90,420-426. Kretschmer, A. E. 1989. In “Persistence of Forage Legumes” ( G . C. Marten, A. G. Matches, R. F. Barnes, R. W. Brougham, R. J. Clements, and G. W. Sheath, eds.), pp. 117-137. American Society of Agronomy, Madison, Wisconsin. Kucey, R. M. N., Chaiwanakupt, P., Arayangkool, T., Snitwongse, P., Siripaibool, C., Wadisirisuk, P., and Boonkerd, N. 1988a. Planr Soil 108,87-92. Kucey, R. M. N., Snitwongse, P., Chaiwanakupt, P., Wadisirisuk, P., Siripaibool, C., Arayangkool, T., Boonkerd, N., and Rennie, R. J. 1988b. Planr Soil 108,33-41. Kueneman, E. A., Root, W. R., Dashiell, K. E., and Hohenberg, J. 1984. Planr Soil 82, 387-396. Kulkarni, K. R., and Pandey, R. K. 1988. In “Sustainable Agriculture: Green Manure in Rice Farming,” pp. 289-299. IRRI, Manila. Kumar Rao, J. V. D. K., and Dart, P. J. 1987. Plant Soil 99,255-266. Kumar Rao, J. V. D. K., Dart, P. J., Matsumoto, T. M., and Day, J . M. 1981. Proc. Inr. Workshop Pigeonpea, (ICRISAT), Paruncheru 1, 190-199. Kumar Rao, J. V. D. K., Dart, P. J., and Sastry, P. V. S. S. 1983. Exp. Agric. 19, 131-141. Kumar Rao, J. V. D. K., Thompson, J. A., Sastry, P. V. S. S., Giller, K. E., and Day, J. M. 1987. Planr Soil 101, 107-1 13. Ladha, J. K., Watanabe, I., and Saono, S. 1988. In “Sustainable Agriculture: Green Manure in Rice Farming,” pp. 165-183. IRRI, Manila. Lal, R. 1989. Adv. Agron. 42,85-197. Langkamp, P. J., Swinden, L. B., and Dalling, M. J. 1979. Ausr, J. Bor. 27, 353-361. LaRue, T. A., and Patterson, T. G. 1981. Adv. Agron. 34, 15-38. Ledgard, S. F., and Peoples, M. B. 1988. In “Advances in Nitrogen Cycling in Agricultural Ecosystems” ( J . R. Wilson, ed.), pp. 351-367. C.A.B. International, Wallingford, Connecticut. Little, D. A., Kompiang, S., and Petheram, R. J. 1989. Trop. Agric. (Trinidad) 66,33-37. MacColl, D. 1989. Exp. Agric. 25,367-374. McClure, P. R., Israel, D. W., and Volk, R. J. 1980. Plant Physiol. 66,720-725. McNeill, A. M., Sheehy, J. E., and Drennan, D. S. H. 1989. J. Exp. Bor. 40, 187-194. Mariotti, A. J., Mariotti, F., and Amarger, N. 1983. In “Nuclear Techniques in Improving Pasture Management,” pp. 61-77. IAEA, Vienna. Minchin, F. R., Witty, J. F., Sheehy, J. E., and Muller, M. 1983. J. Exp. Bor. 34,641-649. Minchin, F. R., Sheehy, J. E., and Witty, J. E. 1986. J. Exp. Bot. 37, 1581-1591. Moong, F. A. 1986. In “Forages in Southeast Asian and South Pacific Agriculture” ( G . J. Blair, D. A. Ivory, and T. R. Evans, eds.), pp. 152-156. ACIAR, Canberra. Munyinda, K., Karamanos, R. E., Legg, J. O., and Sanogho, S. i988. Planr Soil 109,57-63.

NITROGEN FIXATION BY TROPICAL LEGUMES

22 1

Mustapha, M., and Djafar, M. I. 1980. In “Proceeding of Legumes in the Tropics,” pp. 415-424. Univ. Pertanian Malaysia, Serdang, Malaysia. Myers, R. J. K., and Wood, I. M. 1987. In “Food Legume Improvement for Asian Farming Systems” (E. S. Wallis and D. E. Byth, eds.), pp. 46-52. ACIAR, Canberra. Mytton, L. R. 1984. Plant Soil 82,329-336. Nair, P. K. R. 1988. In “Sustainable Agriculture: Green Manure in Rice Farming,” pp. 301-317, IRRI, Manila. Nambiar, P. C. T., Rao, M. R., Reddy, M. S. , Floyd, C., Dart, P. J., and Willey, R. W. 1983. Exp. Agric. 19,79-86. Nambiar, P. C. T., Dart, P. J., Srinivasa Rao, B., and Ravinshankar, H. N. 1984. Oleagineux 39, 150-154. Nangju, D. 1980. Agron. J. 72,403-406. Ndoye, I., and Dreyfus, B. 1988. SoilEiol. Eiochem. 20,209-213. Nelson, R. S ., Ryan, S. A., and Harper, J. E. 1983. Plant Physiol. 72,503-509. Nelson, R. S., Streit, L., and Harper, J. E. 1984. Physiol. Plant. 61, 384-390. Neuhausen, S. L., Graham, P. H., and Orf, J. H. 1988. Crop Sci. 28,769-772. Neves, M. C. P., Didonet, A. D., Duque, F. F., and Dobereiner, J. 1985. J . Exp. Eor. 36, 1179-1 192. Nishio, T., and Fujimoto, T. 1989. Soil Eiol. Eiochem. 21,661-665. Nnadi, L. A., and Balasubramanian, V. 1978. Trop. Agric. (Trinidad)55,23-32. Nnadi, L. A., and Haque, 1. 1988. J. Agric. Sci. 111,513-518. Norhayati, M., Mohd Noor, S., Chong, K., Faizah, A. W., Hemdge, D. F., Peoples, M. B., and Bergersen, F. J. 1988. Plant Soil 108, 143-150. Ofori, F., and Stern, W. R. 1987. Adu. Agron. 41,41-90. Ofori, F., Pate, J. S., and Stern, W. R. 1987. Plant Soil 102, 149-160. O’Hara, G. W., Boonkerd, N., and Dilworth, M. J. 1988. Plant Soil 108,93-110. Pacovsky, R. S., Bayne, H. G., and Bethlenfalvay, G. J. 1984. Crop Sci. 24, 101-105. Palm, 0.. Weerakoon, W. L., De Silva, M. A. P., and Rosswall, T. 1988. Plant Soil 108, 201-209. Pate, J. S., and Atkins, C. A. 1983. In “Nitrogen Fixation. Volume 3. Legumes” (W. J. Broughton, ed.), pp. 245-298. Oxford Univ. Press (Clarendon), Oxford. Pate, J. S . , Atkins, C. A., White, S. T., Rainbird, R. M., and Woo, K. C. 1980. Plant Physiol. 65,96 1-965. Patra, D. D., Sachdev, M. S., and Subbiah, B. V. 1986. Eiol. Fertil. Soils 2, 165-171. Peoples, M. B., and Dalling, M. J. 1988. In “Senescence and Aging in Plants” (L. D. Noodkn and A. C. Leopold, eds.), pp. 181-217. Academic Press, Orlando, Florida. Peoples, M. B., and Gibson, A. H. 1989. In “World Soybean Research Conference, IV” (A. J. Pascale, ed.), pp. 196-21 1. Asoc. Argentina De La Soja, Buenos Aires. Peoples, M. B., Pate, J. S., Atkins, C. A,, and Bergersen, F. J. 1986. Plant Physiol. 82, 946-95 1. Peoples, M. B., Sudin, M. N., and Herridge, D. F. 1987. J . Exp. Eor. 38,567-579. Peoples, M. B., Herridge, D. F., and Bergersen, F. J. 1988. In “Sustainable Agriculture: Green Manure in Rice Farming,” pp. 223-237. IRRI, Manila. Peoples, M. B., Faizah, A. W., Rerkasem, B., and Herridge, D. F. (eds). 1989a. “Methods for Evaluating Nitrogen Fixation by Nodulated Legumes in the Field,” p. 76. Monograph No. 11, ACIAR, Canberra. Peoples, M. B., Hebb, D. M., Gibson, A. H., and Herridge, D. F. 1989b. J. Exp. Eot. 40, 535-542. Pereira, P. A. A., and Bliss, F. A. 1989. Plant Soil 115,75-82. Piha, M. I., and Munns, D. N. 1987a. Plant Soil 98, 169-182.

222

MARK B. PEOPLES AND DAVID F. HERRIDGE

Piha, M. I., and Munns, D. N. 1987b. Agron. J. 79,690-696. Poth, M., La Favre, J. S., and Focht, D. D. 1986. SoilBiol. Biochem. 18, 125-127. Rachie, K. O., and Roberts, L. M. 1974. Adu. Agron. 26, 1-132. Rekib, A., Bhutani, D. K., and Khan, T. A. 1987. Forage Res. 13,63-69. Rennie, R. J . , and Kemp, G. A. 1983. Agron. J. 75,645-649. Rennie, R. J., Rennie, D. A,, Siripaibool, C., Chaiwanakupt, P., Boonkerd, N., and Snitwongse, P. 1988. Plant Soil 112, 183-193. Rerkasem, K., and Rerkasem, B. 1988. Plant Soil 108, 151-162. Rerkasem, B . , Rerkasem, K., Peoples, M. B., Herridge, D. F., and Bergersen, F. J. 1988. Plant Soil 108, 125- 135. Rice, C. W., and Smith, M. S. 1982. SoilSci. SOC.Am . J . 46, 1168-1173. Rinaudo, G., Dreyfus, B., and Dommergues, Y. 1983. SoilBiol. Biochem. 15, 111-113. Rinaudo, G., Alazard, D., and Moudiongui, A. 1988. I n “Sustainable Agriculture: Green Manure in Rice Farming,” pp. 97-109. IRRI, Manila. Rosaiah, G., Kumari, D. S., Satyanarayana, A., and Seenaiah, P. 1987. Indian J. Agric. Sci. 57,271-273. Rose, J. A., Betts, J. H., and Herridge, D. F. 1989. In “World Soybean Research Conference IV”, (A. J. Pascale, ed.), pp. 1176-1 181. Asoc. Argentina De La Soja, Buenos Aires. Roskoski, J. P., and van Kessel, C. 1985. OIKOS 44,306-312. Roskoski, J. P., Montano, J., van Kessel, C., and Castilleja, G. 1982. I n “Biological Nitrogen Fixation Technology for Tropical Agriculture” (P. H. Graham and S. C. Harris, eds.), pp. 447-454. CIAT, Cali, Columbia. Ruschel, A. P., Vose, P. B., Matsui, E., Victoria, R. L., and Saito, S. M. T. 1982. Plant Soil 65,397-407. Sanford, J. 0.. and Hairston, J. E. 1984. Agron. J. 76,623-627. Sanginga, N . , Mulongoy, K., and Ayanaba, A. 1988. Plant Soil 112, 137-141. Sanginga, N., Mulongoy, K., and Ayanaba, A. 1989. Plant Soil 117,269-274. Schubert, K. R. 1986. Annu. Reu. Plant Physiol. 37,539-574. Sellstedt, A., Hogberg, P., and Jonsson, K. 1989. J. Exp. Bor. 40, 1163-1 168. Shearer, G., and Kohl, D. H. 1986. Aust. J . Plant Physiol. 13,699-756. Sheldrake, A. R., and Narayanan, A. 1979. J. Agric. Sci. 92,513-526. Shivaram, S., Rai, P. V., and Hegde, S. V. 1988. Plant Soil 111, 11-16. Singh, S. B., and Verma, B. S. 1985. Indian J . Agron. 30, 397-400. Sisworo, W. H., Mitrosuhardjo, M. M.. Rasjid, H., and Myers, R. J. K. 1990. Plant Soil (in press). Smith, J. H., and Peterson, J. R. 1982. In “Nitrogen in Agricultural Soils. Agronomy No. 22” (F. J. Stevenson, ed.), pp. 791-831. American Society of Agronomy, Madison, Wisconsin. Ssali, H., and Keya, S. 0 1984a. E. Afr. Agric. For. J . 45, 188-193. Ssali, H., and Keya, S. 0. 1984b. E. Afr. Agric. For. J . 45,247-254. Ssali, H., and Keya, S. 0. 1986. Trop. Agric. (Trinidad) 63, 105-109. Stevenson, F. J. (ed.). 1982. “Nitrogen in Agricultural Soils, Agronomy No. 22.” American Society of Agronomy, Madison, Wisconsin. Streeter, J. 1988. CRC Crit. Rev. Plant Sci. 7, 1-23. Streit, L., and Harper, J. E. 1986. Plant Physiol. 81,593-596. Suwanarit, A., Suwannarat, C., and Chotechaungmanirat, S. 1986. Plant Soil 93,249-258. Thairu, D. M. 1972. E. Afr. Agric. For. J . 37,215-219. Thomas, G. W., Blevins, R. L., Phillips, R. E., and McMahon, M. A. 1973. Agron. J. 63, 736-739. Turner, G. L., and Gibson, A. H. 1980. I n “Methods for Evaluating Biological Nitrogen Fixation” (F. J. Bergersen, ed.), pp. 111-138. Wiley, Chichester.

NITROGEN FIXATION BY TROPICAL LEGUMES

223

Upchurch, R. G. 1987. In “Symbiotic Nitrogen Fixation Technology” (G. H. Elkan, ed.), pp. 289-305. Dekker, New York. Urquiaga, S. S., and Boddey, R. M. 1987. Plant Soil 102,291-294. Vallis, I. 1972. Aust. J . Exp. Agric. Anim. Husb. 12,495-501. Vallis, I. 1983. Aust. J . Agric. Res. 34, 367-376. Vallis, I., and Gardener, C. J. 1985. Aust. J . Exp. Agric. 25, 70-75. Vallis, I., and Jones, R. J. 1973. Soil Biol. Biochem. 5, 391-398. Vallis, I., Henzell, E. F., and Evans, T. R. 1977. Aust. J . Agric. Res. 28, 413-425. van Kessel, C., and Roskoski, J. P. 1988. Plant Soil 111, 17-23. van Kessel, C., Roskoski, J. P., Wood, T., and Montano, J. 1983. Plant Physiol. 72,909-910. van Kessel, C., Roskoski, J. P.. and Keane, K. 1988. Soil Biol. Biochem. 20, 891-897. Vincent, J. M. 1965. In “Soil Nitrogen. Agronomy No. 10” (W. V. Bartholomew, and F . E. Clark, eds.), pp. 384-435. American Society of Agronomy, Madison, Wisconsin. Vincent, J. M. 1982. In “Nitrogen Fixation in Legumes” (J. M. Vincent, ed.), pp. 263-284. Academic Press, Sydney. Wahua, T. A. T., and Miller, D. A. 1978. Agron. J . 70,292-295. Wallace, W. 1986. Physiol. Plant. 66, 630-636. Watanabe, I., Yoneyama, T., Padre, B., and Ladha, J. K. 1987. Soil Sci. Plant Nutr. 33, 407-4 I 5. Weber, C. R. 1966. Agron. J. 58,46-49. Wetselaar, R., and Ganry, F. 1982. In “Microbiology of Tropical Soils and Plant Productivity” (Y. R. Dommergues and H. G. Diem, eds.), pp. 1-36. Nijhoff, The Hague. Wetselaar, R., Jakobsen, P., and Chaplin, G. R. 1973. Soil Biol.Biochem. 5,35-40. Whitney, A. S. 1982. In “Biological Nitrogen Fixation Technology for Tropical Agriculture” (P. H. Graham and S. C . Harris, eds.), pp. 361-367. CIAT, Cali, Columbia. Wijeratne, W. B., and Nelson, A. I. 1987. In “Food Legume Improvement for Asian Farming Systems” (E. S. Wallis and D. E. Byth, eds.), Proceeding No. 18. pp. 183-192. ACIAR, Canberra. Wilson, G. F., Kang, B. T., and Mulongoy, K. 1986. Biol. Agric. Hort. 3,251-267. Witty, J. F., and Minchin, F. R. 1988. In “Nitrogen Fixation by Legumes in Mediterranean Agriculture” (D. P. Beck and L. A. Materon, eds.), pp. 331-344. Nijhoff, Dordrecht. Witty, J. F., Rennie, R. J., and Atkins, C. A. 1988. In “World Crops: Cool Season Food Legumes” (R. J. Summerfield, ed.), pp. 715-730, Kluwer Academic, Dordrecht. Wood, I. M., and Myers, R. J. K. 1987. In “Food Legume Improvement for Asian Farming Systems” (E. S. Wallis and D. E. Byth, eds.), Proceeding No. 18. pp. 34-45. ACIAR, Canberra. Wynne, J. C., Ball, S. T., Elkan, G. H., Isleib, T. G., and Schneeweis, T. J. 1982. In “Biological Nitrogen Fixation Technology for Tropical Agriculture” (P. H. Graham, and S. C. H a m s , eds.), pp. 67-75. CIAT, Cali, Columbia. Yaacob, O., Blair, G. J., Chee, Y. K., and Yusof Noor, M. 1980. I n “Proceedings of Legumes in the Tropics,” pp. 303-31 I . Univ. Pertanian Malaysia, Serdang, Malaysia. Yeoh, C. H., and Phang, A. K. 1980. In “Proceeding of Legumes in the Tropics,” pp. 425-433. Univ. Pertanian Malaysia, Serdang, Malaysia. Yoneyarna, T., Yamada, N., Kojima, H., and Yazaki, J. 1984. Plant CeN Physiol. 25, 1561-1565. Yoneyama, T., Narnbiar, P. T. C., Lee, K. K., Srinivasa Rao, B., and Williams, J. H. 1990. Biol. Fertil. Soils.(in press). Zaharah, A. R., Sharifuddin, H. A. H., Razley, M. N., and Mohd Saidi, A. K. 1986a. Pertanika 9 , 4 5 4 9 . Zaharah, A. R., Sharifuddin, A. A. H., and Subramaniarn, R. 1986b. Pertanika 9, 17-22.

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY, VOL. 44

DISTRIBUTION, COLLECTION, AND EVALUATION OF Gossypium A. Edward Percival and RusseMJ Kohel United States Department of Agriculture Agricultural Research Service Southern Crops Research Laboratory College Station, Texas 77840

I . Introduction A. History of Domestication 11. Distribution A. Taxonomy B. Geographic Distribution C. Species D. Evolution 111. Collection A. Source of the Collection B. Collectors C. Plant Explorations IV. Evaluation A. Genetics and Cytology B. Electrophoresis C. Improvement V . Concluding Remarks References

I. INTRODUCTION Cotton is an agricultural and technological, rather than a botanical term, which has been used to describe cultivated species of the genus Gossypiurn. The first word ever defined as meaning cotton is the Sanskrit “karpasa-i.” Even today the word for cotton in modern Hindustani is “kapas.” The word “karapas” in the Bible, Esther 1:6, means cotton and may have contributed to, or derived from, the Sanskrit original (Crawford, 1948). The Spanish, Portuguese, French, and Italian words “algodon,” “algodao,” “cotone,” and “coton” are obviously derivatives of the Arabic “al” or “el-kutum,” as is the English word “cotton,” which also 225

226

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

comes from a corruption of the Arabic word “qutum” or “kutum” (Brown, 1958). Cotton is presently grown between 47”Nand 32”s latitude with over 50% of the production above 30”N latitude (Kohel and Lewis, 1984). Over 32 million ha of cotton were grown worldwide in 1988 (Anonymous, 1988), which indicates the importance of this crop. Cotton, a natural fiber, competes with synthetic fibers in the textile industry, and though competition from synthetic fibers has reduced the relative use of this crop in the recent past, its preeminence as the fiber of choice has been reestablished during the last decade. Although cotton is primarily a fiber crop, it is also important as a food and feed crop. Cottonseed is the world’s second most important oil seed; the oil is used for culinary purposes, and the oil cake residue is a protein-rich feed for ruminant livestock (Kohel and Lewis, 1984). A. HISTORY OF DOMESTICATION Hutchinson (1954) concluded that a form of one of the Asiatic cottons, the South African perennial Gossypium herbaceum L. race africanum, is truly wild and is the modern representative of the wild ancestor of the two cultivated “Old World” diploid cottons. Humans used this wild race africanum and developed the primitive race acerifolium. Its northward spread may have followed the loss of photoperiodism, which is standard for modern cultivated cottons of all species. Botanical relationships and the geographic patterns of ancient trade routes suggested to Hutchinson (1959) that southern Arabia was the locality where domestication as a fiber crop first took place. This seems plausible as the climate there at the beginning of recent times ( 1 1,000 years ago) was more hospitable for humans than it is at present (Fryxell, 1965). From the primitive perennial G. herbaceum, which spread into India, arose the earliest form of G. arboreum L. (Rozi). Presumably cotton was first grown for uses other than textile yarn, such as for wound dressing and wadding. Cotton may have been introduced to the area where the technology for spinning and weaving of flax (Linum usitatissimum L.) already existed (Lee, 1984).Thus, the weaving of cotton followed and first occurred in India, where it developed into a fine craft. The oldest archaeological record of cotton textile dates to 2700 B.C. and was found in excavations at Mohenjo Daro in the valley of the Indus River in what is now Pakistan (Gulatti and Turner, 1928). Knowledge and use of cotton fiber spread from India and Arabia to Greece during the time of Alexander the Great, circa 350 B.C. Cotton culture was spread across North Africa and into Spain by the Moors, and the Crusaders (1096-

DISTRIBUTION AND EVALUATION OF Gossypium

227

1270 A . D . ) introduced Levantine and Occidental cottons to other parts of Europe. The Crusaders also disseminated a knowledge of cotton goods and initiated an industry in the Crusader states of Asia Minor and later a lively trade in cotton goods between the Italian city-states and Asia (Crawford, 1948). A parallel cotton technology developed in the New World with no known connection to that which occurred in the Old World. In the Americas, where allotetraploid cottons were used for lint, domestication occurred after the differentiation of the lint-bearing species (Fryxell, 1965). Stephens and Mosely (1974) examined fiber samples found at the archaeological site of Ancon-Chillon in central coastal Peru, dating from about 2500 to 1750 B.C. (Lee, 1984). It is the general opinion that cotton in America was first used by the peoples that inhabited the coastal areas of Peru, and from there the textile craft spread northward and westward. The pre-Inca and Inca civilizations (1200 ~.c.-1530A . D . ) used G. barbadense L. as their source of lint; though the Inca peoples lived out of the range of cotton culture, they secured it by trade and barter from the Upper Amazon and from the coasts, where it was extensively cultivated (von Hagen, 1961). The Maya civilization (2000 ~ . c . - 1 5 2 7A . D . ) of Guatemala and Yucatan, Mexico, also cultivated cotton and developed a fine textile industry, as did the Aztecs (200 ~ . c . - 1 5 1 9A . D . ) and their predecessors, the Toltecs. The Aztecs obtained most of their cotton material from subjugated tribes that grew cotton on the coastal regions of present-day Mexico. However, it was the weaving craft that spread and not the species of cotton grown, as these peoples of Central America and Mexico cultivated G. hirsutum L. (von Hagen, 1961). Indian tribes of the southwestern United States picked up the weaving craft about 2000 years ago, which reached its peak about 1400 A . D . , and grew cotton as late as 1925 (Jones, 1936). It is evident that cotton growing persisted in the area for a long period of time, as there is still seed available today of a variety of G. hisrutum (Hopi) that was once cultivated by the Indians of Arizona. This variety is adapted to a short growing period (84-100 days), grows at relatively high elevations, and is tolerant of arid conditions (Kent, 1957). The cottons grown commercially in the Americas today are direct descendants of the native varieties found at the time of “discovery” of this hemisphere by Columbus. These are G. hirsutum, native to Mexico and parts of Central America, and G. barbadense, native to South America. The early American colonists grew G. hirsutum in upland sites and called them “Upland” cotton. Types of G. barbadense were found to be adapted to the islands of the coasts of the Carolinas and Georgia and were designated “Sea Island” cottons. These names persist to this day for types of

228

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

both species (Lewis and Richmond, 1968). Sea Island was introduced into Egypt during the last century and resulted in the development of the “Egyptian” long-staple cottons, which were reintroduced to the southwestern United States and resulted in the development of “Pima” cottons. Of the cotton grown in the United States today, 98% is G . hirsuturn and 2% is G . barbadense (Anonymous, 1988). The worldwide distribution of these two species has essentially displaced most of the G . arboreurn and G . herbaceurn previously grown in Asia and the Middle East.

