Advances in Agronomy, Volume 34

ADVANCES IN

AGRONOMY

VOLUME 34

CONTRIBUTORS TO THIS VOLUME M. J . AMBROSE R. B . BEVERLY W. C. GREGORY UMESHC. GUPTA Hu HAN C. L. HEDLEY

W. M. JARRELL PREMP. JAUHAR THOMASA . LARUE JOHN LIPSETT THOMAS G. PATTERSON SHAOQIQUAN JOSE

G . SALINAS

PEDROA . SANCHEZ H. K. SRIVASTAVA J. C. WYNNE

ADVANCES IN

AGRONOMY Prepared in cooperution with the AMERICAN SOCIETY OF AGRONOMY

VOLUME 34 Edited by N. C . BRADY Science and Technology Bureau Agency for International Development Department of State Washington, D . C .

ADVISORY BOARD

H. J . GORZ,CHAIRMAN E. J . KAMPRATH T. M. STARLING

J . B . POWELL J . W. BIGGAR

M . A . TABATABAI M. STELLY, EX

OFFICIO,

ASA Headquarters 1981

ACADEMIC PRESS A Subsidiary of Harcourr Bruce Jovanovich, Publishers

New York

London Toronto Sydney San Francisco

COPYRIGHT @ 1981, BY ACADEMIC PRESS, 1NC. ALL RIGHTS RESERVED. NO PART O F 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 T H E PUBLISHER.

ACADEMIC PRESS,INC.

111 Fifth Avenue, New York. New York 10003

United Kirigdom Edition published b y ACADEMIC PRESS, INC. ( L O N D O N ) LTD. 24/28 Oval Road, London NWI

7DX

LIBRARY O F CONGRESS CATALOG CARD NUMBER:50-5598 ISBN 0-12-000734-7 PRINTED IN THE UNITED STATES O F AMERICA 81 82 83 84

9 8 7 6 5 4 3 2 1

CONTENTS CONTRIBUTORSTO VOLUME 34 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PREFACE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix xi

ADVANCES IN PLANT CELL AND TISSUE CULTURE IN CHINA

Hu Han and Shao Qiquan I. I1. 111. IV . V. VI .

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anther Culture and Crop Improvement ..................... Some Fundamental Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Protoplast Isolation, Culture. and Genetic Manipulation . . . . . . . . Selection of Mutants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Miscellaneous: In Vi'itroPropagation through Plant Tissue Culture References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

2 5 7 9 10 11

HOW MUCH NITROGEN DO LEGUMES FIX?

Thomas A . LaRue and Thomas G . Patterson I. II. I11. IV .

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods of Estimating Fixation by Crops . . . . . . . . . . . . . . . . . . . Estimates for Major Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

15 19 27 34 36

PEANUT BREEDING

J . C . Wynne and W . C . Gregory I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II . Germ Plasm Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III . Economic Importance and Breeding Objectives . . . . . . . . . . . . . . . IV . Breeding and Quantitative Genetics ........................ V . Breeding Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Interspecific Hybridization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V

39 40 44 49 63 65 68 68

vi

CONTENTS

MOLYBDENUM IN SOILS. PLANTS. AND ANIMALS

Umesh C . Gupta and John Lipsett

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Molybdenum Fertilizers. Their Rates and Methods of Application.

13

and Industrial Uses of Molybdenum .....................

75 78 81 85 Responses to Molybdenum on Crops ....................... Factors Affecting the Molybdenum Uptake by Plants . . . . . . . . . . 89 Deficiency and Sufficiency Levels of Molybdenum in Plants . . . 99 Molybdenum Deficiency and Toxicity Symptoms in Plants . . . . . 100 Molybdenum Toxicity and Molybdenum-Copper-Sulfur Interrelationships in Animals ........................... 105 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 109 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

111. Physiological Role of Molybdenum in Plants . . . . . . . . . . . . . . . . IV . Determination of Molybdenum in Soils and Plants . . . . . . . . . . . .

V. VI . VII . VIII . IX .

X.

INTERGENOMIC INTERACTION. HETEROSIS. AND IMPROVEMENT OF CROP YIELD

H . K . Srivastava I . Introduction

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

118

I1. Genetics of Mitochondria and Chloroplasts . . . . . . . . . . . . . . . . . . 119 111. IV. V. VI . VII .

Organelle Involvement in Genetic Phenomena . . . . . . . . . . . . . . . Genetic Implications of Intergenomic Interactions . . . . . . . . . . . . . Molecular-Genetic Aspects of Heterosis .................... Improvement of Crop Yield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

130 147 164 174 182 185

THE DILUTION EFFECT IN PLANT NUTRITION STUDIES

W . M . Jarrell and R . B . Beverly

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I1. System for Expressing Results ............................ 111. Mechanisms

IV . V. VI . VII .

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

Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dilution Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Concentration Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

197 199 200 202 204 216 219

CONTENTS

VIII . Summary and Future Research Needs ...................... References ............................................

vii

221 222

DESIGNING “LEAFLESS” PLANTS FOR IMPROVING YIELDS OF THE DRIED PEA CROP

C . L . Hedley and M . J . Ambrose I. I1. I11. IV . V. VI . VII .

General Introduction ..................................... Comparative Responses of Peas to the Crop Environment . . . . . . Attaining Maximum Biological Yield per Unit Area . . . . . . . . . . . Attaining the Maximum Economic Yield per Unit Area . . . . . . . . Improving the Efficiency of the Pea Fruit . . . . . . . . . . . . . . . . . . . A Plant Ideotype for Improving Yields of Dried Peas . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

225 229 239 252 265 272 274 275

LOW-INPUT TECHNOLOGY FOR MANAGING OXISOLS AND ULTISOLS IN TROPICAL AMERICA

Pedro A . Sanchez and Jose G . Salinas 1. I1. I11. IV . V. VI . VII . VIII . IX .

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Site Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selection of Acid-Tolerant Germplasm ..................... Development and Maintenance of Ground Cover . . . . . . . . . . . . . Management of Soil Acidity .............................. Phosphorus Management ................................. Management of Low Native Soil Fertility . . . . . . . . . . . . . . . . . . . Discussion ............................................ Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

280 293 295 308 334 354 380 390 397 398

CYTOGENETICS OF PEARL MILLET

Prem P . Jauhar I. I1. I11. IV . V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Karyotypic Analysis ..................................... Meiosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Abnormal Meiosis and Its Genetics ........................ Haploidy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

408 410 415 417 424

viii

CONTENTS

VI . Polyploidy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aneuploids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Structural Changes in Chromosomes ....................... B Chromosomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Floral Biology and Hybridization . . . . . . . . . . . . . . . . . . . . . . . . . . Hybridization and Chromosome Relationships . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

VII . VIII . IX . X. XI . XI1 .