II. DISTRIBUTION A. TAXONOMY The genus Gossypium was first named by Linnaeus in the mideighteenth century. Students of the subject have differed as to whether “cotton” (Gossypiurn) should be placed with the Malvaceae or the Bombacaceae family, or the Hibisceae or Gossypieae tribe (Lewis and Richmond, 1968). Edlin (1935) and Prokhanov (1953) classified them as Bombacaceae, whereas Alefeld ( 1 862) had previously divided them between the two families. Mauer (1954)and Kearney (1951)classified Gossypium spp. as Malvaceae. Fryxell (1968) disposed of any lingering taxonomic arguments of classification and removed Gossypium from the tribe Hibisceae. He points to the facts that Gossypieae are unique in possessing lysigenous glands, associated with the capacity to synthesize the sesquiterpene gossypol, and differences in embryo structure, thus establishing the classification of Gossypium as follows: Family-Malvaceae, TribeGossypieae, Genus-Gossypium. According to Fryxell (1979, 1984), a 1928 paper by G. S. Zaitzev, entitled “A Contribution to the Classification of the Genus Gossypium L. ,” is the basis for our current understanding of this genus. It was Zaitzev who first observed that the cottons of the Old World are diploid (2n = 26) and those of the New World are allotetraploid (2n = 52), and each of these could be further subdivided. Zaitzev, he further states, “paved the way for subsequent work and enabled us to advance in our taxonomic understanding of Gossypiurn.”

B. GEOGRAPHIC DISTRIBUTION The genus Gossypium presently contains 432 diverse species. Four species are cultivated and bear spinales seed fibers (lint). These are the two diploid (2n = 2x = 26), Asiatic species G . arboreurn and G . herbaceurn,

DISTRIBUTION AND EVALUATION OF Gossypiurn

229

which are confined to the “Old World,” and the other two are the amphidiploid (2n = 4x = 52) species G . hirsutum and G . barbadense. The latter two have centers of variability in Mexico-Central America and South America, but are now widely distributed in cultivation throughout the world (Hutchinson, 1959). The remaining 39 species are all considered wild and are not cultivated. The seed hairs of the wild species are very short and firmly attached to the seed. Four of the wild species are amphidiploid, one each indigenous to Mexico, the Hawaiian Islands, the Galapagos Islands, and Brazil. The other 35 wild diploid species are found in relatively arid areas of their respective tropical and subtropical regions (Anonymous, 1968). The genus Gossypium is divided into seven genome groups that are diploid and an eighth amphidiploid group that combines the A and D genomes. These groups show cytogenetic separation as well as distinctive geographic distribution: A genome-Two cultivated species from the Far East, Middle East, and Africa. B genome-Six wild species from Africa and the Cape Verde Islands. C genome-Ten wild species from Australia. D genome-Thirteen wild species from Mexico, Peru, and the Galapagos Islands. E genome-Four wild species from the Arabian Peninsula and Northeast Africa. F genome-One wild species from East Central Africa. G genome-One wild species from Australia. AD genome-Six (two cultivated and four wild) species from Mexico, South America, and Hawaiian Islands, the Galapagos Islands, and Brazil (the cultivated two having recently attained worldwide distribution through cultivation) (Fryxell, 1984).

C. SPECIES

The +43 recognized Gossypium species are listed in Table I. Beasley (1940, 1942) established a cytological classification of genomes that is closely related to taxonomic affinities and geographic distribution. However, in recent years opinions have been expressed for a reassessment of the present classifications within the genus. Vollesen (1986) has attempted to address this problem for the species native to Africa and the area of the Arabian Peninsula. However, it may be necessary to do this for the entire genus as recent explorations have uncovered new species and additional existing variabilities and relationships within species groupings.

230

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL Table I The Species of Gossypium'

Species

Genomic group

Distribution

1. Diploids (2n = 2 x G. herbaceum L. G. arboreum L. G. anomalum Wawr. & Peyr. G. rriphyllum (Harv. & Sand.) Hochr. G. capiris-viridis Mauer G. trifurcarum Vollesen G. bricchetrii (Ulbri.) Vollesen G. benadirense Mattei G . srurrianum J. H. Willis G. nandewarense (Derera) Fryx. G. robinsonii F. Muell. G. australe F. Muell. G. costularum Tod. G . cunninghamii Tod. G. nelsonii Fryx. G. pilosum Fryx. G. populifolium (Benth.) Tod. G. pulchellum (C. A. Gardn.) Fryx. G. rhurberi Tod. G. armourianum Kearn. G. harknessii Brandg. G. dauidsonii Kell. G. klofzschianum Anderss. G. aridum (Rose & Standl.) Skov. G. raimondii Ulbr. G. gossypioides (Ulbr.) Standl. G. lobatum Gentry G. frilobum (Moc. & Sess. ex DC.) Skov.

emend. Kearn. G. laxum Phillips G. turner; Fryx. G. schwendimanii Fryx. G. srocksii Mast. ex Hook. G. somalense (Gurke) Hutch. G. areysianum (Defl.) Hutch. G. incanum (Schwartz) Hillc. G. longicalyx Hutch. & Lee G. bickii Prokh.

Old World cultigen Old World cultigen Africa Africa Cape Verde Islands Africa Africa Africa Australia Australia Australia Australia Australia Australia Australia Australia Australia Australia Mexico, United States (Arizona) Mexico Mexico Mexico Galapagos Islands Mexico Mexico Mexico Mexico Mexico Mexico Mexico Mexico Arabia Arabia Arabia Arabia Africa Australia (continued)

23 1

DISTRIBUTION AND EVALUATION OF Gossypium Table I (Conrinued)

Species

Genomic group

2. Allotetraploids (2n G . hirsutum L. G . barbadense L. G . lomentosum Nutt. ex Seem. G . mustelinum Miers ex Watt G . darwinii Watt G . lanceolarum Tod.‘

=

Distribution

4x = 5 2 )

(AD), (AD), (AD)3 (AD), (AD), (AD)?

New World cultigen New World cultigen Hawaii Brazil Galapagos Islands Mexico

Adapted from Endrizzi e t a / . (1984). Dash indicates that genome designation has not been determined. Questions remain concerning the elevation of this variant to species level.

D. EVOLUTION 1. Origin of the Allotetraploids

Skovsted (1937) was the first to advance the hypothesis that the natural amphidiploids G. hirsutum, G. barbadense, and G . tomentosum combine on A genome from a taxon of the Asiatic diploid group and a D genome from a taxon of the American diploid group. He showed that the American amphidiploids consist of a large chromosome set homologous with the 13 chromosomes of similar size in the Asiatic (A) cottons, and 13 smaller chromosomes homologoils with the small chromosomes of the American (D) wild diploid species. Independently, Beasley (January, 1940) and Harland (March, 1940) confirmed Skovsted’s hypothesis by synthesizing amphidiploid hybrids from A and D genomic diploids. Each used G. arboreum (A2) X G . thurberi (D1)to show that the synthetic amphidiploids 2 (A2D1)were morphologically and cytogenetically compatible with the natural amphidiploids. In Beasley’s work, synthetic allotetraploids were produced by doubling the chromosome number in hybrids of G. arboreum x G . thurberi, using colchicine (C22H2506N).In the diploid hybrids between the original species, less than half of the chromosomes paired; those bivalents that were present showed bridges at anaphase, which indicated that structural differences existed between all of the chromosomes of the two species (A2 and D1). After doubling the chromosomes to 2(A2DI),Beasley found mostly pairing, but did observe some multivalents with one or more univalents. The hybrid allotetraploid was then crossed with G. hirsutum (A2D1 X AhDh) and the meiosis of this amphidiploid hybrid observed. In some cells

232

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

all the chromosomes were paired, with one to three or more multivalent associations found. The hybrids were also self-fertile, and fertile when backcrossed to G . hirsutum. Using a similar technique as the preceding, Beasley (1942) synthesized hexaploid hybrids of G . hirsutum x G . herbaceum 2(AhDh) X A, and G . hirsutum X G . thurberi 2(AhDh) X D1. Beasley again found meiotic figures in which some of the chromosomes were in complex associations, indicating overpairing. At the same time, hybrids of G. hirsutum x G . anomalum 2(AhDh) X BI and G . hirsutum X G . sturtianum 2(AhDh) X C1 were near normal in chromosomal behavior; thus, any homologies that might exist between the B1 and CI genomes and chromosomes of the G . hirsutum set must be low. These facts left little doubt that the American 26 chromosome cottons were allotetraploids, with one parent species similar to each of an American D and an Asiatic A 13-chromosome species. Brown (1951) demonstrated that doubling of the chromosome number can occur spontaneously in F I hybrids of Gossypium spp. In addition, incompatibility of the amphidiploid with natural species satisfies the requirement of genetic isolation that would perpetuate it.

2. Genome Parentage of the Allotetraploids Having established that the,New World amphidiploids are of the general constitution that combines one Old World diploid and one New World diploid, work was directed toward pinpointing the taxa of the original parents. Hutchinson et al. (1947), citing Stephens (1942), pointed out that multivalent formation in hybrids in which each genome is represented twice is a more sensitive index of homology than is pairing in hybrids in which it occurs only once. In the former case, normal pairing between true homologues is possible, and any overpairing must indicate fairly close homology. However, in the latter case, in the absence of true homology, even very low affinities may result in pairing. a . A Genome Parent. In some species, translocations may be common and even occur in single populations; this is not the case with Gossypium. Brown and Menzel(l952) reviewed the case of Gossypium translocations known within the species. The American wild diploids (D’s) have shown none when crossed among themselves and with the American allotetraploids. With this in mind, Gerstel (1953a) concluded that translocations in Gossypium would be a useful tool in measuring homology in species hybrids. He found that in triploid hybrids of New World allotetraploids x the two Asiatic diploid species, the most frequent type of meta-

DISTRIBUTION AND EVALUATION OF Gomypiurn

233

~ 131 for G . hirsutum x G . herbaceum and 811 + phase 1’s were 911 + 2 1 + 11v + 1v1+ 131 for G. hirsutum X G . arboreum. The cytological evidence thus suggested that G . herbaceum (A,) is more closely related to the allotetraploid New World cotton than G . arboreum (A*),because G . herbaceum has one less translocated chromosome arm relative to G . hirsutum than does G . arboreum. b . D Genome Parent. Allohexaploid hybrids involving Asiatic diploids, New World diploids, and the natural allotetraploid species were synthesized and analyzed cytologically for multivalent frequencies and genetically for segregation of mutant genes located in the A and D genomes. The pairing relationships and genetic ratios observed in the analysis of the hexaploid hybrids involving the D species showed that those involving G . raimondii (DS) most closely approached autotetraploid behavior, which confirms that this species is the most closely related to the D subgene contributor of the AD allotetraploids (Gerstel, 1953b, 1956, 1963; Sarvella, 1958; Gerstel and Phillips, 1958; Phillips and Gerstel, 1959; Phillips, 1960, 1962, 1963, 1964). Thus, the New World amphidiploids (AD) have been shown to be most closely related to G . herbaceum (A,) and G . raimondii (Ds). However, the D genome group is not as closely related as is the A genome group. This fact would indicate that the chromosome structure in G. raimondii probably evolved at a more rapid pace than it has in G . hirsutum and in G . herbaceum. An alternative possibility to this thesis is that there is an extant species of the D genome that has not yet been found or that is extinct. If found, this D, species would then probably be as closely related as is the A genome species to the AD’S (Phillips, 1963). 3 . Theories of the Origin

Having concluded that A, and DS are the closest derivatives of the ancestral parents of the New World amphidiploids, the question remains as to when and how these two ancestral species were brought together so that hybridization could take place. a . Ancient Origin Theories. Harland (1939), noting the occurrence of New World cottons in widely scattered Micronesian and Polynesian islands, proposed that A and D species came together on a trans-Pacific land bridge in the late Cretaceous or Tertiary (about 60 million years ago). Stebbins (1947) proposed that the natural amphidiploids arose in America before the Eccene (about 70 million years ago) and subsequent to the

234

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

migration of the Asiatic type from the Old to the New World via Behringia. Saunders (1961) also proposed a Tertiary origin supposing that the A and D species involved were sympatric in what is now eastern South America before the rift developed that ultimately led to the formation of the South American and African continents. b. Recent Origin Theories. Davie (1935) proposed that the allotetraploids arose by hybridization in the very recent past. Hutchinson et al. (1947) and Hutchinson (1959) suggested a scenario in which G. arboreum cotton was introduced into the New World by humans over a Pacific route. Sherwin (1970) and Johnson (1975) proposed a polyphyletic and recent origin for the allotetraploids. Though the arguments of the latter two researchers differ, their premise is that G. herbaceum was transported to America by people, or resulting from human activity, where it was cultivated and then hybridized with more than one D genome species.

c . Most Accepted Origin Theory. Reviewing the evidence, Phillips (1963) interpreted the cytogenetic data as establishing that the allotetraploids were of monophyletic origin, and that the event occurred during the Pleistocene (about 1 million years ago). Cytogenetic data showed that the A and D subgenomes of the allotetraploids, primarily the A, had diverged little during their evolution from the progenitor genomes represented in the A and D species. This is assumed to be incompatible with an ancient origin, and since the allotetraploidshave differentiatedinto several distinct species it also suggests that they are not recent in origin. Fryxell (1965, 1979) concurred with this interpretation, and in addition stressed the importance of the several pairs of sibling species and the association of endemic allotetraploids with littoral habitats. Recent data using chloroplast DNA analysis support this last hypothesis. Wendel (1989) observed only 12 mutations out of a total of 3920 restriction sites assayed among seven species of A (diploid) and AD (allotetraploid) genome cottons and indicated that the concomitant low estimate of sequence divergence suggests that the initial hybridization and polyploidization events that led to the evolution of allotetraploid cottons were within the estimates made by Phillips and Fryxell, as given earlier. Wendel also suggests “that although formal methods are lacking for the estimation of divergence time from sequence divergence values calculated from restriction site data, substitution rates have been estimated for several chloroplast genes; these are between 0.12% and 0.16% per million years. It is not clear that nucleotide substitutions accumulate linearly over time or at equivalent rates among plant lineages, nor is it known if substi-

DISTRIBUTION AND EVALUATION OF Gossypium

235

tution rates calculated from information on specific gene sequences are applicable to data derived from restriction enzyme analysis of entire chloroplast genomes. If, however, these potential sources of error prove to be unimportant, then the time of origin of allotetraploid cotton may be estimated. Assuming an average sequence divergence rate of 0.14 per million years, and a sequence divergence estimate of 0.20% 4 0.06% (range 0.15%-0.27%), allotetraploid Gossypium would have originated 1.43 f 0.43 million years ago, in the middle Pleistocene” (Wendel, 1989).

Ill. COLLECTION The U.S. National Cotton Germplasm Collection resides in the Southern Crops Research Laboratory, Southern Plains Area, Agricultural Research Service (ARS), United States Department of Agriculture (USDA), in cooperation with the Soil and Crop Sciences Department, Texas A&M University, College Station, Texas. Regional coordination of its activities is under the auspices of the Technical Committee of Regional Research Project S-77. The Collection is part of the National Plant Germplasm System (NPGS) and, as part of this system, all aspects of the preservation and use of the data and physical germplasm are coordinated through the Cotton Crop Advisory Committee (CCAC). The CCAC functions as an advisory group to provide expert advice to individuals and organizations such as the National Plant Genetics Resources Board (NPGRB), the National Plant Germplasm Committee (NPGC), ARS, State Agricultural Experiment Stations (SAES), and others on technical matters related to cotton germplasm, its breeding, and effective utilization. Information on accessions maintained, and the evaluation information of these, is accessible through the Germplasm Resources Information Network (GRIN) computer system, which is part of the Data Base Management System (DBMS), a part of the Plant Genetics and Germplasm Institute, Beltsville, Maryland. A. SOURCE OF THE COLLECTION The Collection presently maintains 5100 seed accessions of the Gossypium spp. This material has been accumulated through the years and represents a significant base of scientific capital from 76 countries and political jurisdictions. The material was obtained from planned explorations to various parts of the world, by donations from individual collectors,

236

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

and by exchanges with other similar international collections, such as the Institut de Recherche du Coton et des Textiles Exotique, France; Central Institute for Cotton Research, India; Instituto Nacional de Investigaciones Agricolas, Mexico; Cotton Research Institute, Pakistan; Institute of Plant Industry, USSR; Germplasm Resources Research Division, People’s Republic of China; and others. The Collection makes available and preserves the broadest possible genetic base for cotton. It provides source material for basic studies in genetics, cytogenetics, taxonomy, and other disciplines, as well as applied studies in screening for resistance to pests, disease, and environmental stress and in plant productivity. Seeds from the Collection are available to cooperators for research studies of various kinds, within and outside of Regional Research Project S-77. However, activities that focus on maintenance and acquisition continue to be the primary objectives in order to preserve the natural variability of cotton as a resource for continued efforts to modify and improve cotton cultivars (Percival, 1987).

B. COLLECTORS Cotton collecting expeditions have been taking place since the turn of the century. The early collections were for the most part to the purported center of variability of G.hirsutum, that is, southern Mexico and Guatemala. These early expeditions were time-consuming and difficult to arrange, because at that time there were only limited travel facilities in the areas explored. As interest in cotton germplasm collecting and preservation has increased, funding for this type of activity has become available, and today these collections are more easily arranged and carried out. Added interest in obtaining and preserving Gossypium germplasm has also increased the scope, not only of the geographic areas explored, but also of the material collected. The following list includes many of the collectors, and the countries explored, who have participated in this activity since the early 1900s (Anonymous 1974; Percival, 1987).

Collectors 0. F. Cook 0. F. Cook and B. T. Jordan G. N. Collins and C. B. Doyle 0. F. Cook and J. W. Hubbard F. M. Mauer and S. M. Bukasov T. R. Richmond and C. W. Manning S. G. Stephens

Date and Country 1902-1904 Guatemala 1905-1906 Guatemala, Mexico 1906-1907 Mexico 1925 Mexico, Colombia, Ecuador 1929 Mexico, Guatemala, Colombia 1946 Mexico, Guatemala 1946-1947 Mexico, Guatemala, El Salvador

DISTRIBUTION AND EVALUATION O F Gossypiurn Collectors C. W. Manning and J. 0. Ware M. Gutierez C. M. Rick, Jr. H. S. Gentry P. A. Fryxell and W. H. Cross G. Ano and J. Schwendiman G. Ano and J. Schwendiman B. B. Simpson and J. Vreeland J. M. Stewart, L. A. Craven, and P. A. Fryxell G. Ano, J. Schwendiman, and A. E. Percival P. A. Fryxell and S. Koch P. A. Fryxell and C. L. Burandt A. E. Percival and J. M. Stewart J. Schwendiman, A. E. Percival, and J. L. Belot J. M. Stewart, L. A. Craven, and P. A. Fryxell A. E. Percival and F. D. Wilson A. E. Percival, J. M. Stewart, A. Miranda, J. Moreira, and E. Freier

237

Date and Country 1948 Mexico, Guatemala 1960-1961 Argentina, Paraguay 1961 Galapagos Islands 19??Mexico 1976 Honduras, Nicaragua 1980 West Indies, French Guiana 1981 Peru 1983 Peru 1983 Australia 1983 Ecuador, including Galapagos Islands 1983 Mexico 1984 Venezuela 1984 Mexico 1985 Caribbean, South Florida

1985 Australia 1986 Galapagos Islands 1988 Brazil

C. PLANTEXPLORAT~ONS Nine of the plant explorations that have taken place during this decade are reviewed next to give the reader an idea of the scope of these operations. Funding for these expeditions was provided by the USDA, ARS, and the UN, FAO, IBPGR. The information cited was gleaned from the exploration reports of these individuals. I . Simpson and Vreeland in Northern Peru, 1983 Beryl Simpson and James Vreeland, the University of Texas, Austin, had planned a botanical collection to Peru in the summer of 1983. Because they would be in the area where G. raimondii is endemic, and because this species was rumored to have become extinct since last being collected, these scientists were contracted to either verify the rumor or collect seeds of the species. Since herbarium collections of the species had been made in 1979 and 1980, it did not seem plausible that the species was extinct. Collecting during the summer of 1983 was difficult because of the excessive and unseasonable rainfall and flooding of 1982/1983 caused by the

238

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

climatic phenomenon known as “El Niiino.” The roads in this part of the world were severely damaged, and many roads following rivers into the Andes were impassable. Nevertheless, they were able to visit known localities of G. raimondii, plus most localities illustrated in Boza and Madoo (1941) but from which specimens had never been collected, and several apparently previously unexplored neighboring areas. The first sightings of G. raimondii were from the Pan-American Highway where it crosses the Chicama River. These plants, growing on the south bank of the river, were presumably in the same locale as reported by Phillips and Stephens in their unpublished technical report of 1966. From this highway westward along the old road to Cartavio, they found several patches of plants. The plaAt population exhibited variable age distribution. After working the lower Chicama, where they also found several populations of G . barbadense, they traveled from Casca NNW toward Santa Ana. Passing the crest separating the drainage of the Cascas and Santa Ana rivers, they encountered extensive populations of G. raimondii in a region called Pampa Larga (ca. 950 m elevation). Hundreds of plants were seen growing along the valley and continued along the rocky rubble of the Santa Ana River. Descending to 800 m they also found a large population in the Santa Ana Valley. Plants became sparse as the elevation decreased. Short excursions were also made up the Cupinsque River. Because elevations above 300 m were never reached, no evidence of G. raimondii was seen. Traveling to the Huertas River Valley, they found G . raimondii, with populations observed on both sides of the road, starting about 1 km south of Chilete and ending before the town of Huerta. A trip up the Zana River Valley proved fruitless. In the Department of Cajamarca in the province of Hualgayoc, G. raimondii had been previously collected. Nanchoc, a small village that lies along the Zana River, was reached with the aid of a helicopter provided by the Peruvian Military. No G. raimondii was found, and their report indicates that the habitat is such that it is unlikely to grow in the immediate vicinity.