428 434 441 446 451 456 472 473

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

481

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

M. J . AMBROSE (225), Department of Applied Genetics, John Innes Institute, Norwich NR4 7UH, England R. B . BEVERLY (197), Department of Soil and Environmental Sciences, University of California-Riverside, Riverside, California 92521 W. C. GREGORY (39), Crop Science Department, North Carolina State University, Box 5 / 5 5 , Raleigh, North Carolina 27650 UMESH C. GUPTA (73), Research Branch, Agriculture Canada, P. 0. Box 1210, Charlottetown, Prince Edward Island, Canada CIA 7M8 HU HAN ( I ) , Institute of Genetics, Academia Sinica, Beijing, People’s Republic of China C. L. HEDLEY (225), Department of Applied Genetics, John Innes Institute, Norwich NR4 7UH, England W. M. JARRELL (197), Department of Soil and Environmental Sciences, University of California-Riverside, Riverside, California 92521 PREM P. JAUHAR* (407), Department of Botany and Plant Sciences, University of California-Riverside, Riverside, California 92521 THOMAS A. LaRUE (15), Boyce Thompson Institute for Plant Research, Ithaca, New York 14853 JOHN LIPSETT (73), Division of Plant Industry, CSIRO, P. 0. Box 1600, Canberra City, A . C . S . 2601, Australia THOMAS G . PATTERSON (15 ) , Boyce Thompson Institute for Plant Research, Ithaca, New York 14853 SHAO QIQUAN ( I ) , Institute of Genetics, Academia Sinica, Beijing, People’s Republic of China JOSE G. SALINAS (279), Tropical Pastures Program, Centro lnternacional de Agricultura Tropical, Apartado Aereo 67-13, Cali, Colombia PEDRO A. SANCHEZ (279), Soil Science Department, North Carolina State University, Raleigh, North Carolina 27650 H. K . SRIVASTAVAT (1 I7), Department of Biology, University of Valle, Cali, Colombia J . C. WYNNE (39), Crop Science Department, North Carolina State University, Box 5155, Raleigh, North Carolina 27650 *Present address: Division of Cytogenetics and Cytology, City of Hope National Medical Center, Duarte, California 91010. thesent address: National Agricultural Research Project, Gujarat Agricultural University, Anand 388 1 10 (Gujarat), India. ix

This Page Intentionally Left Blank

PREFACE Agronomy is an applied field of endeavor that embodies the combined efforts of soil and crop scientists to increase the yield and production of field crops. It utilizes the methods and inputs of the more basic physical and biological sciences to gain an understanding of plant and soil processes that constrain and/or enhance crop production. It encourages an integration of the best products of these more basic sciences into practical crop production systems. The contributions to this volume illustrate the usefulness of the findings of agronomists in seeking increased crop yields and production. Four of the articles deal with genetics and plant breeding. Research on the cytogenetics of an important food crop of the Old World, pearl millet (Pennisetum), is reviewed. Likewise, the remarkable progress made in recent years in breeding improved varieties of the common peanut (ground nut) is very effectively covered. Recent developments relating to improved crop yield through heterosis and intergenomic interaction are the focus of another article. During the past 30 years, there has been little interaction between scientists in the People’s Republic of China and those in the Western world. The contribution on tissue culture in China discusses an important subject to which the Chinese have made significant contributions during the past two decades. This method of crop regeneration and improvement is also beginning to provide plant breeders with a practical tool that may shorten the time needed to develop new varieties and may simultaneously increase rates of mutation. Soil fertility and management are the subjects of four articles. One deals with factors influencing the amount of nitrogen fixed by legumes, a topic of increasing importance as rising energy costs force concomitant increases in nitrogen fertilizer prices. A second related contribution, which will be equally important to researchers and farmers in large areas of Africa, Latin America, and Asia, focuses on the management of acid soils of the tropics. Means of alleviating the constraints placed on resource-limited farmers in these areas are considered. Soil fertility research methodology is given attention in a contribution concerned with the dilution effect in soil fertility experiments. This review will be helpful in interpreting data from such experiments. A second soil fertility article brings us up-to-date on research concerned with molybdenum in soils, plants, and animals. Knowledge of factors influencing the availability of this element is helpful in both developing and more developed countries. The article on the “leafless” pea crop is of special interest to crop physiologists. It focuses on the comparative responses of ‘‘leafless” and wellleafed peas to environmental and management conditions. The results reviewed will be of interest to scientists concerned with crop-environment-management interactions. xi

xii

PREFACE

As has become the norm for Advances in Agronomy, the subjects covered here are truly international, and the authors selected to review them come from different parts of the world. We sincerely thank these authors for their contributions, which will be of great value to the world’s crop and soil scientists and, ultimately, to the farmers on whom we depend to produce our food and fiber.