2. Stewart, Craven, and Fryxell in Australia, 1983 A 1983 expedition to Western Australia, conducted by James McD. Stewart, Lyn A. Craven, and Paul A. Fryxell, was a collaborative project supported by the USDA through ARS and the National Plant Germplasm Unit, and by the Commonwealth Scientific and Industrial Research Organization (CSIRO) through the Australian National Herbarium. The timing of the trip was set to correspond to the usual period when the Gossypium species of the area have matured some capsules but the plants have not desiccated. The primary purpose of this trip was to document, as far as

DISTRIBUTION AND EVALUATION OF Gossypium

239

possible, the extent of variation within and among Gossypium species of the region and to obtain seeds representative of that diversity for the U.S. Germplasm Collection. Seeds of G. hirsutum, G . australe, G . cunninghamii, G . philosum, G . populifolium, and G . pulchellum were collected during this expedition. It was apparent to the participants that there is extensive diversity in the wet-dry tropics of Australia. The diversity has only begun to be measured because the remoteness of the area makes the logistics of collecting difficult. A previous trip to the area by Stewart suggested that the diversity of the cotton genus was greater in the Kimberley Region of Australia than was previously realized. Collections made on this trip confirmed that the taxonomic understanding of the Gossypium of the area is not complete. Specimens taken at 10-km intervals along the length of the Mitchell Plateau will be useful in deciphering an apparent cline that occurs there. The Gossypium collections north of the Carson River are distinctive and may represent undescribed species. However, the specimens do have similarities to know taxa and will require detailed study to determine their taxonomic position. At the very least, they represent previously unknown variation that will require accommodation in current species descriptions. The location of G . cunninghammii in the Northern Territory of Australia appears disjunct from the species of the Kimberly to which it is related. Quite likely, additional Gossypium diversity will be discovered in these areas once they are penetrated by botanists, as was the case for the Kimberley, where each new area visited yielded something different. 3 . Schwendiman, Ano, and Percival in Ecuador, 1983

Jacques Schwendiman and George Ano, of the Institut de Recherches du Coton et des Textiles Exotiques (IRCT), France, and A. Edward Percival (USDA, ARS), United States, participated in a collecting expedition to Ecuador, including the Galapagos Islands, which was supported by the UN, FAO, IBPGR. They were joined by Andres Brando of the Instituto Nacional de Investigaciones Agropecuarias (INIAP), Ecuador, for part of the collecting. The first phase was collecting in continental Ecuador and began with the intent to travel from Quito south to Puyo in Pastaza Province. This travel was not possible as the road was temporarily blocked by landslides due to the unseasonable weather mentioned earlier. The participants then headed straight south to Azuay and Loja provinces, where the first cotton was collected in Loja. From southern Loja, travel was northwest to El Oro, then north to Guayas, Los Rios, and Manabi provinces. In Loja, El Oro, and Manabi, with few exceptions, only dooryard G. barbadense was

240

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

collected. However, in Guayas and Los Rios, large populations of endemic wild G . barbadense were found. One of the main problems encountered in all of the areas explored was that much of the cotton was not open. It was apparent that the unseasonable rains had greatly delayed boll maturity. On the Galapagos Islands, G . barbadense was found on San Cristobal. Gossypium darwinii was found on Santa Cruz, Eden off of Santa Cruz, Floreana, Espanola, Gardner off of Hood, San Cristobal, and Rabida. Gossypium klotzschianum was found on Santa Cruz and San Cristobal and, as previously reported, it was found growing intermingled with G. hrwinii in extensive populations of both species. The new accessions have added to the germplasm diversity of the collections represented, and the large number of accessions collected from the Galapagos Islands should aid in clarifying questions that have been raised concerning the elevation of G. darwinii to a species level. 4 . Fryxell and Burandt in Venezuela, 1984

Paul A. Fryxell (USDA, ARS) and Charles L. Burandt (Texas A&M University) undertook this collection from January to February, 1984. A rough itinerary of the route followed was Maracaibo, Coro, Maracay, Barquisimeto, Guanare, Merida, Caracas, and back to Barquisimeto and Maracaibo. Collections of cotton seeds were made in natural vegetation, on roadsides, and in dooryards, from sea level to as high as 1800 m elevation. Most of the samples collected were of G . hirsutum, but two dooryard G . barbadense were also found. Considerable variability was found among the collections of G . hirsutum. Many samples were collected opportunistically as they were encountered; others were specifically sought out on the basis of prior information, especially the wild cottons occurring in natural vegetation, sometimes in remote places along the northern coast. The cottons collected were found to be in all stages of development: a few in full foliage and in early stages of flowering with no mature fruits, others still flowering but with both green and open bolls, still others past flowering, and some plants that were merely dry sticks lacking any foliage but with a mature crop of open bolls. The wild cottons observed in natural vegetation formed large but locally restricted populations. There was often one or a few parent plants of apparent great age in each population. Some of the cottons exhibited characteristics (e.g., short brown fiber, small flowers, and fruit) that set them apart from the dooryard and roadside cottons. In the opinion of the collectors, these wild cottons are an indigenous part of the vegetation and not escapes from cultivation.

DISTRIBUTION AND EVALUATION OF Gossypium

24 1

5 . Percival and Stewart in Southern Mexico, 1984 A 1984 cotton collection by A. E. Percival and J. McD. Stewart (USDA, ARS), during the month of September, was a collaborative project with the Secretaria de Agricultura y Recursos Hidraulicos, Instituto Nacional de Investigaciones Agricolas (SARH, INIA), Mexico, represented by Arturo Hernandez and Fernando de Leon. The rough itinerary followed was: Brownsville, Texas, south through the states of Tamaulipas and Veracruz, east to Tabasco, northeast and around the Yucatan Peninsula to Chetumal and Quintana Roo, south through Chiapas, west to the Isthmus of Tehuantepec, and back north to Texas. Seeds were collected of dooryard (one atypical) and wild strains of G. hirsutum, one G . barbadense, and one G . cf. aridum. The only truly wild G . hirsutum cottons collected were G. hirsutum var. yucatanense, from the northern coast of Yucatan. The distribution, growth habit, and morphology clearly indicated that these are wild and well adapted to the ecological niche in which they were found. Interestingly, no dooryard cottons could be classified as yucatanense. Likewise, the majority of feral cottons were associated with human settlement and were of types similar to the dooryard cottons. Just as important as the sites where cotton was found were the observations in areas where cotton was not found. This is the classic story of germplasm loss. The town of Acala, Chiapas, and the valley in which it is located were specifically visited because it was the site of collections of the original germplasm that gave rise to the outstanding Acala cultivars. They found no cotton there, nor at any locations near the road that runs along the length of the valley. One individual in Acala related that promoters from Tapachula, Chiapas, tried to establish commercial cotton production in the area. When insects became a problem, the promoters recommended that all native cotton plants be destroyed to better control the insects. The commercial venture subsequently failed and the collectors found no cotton being grown there today. 6 . Stewart, Craven, and Fryxell in Australia, 1985

As with the 1983 expedition to Australia, the participants in 1985 were James McD. Stewart and Paul A. Fryxell (USDA, ARS) from the United States and Lyn Craven from Australia (Australian National Herbarium, CSIRO). This exploration was based on funding from IBPGR, USDA, and CSIRO. This plant exploration to central, northern, and northwestern Australia collected samples of most of the 12 currently recognized taxa from Austra-

242

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

lia and 8 additional Gossypium variants. These findings have provided new germplasm for study and exploitation, and also made plain the need for taxonomic reinterpretation of the Kimberley cottons. This exploration significantly extended our knowledge of the geographic range of the Australian wild cottons and their range of variation, as was the case with the 1983 collection. The collectors found that in the arid zone of Central Australia, G . nelsonii occurs sympatrically with three other species of Gossypium, including G . australe, with which some workers confuse it. It was established (Stewart et al., 1987) that the two species are indeed distinct in the field and that G . nelsonii occurs over a much wider geographic area than the one site previously reported. The wild Kimberley cottons had previously been allocated to five species: G . costulatum, G . populifolium, G . pilosum, G . pulchellum, and G . cunninghamii. Moreover, they previously were thought to have relatively isolated distributions within the region. It is clear from the results of this collection that this is not an adequate representation of the actual situation, and the descriptions of six new species resulting from this exploration are currently in preparation by the individuals named. In the Kimberley region, Gossypium was found to be far more widespread, abundant, and more variable than previously recognized. An exception is G . cunninghamii, which occurs on the Cobourg Peninsula, outside the Kimberley, and thus is isolated from the others. However, even this species was found to be more widespread and abundant than previously known. The climate in which these species are found is tropical with alternating wet and dry seasons, and the plants are long-lived perennials that have adapted to a fire-mediated ecology by regrowing annual stems from woody rootstocks. In the absence of fire for one or more years, the stems occasionally survive the dry season and persist, especially in the erect-growing type of plants. The variability among the many populations sampled appeared to be complex, with the morphological characters recombining in various ways. It seems clear that this group of wild cottons is in an early and active stage of speciation. Thus, this exploration provided materials to begin an analysis that will lead to a more satisfactory interpretation of the variability and the recognition of newly discovered species. 7 . Schwendiman, Percival, and Belot in the Caribbean, 1985

A 1985 exploration, financed by IBPGR, included the same IRCT and USDA, ARS personnel of the 1983 exploration to Ecuador mentioned previously. It was conducted from the last of February to the first of April

DISTRIBUTION AND EVALUATION OF Gossypium

243

and included the following localities listed in the order collected: Trinidad and Tobago; Curacao, Bonaire, and Aruba (Netheland Antilles); Jamaica; Grand Cayman (British West Indies); southern Florida (United States); the Dominican Republic; and Puerto Rico. The period for collecting seeds was optimal, as with few exceptions the cotton found was in the late flowering, open mature boll stage. Accessions of dooryard, feral, and wild G . hirsutum and dooryard G . barbadense were collected. The distribution, growth habit, and morphology of the wild cottons indicated that they are truly wild and adapted to the ecological niches in which they were found. Wild types were found on Curacao, Bonaire, Jamaica, southern Florida, the Dominican Republic, and Puerto Rico. The feral and dooryard types were found on all the islands and in Florida and were associated with human settlements or disturbances. With the exception of those populations that appeared to be of a truly wild nature on the islands of Curacao and Bonaire, the cottons found appeared to be plentiful and in no danger of being eliminated. However, the one wild population on Curacao and the one on Bonaire could be lost to developments in the areas where they are established. 8. Percival and Wilson in the Galapagos Islands, 1985 A 1985 exploration by A. E. Percival and F. D. Wilson (USDA, ARS) was conducted to collect the western and northern islands of the Galapagos Archipelago that had not been collected during the 1983 expedition to these islands. This collection was a collaborative project with Instituto Nacional de Investigaciones Agropecuarias (INIAP), Ecuador, which was represented on the exploration by Gelasio Basante. The following islands were collected and explored during September: Santa Cruz (Indefatigable)-Puerto Ayora, road from Las Gemelas to Baltra crossing, and Turtle Beach; Marchena (Bindloe)-Black Beach and Point Mejia; Pinta (Abingdon)-Cape Chalmers and north of Cape Chalmers; Isabela (A1bemarle)-Point Vincente Roca, Banks Bay, Black Cove, Tagus Cove, Urvina Bay, Elizabeth Bay, Iguana Cove, San Pedro Cove, and the road from Villamil to Santo Tomas; and Fernandina (Narborough)-two locations between Point Espinosa and Cape Douglas. Gossypiurn darwinii was collected from Santa Cruz, Marchena, and Isabela, and G. klotzschianum was collected from Santa Cruz and Isabela. The G . klotzschianum collected from Isabela was unique, as it had not been found on this island during previous expeditions. It was found at two locations, and in each only a few small plants were seen growing and only a few seeds

244

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

were gathered. It was not possible to determine if there might be other larger populations of the species on this island or whether the few plants growing resulted from being recently introduced. 9 . Percival and Stewart in Brazil, 1988 A 1988 USDA, ARS exploration by A. E. Percival and J. McD. Stewart to northeast Brazil during the month of September was a collaborative project with the Centro Nacional de Recursos Geneticos, Empresa Brasileira de Pesquisa Agropecuaria (CENARGEN, EMBRAPA), Brazil. Antonio Miranda, Jose de Alencar, and Elusio Freire represented EMBRAPA. The area collected involved portions of the states of Bahia, Ceara, Pernambuco, Piaui, and Rio Grande do Norte. With the exception of a small area on the northern coast around Touros, Rio Grande do Norte, all of the area collected is a tropical semiarid region with a wet-dry season. Seeds of G. hirsutum, G. mustelinum, and G . barbadense were collected. The endemic allotetraploid wild species G. mustelinum was collected at four sites from where it had previously been reported (Pickersgill et al., 1975) and from two new sites. Except for variation in the ages of some of the plants at each site, little morphological variation was noted, and all of the sites were next to or near water drainages, indicating that the species has adapted to take maximum advantage of the limited rainfall of the area. It may have been indeed fortunate that this collection was made at this time. The area collected is an area almost exclusively devoted to the production of “Moco” cotton, with limited corporate production. Moco cotton (G. hirsutum var. marie-galante) is morphologically variable and has characteristics suggesting introgression from G. barbadense and G . mustelinum. Moco is grown as a perennial and plants are ratooned (cut back) each season. Once fields are established, planting involves only replacement of plants that may have died. Moco growers are mostly small farmers who grow the crop with limited or no technical agricultural input. Many of the fields are established on rocky hillsides, almost exclusively adaptable to a crop such as Moco cotton that can survive the dry season, where the plants can grow among the rocks and boulders. There are native insect pests, such as boll worms, that damage the crop, but not to the extent that it was not economical to grow. However, with the invasion of the area in 1985 by the boll weevil (Anthonomous grandis Boheman), production in parts has been so reduced that it is no longer economical to harvest what little crop is produced. Breeding schemes are under way by EMBRAPA personnel to reduce the impact of this insect,

DISTRIBUTION AND EVALUATION OF Gossypium

245

some of which involve developing early and/or resistant Moco-type cultivars. Given the environmental conditions and nature of agricultural practices of cotton production in the area, it remains to be seen whether or not this will succeed. Regardless of the outcome of the breeding efforts to produce boll weevil-resistant and adapted varieties, the germplasm base of the material presently grown will change in the not to distant future. Moco cotton will either be eradicated in parts or all of the area, or adaptable varieties will be successfully produced with introduced germplasm from other G . hirsutum types. In either case, this will permanently alter the present Moco germplasm base of the area. It is satisfying to note that this collection secured representative cotton germplasm from this area of the world. Some of this material may in future prove valuable as it appears to be variable for many lint quality and agronomic characters.

IV. EVALUATION A. GENETICS A N D CYTOLOGY

1 . Qualitative Mutants Studies dealing with the genetic mutants found in Gossypium have tended to reflect the commercial utility or importance of the particular species under scrutiny. Though there are exceptions, these studies have concentrated within the cultivated species and only involved the wild species when utilitarian characters have been found that have some degree of benefit for the improvement of the cultivated species. Within the cultivated species, the degree of interest has also been determined by the commercial importance of each of these. Thus, the progress made in identifying mutant genes has been greatest in the order G . hirsuturn, G . barbadense, and G . arboreurn. Kohel (1973) has addressed the genetic nomenclature for Gossypium and Endrizzi et al. (1984) give a current listing of the genetic mutants and linkage relationships for the genus. Endrizzi et al. (1984) list 112 mutant genes in the 26-chromosome allotetraploids, with most of these being identified in G. hirsutum. Some of these have been transferred to G . barbadense, and some have been transferred from the wild allotetraploids, diploid, and cultivated diploid species. Sixty-one mutant loci have been identified with 16 linkage groups. Eleven of these linkage groups have been identified with chromosomes, 2 have been associated with specific subgenomes, and 3 remain to be associated. In the diploid 13-chromosome Asiatic cottons, 86 currently recognized

246

A. EDWARD PERCIVAL AND RUSSELL J . KOHEL

mutant loci are listed. They also point out that the last summary of linkages in the Asiatic cottons was that made by Knight (1954), who lists 17 mutant genes with 7 linkage groups. Of these 7 linkage groups, 2 are associated with 2 linkage groups in the allotetraploids. 2 . Translocations

A translocation occurs when two nonhomologous chromosomes reciprocally exchange pieces of their chromosomes. They are useful for assigning genes to specific chromosomes and for generating duplication deficiencies. They have also been used to detect incipient differentiation between the subgenomes of the A and D allotetraploid species, principally in G. hirsutum and G . barbadense (Menzel and Brown, 1952; Menzel et al., 1978, 1982; Brown, 1980). 3 . Monosomes

Plants that are deficient for one chromsome of a pair are referred to as monosomics. Unlike diploids, allopolyploids such as G. hirsutum can transmit the haplo-deficiency to subsequent generations. Thus, chromosome pairing in monosomic plants of allotetraploid cottons will normally consist of 25 bivalents and one univalent, with 26 different monosomics theoretically possible. Thus far, 15 of the 26 chromosomes of G. hirsutum have been identified. Because of the chromosome imbalance, a syndrome of morphologicalcharacters is associated with each monosomic type so far identified, which can be used to easily identify the monosomic plants from their disomic sibs. Some monosomic plants can even be identified in the seedling stage (Endrizzi and Ramsey, 1979; Endrizzi et a f . , 1984). Monosomes are used in identifying the chromosomes in translocations and have been used extensively in assigning genetic factors to specific chromosomes. They are ideal for separating duplicate linkage groups so that gene distances or linkage values for each can be determined. They are used to create chromosome substitution lines, and these lines in turn can be used to study the genetic effect of individual chromosomes on plant traits and for estimatingthe number of genes, gene interactions, and linkage relationships controlling plant characters (Endrizzi et al., 1984). 4 . Monotelodisomes and Monodisomes

Telocentric chromosomesare created by the misdivision of univalents at the first or second meiotic division and are recovered as monotelodisomic

DISTRIBUTION AND EVALUATION OF Gossypiurn

247

plants. A monotelodisomic plant has 25 bivalents, plus a heteromorphic bivalent. The heteromorphic pair consists of a telocentric chromosome plus an entire homologous chromosome. Once the association of a marker gene to a specific chromosome has been made by the monosome test, the monotelodisomic can then be used to determine the arm location of that marker gene as well as its linkage distance to the centromere. Isochromosomes are chromosomes with two homologous or identical arms, and they arise by misdivision of univalent, plus misdivision of telocentric, chromosomes. They may also be used to determine gene location on chromosome arms. However, these are not as useful as telocentrics for gene mapping studies, as they may give faulty rates of recombination frequencies because of exchanges in the trisomic arms (Endrizzi and Kohel, 1966; Endrizzi and Bray, 1980; Endrizzi et a f . , 1984).

B. ELECTROPHORESIS I . Protein and Isozyme Analysis Electrophoresis began to be employed in systematics in the early 1970s (Avis, 1975). As the reader is aware, the advent of electrophoresis (Ornstein, 1964) and its application in the struggle to measure variation in organisms bypass the problems of sterility barriers and incompatibility. Its application within the genus Gossypium was first employed by two groups, B. L. Johnson and M. M. Thein at Riverside, California, and J. P. Cherry, F. R. H. Ketterman, and J. R. Endrizzi at Tucson, Arizona. Both groups employed proteins extracted from Gossypium seeds, the premise being that proteins from dormant tissue reflect a more stable genomic state. The reserve proteins of seed form a class in which vital processes are not canalized and are buffered against environmental shock. Thus, these proteins should provide a better sample of those loci that may vary without impairment of survival ability (Cherry et al., 1970; Johnson and Thein, 1970). By using seed, populations can be sampled easily and quickly, and seed is less affected by variation in nutrition and environment during development (Dunnill and Fowden, 1965). Johnson and Thein (1970) extracted protein from seed of 25 diploid species of Gossypium for analysis and surveyed the banding patterns of these samples. In general they found that these species fell into the genomic groups to which they had been assigned on morphological and cytological evidence. However, they suggest that the evidence presented (correlation of banding patterns) should split the D genome group: one subgroup (DB) comprising the species G . ruimondii, G. lobatum, G . aridum, G . laxum, and G . gossypioides; and the other subgroup (DE) the

248

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

species G . thurberi, G . trilobum, G . davidsonii, G . klotzschianum, G . armourianum, and G . harknessii. Their conclusion is that the DB subgenomic group evolved from B genomic group and that the DE subgenomic group evolved from the E genomic group, and that all genomic groups are derived from a common primordial population occupying Central Africa prior to continental drift. Further work by Johnson (1973, again using banding patterns obtained from crude seed proteins of allotetraploid cottons that included G . hirsutum f. palmeri, elevates “palmeri” to a species level and fits a postulated parentage of G . herbaceum (A,) and G . trilobum (Ds). Johnson (1975) also indicates that G . hirsutum originated from a combination of “ G . palmeri” and one or more other AIAIDSDstype, whereas G . barbadense is the true descendant of G . herbaceum x G . raimondii (AIAID~Ds). In addition, the cultivated varieties of G . hirsutum represent various degrees of introgression involving G . barbadense and the preceding G . hirsutum complex. Cherry et al. (1970) extracted crude proteins from seeds of 26 species and 10 varieties of the genus Gossypium. They found that the species fall into the genomic groups to which they had been previously assigned on geographical, morphological, and cytological grounds. They found the greatest variability within the D genomic group, but no pattern that would allow the splitting of it into two subgroups, as suggested by Johnson and Thein (1970). Within the groups, pairs of “sibling species” were supported by their analysis, similar to those pointed out by Fryxell(1965) to exist in all genomic groups. Further analysis by Cherry et al. (1971, 1972) used three enzyme systems (esterases, leucine aminopeptidases, and catalases). The results reported again support, in broad outline, the present classification of the genus. The major value of this early work using gel electrophoresis of plant proteins may be in having introduced the technique to the genus. Questions persist concerning the origins of Gossypium, and though there is both support for the present classification and possible evidence for some reclassification, the early application of the method is questionable. These early works employed the gel electrophoresis technique in a gross manner to measure qualitative differences, and the application of the technique in this manner may have value only in a supportive role, for the conclusions are contradictory. When contradictory conclusions are reached, using the general application of the method, judgment should be reserved until specific protein surveys can be conducted as outlined by Lewontin (1974). Only recently have improved methods using this technique been employed within the genus. The suspected close relationship of G . davidsonii and G . klotzschianum

DISTRIBUTION AND EVALUATION OF Gossypiurn

249

was confirmed through allozyme (different enzyme forms produced by different alleles at the same locus) analysis of 33 populations for 41 loci, indicating that the allelic composition of G . klotzschianum represents a subset of G . dauidsonii, thus making the former a derivative species of the latter (Wendel and Percival, 1989). One hundred and three accessions of G . arboreurn and 3 1 of G . herbaceum were examined for allelic variation at 40 allozyme loci. Expectations, based on morphology and other chemical data, were that the two species would show a close relationship. However, the results indicated that these two species are highly differentiated, at least with respect to their allozyme composition (Wendel et al., 1989). Allozyme analysis was also performed on 153 accessions representing the spectrum of G . barbadense diversity. Materials from northwest South America contained the greatest diversity, suggesting that this region is the ancestral home of this species. The data also indicate separate diffusion pathways from the center of origin into southern, eastern, and northern (east of the Andes) South America. The Caribbean island and Central American forms appear to be derived from the northern forms. These pathways support previous historical determinations and fit the morphological evidence. The Pacific island forms have closer relationship to the accessions from eastern South America, which is opposite to their geographic proximity (Percy and Wendel, 1989). A study conducted by Saha and Stelly (1987) to develop suitable isozyme techniques for Gossypium spp. and to survey the genetic variation used profiles of five isozymes among 60 allotetraploid cultivars. Results of this study indicated that there was considerable variation between G . hirsutum and the other allotetraploid cultivars used, suggesting that interspecific genetic introgression from other allotetraploids would improve the genetic diversity of Upland cotton. Using this same technique, a further study by Saha et al. (1988) again revealed the close phylogenetic relationship between G . hirsutum and G . lanceolatum ( G .hirsutum f. palmeri) and that the relationship between G . barbadense and G . darwinii, though more distant, nevertheless was close enough to question elevating the latter to a species level. 2 . Endonuclease Restriction Analysis

Organelle DNA analysis is also valuable in phylogenetic studies associated with the evolution of Gossypium, because organelle DNA is highly conserved, making it comparable to asexual genomes without recombination (Takahata and Slatkin, 1983). Chloroplast DNA has been used in Gossypium systematics.

250

A. EDWARD PERCIVAL AND RUSSELL J . KOHEL

Wendel (1989) has reexamined the evolution of allotetraploid cottons using variation in restriction enzyme cleavage sites in the chloroplast genome (cpDNA). The data obtained suggest that the hybridization of ancestral diploid species to form the allotetraploid species is within the estimates suggested by earlier workers and occurred “relatively recently” within the last 1-2 million years. Chloroplast DNA analysis used by Altman and Thomas (1985) indicates that the elevation of G . lanceolatum to species rank is not supported from the results; however, the elevation of G . darwinii to species rank is. Results comparing the banding patterns of the diploids G . arboreum and G . herbaceum indicated a very close relationship of these two species, which is contrary to results mentioned previously. Similar chloroplast DNA analysis of eight species (two allotetraploid and six diploid) by Wilkins and Galau (1985) indicated that their results were consistent with the current understanding of Gossypium evolution. They also found that the allotetraploids appear to have diverged earlier than their nuclear cpDNA’s, suggesting that introgressive hybridization may have occurred in their evolution. All of these studies point to the fact that continued use of the various applicationsof gel electrophoresis will greatly aid in the correct determination of the variation that exists within Gossypium and the systematics of the genus. C. IMPROVEMENT 1 . Sources of Variability

a . Wild and Diploid Germplasm. Nonfiber-producing cottons include most of the wild diploid species of Gossypium. Seeds of some of these species have hairs, but none bear usable or spinnable fiber. The seed hairs that may be present are too short and too firmly attached to the seed to be of any potential utility. Being diploids, these species are also too distantly related to cultivated allotetraploid cotton to be directly useful in conventional breeding programs. The fiber-producing cultivated Asiatic diploids also fall into this category. Nevertheless, they are potential sources of useful genes that have been, and can be, transferred to cultivated cottons using special techniques (Stewart and Hsu, 1978; Stewart, 1979). b . Wild and Allotetraploid Germplasm. The fiber-producing cottons include the two cultivated allotetraploid species G. hirsutum and G . barbadense, and may also include G . lanceolatum. The inclusion of G . lan-

DISTRIBUTION AND EVALUATION OF Gossypium

25 1

ceolatum is questioned, as the elevation of this variant to species level from G . hirsutum has been questioned as needing experimental verification (Endrizzi et al., 1984). The other three allotetraploid species that are a source of potential germplasm are the wild G . tomentosum, G . darwinii, and G . mustelinum. The cultivated species have a wide range of variability in terms of cultivars, strains, feral types, and genetic mutants, followed by G . darwinii, which has less variability and is limited to its geographic distribution on each of the Galapagos islands (Fryxell, 1984). The remaining two species, G . tomentosum and G . mustelinum, have little observed variability, probably because few accessions of these have been collected, and because they have only been found in limited geographic locations. The transference of desirable characters between the allotetraploids is more straightforward, but it is also difficult. Hybrids between the allotetraploids break down in the F2 generation. The viable offspring tend to assimilate back to each of the two parent types, with the true recombinants being weak or not able to survive, thus large populations are required to transfer the desired character. c . Germplasm Modification. As stated previously, much of the cultivated cotton acreage grown throughout the world is in the temperate zone. Thus cotton, which is native to tropical and semitropical areas of the world, has had to be changed to, or selected for, photoperiodic neutrality. Since most of the wild species, and many of the primitive and/or feral forms of the cultivated species, fail to flower in a long-day regimen, it is necessary to circumvent this problem if one is to use a crossing program in plant improvement. This may be accomplished by (a) crossing in a tropical environment where day length is not a problem, (6) crossing in a greenhouse during the short-day winter months, or ( c ) introducing genes for day-neutralism into the germplasm accessions with a backcross scheme so that the genotype is disturbed as little as possible, thus making the feral cotton suitable in temperate areas. However, the ease of transference of desirable traits is related in part to the mode of inheritance of the desired trait, and in part to the closeness of the relationship of the materials involved (Fryxell, 1984). d . Germplasm Utilization. The introduction of desirable germplasm into agronomically acceptable cotton cultivars is an ongoing and dynamic enterprise in most cotton breeding programs. However, the transfer of desirable characters from exotic intraspecific and interspecific sources, though continuous, has primarily been done in state and federal breeding programs. Some examples of these have been reviewed by Fryxell(l976, 1984) and are summarized here.

252

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

Cotton rust, Puccinia cacabata Arth. & Holw., is a disease that occurs intermittently in northern Mexico and the southwestern United States. L. M. Blank of Arizona developed an inoculation technique that permitted him to survey a wide range of cotton germplasm for resistance to this disease. He found no resistance in G. hirsutum, but found strong resistance in the wild species G. anomalum and in the cultivated G. arboreum. By creating synthetic hexaploids and allotetraploids, and backcrossing these to G. hirsutum, he created acceptable agronomic lines that were resistant to cotton rust. Bacterial blight (Xanthomonas maluacearum) resistance is conditioned by a multiplicity of genes (16 have been isolated) as the disease itself is variable and exists as differing races with varying degrees of virulence. R. L. Knight isolated blight resistance from G. hirsutum, G. barbadense, G . arboreum, G . herbaceum, and G . anomalum (Brinkerhoff, 1970). However, the most blight-resistant genes were found within the species G. hirsutum, though as many as 16 backcrosses were needed to recover the acceptable fiber properties of the recurrent parent. Fiber strength is a polygenic character, and high fiber strength comes from a variety of sources, principally by introgression from the primitive cultivar Hopi and from a hybrid made by J. 0. Beasley in the 1930s between the diploid species G. thurberi and G . arboreum. Doubling the hybrid to the allotetraploid level allowed it to be crossed and backcrossed, with relative ease, to G. hirsutum. Cotton varieties that are hairy impart resistance to insects like the jassids (Empoasca spp.), which are important pests in Africa and parts of Asia. The presence in the plant of the single major gene T I (Lee, 1985) is responsible for this desired phenotype. Conversely, the single major gene TISm controls the smoothleaf character and would be beneficial to have in varieties where dense pubescence is not desirable. Varieties with smooth leaf characteristics help control insets that require plant hairs for egg laying. Okra leaf shape is conditioned by the gene Lo, which is desirable in areas that have relatively high humid conditions as harvesting is approached. This leaf type has been found to reduce losses from boll rot organisms and effects earlier maturity because of the more open plant canopy in varieties that have this leaf characteristic (Jones, 1970). Varying degrees of pest resistance and/or plant modification have been obtained using other monogenic inherited characters such as red plant color, bract genes, nectariless genes, leaf shape genes, and dwarf genes and polygenic characters that control plant allelochemistry , fiber properties, water use efficiency, nematode resistance, and boll types (Fryxell, 1976; Niles, 1980).

DISTRIBUTION AND EVALUATION OF Gossypium

253

The potential for the economic development of hybrid cotton cultivars was made possible by Meyer (1973a, 1973b, 1975). She showed that the combination of the G. hirsutum nuclear genome with G. harknessii cytoplasm produces male sterility (A line); fertility could then be restored to male-sterile lines by introducing either one dominant gene or a homozygous recessive gene introgressed from G. harknessii into G . hirsutum (R line). Pure-breeding male-fertile lines, which restore male fertility to malesterile lines, can be isolated from certain stocks with G. harknessii cytoplasm. The male-sterile lines are maintained by crossing with normal, fertile G. hirsutum stocks (B lines) (Niles and Feaster, 1984).

V. CONCLUDING REMARKS Cotton is of enormous importance to the world today. It is not only important economically in international trade, but is also used to clothe a substantial portion of the world’s population. Cotton is both comfortable and utilitarian in nature. As a natural fiber and feed source, cotton is a renewable agricultural resource, which may help to keep it competitive with synthetic fibers from an environmental and ecological standpoint. Continued research with Gossypium germplasm is essential, as this is a complex genus. Though much has been accomplished in the understanding of the genus, much remains to be done. The genus contains differing ploidy levels that yield a high degree of variability, from highly improved allotetraploid species to wild diploid forms, and this variability has only begun to be tapped as a source of beneficial characteristics. In addition, much variability may yet be found. A new species has been discovered every two to three years since the late 1950s, and as old areas are explored again, because of better access, and new areas become accessible, more species are likely to be found that provide additional variability and a better understanding of the genus.

REFERENCES Alefeld, F. G . C. 1862. Osterr. Bot. Zeitschr. 12, 144-148. Altman, D. W., and Thomas, M. D. 1985. Proc. Beltwide Cotron Prod. Res. Conf., New Orleans, Louisiana, pp. 47-49. Anonymous. 1968. Southern Coop. Series Bull. no. 139. Anonymous, 1974. USDA Rep. ARS-H-2. Anonymous. 1988. “Agricultural Statistics 1988.” United States Department of Agriculture, Washington, D.C.

254

A. EDWARD PERCIVAL AND RUSSELL J . KOHEL

Avis, J. C. 1975. Syst. 2001.23,465-481. Beasley, J. 0. 1940. Am. Nat. 74,285-286. Beasley, J. 0. 1942. Genetics 27,25-54. Boza, B., and Madoo, R. M. 1941. Ministerio de Fomento. Dir. Agric. Ganad. Bol. 22, 1-29. Brinkerhoff, L. A. 1970. Annu. Rev. Phytopathol. 8,85-110. Brown, H. B. 1958. “Cotton.” McGraw-Hill, New York. Brown, M. S. 1951. Evolution 5,25-41. Brown, M. S. 1980. J. Hered. 71,266-274. Brown, M. S., and Menzel, M. Y. 1952. Bull. Torrey Bot. Club 79, 110-125. Cherry, J. P., Katterman, F. R. H., and Endrizzi, J. E. 1970. Evolution 24(2), 431-447. Cherry, J . P., Katterman, F. R. H., and Endrizzi, J. E. 1971. Can. J. Genet. Cytol. 13, 155-158. Cherry, J. P., Katterman, F. R. H., and Endrizzi, J. E. 1972. Theor. Appl. Genet. 42, 2 18-226. Crawford, M. D. C. 1948. “The Heritage of Cotton.” Fairchild, New York. Davie, J. H. 1935. Genetica 17,487-498. Dunnill, P., and Fowden, L. 1965. Phytochemistry 4,933-944. Edlin, H. L. 1935. New Phytol. 34, 1-143. Endrizzi, J. E., and Bray, R. 1980. Genetics 94,979-988. Endrizzi, J. E., and Kohel, R. J. 1966. Genetics 54,535-550. Endrizzi, J. E., and Ramsey, G. 1979. Can. J. Genet. Cytol. 21,531-536. Endrizzi, J. E., Turcotte, E. L., and Kohel, R. J. 1984. I n “Cotton” (R. J. Kohel and C. F. Lewis, eds.), Agronomy, No. 24, pp. 81-129. American Society of Agronomy, Madison, Wisconsin. Fryxell, P. A. 1965. Adv. Front. Plant Sci. 10,31-56. Fryxell, P. A. 1968. Bot. Gaz. (Chicago) 129(4), 296-308. Fryxell, P. A. 1976. USDA Rep. ARS-S-137. Fryxell, P. A. 1979. “The Natural History of the Cotton Tribe.” Texas A & M Univ. Press, College Station, Texas. Fryxell, P. A. 1984. I n “Cotton” (R. J. Kohel and C. F. Lewis, eds.), Agronomy, No. 24, pp. 27-56. American Society of Agronomy, Madison, Wisconsin. Gerstel, D. U. 1953a. Evolution 7,234-244. Gerstel, D. U . 1953b. Genetics 38,664-665. Gerstel, D. U. 1956. Genetics 4 1 , 3 1 4 4 . Gerstel, D. U. 1963. Second Int. Wheat Genet. Symp. Hereditas Suppl. 2,481-504. Gerstel, D. U., and Phillips, L. L. 1958. Cold Spring Harbor Symp. Quant. Biol. 23,225-237. Gulatti, A. M., and Turner, A. J. 1928. Indian Central Cotton Comm., Tech. Lab. Bull. N o . 17. Harland, S. C. 1939. “The Genetics of Cotton.” Jonathan Cape, London. Harland, S. C. 1940. Trop. Agric. (Trinidad) 17,53-55. Hutchinson, J. B. 1954. Heredity 8,225-241. Hutchinson, J. B. 1959. “The Application of Genetics to Cotton Improvement.” Cambridge Univ. Press, London. Hutchinson, J. B., M o w , R. A., and Stephens, S. G. 1947. “The Evolution of Gossypium and the Differentiation of the Cultivated Cottons.” Oxford Univ. Press, London. Johnson, B. L. 1975. Bull. Torrey Bor. Club 102,340-349. Johnson, B. L., and T. M. M. 1970. A m . J. Bot. 57(9), 1081-1092. Jones, J. E. 1970. Proc. Beltwide Cotton Prod. Res. Con$ Houston, Texas, p. 5 5 . Jones, V. H . 1936. Univ.New Mexico Bull. No. 296, Anthropol. Ser. 1(5), 51-64. Kearney, K. P. 1951. A m . Midl. Nat. 46,93-131.

DISTRIBUTION AND EVALUATION OF Gossypium

255

Kent, K. P. 1957. Trans. A m . Philos. Soc. 47,459-732. Knight, R. L. 1954. “Abstract Bibliography of Cotton Breeding and Genetics, 1900-1950.” Commonwealth Agriculture Bureau, Farnham Royal, Cambridge, England. Kohel, R. J. 1973. J. Hered. 64,291-295. Kohel, R. J., and Lewis, C. F. 1984. In “Cotton” (R. J. Kohel and C. F . Lewis, eds.), Agronomy, No. 24, p. 12. American Society of Agronomy, Madison, Wisconsin. Lee, J. A. 1984. In “Cotton” (R. J. Kohel and C. F. Lewis, eds.), Agronomy, No. 24, pp. 1-25. American Society of Agronomy, Madison, Wisconsin. Lee, A. 1985. J . Hered, 76, 123-126. Lewis, C. F., and Richmond, T. R. 1968. In “Advances in Production and Utilization of Quality Cotton: Principles and Practice” (F. C. Elliot, M. Hoover, and W. K. Porter, eds.), pp. 1-21. Iowa State Univ. Press, Ames, Iowa. Lewontin, R. C. 1974. “The Genetic Basis of Evolutionary Change.” Columbia Univ. Press, New York and London. Mauer, F. M. 1954. “Origin and Systematics of Cotton.” Akad. Nauk Uzbek S.S.R., Tashkent. Menzel, M. Y., and Brown, M. S. 1952. Genetics 39,678-692. Menzel, M. Y., Brown, M. S., and Naqi, S. 1978. Generics 90, 133-149. Menzel, M. Y., Haskenkampf, C. A., and Stewart, J. M. 1982. Genetics 100,89-103. Meyer, V. G. 1973a. Proc. Beltwide Corron Prod. Res. Conf., Phoenix, Arizona, p. 65. Meyer, V. G. 1973b. Crop Sci. 13,778. Meyer, V. G. 1975. J. Hered. 66,23-27. Niles, G . A. 1980. In “Breeding Plants Resistant to Insects” (F. G. Maxwell and P. R. Jennings, eds.), pp. 337-369. Wiley, New York. Niles, G. A., and Feaster, C. V. 1984. In “Cotton” (R. J. Kohel and C. F. Lewis, eds.), Agronomy, No. 24, pp. 201-231. American Society of Agronomy, Madison, Wisconsin. Ornstein, L. 1964. Ann. N . Y . Acad. Sci. 121,321-349. Percival, A. E. 1987. South. Coop. Ser. June 1987, Bull. N o . 321. Percy, R. G . , and Wendel, J. F. 1989. Theor. Appl. Genet. (in press). Phillips, L. L. 1960. Genetics 46,77-83. Phillips, L. L. 1962. Am. J . Bot. 49,51-57. Phillips, L. L. 1963. Evofurion 17(4), 460-469. Phillips, L. L. 1964. Am. J . Bor. 51,324-329. Phillips, L . L., and Gerstel, D. U. 1959. J . Hered. 50, 103-108. Pickersgill, B., Barrett, S. C. H., and de Andrade-Lima, D. 1975. Biotropica 7(1). 42-54. Prokhanov, Y. I. 1953. Bot. Mat. Hebariya Bot. Inst. V . L . Kamarov, Akad. Nauk U.S.S.R. 15, 159-176. Saha, S., and Stelly, D. M. 1987. Agron. Abstr., p. 78. Saha, S., Stelly, D. M., and Percival, A. E. 1988. Proc. Beltwide Cotton Prod. Res. Conf., New Orleans, Louisiana, p. 97. Sarvella, P. 1958. Genetics 43,601-619. Saunders, J. H. 1961. “The Wild Species of Gossypium.” Oxford Univ. Press, London. Sherwin, K. H. 1970. Res. Rec. Uniu. Mus. S . Ill. Univ. Mesoamerican Stud. 6, 1-33. Skovsted, A. 1937. J. Genef.34,97-134. Stebbins, G . L 1947. Evol. Monogr. 17,213-221. Stephens, S . G. 1942. J. Genet. 44(2,3), 272-295. Stephens, S. G., and Mosely, M. E. 1974. Am. Anriq. 39, 109-122. Stewart, J. M. 1979. I n “Plant Tissue Culture: Symposium of the Southern Section of the American Society of Plant Physiology.” ( J . T. Barber, ed.), pp. 44-46. Tulane Univ., New Orleans, Louisiana.

256

A. EDWARD PERCIVAL AND RUSSELL J. KOHEL

Stewart, J. M., and Hsu, C. L. 1978. J . Hered. 69,404-408. Stewart, J. McD., Fryxell, P. A., and Craven, L. A. 1987. Brunomia 10,215-218. Takahata, N., and Slatkin, M. 1983. Genet. Res. 42,257-265. Vollesen, K. 1986. Kew Bull. 42(2), 337-349. von Hagen, V. W. 1961. “The Ancient Sun Kingdoms of the Americas.” World, Cleveland and New York. Wendel, J. F. 1989. Proc. Narl. Acad. Sci. U . S . A . 86,4132-4136. Wendel, J. F., and Percival, A. E. 1989. Plant Syst. Euol. (in press). Wendel, J. F., Olson, P. D., and Stewart, J. MacD. 1989. Am. J . Bor. 76, 1797-1808. Wilkins, T. A., and Galau, G. A. 1985. Proc. Beltwide Cotton Prod. Res. Conf., New Orleans, Louisiana, p. 73. Zaitzev, G. S. 1928. Bull. Appl. Bot. Genet. Plant Breed. 18, 1-65.

ADVANCES IN AGRONOMY. VOL. 44

BREEDING WHEAT FOR RESISTANCE TO Septoria nodorum AND Septoria tritici Lloyd R. Nelson’ and David Marshall2

’ Texas A&M University Agricultural Research and Extension Center at Overton Overton, Texas 75684 Texas A&M University Research and Extension Center at Dallas Dallas, Texas 75252

I . Introduction 11. Identification of Resistance A. Disease and Pathogen Assessment B. Components of Resistance C. Yield Reduction 111. Pathogen Variation IV. Genetics of Resistance V. Sources of Resistance VI. Discussion and Conclusions References

I. INTRODUCTION Breeding for resistance has been the most wide.j used method of disease control for the foliar diseases of wheat (Triticum aestiuum L.). The Septoria diseases have been no exception. There are three recognized Septoria diseases of wheat, each characterized by its causal fungus. On a worldwide basis, the two most important are septoria nodorum blotch (SNB) and septoria tritici blotch (STB). The third, septoria avenae blotch, is found infrequently and is generally considered a minor disease (Cunfer, 1987), although there are exceptions (Shearer and Calpouzos, 1973; Haugen et al., 1985). SNB (aka septoria leaf and glume blotch, or glume blotch) is caused by Leptosphaeria nodorum E. Mueller (anamorph Septoria nodorum (Berk.) Berk.). Some authors (Castellani and Germano, 1977; Bisset, 1982) have described the asexual stage of the fungus as Stagnospora nodorum (Berk.) Cast. & Germ. Additionally, others (Eriksson, 1967; Hedjaroude, 1968) have argued that the sexual stage should not be Lepto257 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

258

LLOYD R. NELSON AND DAVID MARSHALL

sphaeria, but rather Phaeosphaeria nodorum (Mueller) Hedja. STB (aka septoria leaf blotch or speckled leaf blotch) is caused by Mycosphaerella graminicola (Fuckel) Schroeter anamorph Septoria tritici Rob. & Desm.). Many aspects of the Septoria diseases of wheat have been discussed in at least four literature reviews (Shipton et a / . , 1971; Berggren, 1981; King et al., 1983; Karjalainen, 1985) and three international workshops (Cunfer and Nelson, 1976; Scharen, 1985; Fried, 1989). Eyal et a / . , (1987) have described methodology of disease management of the septoria diseases. This review will focus explicitly on breeding for resistance to SNB and STB. We will cite research from bread wheat (Triticum aestiuum L.) and durum wheat (T. durum L.). The relative emphasis that different breeding programs place on resistance to SNB and STB typically depends on the importance of the diseases in the geographic area of interest. The need for Septoria resistance must be viewed in relation to other breeding objectives. Thus, in some breeding programs it may be sufficient to simply avoid extremely susceptible germplasm, whereas other programs may require a high level of resistance (Russell, 1981). Undoubtedly, what most breeders and pathologists want is a form of resistance that is stable over time, easy to transfer across genotypes, easy to identify in segregating progeny, effective under disease-conduciveconditions, and nondetrimental to yield potential under disease-free conditions. In this review, we will attempt to show how resistance to SNB and STB has been identified, relevant features concerning pathogen variation, what is known about the genetics of resistance, and finally a reference list containing sources of resistance to SNB and STB.