N. C . BRADY

ADVANCES IN

AGRONOMY

VOLUME 34

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY, VOL. 34

ADVANCES IN PLANT CELL AND TISSUE CULTURE IN CHINA Hu Han and Shao Qiquan Institute of Genetics, Academia Sinica, Beijing, People's Republic of China

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

................... Wheat ....................................................... Rice . .............................................. Corn .......................... Rubber Tree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Culture Media . . . . . . . . . . . .

11. Anther Culture and Crop Improvement

111.

IV .

V. VI .

A. B. C. D. E. Some Fundamental Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Androgenesis of Gramineous Species in Anther Culture B. Albinism of Pollen Culture-Derived Plants ................................ C. Genetics and Cytology of Progenies of Regenerated Plants . . . . . . . . . . . . . . . . . . . Protoplast Isolation, Culture, an A. Protoplast Isolation . . . . . . . B. Protoplast Culture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. 'Genetic Manipulation . . . . . . . . . . Selection of Mutants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Miscellaneous: In V i m Propagation through Plant Tissue Culture . . . . . . . . . . . . . . . . A. Tissue Culture of Drug-Producing Plants.. ................................ B. In Virro Propagation by Plant Tissue Culture . . . . . . . . . . . References . . . . . . . . . . . . . . . ............................

1

2 2 2 4 4

5 5 5

6 6

I I 8 8 9 10 10 10 11

1. INTRODUCTION Plant cell and tissue culture is an old as well as a new area of research in biological sciences in China. It is old because investigations in this field were initiated several decades ago. It is new because extensive research has been carried out and much progress has been made during the last decade, especially in the last few years. In the past foreign scientists were not aware of the research work done in China. As a result of the Sino-Australian Symposium on Plant Tissue Culture held in May of 1978 in Beijing and the participation of Chinese scientists in international meetings in recent years, this situation has gradually improved. 1

Copyright 0 1981 by Academic Ress, Inc. All rights of repduction in any form ~ ~ S E N U I . ISBN 0-12-ooO734-7

2

HU HAN AND SHAO QIQUAN

II. ANTHER CULTURE AND CROP IMPROVEMENT During the last ten years, much progress has been made in plant anther culture, including development of media, standardization of culture conditions, and determination of the response of genotypes to anther culture. In more than a dozen species, pollen culture-derived plants were first obtained by Chinese scientists (Table I). Emphasis has been placed on using anther culture techniques in crop plant improvement. A. W H E A T

Plants regenerated from anther culture have been more easily obtained from materials of hybrid origin than from pure varieties, and significant differences in plant regeneration frequency have been noted among various hybrid combinations. It is now possible to generate plants through pollen culture of winter and spring wheat varieties as well as their hybrid combinations. Using materials with high induction frequency of callus, several dozen green plantlets were regenerated under appropriate conditions from 100 inoculated anthers. Gangsu Academy of Agricultural Sciences is now capable of producing up to 600 pollen-derived and colchicine-doubled plants of wheat within a year. B . RICE

The induction frequency of rice plants regenerated through anther culture has markedly improved in recent years. For example, Zhang (1980) has been able to obtain 5106 clumps of green seedlings from 125 combinations of Oryza sutivu subspecies japonica intervarietal crosses and of indica -japonica intersubspecific crosses, resulting in thousands of pedigree lines with stable characteristics available for selection and evaluation in 1 year. Thus anther culture can now be used as a tool in applied plant breeding programs. Up to now several rice varieties developed through anther culture techniques in rice have been released for commercial production. Hua Yu No. 1, developed cooperatively by the Institute of Genetics, Academia Sinica and the Institute of Rice Research, Tianjing, is characterized as high yielding (about 7500 kg/ha), resistant to bacterial leaf blight, and widely adaptable, and has been grown in suburbs of Tianjing and Beijing covering an area of thousands of hectares. Chen Ying et al. (1980b) were able to isolate microspores of rice using a float-