II. IDENTIFICATION OF RESISTANCE A plant interacting with a pathogen can express several forms of reaction. Besides immunity, defined as absolute resistance (Robinson, 1969), resistance, tolerance, and susceptibility are possible reactions of the host in response to the attack by the pathogen. A susceptible plant possesses all necessary qualities to be a fit host for the pathogen. Resistance is defined as the ability of the host to hinder the growth of the pathogen (Robinson, 1969). Susceptible plants do not hinder the growth of the pathogen. Furthermore, resistance can be differentiated into subclasses: complete resistance, in which the spore production is completely inhibited; and incomplete resistance, which allows for some spores to be produced (Parlevliet, 1979). Incomplete resistance, also referred to as partial resistance, results

WHEAT RESISTANCE TO Seproria DISEASES

259

when the host is able to cause a reduction in spore production of the pathogen despite a relative “susceptible” reaction type. This reduction in spore production can decrease the rate of disease development and has been referred to as “slow-septoring” (Broennimann, 1982) in the case of SNB. This term is in imitation of the term “slow-rusting,’’ describing the reduced development rate of rust pathogens. Zadoks and Schein (1979) defined partial resistance simply as an intermediate level between susceptibility and resistance. Besides these terms, which evaluate the type of reaction on the basis of the visual disease development and ability to influence spore production, another form of interaction is tolerance. An infected host showing the same disease development over time, but less reduction in yield than a susceptible cultivar, is said to be tolerant of the pathogen.

A. DISEASE A N D PATHOGEN ASSESSMENT Early research in STB resistance breeding relied on identification of less severely affected germplasm in the field (Mackie, 1929). This work, as well as early work on SNB (Dantuma, 1955), showed that immunity to either pathogen could not be found, but that some varieties were more resistant than others. Renfro and Young (1956) outlined greenhouse and field methods, based on disease severity, that could be used to differentiate resistant and susceptible germplasm to STB. They found that ‘Red Chief and ‘Nabob’ had some resistance relative to ‘Westar,’ ‘Triumph,’ ‘Early Blackhull,’ and ‘Wichita,’ which were susceptible. The percentage of seedling leaves showing infection and the mean number of lesions per leaf were used by Hilu and Bever (1957) to identify the bread wheat CI 12557 as susceptible to STB and the durums ‘Kubanka’ and ‘Amautka’as intermediate and resistant, respectively. Perhaps the most common method of identifying resistance to SNB and STB has been the assessment of disease severity (percentage leaf or head area affected) in the field. Many breeding programs throughout the world have and continue to use this method to identify resistant germplasm. Also, most studies have shown good correlation between assessments made following artificial inoculation and those under natural inoculum conditions (Bayles et al., 1985). One of the earliest scales for assessing SNB severity was derived by Doling (1961), who first categorized 93 varieties as to either severe infection, moderate infection, slight infection, trace infection, or no infection, based on visual assessment of severity in the field. He further assessed the glume blotch phase of the disease and developed a glume blotch index for varieties based on weighing factors

260

LLOYD R. NELSON AND DAVID MARSHALL

from either a slight, moderate, or severe visual rating. A similar scale for SNB was also used by Nelson et al. (1974) in Georgia. Cooke and Jones (1971) used similar severity scales to assess SNB and STB severity on 1 1 winter and spring wheats in the field. Rosielle (1972) used a six-category scale to score bread and durum wheats for resistance to STB. The categories accounted for lesion size and pycnidial density with both increasing as the category score increased. Rosielle noticed, however, that some varieties departed from the scale by exhibiting extensive necrotic lesions with few or no pycnidia. A positive correlation between percentage severity of SNB on flag leaves to that on the heads was identified under field conditions (Scott, 1973). Gough and Smith (1976) used visual severity estimates of STB in the field to characterize germplasm in Oklahoma. Eyal et al. (1983) developed a relation between pycnidial density on the four uppermost leaves and the ratio between height of disease divided by plant height (called septoria progress coefficient) for STB. Using this method, they categorized several wheats into four categories ranging from highly resistant to highly susceptible to STB. The majority of greenhouse/growth chamber studies have used percentage leaf area necrosis and/or pycnidial density to quantify resistance (Shaner and Finney, 1982). In most cases, seedling and adult plant reaction to SNB and STB are positively correlated, with some exceptions (Mullaney et al., 1983; Scharen and Eyal, 1983). In greenhouse tests, Rillo and Caldwell (1966) differentiated six severity classes to STB in both seedling and adult plants, where 0 represented immunity and 5 was fully susceptible. Using this method, they identified 'Bulgaria 88' as a promising source of resistance. Later, Rillo et al. (1970) distinguished just three classes of host reaction, where resistant reactions had small lesions with pycnidia absent or few; intermediate was medium to large lesions, commonly containing pycnidia; and susceptible was large lesions with abundant pycnidia. Eyal et al. (1973) utilized pycnidial density classes of wide ranges to evaluate bread and durum wheat seedlings for resistance to a number of S . tritici isolates. Lesion size was used by Krupinsky et al. (1977) to screen 6161 wheat genotypes of T. aestivum and other Triticum spp. with mixed inoculum of SNB and STB. In five experiments, only 4 bread wheats and 16 other wheats, primarily T. timopheevi and T. dicoccum, averaged 10%or less necrosis. Scharen and Krupinsky (1978) assessed segregating generations of seedlings for percentages of plants with either no symptoms, typical lesions, chlorosis without lesions, and general necrosis to derive a damage index for SNB. Eyal and Brown (1976) digitized photographic transparencies on leaves representing a range of pycnidial densities to determine more precisely the pycnidial density on leaves. Later, Eyal and Scharen

WHEAT RESISTANCE TO Seproria DISEASES

26 1

(1977) devised a quantitative technique for evaluating germplasm for SNB resistance and found that the resistant variety ‘Manitou’ required about 2.1 times more pycnidiospores to produce a single lesion than the number required for the susceptible variety ‘Fortuna.’ However, Shearer (1978) estimated that it took nearly lo4 more pycnidiospores of STB to produce a lesion on the susceptible variety ‘Kondut’ compared to the susceptible ‘Federation.’ Mullaney et al. (1982) used photographs of seedling leaves to assess total leaf area, necrotic leaf area, and number of lesions per leaf. Baker (1970) used excised leaves on benzimidazole-supplemented agar and assessed the percentage of lesion-bearing tissue for SNB and STB. She used this technique to compare varieties to ‘Cappelle-Desprez,’ then classified as highly susceptible to SNB and tolerant to STB. Detached seedling leaves were also used by Benedikz et al. (1981) to assess lesion length for determination of resistance to SNB. Griffiths et al. (1985) used detached wheat leaves infected with S. nodorum to correlate ergosterol content with lesion severity. Resistance to SNB can also depend on the growth stage of the host. Mullaney et al. (1983) found seedlings of the spring durum wheat cultivar ‘Giorgio 396’ to be resistant whereas the mature plants showed the same disease development and yield reduction as a susceptible cultivar. On the contrary, Kajalainen (1985) found high correlation between seedling and field resistance ( r = 3 2 , p < .001). In a second experiment involving detached seedling leaves on benzimidazole agar, the correlation also was high (r = .62, p < .001). Similar results were reported by Rufty et al. (1981a), who found a high correlation (r = .64,p < .01) for the percentage of leaf necrosis between seedlings and mature plants. However, both authors reported deviations from this pattern, indicating the existence of resistance genes, which are effective only in the seedling or later growth stages. Roettges (1986) reported a high correlation (r = .93, p < .001) between the resistance index (sum of ranks over all recorded parameters) of six cultivars and their field resistance measured as reaction to natural infections of SNB over several years. Various methods have been used to assess the level of seedborne infection of SNB (S. tritici is not known to be seedborne). Cunfer (1978) determined the incidence of seedborne SNB by germinating suspected seed for 7 days, then freezing the seedlings and subsequently observing for lesions. Mathur and Lee (1978) developed the near ultraviolet (NUV) light fluorescence method on oxgall agar to detect incidence of seedborne SNB. Cunfer and Johnson (1981) used the oxgall agar-NUV fluorescence method to conclude that visual assessment of SNB on glumes in the field was not correlated to actual seed infection by S. nodorum.

262

LLOYD R. NELSON AND DAVID MARSHALL

B. COMPONENTS OF RESISTANCE

It is our view that “components of resistance” and “components of partial resistance” are equivalent for the septoria diseases of wheat. Because no immunity exists, any restriction or delay in pathogen development is a form of resistance. Following Vanderplank’s (1963) epidemiological interpretation of disease progress, some investigators began researching the components of resistance to SNB and STB. Shaner et al. (1975) identified a reduced rate of STB development in several winter wheat varieties. To do this, they evaluated STB severity on the four top leaves of mature plants in the field and rated pycnidial density on an A-to-E scale, where A represented the absence of pycnidia and E was equivalent to about 12 pycnidia per square millimeter. Brokenshire (1976) evaluated seedling and adult plants in the greenhouse and field for STB resistance based on incubation period (time from inoculation to symptom expression), latent period (time from inoculation to appearance of pycnidia), sporulation (area occupied by pycnidia), and disease severity. Good agreement was found between seedling reactions in the greenhouse and adult-plant reactions in the field. Gough (1978) found that pycnidia produced on the susceptible varieties ‘TAM W-101,’ ‘Improved Triumph,’ and ‘Triumph 64’ yielded over two times the number of pycnidiospores than those produced on the resistant variety ‘Oasis.’ Parlevliet (1979) stated that partial resistance resulted in delayed progress of disease epidemics by reduction of the infection rate r, the measure for the multiplication rate of the pathogen. He divided the rate-reducing resistances into four components that collectively influenced the disease development by the pathogen: 1. Infection frequency (measured as number of sporulating lesions per amount of spores applied) can be used to assess resistance to the initial infection and also to colonization of the pathogen. 2. Latent period (time between inoculation and first appearance of mature pycnidia) is an additional assessment factor to characterize resistance to colonization. 3. Spore production (number of spores produced per unit leaf area or fruiting body) can be used to measure resistance to fungal reproduction. 4. Infectious period (time during which pycnidia are sporulating) is an additional assessment character of fungal reproduction.

Parlevliet referred to R. R. Nelson’s statement that many resistances that reduced the infection rate had polygenic inheritance. He also pointed out ample evidence for mono- or oligogenic inheritance, which he demonstrated by several examples for small grains, tomato, and maize.

WHEAT RESISTANCE TO Septoria DISEASES

263

Jeger (1980) and Jeger et al. (1983) studied 11 different components of resistance to SNB in seedlings of 41 wheat cultivars and related Triticum species. Besides infection frequency, latent period, and spore production, they also recorded incubation period (time between inoculation and first appearance of symptoms), lesion size, lesion cover (percentage of leaf area covered by lesions), and percentage of necrotic leaf tissue. In addition, the three latter components and spore production were assessed again when the infected leaf had turned completely yellow. Factor analysis revealed that four independent factors conditioned resistance. Necrosis and reduced spore production were two of these factors. They were interpreted as resistance to the action of the toxin and to fungal reproduction, respectively. Lesion size, lesion cover, and latent period together accounted for the third factor (resistance to the growth of the pathogen). The fourth factor was due to combined effects of infection frequency, incubation period, and lesion cover, which is associated with resistance to colonization by the pathogen. Analysis of principal components resulted in four classes that were dominated by a different combination of components. The first principal component was influenced by latent period, lesion size, and lesion cover. The second principal component was dominated by unit spore production and necrosis. Incubation period, lesion size, infection frequency, and necrosis accounted for the third principal component. All components characterized the fourth principal component, which could not reasonably be interpreted. From the results of the two analyses, the authors suggested the presence of four independent components corresponding to Parlevliet’s (1979) classification of resistance to establishment, growth, and reproduction of the fungus and one additional component for pathogen-induced necrosis. A study of 14 components on 10 cultivars by Lancashire and Jones (1985) agreed with the findings of Jeger et al. (1983). The principal components analysis of their data resulted in four components, of which the first (characterized by resistance to growth within the host) was most predominant, accounting for 49% of the variation. The second component conditioned resistance to infection and decrease of sporulation. Lesion length was responsible for the third character, which was suggested to be related to resistance to the toxin. The fourth component was influenced mainly by lesion width, which accounted for resistance to mycelium growth. A cluster analysis of the 10 cultivars allowed for classification of three groups on the basis of their components of resistance. In one case, related cultivars were classified into one group; however, another pair of related cultivars was classified into different groups. Lancashire and Jones (1985) developed an “r-index” for comparing the relative ability of variet-

264

LLOYD R. NELSON AND DAVID MARSHALL

ies to slow the rate of SNB development. Contrary to Jeger et al. (19811, Griffiths and Jones (1987) found that incubation period did not correlate with field assessments of SNB. They did find that sporulation was a good indicator of field-assessed disease ratings. Roettges (1986) studied nine components of resistance in the seedling stage of seven winter wheat cultivars and two advanced breeding lines. He found cultivar differences for every component (incubation and latent period, necrosis, infection frequency, lesion area, percentage of lesions with pycnidia, and spore production perpycnidium, cm2leafarea, and cm2 lesion area). Most of the correlations between components were large and significant, ranging from r = .63 to .90. Only small negative correlations were found between latent period and necrosis (r = -.50,p > .05) and between latent period and spore production per unit leaf area ( r = -.58,p > .05). No correlation was found between lesion area and number of spores or pycnidia per square centimeter. A ranking of cultivars within each component indicated different ranks for other components. This should not be accepted as proof of different degrees of resistance for particular components, however, because the ranking was not based on the significant differences. Variation within the ranking occurred only within two groups roughly classified as “resistant” and “susceptible.” Only ‘Kanzler’ had a susceptible reaction concerning components of the vegetative stage, and a more resistant reaction to the reproductive stage of the fungus. Similar results were obtained by Nelson and Bruno (1985) who showed significant differences in 21 winter wheat cultivars for some of the measured components of resistance. Besides variation due to environmental effects, significant cultivar differences for incubation period, latent period, necrosis, and spore production were observed. Trottet and Benacef (1989) reported that there are several mechanisms of resistance whose expression depends on growth stage, and that great differences can be found between leaves and heads for disease development. Nelson and Crowder (1989) compared winter wheats at different growth stages for components of resistance. They reported that adult plants were more susceptible than seedling plants and that lower leaves were likewise more susceptible than upper or flag leaves. For percentage necrosis, they found that ‘Oasis’ had resistance in the adult stage that was not expressed in the seedling stage. Wilkinson and Murphy (1986) investigated the combining ability of winter wheat for four components of resistance using a diallel analysis. They found significant general combining ability (GCA) for incubation period, infection frequency, and latent period. No significant GCA was observed for spore production. Specific combining ability (SCA) affected only incubation period. Reciprocal crosses did not show significant differences for

WHEAT RESISTANCE TO Septoria DISEASES

265

any of the components. Except for some observed degree of dominance for susceptibility in the resistant x susceptible crosses, primarily additive effects were observed. In a similar study involving a diallel cross, Stooksbury et al. (1987) reported highly significant results for both GCA and SCA effects for incubation and latent period. Bruno and Nelson (1990) conducted a diallel analysis to investigate gene effects for six components of resistance of wheat to SNB. The components were incubation period, latent period, percentage of diseased leaf tissue, initial spore production at the end of the latent period, total spore production at 100% necrosis, and maturation period. Maturation period was defined as the time between the first visual disease symptoms and the appearance of mature pycnidia. Significance was observed for incubation period, latent period, maturation period, and total spore production; however, percentage necrosis and initial spore production were not significantly different. Cunfer et al. (1988) rated seven winter wheat varieties for incubation period, latent period, lesion development, and sporulation of SNB and found good correlation among the resistance components. After studying the effects that resistance components have in slowing SNB development, Leonard (1988) suggested that a longer latent period was more effective than a reduction in spore production on susceptible varieties, but that the opposite was true for more resistant varieties. Rapilly (1988) used a model of SNB development to conclude that the rate of necrosis on leaves was more important than incubation period, which in turn was more important than latent period in reducing disease spread. Baker and Smith (1979) investigated the number and size of lesions on adult plants caused by SNB in three winter wheats with different degrees of resistance. Only the intermediate and susceptible varieties were noted for a sharp increase in number and size of lesions at the end of the season. The susceptible genotype began to develop a larger number of lesions about 5 weeks earlier than the intermediate. They differed from each other more in lesion number than in lesion size. The authors noted a more uniform and more dense canopy of the susceptible variety, which might have influenced the microclimate (higher humidity) promoting fungal development. The effects of postinoculation wet periods on the number of lesions and necrosis on seedlings of eight wheat varieties were investigated by Eyal et al. (1977). The data showed that an increase in the postinoculation wet period increased symptom development (more lesions and higher percentage of necrosis). Winter wheats were less affected than spring wheats, but varietal differences were observed for both groups. The infection frequency also has been studied in relation to inoculum concentration and postinoculation wet and dry periods (Jeger et al., 1984).

266

LLOYD R. NELSON AND DAVID MARSHALL

Seedlings of two spring wheat varieties ('Kolibri' and 'Maris Butler') showed an increasing number of lesions on the leaves with an increase of the spore concentration of the inoculum, independent of subsequent wet and dry periods. The minimum concentration for symptom development was lo3 spores per milliliter, which is in agreement with other reports (Broennimann, 1968a; Roettges, 1986). Increasing the dry period resulted in a curvilinear decrease in the number of lesions at two temperatures (15°C and 20°C). In most cases, increasing the wet period caused a linear increase in the number of lesions, except for Kolibri. For this variety, the number of lesions leveled off after about 10 hr of wetness at 15°C. The latent period was studied by Shearer and Zadoks (1972,1974) under controlled and field conditions on seedlings of a susceptible winter wheat. Under controlled conditions, the effects of eight temperature treatments (5"-25"C),three moisture treatments, and two spore concentrations of the inoculum were studied. The shortest latent period of 6 days was observed on plants continuously kept in saturated air at 23°C. If these moisture conditions were interrupted for 12 hr with exposure to 85-90% relative humidity (r.h.), the formation of mature pycnidia was delayed by 5.6 days. Leaves continuously exposed to air with 85-90% r.h. became infected but did not produce spores at all, unless they were exposed for a short period (24 hr) to saturated air. The lower spore concentration, 5 X lo4spores ml', delayed spore production on the average by 2.4 days. The response for different temperature regimes was a parabolic curve for all moisture and inoculum levels, with the shortest latent period between 20°C and 23°C. The results from field conditions were in accordance with the controlled conditions. The latent period was reduced by an increase in temperature and/or duration of leaf wetness. A multiple regression analysis showed significant effects of both parameters as well as their interaction. In an attempt to predict the latent period from the recorded data, the duration of leaf wetness and minimum temperature were most useful. When measuring latent period in our research, we normally place plants in an incubation chamber overnight to stimulate formation of pycnidia. Aust and Hau (1981) studied the compensative effects of temperature, postinoculation wetness, and inoculum density on the latent period of the spring wheat Kolibri. The temperature conditions ranged from 12°C to 25"C, the postinoculation wetting periods from 15 to 72 hr, the postinfection wet period from 2 to 8 hr, and the spore density from lo4 to lo7 spores ml-', The shortest latent period was 7.7 days for a postinoculation wetting period of 72 hr, postinoculation period of 8 hr, spore concentration of lo7 spores ml-', and temperature of 20°C. Longer wetting periods and/or higher spore concentrations were shown to compensate for suboptimal temperatures.

WHEAT RESISTANCE TO Seproria DISEASES

267

C. YIELDREDUCTION Another method used to identify resistance to SNB and STB has been by measuring the amount of yield loss, from either small plots or plant components. Some researchers (Broennimann, 1968b; Ziv and Eyal, 1976; Ziv et al., 1981) have used this type of assessment to measure disease tolerance. It is our opinion that tolerance should be used to describe the phenomenon whereby plants exhibit no significant yield reduction under the same disease severity and rate of disease development compared to susceptible plants. Thus, germplasm labeled as tolerant may actually yield better than a susceptible type by inducing some sort of restriction in pathogen development, rather than by being truly tolerant. Broennimann (1968a) was among the first to show that thousand-kernel weight (TKW) was strongly affected by SNB and that TKW could be used to assess for resistance. Doling (1961) and Cooke and Jones (1970) found an association between SNB severity and the degree of grain shriveling. Cooke and Jones (1971) subsequently found a good correlation between SNB and STB severity assessments with mean yield per head, TKW, and a sieving index. Variable results were found by Sharp et al. (1972) when they compared the SNB field rating of 30 spring wheats to their respective losses in grain yield per head, TKW, and kernels per head. They noticed that SNB had little effect on the TKW of cultivars ‘MT 6903’ and ‘Centana,’ but that disease did reduce the varieties’ kernels per head. Scott (1973) found a strong positive correlation between leaf and head infection by SNB and yield loss. Eyal and Ziv (1974) used STB progress curves from field trials to relate with losses in plot yield and TKW. In the STB-resistant varieties ‘Elite Lepeuple,’ ‘Maris Dove,’ ‘Maris Ensign,’ ‘Chalk,’ and ‘Tommy,’ a good correlation was found between incubation period, latent period, disease severity, and sporulation with TKW (Brokenshire, 1976). Across many winter wheat cultivars, field ratings for leaf and head severity of SNB were negatively correlated with plot yield, TKW, and test weight (Nelson et al., 1976). Gough and Merkle (1977) suggested that the effects of STB may be more highly manifested as root mass reductions rather than reduction in grain yield, and thereby may be a better indicator of resistance. Results from studies conducted by Scott and Benedikz (1977) indicated that yield loss was positively correlated with SNB severity but that disease was more easily measured and with less variability than yield loss. Rosielle and Brown (1980) used results from SNB-inoculated and uninoculated plots of segregating wheat populations to suggest that selection for yield or seed weight in inoculated plots would aid in the identification of resistant plants more so than selection for seed weight percentage. Nass

268

LLOYD R. NELSON AND DAVID MARSHALL

and Johnston (1985) found that resistance to SNB could be identified in the field by comparing the yields of fungicide-treatedto untreated plots. They indicated that SNB severity, plot yield, and TKW were good selection criteria for SNB resistance. Zilberstein et al. (1985) suggested that postanthesis chemical desiccation of field plots could be used as a screening method to identify germplasm that is able to endure severe STB infection. Kelaniyangoda (1987) identified SNB-resistant varieties by using the correlation between field severity rating and yield loss.

111. PATHOGEN VARIATION Significant variability for pathogenicity and related fitness factors in the S . nodorum and S.tritici populations can influence how to best breed for resistance to these two pathogens. Much of the early work on Septoria variability centered on determining which plant species the fungi could infect (Beach, 1919; Weber, 1922). Broennimann (1968a) found that individual isolates of S . nodorum varied in regard to spore production, but that small differences were evident in terms of their pathogenicity. Krupinsky et al. (1973) found that S. nodorum isolates varied in pathogenicity (aggressiveness) as measured by changes in the plants’ photosynthetic rate. Krupinsky et al. (1989) studied 27 isolates from 1 1 different species of grass. The isolates differed in their ability to cause SNB symptoms on wheat and this difference was interpreted to be due to aggressiveness. Scharen and Krupinsky (1970) showed the wide range of variability present in the pycnidiospores from a single pycnidium. However, Griffiths and Ao (1980) found that even though S . nodorum isolates were highly variable for several cultural characteristics, variation for pathogenicity was narrower. Septoria nodorum isolates in Florida were found to be quite variable in terms of pathogenicity on eight wheat varieties (Allingham and Jackson, 1981). Several studies have examined the effects of cross-inoculationreisolation experiments to determine host range and to detect physiological specialization. Shearer and Zadoks (1972) suggested that S.nodorum could become more specialized on a host following repeated passages. Similarly, Ao and Griffith (1976) found that virulence of S.nodorum and S . tritici may change after a single passage through an alternate host and that direction and magnitude of change are similar for both fungi. An S . nodorum isolate from barley was found to gain pathogenicity toward wheat and lose pathogenicity to barley after several passages through wheat (Fitzgerald and Cooke, 1982). However, Cunfer (1984) suggested that the S.

WHEAT RESISTANCE TO Septoria DISEASES

269

nodorum found on barley in the southeastern United States was largely restricted to barley. Osbourn et al. (1986) found some changes in S. nodorum isolates with passages through barley and wheat, but concluded that the changes may have resulted from cross-contamination with subsequent selection. Eyal et al. (1973) used pycnidial density classes and percentage leaf area covered by pycnidia to suggest that true physiological specialization existed in STB isolates in Israel. Research in Argentina (Perello et al., 1989) had similar results, and they indicated that pycnidial coverage was a better parameter than necrotic lesion area to differentiate between isolates. Although some minor statistically significant differences were found between isolates and varieties, much greater interactions were evident between isolates from bread and durum wheats. Rufty et al. (1981b) assessed lesion percentage on seedling leaves and identified significant SNB isolate X wheat variety interactions, but indicated that the magnitude of the differences was small and therefore classification of the isolates into physiological races was inappropriate. A similar level of pathogenic variation in S. nodorum was found by Scharen and Eyal (1983). Specialization in S. nodorum isolates was evident among several wild species of Triticum (Yechilevich-Auster et al., 1983). Eyal et al. (1985) tested 97 isolates of S. tritici on varieties of bread wheat, durum wheat, and triticale and found significant isolate X variety interactions. They also suggested that seedlings exhibiting necrotic leaf area of below 16.6% could be considered resistant, whereas those with greater amounts of necrosis were susceptible. Similar results of Eyal and Levy (1987) suggested geographic distribution of specific virulences in S. tritici. Likewise, Scharen et al. (1985) found significant S. nodorum isolate X variety interactions with bread wheats, durum wheats, and triticales. Using percentage necrosis on seedlings, they suggested that 17.9%necrosis was the separation point between susceptibility and resistance to S. nodorum. Marshall (1985) assessed S. tritici severity in the field over 13 locations in the United States in addition to greenhouse assessments of leaf severity and pycnidial density and found wide variation for aggressiveness but no significant isolate X variety interactions. Saadaoui (1987) used bread and durum wheat cultivars and found physiological specialization in S. tritici when 20% necrotic leaf area with few pycnidia was used as the separation point between susceptibility and resistance. Van Ginkel and Scharen (1988) suggested that S. tritici isolates were specifically adapted to either bread or durum wheat (species specificity), but that cultivar specificity was not significant. Silfhout et al. (1989) reported that certain S. tritici isolates were able to overcome the resistance of cultivars, and that these isolates. must have one or more genes for specific virulence that do not occur in

270

LLOYD R. NELSON AND DAVID MARSHALL

other isolates. Cordo et al. (1989) reported that isolates of STB obtained from wheat leaves have the characteristics of being heterogeneous in virulence. They can originate from pycnidia that are homogeneously virulent, avirulent, or show a mixed virulence. It is clear that genus and species (Triticum) specificity exists in both S. nodorum and S. tritici. This may be regarded as genome specificity. Specificity at the varietal level is less clear, particularly because “where you draw the line” between resistance and susceptibility greatly influences isolate by cultivar interactions. However, there is wide variability in both pathogens for pathogenic aggressiveness, which can vary with geographic location. Therefore, it seems reasonable that breeding programs should determine the locations within their geographic area of interest where the most aggressive isolates are indigenous. Single-spore cultures from the most pathogenic isolates should be used in greenhouse/growth chamber studies. Because resistance to Septoriu diseases is relative and not absolute (immunity), perhaps the most effective method to identify resistance would be to use a single, highly pathogenic isolate of S. nodorum or S. tritici (Parlevliet, 1983).

IV. GENETICS OF RESISTANCE Mackie (1929) was the first to report that S. tritici resistance could be bred into wheat and that resistance was simply inherited and recessive. Narvaez and Caldwell(l957) crossed the S. tritic&resistant‘Nabob’ with the susceptibles ‘Knox’ and ‘Vermillion’ and found that resistance was controlled by two independent, partially dominant genes with additive effect. They also found that in the ‘Lerma 52’ and ‘P14’ X ‘Lee’ and ‘Mayo 54’ crosses, resistance was controlled by a single dominant gene. Efforts to find resistance to STB were accelerated in the late 1960s and early 1970s by the generalized susceptibility of high-yielding spring wheats coupled with conducive environmental conditions that resulted in severe STB epidemics (Saari and Wilcoxson, 1974; Marshall, 1989). Rillo and Caldwell(l966) identified ‘Bulgaria 88’ as a source of resistance to S. tritici and suggested that a single, dominant gene conditioned resistance. The resistance in ‘Bulgaria 88’ was transferred to the winter wheat variety ‘Oasis’ (Patterson el al., 1975).STB resistance in an Agropyron-wheat derivative was found to originate on a singleAgropyron chromosome that was transmitted 25% of the time through the male and female gametes (Rillo et al., 1970). Rosielle and Brown (1979) used ‘Gamenya’ as the susceptible parent in crosses with the STB-resistant wheats ‘Seabreeze,’ ‘Veranopolois,’ and

WHEAT RESISTANCE TO Septoria DISEASES

27 1

‘IAS-20.’ They found that the resistance in ‘Seabreeze’ was conditioned by at least three recessive genes, whereas a single gene probably conferred resistance in ‘Veranopolis’ and ‘IAS-20.’ Wilson (1979) reported that ‘Israel 493’ and ‘Veranopolis’ each contained a single gene for resistance to STB and that they were independent of each other. Danon et al. (1982) concluded that resistance to STB in ‘Bezostaya 1,’ ‘Colotana,’ ‘Fortaleza1,’ ‘Polk/Waldron,’ ‘Sheridan,’ and ‘Oasis’ was conferred by only one or a few genes. Lee and Gough (1984) crossed the STB-resistant ‘Carifen 12’ with the susceptible wheats ‘TAM W-101’ and ‘Triumph 64’ and found that a single, dominant gene conferred resistance. Both additive and dominance gene effects for resistance to STB were significant in several durum wheat crosses (Van Ginkel and Scharen, 1987). Broennimann (1975) found that resistance to SNB was inherited mainly additively and polygenically. Kleijer et al. (1977) crossed the monosomic set of ‘Chinese Spring’ with the SNB-resistant variety ‘Atlas 66’ and found that a single, dominant gene for resistance was located on chromosome 1B. Nelson (1980) identified ‘Oasis’ and ‘Blueboy 11’ as good sources of resistance to SNB because of the varieties’ significant general combining ability and specific combining ability. Nelson and Gates (1982) further indicated that SNB resistance in ‘Oasis’ and ‘Blueboy 11’ was additive and inherited in a very complex manner. Mullaney et al. (1982)used generation mean analysis to suggest that SNB resistance in ‘Frondoso’ and ‘Fronthatch’ was pol ygenically controlled and explained principally by additive gene effects. Because additive gene action may be quite prevalent in sources for resistance to SNB, Karjalainen (1985)proposed that transgression breeding may be a useful method for improving resistance. Resistance to SNB on leaves appeared to be independent of resistance on the heads (Fried and Meister, 1987), and overall heritability or resistance was quite low. Rapilly et al. (1988) suggested that resistance to S. nodorum as measured by incubation time and rate of spread of leaf necrosis was polygenically inherited and possibly located on chromosomes 3A, 2B, and 5B. Similarly, Ecker et al. (1989) found that resistance to SNB as assessed by infection efficiency, disease severity, lesion size, and latent period was controlled by three to four quantitative genes, with mainly additive resistance. Frecha (1973) analyzed the resistance of ‘Atlas 66’ to SNB and reported it to be inherited by a single dominant gene for resistance. However, Kleijer et al. (1977) could not reproduce the data using chromosome substitutions to locate the gene. Although only 1 substitution line (1B substitution) differed significantly from the other 20 lines, they were not able to obtain a clear 3 : 1 (resistant : susceptible) segregation, since too few plants showed a resistant reaction. On the basis of these results, they suggested the

272

LLOYD R. NELSON AND DAVID MARSHALL

presence of one or more modifier genes in ‘Chinese Spring’ (the monosomic, susceptible parent) or the same chromosome. Several studies have indicated that resistance to Septoriu diseases tends to be found in wheat lines that were tall and/or late maturing. Tavella (1978) found positive correlations between low S. tritici field ratings and tall, late varieties. Rosielle and Brown (1979) suggested that even though there appeared to be a correlation between tall, late plants and resistance to STB, it should not be a major obstacle to selection. A low correlation between plant height and S. tritici severity suggested that there is probably no linkage or pleiotropy between shortness and susceptibility (Danon et ul., 1982). Scott et al. (1982) suggested that breeding S.nodorum-resistant germplasm of any height should be possible because they found genes conferring resistance that were independent of plant height and heading day. The presence of genes for tolerance that were independent from tallness was reported by Broennimann and Fossati (1977). Using induced mutation, the authors were able to recover a tolerant mutant from a short, susceptible genotype that was only slightly taller than the original genotype. In addition, they were able to induce short, tolerant mutants in tolerant but tall lines. Unfortunately, in both cases the mutants had low yields and could only be used in breeding programs.

V. SOURCES OF RESISTANCE In Table I, we have attempted to list those publications in which authors have identified sources of resistance to S. nodorum and/or S. tritici. Even though we have attempted to be thorough, we apologize in advance if any sources have been inadvertently omitted.

VI. DISCUSSION AND CONCLUSIONS Compared to other wheat diseases, such as leaf or stem rust, little progress in host plant resistance has been accomplished with the septoria diseases. Genes for immunity or complete resistance to either SNB or STB have not be discovered or reported in the literature. The complexity of the disease organism and the inheritance of resistance to SNB and STB have not been and are not clearly understood. Most of the resistant genotypes were developed in plant breeding programs located in environments that

WHEAT RESISTANCE TO Septoria DISEASES

273

Table I Published Sources of Resistance to Septoria nodorum and S. fritici as listed in the Literature Septoria Baker (1970) Benedikz et a / . (1981) Bruno and Nelson (1990) Cooke and Jones (197 I ) Doling (1961) Eyal et a/. (1987) Jeger e f a/ (1983) Krupinsky et a / . (1977) Nass and Johnson (1985)

nodorum Nelson (1979) Nelson (1980) Nelson and Crowder (1989) Rufty e f al. (1981b) Scharen and Eyal(1980) Scharen et a / . (1985) Scott (1973) Sharp et a / . (1972) Wilkinson et a / . (1990)

Septoria tritici Baker (1970) Eyal et a / . (1987) Krupinsky et a / . (1977) Bayles et a!. (1985) Brokenshire (1976) Narvaez and Caldwell(1957) Renfro and Young (1956) Cooke and Jones (1971) Rosielle (1 972) Danon e t a / . (1982) Eyal et a / . (1983) Shaner and Finney (1982) Eyal et a / . (1985) Shaner et a / . (1975)

were conducive to disease epidemics, and therefore susceptible germplasm was discarded annually during normal selection procedures in the breeding program. Advances in the selection of resistant germplasm were very slow because of escapes caused by weather patterns that limited the septoria disease epidemic. Tall or late-maturing genotypes often were rated incorrectly as resistant. Other diseases such as rusts or powdery mildew often were of greater importance, and therefore the more septoriaresistant germplasm was discarded. Indentification of genes for resistance was difficult at best, because resistance was often inherited in an additive manner. During the past 10 years much has been learned about these diseases. We now use several new techniques to screen germplasm both as seedlings and as adult plants. We also know that exact environmental conditions must be maintained to conduct any genetic studies. Comprehension of components of partial resistance has begun to simplify the inheritance of resistance in some ways. At least we have begun to understand why and how some cultivars are resistant. On the other hand, utilizing or manipulating certain types of components, such as latent period, in plant breeding programs is very difficult, is time-consuming, and limits the germplasm that can be screened.

274

LLOYD R. NELSON AND DAVID MARSHALL

The future remains clouded. Many sources of resistance are available for wheat breeding programs; however, the indentification of segregating germplasm for one component or several components of resistance will be necessary to achieve high levels of resistance. This may not be practical or cost-effective in many wheat breeding programs. There is likely much more variation in both the S . nodorum and S . tritici organisms than has been reported in the literature. If this is the case, the septoria organisms may be able to adapt to and/or attack at least some components of resistance. This is certainly not as likely as with major genes for resistance on other wheat diseases. The use of multiple isolates of either S . nodorum or S . tritici may be useful in identifying components of partial resistance that are resistant to all isolates. This may not be the case, however, if some genes are resistant to only some isolates. In this case, progress for selection of these components would be difficult or perhaps useless. Great advances have been made in the production of inoculum that can be used in either the field or laboratory. Therefore it is relatively easy to produce epidemics in our breeding programs that create selection pressures that are useful in selecting superior genotypes. The use of biotechnology may offer even greater potential. Since septoria diseases produce toxins, selection at the cellular level of resistant types could be worthwhile and new and improved resistance may be found or developed. The use of RFLP (restriction fragment length polymorphism) technology also offers advances. particularly to determine if pathogenic differences between isolates of S . nodorum and/or S . tritici exist. If pathogenic differences in isolates do exist, and they can be fingerprinted, then the movement of an isolate in the field can be followed and the epidemiology of the organism will be better understood.

REFERENCES Allingham, E. A., and Jackson, L. F. 1981. Phytopathology 71, 199. Ao, H. C., and Griffiths, E. 1976. Trans. Br. Mycol. Soc. 66,337-340. Aust, H. J . , and Hau, B. 1981. Z . Pjlanzenkrank. (Pjlanzenpathol.) PJlanzenschutz 88, 655-664. Baker, C. J. 1970. Trans. Br. Mycol. Soc. 54,498-504. Baker, E. A., and Smith, I . M. 1979. Trans. Br. Mycol. SOC.73,57-63. Bayles, R. A., Parry, D. W., and Priestley, R. H. 1985. J . Nafl. Inst. Agric. Bot. 17,21-26. Beach, W. S. 1919. A m . J . Bot. 6, 1-33. Benedikz, P. W., Mappledoram, C. J . , and Scott, P. R. 1981. Trans. Br. Mycol. Soc. 77, 667-669. Berggren, B. 1981. Inst. Plant For. Prot., Swed. Univ. Agric. Sci., Uppsala, Rep. N o . 19. Bisset, J. 1982. Natl. Mycol. Herb., Biosys. Res. Inst., Agric. Can. Rep. N o . 240. Broennimann, A. 1968a. Phytopathol. Z . 61, 101-146.

WHEAT RESISTANCE TO Septoria DISEASES

275

Broennimann, A. 1968b. Phytopathol. Z. 62, 190-194. Broennimann, A. 1975. Phytopathol. Z. 75, 138-160. Broennimann, A. 1982. Neth. J . Agric. Sci. 30,47-69. Broennimann, A. and Fossati, A. 1977. IAES-SM-214/17, 403-408 (Zurich-Reckinholz, Switzerland). Brokenshire, T. 1976. Ann. Appl. Biol. 82,415-423. Bruno, H. H., and Nelson, L. R. 1990. Crop Sci. 30,54-59. Castellani, E., and Germano, G. 1977. Ann. Fac. Sci. Agric., Univ. Torino 10, 1-132. Cooke, B. M., and Jones, D. G. 1970. Trans. Br. Mycol. Soc. 54,395-404. Cooke, B. M., and Jones, D. G. 1971. Trans. Br. Mycol. SOC.56, 121-135. Cordo, C. A., Perello, A., Arriaga, H., and Pascar, A. 1989. In “Proceedings of Septoria of Cereals Workshop” (P. M. Fried, ed.), pp. 55-58. Swiss Federal Research Station, CH-8046 Zurich-Reckinholz, Switzerland. Cunfer, B. M. 1978. Phytopathology 68,832-835. Cunfer, B. M. 1984. Ann. Appl. Biol. 104,61-68. Cunfer, B. M. 1987. In “Wheat and Wheat Improvement.” (E. G. Heyne, ed.), 2nd Ed., pp. 528-541. American Society of Agronomy, Madison, Wisconsin. Cunfer, B. M., and Johnson, J. W. 1981. Trans. Br. Mycol. SOC.76,205-211. Cunfer, B. M., and Nelson, L. R. 1976. Proc. Sept. Dis. Wheat Wkshp., Univ. Georgia, Spec. Pub/. N o . 4. Cunfer, B. M., Stooksburg, D. E., and Johnson, J. W. 1988. Euphytica 37, 129-140. Danon, T., Sacks, J. M., and Eyal, Z. 1982. Phytopathology 72, 1037-1042. Dantuma, G. 1955. Euphytica 4,94-95. Doling, D. A. 1961. Plant Pathol. 10,24-27. Ecker, R., Dinoor, A., and Cahaner, A. 1989. Plant Breed. 102, 113-121. Eriksson, 0. 1967. Ark. Bot. 6, 381-440. Eyal, Z., and Brown, M. B. 1976. Phytopathology 66, 11-14. Eyal, Z., and Levy, E. 1987. Euphytica 36,237-250. Eyal, Z., and Scharen, A. L. 1977. Phytopathology 67,712-714. Eyal, Z., and Ziv, 0. 1974. Phytopathology 64, 1385-1389. Eyal, Z., Arniri, Z., and Wahl, I. 1973. Phytopathology 63, 1087-1091. Eyal, Z., Brown, J. F., Krupinsky, J. M., and Scharen, A. L. 1977. Phytopathology 67, 874-878. Eyal, Z., Wahl, I., and Prescott, J. M. 1983. Euphytica 32,439-446. Eyal, Z., Scharen, A. L., Huffmann, M. D., and Prescott, J. M. 1985. Phytopathology 75, 1456- 1462. Eyal, Z., Scharen, A. L., Prescott, J. M., and van Ginkel, M. 1987. “The Septoria Diseases of Wheat: Concepts and Methods of Disease Management.” Mexico. D. F. CIMMYT, Mexico. Fitzgerald, W., and Cooke, B. M. 1982. Plant Pathol. 31,315-324. Frecha, J. H. 1973. Biol. Genet. Inst. Fitotec. Casrelar 8,29-30. Fried, P. M. 1989. “Proceedings of Septoria of Cereals Workshop” Swiss Federal Research Station, CH-8046 Zurich-Reckinholz, Switzerland. Fried, P. M., and Meister, E. 1987. Phytoparhology 77, 1371-1375. Gough, F. J. 1978. Phytopathology 68, 1343-1345. Gough, F. J., and Merkle, 0. G. 1977. Plant Dis. Rep. 61,597-590. Gough, F. J., and Smith, E. L. 1976. Plant Dis. Rep. 60,698-700. Griffiths, E., and Ao, H. C. 1980. Ann. Appl. Biol. 94,294-296. Griffiths, H. M., and Jones, D. G. 1987. Ann. Appli. Biol. 110,603-610. Griffiths, H. M., Jones, D. G., and Akers, A. 1985. Ann. Appli. Biol. 107,293-306. Haugen, L . G., Wilcoxson, R. D., and Baumer, J. S. 1985. Plant Dis. 69, 162-163.

276

LLOYD R. NELSON AND DAVID MARSHALL

Hedjaroude, G. A. 1968. Sidowia 22,57-107. Hilu, H. M., and Bever, W. M. 1957. Phypathology 47,474-480. Jeger, M. J. 1980. Prof. Ecol. 2,265-269. Jeger, M. J., Griffiths, E., and Jones, D. G. 1981. Ann. Appl. Biol. 99,29-34. Jeger, M. J., Jones, D. G., and Griffiths, E. 1983. Euphytica 32,575-584. Jeger, M. J., Griffiths, E. and Jones, D. G. 1984. Ann. Appli. Biol. 106,55-63. Kajalainen, R. 1985. J . Agric. Sci. Fin. 57, 1-66. Kelaniyangoda, D. B. 1987. J. Phytopathol. 119,279-288. King, J. E., Cook, R. J., and Melville, S. C. 1983. Ann. Appli. Biol. 103,345-373. Kleijer, G., Broennimann, A., and Fossati, A. 1977. J . Plant Breed. 78, 170-173. Krupinsky, J. M., Scharen, A. L., and Schillinger. J. A. 1973. Physiol. Plant Pathol. 3, 187- 194. Krupinsky, J. M., Craddock, J. C., and Scharen, A. L. 1977. Plant Dis. Rep. 61,632-636. Krupinsky, J. M., Craddock, J. C., and Scharen, A. L. 1989. In “Proceedings of Septoria of Cereals Workshop” (P. M. Fried, ed.), pp. 13-15. Swiss Federal Research Station, CH-8064 Zurich-Reckenholz, Switzerland. Lancashire, P. D., and Jones, D. G. 1985. Ann. Appli. biol. 106,541-553. Lee, T. S., and Cough, F. J. 1984. Plant Dis.68,848-851. Leonard, K. J. 1988. J . Plant Dis. Prof. 95,337-346. Mackie, W. W. 1929. Phytoparhology 19, 1139-1 140. Marshall, D. 1985. Phytopathology 75, 1319. Marshall, D. 1989. In “Spatial Components of Plant Disease Epidemics” (M. J. Jeger, ed.), pp. 182-203. Prentice-Hall, Engelwood Cliffs, New Jersey. Mathur, S. B., and Lee, S. L. N. 1978. Seed Sci. Technol. 6,925-926. Mullhaney, E. J., Martin, J. M., and Scharen, A. L. 1982. Euphytica 31,539-545. Mullaney, E. J., Scharen, A. L., and Bryan, M. D. 1983. Can. J . Bot. 61,2248-2250. Narvaez, I., and Caldwell, R. M. 1957. Phytopathology 47,529-530. Nass, H. G., and Johnston, H. W. 1985. Can. J . Plant Pathol. 7,91-97. Nelson, L. R. 1979. Plant Dis. Rep. 63,699-701. Nelson, L. R. 1980 Crop Sci. 20,447-449. Nelson, L. R., and Bruno, H. 1985. In “Proceedings of the Southern Small Grain Workers Conference,” p. 32 Tallahassee, Florida. Nelson, L. R., and Crowder, J. 1989. I n “Proceedings of Septoria of Cereals Workshop” (P. M. Fried ed.), pp. 142-144. Swiss Federal Station for Agronomy CH-8046 ZurichReckenholz, Switzerland. Nelson, L. R., and Gates, C. E. 1982. Crop Sci. 22,771-773. Nelson, L. R., Morey, D. D., and Brown, A. R. 1974. Plant Dis. Rep. 58,21-23. Nelson, L. R., Holmes, M. R., and Cunfer, B. M. 1976. Phytopathology 66, 1375-1379. Osbourn, A. E., Scott, P. R., and Caten, C. E. 1986. Plant Pathol. 35, 135-145. Parlevliet, J. E. 1979. Annu. Rev. Phytopathol. 17,203-222. Parlevliet, J. E. 1983. Phyroparhology 73,379. Patterson, F. L., Roberts, J. J., Finney, R. E., Shaner, G. E., Gallun, R. L., and Ohm, H. W. 1975. Crop Sci. 15,736-737. Perello, A. E., Cordo, C. A., Aniaga, H. 0. and Alippi, H. E. 1989. In “Proceedings of Septoria of Cereals Workshop” (P. M. Fried, ed.), pp. 42-46. Swiss Federal Research Station, CH-8046 Zurich-Reckenholz, Switzerland. Rapilly, F. 1988. J . Plant Dis. Prof. 95,419-427. Rapilly, F., Auriau, P., Cauderon, Y ., Richard, H., and Depatureaux, C. 1988. Agronornie 8, 647-652. Renfro, B. L., and Young, H. C., Jr. 1956. Phyropathology 46,23-24. Rillo, A. O., and Caldwell, R. M. 1966. Phytopathology 56,897.