PLANT CELL AND TISSUE CULTURE IN CHINA

3

Table I Species in Which Regenerated Plant Was First Obtained in China Species

Year"

References

Wheat (Triticum aesrivum)

1971

Triticale (Triticale) Wheat-wheat grass hybrid (Triticum aestivum X Agropyron glaucum) Maize (Zea mays L.) Boom corn (Sorghum vulgare) soybean (Glycine max) Alfalfa (Medicago denticulata) Rubber tree (Hevea brasiliensis) Poplar (Populus nigra L.) Pepper (Capsicum annum L.) Brassica pekinesis Chinese cabbage (Brassica chinesis) Sugar beet ( 2 n =4x=36) (Beta vulgaris L.)

1971

Ouyang et al. (1973) Wang et al. (1973) Chu ef al. (1973) Sun et al. (1973)

1973

Wang et al. (1975a,b)

1975

Ku et al. (1975)

1978

Zhao (1978)

1980

Yin et al. (1980)

1979

Xu (1 979)

1977

Chen et al. (1978)

1974

Wang et al. (1975a,b)

1971

Wang et al. (1973)

1973 1977

Teng and Kuo (1977) Chung et al. (1977)

1973

1979

Breeding Lab. Heilingchiang Sugarbeet Institute (1973) Shao (1979)

1979

Chen et al. (1980~)

I979

Sun (1979)

1978

He Ze Chinese Medicine Laboratory ( 1978)

Sugar beet (2n=2x=18) (Beta vulgare) Sweet orange (Citrus microcarpa) Flax (Linum usitatissimum) Rehmannia glurinoso

"Indicates the year the plants were obtained.

ing culture method-subjected pollen grains liberating continuously from dehiscing anthers-and obtained many calli and green plantlets. More recently green seedlings were directly induced from pollen grains on media without phytohormones.

4

HU HAN AND SHAO QIQUAN

C. CORN

It is valuable and beneficial to develop corn inbred lines through anther culture. Much investment of effort has gone into anther culture investigations of corn inbred lines in China. Different starting materials have been used, resulting in 25 inbred lines obtained by 14 institutes and universities in the past 2 years. Meanwhile a test crossing of some inbred lines is under way. Uniform canopy structure was demonstrated in these inbred lines. The comparison of corn inbred lines and the parent varieties “525” and “Gui No. 622” for plant height, ear length, 100 grain seed weight, and seed weight per ear revealed that the coefficient of variation was much lower in the pollen-derived inbred lines (Chen et al., 1979a). For example, a corn inbred line, Qunhua No. 105, has recently been produced from hybrid corn Qundan No. 105 through anther culture (Wu er al., 1980). Results of the test cross with this line showed that of the ten hybrid combinations tested, nine (90%) were superior than the control. Therefore, Qunhua No. 105 is a promising corn inbred line. Using corn varieties“Ba Tang Bai” and “Qundan No. 105” as parental materials, Cao (personal communication) and Wu and Zhong (1980) have been able to obtain maize cell clones through anther culture techniques in recent years. Gu (personal communication) and Cao (personal communication) have subcultured the cell clones derived from “Ba Tang Bai” every 4 weeks, and now these clones have gone through 30 generations of subculturing. The clones can be grouped into four types according to the degree of their totipotency. The No. 1 clone has intensive capacity of differentiation and regeneration, and green plantlets have been readily regenerated from it. Cytological studies of 427 cells of No. 1 clone and of 645 root tip cells of 35 plants regenerated from No. 1 clone revealed that 89.7% of the former and 87.4% of the latter were haploids. This type of cell clone, which has such high regeneration capacity and genetic stability, is a rare material in plant tissue culture. D. RUBBERTREE