.

WHEAT RESISTANCE TO Septoria DISEASES

277

Rillo, A. O., Caldwell, R. M., and Glover, D. V. 1970. Crop Sci. 10,223-227. Robinson, R. A. 1969. Reu. Appl. Mycol. 48,593-606. Roettges, E. 1986. “Diplomarbeit” Univ. of Goettingen, Federal Republic of Germany. Rosielle, A. A. 1972. Euphyrica 21, 152-161. Roseille, A. A., and Brown, A. G. P. 1979. Euphyrica 28,385-392. Rosielle, A. A., and Brown, A. G. P. 1980. Euphyrica 29,337-346. Rufty, R. C., Hebert, T. T., and Murphy, C. F. 1981a. Phyropathology 71,593-596. Rufty, R. C., Hebert, T. T., and Murphy, C. F. 1981b. Plant Dis.65,406-409. Russell, G. E. 1981. “Plant Breeding for Pest and Disease Resistance.” Butterworth, Boston. Saadaoui, E. M. 1987. Plant Dis.71, 153-155. Saair, E. E., and Wilcoxson, R. D. 1974. Annu. Rev. Phyroparhol. 12,49-68. Scharen, A. L. 1985. Proc. Sepr. Cereals Wkshp. Bozeman, Montana. USDA-ARS N o . 12. Scharen, A. L., and Eyal, Z. 1980. Plant Dis.64,492-496. Scharen, A. L., and Eyal, Z. 1983. Phyroparhology 73, 143-147. Scharen, A. L., and Krupinsky, J. M. 1970. Phyropathology 60, 1480-1485. Scharen, A. L., and Krupinsky, J. M. 1978. Phyroparhology 68,245-248. Scharen, A. L., Eyal, Z., Huffman, M. D., and Prescott, J. M. 1985. Phyroparhology 75, 1463- 1468. Scott, P. R. 1973. Ann. Appli. Biol. 75,321-329. Scott, P. R., and Benedikz, P. W. 1977. Ann. Appl. Biol. 85,345-358. Scott, P. R., Benedikz, P. W., and Cox, C. J. 1982. Plant Parhol. 31,45-60. Shaner, G . , and Finney, R . L. 1982. Phytoparhology 72, 154-158. Shaner, G., Finney, R. L., and Patterson, F. L. 1975. Phyroparhology 65,761-766. Sharp, E. L., Broennimann, A., and McNeal, F. H. 1972. Planr Dis. Rep. 56,761-764. Shearer, B. L. 1978. Nerh. J . Plant Parhol. 84, 1-12. Shearer, B. L., and Calpouzos, L. 1973. Plant Dis. Rep. 57,99-103. Shearer, B. L., and Zadoks, J. C. 1972. Neth. J . Plant Parhol. 78,231-241. Shearer, B. L., and Zadoks J. C. 1974. Nerh. J . Plant Pathol. 80,48-60. Shipton, W. A., Boyd, W. R. J., Rosielle, A. A., and Shearer, B. I. 1971. Bor. Rev. 37, 23 1-262. Silfhout, C. H. van, Arama P. F., and Kema, G. H. J. 1989. I n “Proceedings of Septoria of Cereals Workshop” (P. M. Fried, ed.), pp. 36-38. Swiss Federal Research Station, CH-8046 Zurich-Reckenholz, Switzerland. Stooksbury, D. E., Johnson, J. W., and Cunfer, B. M. 1987. Plant Dis.71, 1109-1112. Tavella, C. M. 1978. Euphyrica 27,577-580. Trottett, M., and Benacef, N. 1989. I n “Proceedings of Septoria of Cereals Workshop” (P. M. Fried, ed.), pp. 47-50. Swiss Federal Research Station, CH-8046 ZurichReckenholz, Switzerland. Vanderplank, J. E. 1963. “Plant Diseases: Epidemics and Control.” Academic Press, New York. Van Ginkel, M., and Scharen, A. L. 1987. Phytoparhology 77, 1629-1633. Van Ginkel, M., and Scharen, A. L. 1988. Phyroparhology 78,762-766. Weber, G . F. 1922. Phytoparhology 12,539-589. Wilkinson, C . A., Murphy, J. P., and Rufty, R. C.. 1990. Plant Dis.7 4 , 4 7 4 0 . Wilson, R. E. 1979. Aust. Plant Parhol. 8, 16-18. Yechilevich-Auster, M., Levi, E., and Eyal, Z. 1983. Phytopathology 73, 1077-1083. Zadoks, J. C., and Schein, R. D. 1979. “Epidemiology and Plant Disease Management.” Oxford Univ. Press, New York. Zilberstein, M., Blum, A,, and Eyal, Z. 1985. Phytopathology 75,226-230. Ziv, O., and Eyal, Z. 1976. Phyroparhology 66,485-488. Ziv, O . , Sacks, J. M., and Eyal Z. 1981. Phyroparhology 71, 119-123.

This Page Intentionally Left Blank

INDEX A

Absorption nitrogen fixation by legumes and, 174 root growth models and, 118-120, 126, 127 Acacia, forage tree legumes and, 29, 30, 36 Acetylene reduction assay, nitrogen fixation by legumes and, 174, 175, 177, 187 Adaptation cowpea and, 145 genetic resources in cereals and, 94,97, 100, 104, 106, 107 Gossypium and, 241,243-245 multilocation trials and, 72, 74-76 nitrogen fixation by legumes and, 182, 190 time of seedling emergence and, 17, 18 wheat resistance to Septoria and, 269 Additive genes, wheat resistance to Septoria and, 271 Additive Main effect and Multiplicative Interaction, 59,71, 72,76-82 Aegilops, genetic resources in cereals and, 97-99, 101 Aeration root growth models and, 124, 129, 130 time of seedling emergence and, 5 Age cowpea and, 140 forage tree legumes and, 41,44 Gossypium and, 238,240,244 nitrogen fixation by legumes and, 172, 175

root growth models and, 130 components, 118, 120, 124, 125 features, 117 time of seedling emergence and, I6 Agricultural Research Service, Gossypium and, 235,237-244 Agropyron genetic resources in cereals and, 97,98 wheat resistance to Septoria and, 270

Alleles genetic resources in cereals and, 101 Gossypium and, 249 Allotetraploids, Gossypium and, 227, 244, 253 distribution, 231-235 evaluation, 245, 246,248-252 Allozymes, Gossypium and, 249 Aluminum, root growth models and, 129 Amides, nitrogen fixation by legumes and, 168, 169 Amino acids, nitrogen fixation by legumes and, 169, 170 Ammonia, nitrogen fixation by legumes and, 167-169, 174 Amphidiploidy, Gossypium and, 229, 231-233 Aphids, cowpea and, 139, 140, 149 Aphis craccivora, cowpea and, 139, 140 Aridity, Gossypium and, 227, 229,242

B

Backcross cowpea and, 143 genetic resources in cereals and, 106 Gossypium and, 232,251,252 Bacteria cowpea and, 134, 136, 138, 148, 149 Gossypium and, 252 Barley genetic resources in cereals and, 87, 89, 91-98, 100-103, 105 multilocation trials and, 63 wheat resistance to Septoria and, 268 Bean, nitrogen fixation by legumes and, 204,206,207,212 Beet, time of seedling emergence and, 4,5,7 Biochemistry cowpea and, 142 nitrogen fixation by legumes and, 207

279

280

INDEX

time of seedling emergence and, I5 Biology cowpea and, 141, 145 forage tree legumes and, 47 multilocation trials and, 56,62-66, 81,82 nitrogen fixation by legumes and, 156 root growth models and, 114 Blight cowpea and, 144, 146 Gossypium and, 252 Boll weevil, Gossypium and, 244,245 Bradyrhizobium japonicum, nitrogen fixation by legumes and, 202,204,212 Branching rate, root growth models and, 124, 125 Breeding cowpea and, 139, 141, 143, 150 genetic resources in cereals and, 87, 88, 90,93,96-107 Gossypium and, 244,251, 253 multilocation trials and, 5 5 , 56, 81 AMMI analysis, 77 joint linear regression, 61,64, 67,68 multivariate analyses, 74 variance, 60 nitrogen fixation by legumes and, 205-207,209 wheat resistance to Septoria and, see Wheat resistance to Septoria Buckwheat, time of seedling emergence and, 17, 18,21 Bulk density, root growth models and, 124, I29

C

Calcium cowpea and, 146 forage tree legumes and, 41 root growth models and, 123, 129 Calliandra, forage tree legumes and, 29 Calliandra calothyrsus, forage tree legumes and management, 38-40,42 nitrogen recycling, 47 nitrogen yields, 45 performance, 30 Callus, cowpea and, 142, 143

Calories, nitrogen fixation by legumes and, 177, 178 Canopy cowpea and, 138 forage tree legumes and, 32-34,43,44 Gossypium and, 252 root growth models and, 117 time of seedling emergence and, 17 Carbohydrate forage tree legumes and, 36 nitrogen fixation by legumes and, 206 root growth models and, 113, 117, 122, 127-129 time of seedling emergence and, 18 Carbon forage tree legumes and, 47 nitrogen fixation by legumes and, 162 root growth models and, 128 time of seedling emergence and, 9 Carrot genetic manipulation and, 143 multilocation trials and, 72 time of seedling emergence and, 1, 5.6, 11, 14, 15 Cellular selection, cowpea and, 143, 145 Cereal genetic resources in, 87-89, 105-107 barley, 93-95 documentation, 102-105 Ethiopia, 95, 96 primitive forms, 96-102 wheat, 89-93 nitrogen fixation by legumes and assessment, 176 contribution to production, 191, 194, 196- I99 enhancement, 213, 215 production systems, 177, 178, 192-184 root growth models and, 118, 120, 130 Characterization, genetic resources in cereals and, 88,89, 100, 104 Chenopodium, time of seedling emergence and, 14, 15 Chimaeras, cowpea and, 147, 148 Chlorophyll genetic resources in cereals and, 91 time of seedling emergence and, 18 Chloroplasts cowpea and, 146, 147 Gossypium and, 234,235, 249, 250

28 1

INDEX Chromatography, nitrogen fixation by legumes and, 174 Chromosomes cowpea and, 145 genetic resources in cereals and, 93 Gossypium and, 231-235, 245-247 wheat resistance to Septoria and, 270-272 Climate forage tree legumes and, 39.49 genetic resources in cereals and, 94,95, 101 Gossypium and, 238 multilocation trials and, 80 nitrogen fixation by legumes and, 182 time of seedling emergence and, 19 Clinal pattern, genetic resources in cereals and, 90,96 Cluster analysis, multilocation trials and, 71,74-76,81 Cocksfoot, time of seedling emergence and, 11, 15 Codariocalyx gyroides, forage tree legumes and, 38,39 Coevolution, genetic resources in cereals and, 89 Collection, Gossypiurn and, 235-245 Commonwealth Scientific and Industrial Research Organization, Gossypium and, 238, 241 Compaction, root growth models and, 124 Compensation nitrogen fixation by legumes and, 156 wheat resistance to Seproria and, 266 Competition forage tree legumes and, 33,42,44 multilocation trials and, 57 nitrogen fixation by legumes and, 174, 182,216 time of seedling emergence and, 2, 10, 1 1 . 14-21 Composite cross, genetic resources in cereals and, 105 Conservation, genetic resources in cereals and, 88, 105, 106 Core collections, genetic resources in cereals and, 104 Cotton, see Gossypium Cover crops, nitrogen fixation by legumes and, 200

Cowpea genetic manipulation of, 133-139, 149, 150 embryo culture, 145, 146 fungal pathogens, 141, 142 insects, 139-141 somaclonal variation, 144, 145 somatic hybridization, 146-148 tissue culture technology, 142-144 transformation, 148, 149 nitrogen fixation by legumes and contribution to production, 195-197 enhancement, 210, 215 production systems, 178, 181-183 Crop root growth models, see Root growth models Cross-pollination, cowpea and, 145 Crossover interactions, multilocation trials and, 68-10 Cutting, forage tree legumes and, 38-41, 43,44 Cytogenetics, Gossypium and, 231,236 Cytology, Gossypium and, 229, 233,247, 248 Cytoplasm cowpea and, 145-147 Gossypium and, 253

D

Decay forage tree legumes and, 47,48 nitrogen fixation by legumes and, 191 root growth models and, 113, 114, 117, 118 Decomposition, nitrogen fixation by legumes and, 191, 192, 197-199 Deforestation, genetic resources in cereals and, 96 Degree days, root growth models and, 128, I30 Denitrification, fixation by legumes and, 213 Density forage tree legumes and, 41-44,46 multilocation trials and, 55 time of seedling emergence and, 10-17, 19,20

282

INDEX

Detrended correspondence analysis, multilocation trials and, 80 Dicotyledon cowpea and, 148 root growth models and, 120 time of seedling emergence and, 6 Diffusion Gossypium and, 249 root growth models and, 114, 125, 126 Digitalis purpurea, time of seedling emergence and, 14, 16 Diploidy, Gossypium and, 226, 253 distribution, 229, 231-234 evaluation, 245,247, 250 Discriminant analysis, multilocation trials and, 71 Disease cowpea and, 137-142, 149 forage tree legumes and, 28 genetic resources in cereals and, 88.91, 94,95,97,99, 102-104 Cossypium and, 236,252 nitrogen fixation by legumes and, 187, I97 wheat resistance to Septoria and, see Wheat resistance to Septoria Distribution, Gossypium and, 228-235 DNA genetic resources in cereals and, 93, 106 Gossypium and, 234,235,249,250 Documentation, genetic resources in cereals and, 102-106 Domestication genetic resources in cereals and, 96 Gossypium and, 226-228 Dominance multilocation trials and, 63, 64 time of seedling emergence and, 2, 17 wheat resistance to Seproria and, 265, 270,271 Dormancy, time of seedling emergence and, 2, 7, 14 Drought forage tree legumes and, 34 genetic resources in cereals and, 91,97, 100, 103 E

Ecology cowpea and, 139

genetic resources in cereals and, 89, 104 Gossypium and, 241-243,253 multilocation trials and, 80 Economics cowpea and, 140 forage tree legumes and, 49 genetic resources in cereals and, 91,95, 100, 104, 107 Gossypium and, 244, 253 nitrogen fixation by legumes and, 176, 178, 186, 187 time of seedling emergence and, 2 Electrophoresis genetic resources in cereals and, 100, 101,106 Ethiopia, 96 wheat, 92,93 Gossypium and, 247-250 Embryo cowpea and, 143-146 Gossypium and, 228 time of seedling emergence and, 6 , 7 Endonuclease restriction, Gossypium and, 249,250 Environment cowpea and, 143, 149 forage tree legumes and, 36,49 genetic resources in cereals and, 88, 92-94,97,98, 100-102, 104-107 Gossypium and, 236, 245,247, 251, 253 multilocation trials and, 5 5 , 56, 80-82 AMMI analysis, 76-90 crossover interactions, 68-70 joint linear regression, 61-68 multivariate analyses, 70-76 variance, 57-60 nitrogen fixation by legumes and, 156 assessment, 158, 172 contribution to production, 199 enhancement, 204,206,211,214 production systems, 182, 190 root growth models and, 124, 127, 130, 131 time of seedling emergence and, 6-8, 19 wheat resistance to Seproria and, 264, 270,272, 273 Enzymes genetic resources in cereals and, 94 Gossypium and, 235,248,250 nitrogen fixation by legumes and, 174

283

INDEX Epidemics, wheat resistance to Septoriu and, 273,274 Erosion forage tree legumes and, 33 nitrogen fixation by legumes and, 184, 187 Ethiopia, genetic resources in cereals and, 90,91,94-96 Evolution cowpea and, 145, 149 genetic resources in cereals and, 89,97, I02 Gossypium and, 231-235.248-250 nitrogen fixation by legumes and, 174, 206 Expected mean squares, multilocation trials and, 60 F

Factor analysis, multilocation trials and, 71.75 Fecundity cowpea and, 146 time of seedling emergence and, 2, 12, 13, 15 Feedback root growth models and, 127 time of seedling emergence and, 17 Fertility genetic resources in cereals and, 90 Gossypium and, 232, 253 mu1tilocation.trials and, 57 nitrogen fixation by legumes and, 214 root growth models and, 116, 123, 124 Fertilization, cowpea and, 145, 146 Fertilizer forage tree legumes and management, 41 nitrogen recycling, 48, 49 nitrogen yields, 46 performance, 34,35 species, 28 multilocation trials and, 55 nitrogen fixation by legumes and, 156, 157 assessment, 175, 176 contribution to production, 193, 196, 200,201 enhancement, 202,203,206,212, 213

production systems, 180, 181, 183 root growth models and, 117 time of seedling emergence and, 4 Fitness, wheat resistance to Septoriu and, 268 Fixation, nitrogen, see Nitrogen fixation Flooding Gossypium and, 237 nitrogen fixation by legumes and, 187 root growth models and, 130 Flowering cowpea and, 136, 137, 145 genetic resources in cereals and, 94 Gossypium and, 240,243 nitrogen fixation by legumes and, 208 Fluorescence, wheat resistance to Septoriu and, 261 Food legumes, nitrogen fixation and, 177- 184 Forage, nitrogen fixation by legumes and, 156 assessment, 170, 171 contribution to production, 190, 200-202 production systems, 184, 185, 188, 189 Forage tree legumes, 27, 28,49, 50 animal feed, 34-36 management, 36-38 cutting, 38-41 density, 41-44 nitrogen recycling, 46-49 nitrogen yields, 45,46 performance, 29-34 species, 28, 29 Fungus cowpea and, 136, 138, 139, 141, 142, 145, 149 time of seedling emergence and, 8 wheat resistance to Septoriu and, 257, 262-265,268 Fusion, cowpea and, 146, 147

G

Gene pool, genetic resources in cereals and, 101, 103, 107 General combining activity, wheat resistance to Septoriu and, 264, 265, 27 1 Genetic diversity cowpea and, 138, 139

284

INDEX

genetic resources in cereals and, 92-94, loo, 102,106 Gossypium and, 249 Genetic erosion, genetic resources in cereals and, 96, 105 Genetics cereals and, see Cereals, genetic resources in cowpea and, see Cowpea, genetic manipulation of Gossypium and collection, 236 distribution, 232, 233 evaluation, 245-247, 249,25 1,253 multilocation trials and, 60-62, 79 nitrogen fixation by legumes and, 210 root growth models and, 131 time of seedling emergence and, 2 , 7 , 8 wheat resistance to Septoria and, 269-273 Genomes cowpea and, 145, 147 genetic resources in cereals and, 92,98 Gossypium and distribution, 229-235 evaluation, 247-250,253 Genotype cowpea and, 142, 143, 145 genetic resources in cereals and, 91,92, 101,104

Gossypium and, 251 multilocation trials and, 55, 56,80-82 AMMI analysis, 76-80 crossover interactions, 68-70 joint linear regression, 61-68 multivariate analyses, 70-76 variance, 57-60 nitrogen fixation by legumes and, 158, 190,205,206,208,216 stability, multilocation trials and, 61 wheat resistance to Septoria and, 258, 260,272-274 Geography genetic resources in cereals and, 89,90, 92,94,95, 102, 105 Gossypium and, 226 collection, 236, 242 distribution, 228, 229 evaluation, 248,249,251 multilocation trials and, 82

wheat resistance to Septoria and, 258, 269 Germination root growth models and, 123 time of seedling emergence and, 2-9 GermpIa sm cowpea and, 141 genetic resources in cereals and, 88-91, 93,95-98, 103-107 Gossypium and, 253 collection, 239-242, 245 distribution, 236 evaluation, 250-252 wheat resistance to Septoria and, 258-261,267,268,272-274 Gliadin, genetic resources in cereals and, 96,98 Gliricidia, forage tree legumes and, 30, 35,47 Gliricidia sepium, forage tree legumes and, 28-30, Glutamate synthase, nitrogen fixation by legumes and, 167, 169 Glutamine, nitrogen fixation by legumes and, 167, 168 Glutamine synthetase, nitrogen fixation by legumes and, 167, 169 Gluten, genetic resources in cereals and, 92 Glutenins, genetic resources in cereals and, 101 Gossypium, 225, 226, 253 collection, 235 plant explorations, 237-245 source, 235, 236 distribution evolution, 231-235 geography, 228,229 species, 229-231 taxonomy, 228 evaluation electrophoresis, 247-250 genetics, 245-247 improvement, 250-253 history of domestication, 226-228 root growth models and, 114, 115 Grass forage tree legumes and management, 42 nitrogen recycling, 47

285

INDEX nitrogen yields, 46 performance, 31,33,34 nitrogen fixation by legumes and, 188, 189, 191,201,202 root growth models and, 128, 129 time of seedling emergence and, 15, 21 Grazing forage tree legumes and, 27,28 animal feed, 36 management, 36 nitrogen recycling, 47 nitrogen yields, 46 performance, 33,34 nitrogen fixation by legumes and, 174, 177,200 Green grams nitrogen fixation by legumes and, 182, 195, 196 time of seedling emergence and, 17, 18,21 Green manure forage tree legumes and, 49 nitrogen fixation by legumes and, 176, 177, 184, 186-200