Plants of rubber tree (Hevea brasiliensis) were regenerated in 1977 from anthers cultured in vitro (Chen, 1978). In the last 2 years Chen et al. (1979b) have studied the relationship between callusing anthers and the formation of pollen embryoids of Hevea brasiliensis. They found that after 20-30 days of culture 80% cells at metaphase of mitotic division were observed to be diploids (2n = 36), while after 50 days of inoculation only 10% were diploids and 69% were haploids. The result shows that after 50 days of culture the callus derived from

PLANT CELL AND TISSUE CULTURE IN CHINA

5

anther walls was degenerating, whereas the haploid tissue of microspore derivation was dividing vigorously. In order to obtain haploid embryoids or calli it is appropriate to transfer callusing anthers to differentiation medium after 50 days of inoculation. A total of 238 cells from 46 embryoids were observed at mitotic metaphase. Of these, 4% had 9 chromosomes, 80% had 18 chromosomes, 15% had 27 chromosomes, and 1% had 32, 36, or 45 chromosomes. The preponderance of haploids originating from embryoids was demonstrated. Root tips of 18 plantlets were also analyzed cytogenetically and similar results were obtained (Chen et a f . , 1979b). E. CULTURE MEDIA

Several culture media developed first by Chinese scientists are now widely used within the country as well as abroad. For example, the potato medium (First Group, 1976; Ouyang e t a l . , 1977; Chuang e t a l . , 1978), which contains 10-30% potato extracts, works as well or even better than the Miller’s and MS’s media. The use of potato extract as a medium component has greatly improved callus induction. N6 is another efficient medium developed by Chu et al. (1975) in which ammonium salts in optimum concentration are added along with nitrate salts. The frequency of anther callus induction on N6 medium is higher than that on Miller’s and MS media. It was as high as 16% (even 50% in some cases) for intervarietal hybrids of rice.

Ill. SOME FUNDAMENTAL PROBLEMS A. ANDROGENESIS OF GRAMINEOUS SPECIES IN ANTHER CULTURE

The processes of androgenesis during anther culture, particularly of gramineous species such as wheat, triticale, rye, barley, rice, and corn, were studied in the last few years. In addition to A and B pathways, others (such as C, D, and E pathways) were discovered (Sun, 1978). It is obvious that the abnormal mechanisms occur in early stages of pollen grain development during anther culture of wheat under different incubation conditions, particularly at low temperature (Tseng and Ouyang, 1980). Results obtained are useful in establishing a scientific basis not only for improving the frequency of callus induction and plant regeneration, but also for the elucidation of the causes of spontaneous chromosome doubling and chromosome aberrations in some plants, as well as for understanding the mechanisms of microspore differentiation.

6

HU HAN AND SHAO QIQUAN

B. ALBINISM OF POLLEN CULTURE-DERIVED PLANTS

In gramineous species, more than one-third of the plantlets regenerated from anther culture are albinos. For the effective use of anther culture in haploid breeding it is important to discover the causes of albinism and seek remedies to overcome this problem. Systematic studies of this problem should also contribute to the understanding of the genetics, physiology, and cytology of chloroplast development. Systematic investigations on the occurrence of albino plantlets in rice revealed that the viability of the developing plastid in chloroplast was the direct cause of albinism. A barrier to the nucleic acid translation system of the proteins needed for the development of a plastic lamellar system (a barrier to gene expression) is perhaps present (Wang et al., 1978; Liang et al., 1978). Recently components of soluble protein and ribosomal RNA from green and albino pollen culture-derived plantlets of rice have been analyzed by polyacrylamide gel electrophoresis. It was found that little or no band 3 (Fraction I protein), 23 S RNA, and 16 S RNA are present in albino plantlets (Sun et al., 1978). Together with the evidence obtained from other investigations, it was suggested that the albinism is caused by the impairment function of DNA.

c. GENETICSA N D CYTOLOGY OF PROGENIES OF REGENERATED PLANTS 1 . Genetic Stability

Genetic stability of plants derived from tissue and anther culture has been investigated by numerous scientists in China as well as abroad. These investigations revealed that the progenies of tissue and anther culture-derived plants of crop species are highly stable. Anther and tissue culture-derived progenies of rice, tobacco, and wheat were compared with their parental varieties by several research groups. The coefficient of variation (CV) values for several agronomically important traits were similar in two groups or sometimes small in pollen culture-derived progenies. Ninety percent of the diploid lines obtained from doubled haploids of anther culture derivation were homozygous, whereas only 10% showed segregation (Hu et al., 1980). Cytological examination of root tip chromosome number of 54 H I (regenerated) plants of wheat showed that about 90% of the regenerated plants were either haploids or homozygous diploids. The analysis of PMCs of 72 H I wheat plants (Hu et al., 1980) showed that 87.5% of these regenerated plants were either 3x or 6x. Similar results were obtained from examination of somatic chromosome numbers of field-grown, pollen culture-derived plants of wheat. Over 80% were

PLANT CELL AND TISSUE CULTURE IN CHINA

7

either haploids (3x) or homozygous diploids (6x). These observations favor the use of the anther culture technique in applied plant breeding programs.

2 . Variation

,

Cytological investigations of H plants revealed that approximately 10% have deviant chromosome numbers, such as monosomics, nullisomics, aneupolyhaploids, pentaploids, and octoploids. These plants with variant chromosome numbers have proved useful in chromosome and genome engineering. The mechanisms underlying the origin of these variant plants have been discussed by D’Amato (1978). Variability of chromosome number of regenerated wheat plants has also been analyzed in our lab. Mixoploids and cells showing tripolar mitosis, which could result in heteroploids and aneuploids, were observed in wheat root-tip and calk dicentric chromosomes and chromosome fragments in pollen calli, suggesting that in development of anthers cultured in vitro, chromosomevariation may occur (Hu et al., 1980). Moreover, in early stages of pollen callus formation in virro, nuclear asynchronous divisions, fusion of different nuclei, and endomitosis were occasionally observed. These abnormalities in mitosis may lead to the variation in the chromosome number of plants regenerated from anther culture. From the foregoing, we consider that either the stability or the variability of the chromosome numbers of progenies of plants derived through anther culture is useful in crop improvement as well as in genetic studies.

IV. PROTOPLAST ISOLATION, CULTURE, AND GENETIC MANIPULATION A. PROTOPLAST ISOLATION

Studies on protoplast isolation were initiated in China in 1973. Mesophyll cells and cells from cultured calli, especially of cereal crops, have been often used. In order to overcome the difficulty of inducing differentiation in long-term subcultures, stem tips of hybrid rice seedlings were cut and incubated, and spherical protoplasts were then released (Guangdong Institute of Botany, 1978). Potrykus et al. (1977) also reported that differentiation is more easily induced in corn protoplasts obtained in this manner than those from calli cultured for a long time. A complex enzyme extracted from Trichoderma viride EA3-867used in protoplast isolation has been as effective as cellulase Onozuka R-10made in Japan (Hsu, 1978).

8

HU HAN AND SHAO QIQUAN

B. PROTOPLAST CULTURE

Plants have been regenerated successfully from protoplasts of tobacco, petunia, dnd carrot (Li et al., 1978a,b; Wu et al., 1977). Efforts to generate plants from protoplasts derived from cereals, legumes, and other species have been less successful (Yen and Li, 1979; Li et al., 1978a,b; Cytology Lab., 1977; Tsai et al., 1978). Recently a two-layer culture method, i.e., liquid in the upper layer and solids in the lower, has been developed, and plants have been regenerated from mesophyll protoplasts of Nicotiana rustica x N . alata (Hsia et al., 1979). Moreover, typical division of cells regenerated from wheat and barley mesophyll protoplasts has been observed at about 0.1% frequency (Li, 1979a,b). Of the nutrients explored, vitamin and other organic compounds are very important to protoplast division, while glucose and low levels of other sugars have favorable effects on wheat protoplast culture. When protoplasts were cultured under lower osmotic conditions, e.g., 0.4 M glucose, budding, bulbing, and anuclear subprotoplasts were generated probably due to rapid swelling of protoplasts and incomplete cell wall formation (502 Group, 1974). More recently somatic hybrid plants have been regenerated through fusion of protoplasts from somatic cells of N . tabacum and those of N . rustica. Examination of chromosome number and identification of peroxidase isoenzymes revealed that the plants are of somatic hybrid origin (Wang et al., 1981; Gong et al., personal communication). C . GENETICMANIPULATION