H

Habitat cowpea and, 149 genetic resources in cereals and, 92,96, 103, 104 Gossypium and, 234,238 Harvest cowpea and, 138, 140 forage tree legumes and, 29, 36, 37,40, 41,43,45-47 Gossypium and, 244,252 nitrogen fixation by legumes and, 56 assessment, 172 contribution to production, 191 193, 197,200 time of seedling emergence and, I 114, 15, 17 Harvest index, nitrogen fixation by legumes and, 191, 193, 198 Hierarchy multilocation trials and, 75,76 time of seedling emergence and, 2, 14, 15

Homology genetic resources in cereals and, 98 Gossypium and, 23 I, 232, 247 Hordeum, time of seedling emergence and, 14, I5 Hordeum murinum, genetic resources in cereals and, 100 Hordeum spontaneum, genetic resources in cereals and, 93, 100, 101 Hordeum uulgare, genetic resources in cereals and, 92,96 Humidity cowpea and, 133, 135, 141 forage tree legumes and, 28 genetic resources in cereals and, 91 Gossypium and, 252 nitrogen fixation by legumes and, 156, I86 time of seedling emergence and, 16 Hybrids cowpea and, 143-148, 150 genetic resources in cereals and, 90, 96,98 Gossypium and, 231-234,250-253 multilocation trials and, 77 nitrogen fixation by legumes and, 206, 213 Hydrogen, nitrogen fixation by legumes and, 175, 206 Hysteresis, root growth models and, 130 I

ICARDA, genetic resources in cereals and, 91,94, 102, 103 Immobilization, nitrogen fixation by legumes and, 191, 213 Immunity, wheat resistance to Septoria and, 259,260, 262 Impedance root growth models and, 129 time of seedling emergence and, 5 , 8 Incompatibility, Gossypium and, 247 Incubation, wheat resistance to Septoria and, 262-267,271 Inhibition forage tree legumes and, 42 nitrogen fixation by legumes and, 174, 202,206

286

INDEX

time of seedling emergence and, 3 , 7 , 8 wheat resistance to Septoria and, 258 Inoculation Gossypium and, 252 nitrogen fixation by legumes and, 202, 203,210-213,216 wheat resistance to Septoria and, 259, 260,263,265-268,274 Insects cowpea and, 136-143, 148-150 forage tree legumes and, 28, 50 Gossypium and, 236,241,244,252 nitrogen fixation by legumes and, 197 International Board for Plant Genetic Resources genetic resources in cereals and, 88, 103 Gossypium and, 231, 239,241, 242 International Institute of Tropical Agriculture, cowpea and, 137-142, 149 International Wheat and Maize Improvement Center, genetic resources in cereals and, 93, 94 Irradiance, forage tree legumes and, 32,44 Irrigation, nitrogen fixation by legumes and, 182,214 Isoenzymes, nitrogen fixation by legumes and, 207 Isolation, cowpea and, 146, 147 Isotopes, nitrogen fixation by legumes and, 158-165 Isozymes, genetic resources in cereals and, 88 J

Joint linear regression, multilocation trials and, 61-68 L

Landraces, genetic resources in cereals and, 89-94,98, 102, 106, 107 Latency, wheat resistance to Septoria and, 263,265-267,271,273 Leaf Area Index, forage tree legumes and, 43 Legumes forage tree, see Forage tree legumes nitrogen fixation by, see Nitrogen fixation by legumes

Lesions, wheat resistance to Septoria and genetics, 271 identification, 259-266 pathogen variation, 269 Lettuce, time of seedling emergence and, 6,7,9 Leucaena leucocephala, forage tree legumes and, 49 animal feed, 34, 35 management, 38-40,42 nitrogen recycling, 47,48 nitrogen yields, 45 performance, 29,30 Light forage tree legumes and, 30,32,33, 42,44 time of seedling emergence and, 7, 16, 17, 19 Linear regression, multilocation trials and, 61-68,81 Linkage genetic resources in cereals and, 94 Gossypium and, 246 Livestock, nitrogen fixation by legumes and, 184,200-202 Longevity, time of seedling emergence and, 19,20 Lysine, genetic resources in cereals and, 92,99, 102 M

Maize forage tree legumes and, 49 genetic manipulation and, 135, 144, 145 multilocation trials and, 73, 75, 79 nitrogen fixation by legumes and contribution to production, 196, 197, 199 enhancement, 216 production systems, 182, 184 Maruca testulalis, genetic manipulation and, 139-141, 149 Megalurothrips sjostedti, genetic manipulation and, 139, 140 Meiosis, Gossypium and, 231,232,246 Mildew cowpea and, 144 genetic resources in cereals and, 90,94, 96, 100 wheat resistance to Sepioria and, 273

287

INDEX Millet genetic manipulation and, 135 multilocation trials and, 67 nitrogen fixation by legumes and, 182 Mimosine toxicity, forage tree legumes and, 35,36 Mineralization, nitrogen fixation by legumes and, 162, 191, 192, 197, 198 Mitochondria, cowpea and, 146, 147 Mitosis, cowpea and, 144, 145, 147 Models, root growth, see Root growth models Moisture cowpea and, 136 forage tree legumes and, 30,42 genetic resources in cereals and, 97 nitrogen fixation by legumes and, 182, 197 root growth models and, 130 time of seedling emergence and, 3-6, 8,9 wheat resistance to Septoria and, 266 Moisture stress forage tree legumes and, 36 nitrogen fixation by legumes and, 214 Monocotyledons cowpea and, 148 time of seedling emergence and, 6 Monosomes Gossypium and, 246,247 wheat resistance to Septoria and, 271 Monotelodisomes, Gossypium and, 246, 247 Morphology cowpea and, 145 genetic resources in cereals and, 88.91, 92,94-96,98-100, 104 Gossypium and, 231,241-244,246-249 multilocation trials and, 75 nitrogen fixation by legumes and, 166 root growth models and, 120, 130 time of seedling emergence and, 17-19 Mortality root growth models and, 128 time of seedling emergence and, 2, 20 Multilocation trials, 55, 56, 80-82 AMMI analysis, 76-80 crossover interactions, 68-70 genetic resources in cereals and, I 0 4 joint linear regression limitation, 61-65

risk, 67,68 yield stability, 65, 66 multivariate analyses, 70, 71 cluster analysis, 75,76 factor analysis, 75 principal components, 71-74 principal coordinates, 74,75 variance, 57 components, 59-61 limitations, 58, 59 Multiple linear regression model, multilocation trials and, 80 Multivariate analysis genetic resources in cereals and, 91 multilocation trials and, 70-76 Mutagenesis, nitrogen fixation by legumes and, 206 Mutation cowpea and, 145 Gossypium and, 233,245, 246 nitrogen fixation by legumes and, 207 wheat resistance to Septoria and, 272

N

Natural selection, genetic resources in cereals and, 100 Necrosis, wheat resistance to Septoria and genetics, 271 identification, 260, 261, 263-265 pathogen variation, 269 New squares, multilocation trials and, 60 Nitrate, fixation by legumes and assessment, 168-171, 175, 176 enhancement, 202,204,212-215 Nitrate reductase, fixation by legumes and, 207 Nitrate reduction, fixation by legumes and, 168, 169 Nitrate tolerance, fixation by legumes and, 206-210 Nitrogen cowpea and, 134, 135 forage tree legumes and, 41,42,50 animal feed, 34, 35 management, 43 nitrogen recycling, 46-49 nitrogen yields, 45,46 species, 28, 29 multilocation trials and, 72, 73

288

INDEX

root growth models and, 124, 128-130 time of seedling emergence and, 4, 18 Nitrogen fixation cowpea and, 133, 135 forage tree legumes and, 28,29,34,42, 46,48 by legumes, 156-158,216 acetylene reduction assay, 174, 175 balance, 174 crops, 191-200 direct transfer, 190, 191 enhancement, 202-2 16 fertilizer, 175, 176 isotopic techniques, 158-165 livestock, 200-202 N-difference method, 165, 166 nodules, 176, 177 production, 177-190 ureide method, 167-173 Nitrogenase, fixation by legumes and, 174, 175 Nodulation cowpea and, 134 nitrogen fixation by legumes and, 156, 157 assessment, 165, 167-170, 173, 175-1 77 contribution to production, 191-193 enhancement, 202,204-208,211-214 Nopaline synthase promoter, cowpea and, 148 Nucleotides genetic resources in cereals and, 93 Gossypium and, 234 Nucleus cowpea and, 146, 147 Gossypium and, 253 Nutrient uptake, root growth models and, 113, 117, 126, 127 Nutrition forage tree legumes and, 28, 30.41 Gossypium and, 247 nitrogen fixation by legumes and, 156, 172, 177, 184, 204 root growth models and, 128 0

Onion, time of seedling emergence and, 4 , 6

Ootheca mutabilis, genetic manipulation and, 139-141 Osmotic effects, time of seedling emergence and, 4 , 6 , 7 Oxygen nitrogen fixation by legumes and, 175 root growth models and, 129 time of seedling emergence and, 3 Oxygen diffusion rate root growth models and, 123 time of seedling emergence and, 3

P Panicurn maximum, forage tree legumes and, 33, 35,42 Partitioning nitrogen fixation by legumes and, 173, 191,206 root growth models and, 122, 123, 128, 129 time of seedling emergence and, 16 Pathogens cowpea and, 141, 142, 149 root growth models and, 114 wheat resistance to Septoria and, 274 identification, 258,259,262, 263,267 variation, 268-270 Pennisetum, forage tree legumes and, 32, 33,42 Peroxide, time of seedling emergence and, 3

PH cowpea and, 146 forage tree legumes and, 30 root growth models and, 123 Phaseolae, cowpea and, 134, 141 Phaseolastrae, cowpea and, 135, 144 Phaseolinae, cowpea and, 135, 144, 146 Phaseolus, genetic manipulation and, 148 Phaseolus vulgaris, genetic manipulation and, 146 Phenotype genetic resources in cereals and, 90,95, 96, 104 Gossypium and, 252 nitrogen fixation by legumes and, 206 Phosphorus forage tree legumes and, 41 root growth models and, 124

289

INDEX time of seedling emergence and, 4 Photosynthate, root growth models and, 127, 128 Photosynthesis root growth models and, 117, 125 time of seedling emergence and, 9, 16, 17,28 wheat resistance to Septoria and, 268 Physiology cowpea and, 142 genetic resources in cereals and, 91, 98, 100

multilocation trials and, 80, 82 nitrogen fixation by legumes and, 172 root growth models and, 113, 114, 118, I19 time of seedling emergence and, 15 wheat resistance to Septoria and, 268, 269 Phytoalexins, cowpea and, 142 Planlago, time of seedling emergence and, 16, 19 Plasmogamy, cowpea and, 146, 147 Ploidy cowpea and, 145 genetic resources in cereals and, 92.93, 95,96,99, 100 Gossypium and, 253 Point mutation, cowpea and, 145 Polyacrylamide gel electrophoresis, genetic resources in cereals and, 92.98 Polymorphism, genetic resources in cereals and, 89,95 Population cowpea and, 142, 144, 145, 147 forage tree legumes and, 46 Gossypium and, 232,238,240,242, 243, 247-249,253 multilocation trials and, 65, 72 nitrogen fixation by legumes and, 156, 202,209,210,213,216 time of seedling emergence and, see Seedling emergence, time of Potassium, root growth models and, 126 Potato genetic manipulation and, 144, 145, 149, 150 root growth models and, 128 Predictive criteria, multilocation trials and, 56,77-80,82

Primary evaluation, genetic resources in cereals and, 88.89 Principal components analysis, multilocation trials and, 71-74, 76, 80-82 Principal components axes, multilocation trials and, 73,77-80 Principal coordinates analysis, multilocation trials and, 71, 74, 75 Protein cowpea and, 135, 138 forage tree legumes and, 28,35,36, 41,46 genetic resources in cereals and, 88, 92, 93,96,97,99-102 Gossypium and, 226, 247 nitrogen fixation by legumes and, 156, 177, 178, 184,202 Protoplast, cowpea and, 142-144, 146-148 Pycnidia, wheat resistance to Septoria and identification, 260, 262, 264-266 pathogen variation, 268-270

Q Qualitative interactions, multilocation trials and, 68 Quantitative interactions, multilocation trials and, 68 Quantitative traits, genetic resources in cereals and, 92

R Rain Forest Belt, cowpea and, 135, 138, 139, 141, 145, 149 Rainfall cowpea and, 135, 137 Gossypium and, 237,244 nitrogen fixation by legumes and, 213 root growth models and, 114 Reciprocal average, multilocation trials and, 80 Reciprocal crosses, wheat resistance to Septoria and, 264 Recombination cowpea and, 145, 147 genetic resources in cereals and, 98 Gossypium and, 242, 247, 249,251

290

INDEX

Recycling, forage tree legumes and, 46 Regrowth, forage tree legumes and, 40, 42-44 Relative efficiency, nitrogen fixation by legumes and, 206 Relative growth rate, time of seedling emergence and, 14, I5 Replicates genetic resources in cereals and, 89, 91,97 multilocation trials and, 56, 57, 76,77 Resistance cowpea and, 139, 141, 144, 148-150 genetic resources in cereals and, 88.94, 97-99 Gossypium and, 236,245,252 of wheat to Septoria, see Wheat resistance to Septoria Respiration root growth models and, 122, 123, 127- I29 time of seedling emergence and, 16, 18 Restricted maximum likelihood, multilocation trials and, 60, 61 Restriction fragment length polymorphism genetic resources in cereals and, I 0 6 wheat resistance to Seproria and, 214 Rhizobia, nitrogen fixation by legumes and, 202-205,207,210,212, 213 Rhizobium cowpea and, 134 nitrogen fixation by legumes and, 172, 175,206,210-213 Rhizosphere, root growth models and, 113, 126, 127 Rhizotron, root growth models and, 115, 116, 123 Rice genetic manipulation and, 144 genetic resources in cereals and, 87 nitrogen fixation by legumes and, 163, 182, 187, 198, 199 Risk, multilocation trials and, 67 Root cowpea and, 142 elongation rate, root growth models and, 124, 125 forage tree legumes and, 36,47 genetic resources in cereals and, 107 length density, 114, 116-1 18, 122, 123

nitrogen fixation by legumes and assessment, 160, 166, 167, 175, 176 contribution to production, 191-193, 197 enhancement, 204,207 production systems, 187 sink, 114, 126 time of seedling emergence and, I 1 Root growth models, 113, 114 components classification, 118-121 parameters, 120, 122, 123 soil, 123-125 uptake functions, 125-127 early models, 114, 115 existing models, 128-130 features, 115-1 18 limitations, 130, 131 Rotation forage tree legumes and, 34 nitrogen fixation by legumes and, 193, 197,21I , 216 Rust genetic resources in cereals and, 94, 98,99 Gossypium and, 252 wheat resistance to Septoria and, 259, 272,213 5

Salinity, genetic resources in cereals and, 91.97 Sand, root growth models and, 129 SDS-PAGE, genetic resources in cereals and, 101 Seed cowpea and, 134, 138, 139 genetic resources in cereals and, 88,90 Gossypium and, 227,228 collection, 235, 237,239, 243 evaluation, 248, 250 nitrogen fixation by legumes and, 156, 157 contribution to production, 191, 193-195, 197, 198 enhancement, 204,208,209,212,214, 215 wheat resistance to Septoria and, 261, 264

INDEX Seedling emergence, time of, 1 , 2, 20, 21 factors, 2,3, 8, 9 seed attributes, 6-8 sowing depth, 6 temperature, 4 , 5 water, 3 , 4 importance of variation, 9, 10 longevity, 19, 20 organ composition, 17-19 partitioning, 16, 17 total plant growth, 10-15 Selection cowpea and, 143, 145 nitrogen fixation by legumes and, 205, 206,209 wheat resistance to Septoria and, 268, 269, 272-274 Septoria nodorum, wheat resistance to, 257,268-270,272-274 Septoria nodorum blotch, wheat resistance to, 257,258,272 genetics, 271 identification, 259-264,267, 268 pathogen variation, 268, 269 Septoria tritici genetic resources in cereals and, 91 wheat resistance to, 258, 272-274 genetics, 270, 272 identification, 260, 263 pathogen variation, 269, 270 Septoria tritici blotch, wheat resistance to, 257,258,272 genetics, 270-272 identification, 259-262, 267, 268 pathogen variation, 269, 270 Sesbania, forage tree legumes and, 29 Sesbania cannabina, forage tree legumes and, 45 Sesbania formosa, forage tree legumes and, 30 Sesbania grandifiora, forage tree legumes and, 35, 38,40,45,47 Sesbania sesban, forage tree legumes and, 30, 38,41,45 Seteria, forage tree legumes and, 32, 35 Sexual incompatibility, cowpea and, 145, 146 Shading multilocation trials and, 57 nitrogen fixation by legumes and, 184

29 1

Soil cowpea and, 134-136 forage tree legumes and, 47,49 animal feed, 36 management, 41-43 nitrogen yields, 46 performance, 30 genetic resources in cereals and, 91 nitrogen fixation by legumes and, 156, 157,216 assessment, 158-166, 174-177 contribution to production, 191, 193, 197, 198,200,201 enhancement, 202,204,208-210, 212-21 6 production systems, 182, 187, 190 root growth models and, 113-1 15, 128-1 30 components, 118, 120, 122-127 features, 116, 117 time of seedling emergence and, 3.4, 6-9 Soil moisture forage tree legumes and, 30 nitrogen fixation by legumes and, 213 Somaclonal variation, cowpea and, 143-145, 147 Somatic crossover, cowpea and, 145 Somatic hybridization, cowpea and, 143-148 Sorghum genetic manipulation and, 135, 144 nitrogen fixation by legumes and, 182 Sowing cowpea and, 137 forage tree legumes and, 38 nitrogen fixation by legumes and, 200, 204,209,210 time of seedling emergence and, 4,6,9, 10, 14, 17, 20 Soybean multilocation trials and, 72-74,77-79 nitrogen fixation by legumes and, 156, 157 assessment, 163, 165, 170, 171, 173 contribution to production, 193, 194, 196, 197 enhancement, 202,204,207,208, 212-215 production systems, 176, 178-182, 184

292

INDEX

root growth models and, 117 Specific combining ability, wheat resistance to Septoria and, 264, 265, 27 1 Spore production, wheat resistance to Septoria and, 258,259,262-268 Statistical analysis of multilocation trials, see Multilocation trials Sterility, Gossypium and, 247,253 Stress genetic resources in cereals and, 88,91, 96, 100, 106 Gossypium and, 236 multilocation trials and, 80 nitrogen fixation by legumes and, 175 root growth models and, 128, 130 Subterranean clover, time of seedling emergence and, 9, 10 Subtropical agriculture, nitrogen fixation by legumes and, see Nitrogen fixation by legumes Sugar cane, genetic manipulation and, 144, 145, 150 Sum of squared difference, multilocation trials and, 77,78 Suppression, time of seedling emergence and, 17, 18 Survival Gossypium and, 247 time of seedling emergence and, 19, 20 Susceptibility, wheat resistance to Septoria and, 273 genetics, 270-272 identification, 258-262, 264,265 pathogen variation, 269,270 Symbiosis cowpea and, 134 forage tree legumes and, 46,48, 50 nitrogen fixation by legumes and, 156, 157

assessment, 158, 161, 165, 169, 173, 175, 176 enhancement, 206-21 1 production systems, 179 Synchrony cowpea and, 138 time of seedling emergence and, 6 , 9 , 2 0 Systematic evaluation, genetic resources in cereals and, 89

T

Taxonomy cowpea and, 135, 136 genetic resources in cereals and, 97 Gossypium and, 229,236,239,242 Telocentrics, Gossypium and, 246, 247 Temperature forage tree legumes and, 48 genetic resources in cereals and, 97, 103 nitrogen fixation by legumes and, 197, 213 root growth models and, 123-125, 128- 130 time of seedling emergence and, 3-6, 16 wheat resistance to Septoria and, 266 Thrips, cowpea and, 139, 140, 149 Tillage, nitrogen fixation by legumes and, 177,213-216 Tissue culture technology, cowpea and, 142-149 Tobacco cowpea and, 146, 147 time of seedling emergence and, 7 Tolerance genetic resources in cereals and, 96 Gossypium and, 227 nitrogen fixation by legumes and, 208 wheat resistance to Septoria and, 258, 259,261,267,272 Totipotency, cowpea and, 142, 143 Toxicity, cowpea and, 135, 145 Transformation cowpea and, 143-145, 148, 149 nitrogen fixation by legumes and, 197 Transgression, wheat resistance to Septoria and, 271 Translocation, Gossypium and, 232,233, 246 Transpiration forage tree legumes and, 42 nitrogen fixation by legumes and, 170 root growth models and, I18 Tree legumes, forage, see Forage tree legumes Triticum genetic resources in cereals and, 98-101 wheat resistance to Septoria and, 260, 261,263,269,270

293

INDEX Triticum aestivum genetic resources in cereals and, 91,92 wheat resistance to Septoria and, 257, 260 Triticum dicoccoides, genetic resources in cereals and, 97-99, 101 Tropical agriculture, nitrogen fixation by legumes and, see Nitrogen fixation by legumes Tropical environments forage tree legumes and, see Forage tree legumes Gossypium and, 244 multilocation trials and, 79 Turgor pressure, root growth models and, 123, 124, 129 U

United States Department of Agriculture genetic resources in cereals and, 90, 92-94, 105 Gossypium and, 235,237-241,243,244 Ureide method, nitrogen fixation by legumes and assessment, 167-173, 177 enhancement, 202,209,210,215 V

Variability cowpea and, 144, 147 genetic resources in cereals and, 88,89, 91-101, 104-106 Vigna unguiculata, see Cowpea Virulence Gossypium and, 252 wheat resistance to Septoria and, 268-270 Virus, cowpea and, 136, 138, 148 W

Water forage Cree legumes and, 46 nitrogen fixation by legumes and, 165, 182.204

root growth models and, 113-115, 128-130 classification, 118, 119 soil, 124, 125 time of seedling emergence and, 2-4, 7-9 Weeds cowpea and, 136 genetic resources in cereals and, 96 nitrogen fixation by legumes and, 184, 186, 187 time of seedling emergence and, 14, 15 Weighted average, multilocation trials and, 80 Wheat genetic manipulation and, 144, 149 genetic resources in cereals and, 87, 89-93,95-101, 103 multilocation trials and, 75 nitrogen fixation by legumes and, 182 resistance to Septoria and, 257,258, 272-274 assessment, 259-261 components, 262-266 genetics, 270-272 identification, 258,259 pathogen variation, 268-270 sources, 272 yield reduction, 267,268 root growth models and, 120, 121, 125, 128 time of seedling emergence and, 3 X

Xylem nitrogen fixation by legumes and assessment, 167-173 enhancement, 202,208-210 root growth models and, 130 Y

Yield cowpea and, 138, 140, 141, 143, 149 forage tree legumes and, 31,33,34, 37-50

294 genetic resources in cereals and, 87, 90,97 multilocation trials and, 55, 56, 81 AMMI analysis, 76, 77 joint linear regression, 61,63-68 multivariate analyses, 75

INDEX variance, 58,60 nitrogen fixation by legumes and, 166, 176, 182, 198,200,208,213-215 time of seedling emergence and, 14 wheat resistance to Septoria and, 261, 267,268,270