Homologous fusion of mesophyll protoplasts of wheat, corn, and other species, and heterologous fusion of those of wheat and Viciafaba by adding NaN03 to culture medium were first obtained at a low frequency of approximately 4% (Sun, personal communication). Using Kao’s fusion technique with high pH, calcium ion solution, and PEG,interspecific fusion was induced between protoplasts from wheat yellowing leaves and those from green leaves of petunia. The fusion occurred at a frequency of 25%, and 10% of the fusion bodies were of heterologous origin. Nuclear staining confirmed that the fusion bodies were heterokaryons. Some of the fused protoplasts regenerated cell wall, underwent cell division, and developed into small calli after being transferred to fresh medium, but they could not be identified as hybrids due to lack of a selection system. In chloroplast transplantation research, chloroplasts of wheat and spinach were introduced into carrot callus protoplasts with PEG used as inducing agent and

PLANT CELL AND TISSUE CULTURE IN CHINA

9

successful transfers occurred at a frequency of 2-5% (C. Ma, personal communication). The characteristics of rape mosaic virus (YMV 15) were studied by introducing the virus into tobacco (N. tubacum var. Sumsan) mesophyll protoplasts. It was found that YMVl5 is serologically related to TMV, but the infection pathway is different. When polyomithine was added, 46-96% of tobacco protoplasts were infected by YMVI5, and each infected protoplast contained (1-3) X 105 virus particles (Tian et al., 1977). Crown gall cell is useful in plant genetic engineering. In general plant tumor cells induced by different Agrobacterium tumefaciens strains cannot grow on medium containing o-lactose or D-galactose as the sole carbon source, since both sugars are toxic to these tumor cells. However, Li Xiang Hui and Schieder ( 1981) recently discovered that tobacco tumor cells induced by A . tumefaciens strain B6S3 are different from other tumor cells and can grow on medium supplemented with D-lactose as the sole carbon source. This characteristic might provide a selective system for protoplast fusion and transformation experiments. Somatic hybrids have been obtained through fusion of protoplasts from B6S3 tobacco tumor cells and those from mesophyll of N. tubacum cv. Xanthi. The hybrid characteristics were demonstrated by the presence of octopine identified in the hybrid cells. It now appears possible to transfer T:DNA segments of the tumor cells to normal ones through protoplast fusion (X. H. Li, personal communication).

V. SELECTION OF MUTANTS Cell cultures are also used as means of obtaining variant strains. Investigations on selection of mutants at plant cell levels got under way recently in China. This appears to be a promising field of research. A chlorophyll mutant of rice (HY 101) was obtained that segregated at a frequency of 19% from an H2strain. The anther callus of variety No. 8126 was treated with ethyl methane sulfonate (EMS) at an early stage in culture (Hu er al., 1981) and the aforementioned H 2 strain was obtained from this callus. The mutant plants are yellowish-green in color and stable in characteristics. The reciprocal F, s obtained from crosses between the mutant and normal rice plants were green. The F2population segregated in a 3 : 1 ratio of green yellowish-green plants. Green and yellowish-green plantlets appeared at a ratio of 1 : 1 in a plant population obtained from hybrids through anther culture. Chen et al. (1980a) were able to obtain a BUdR-resistant mutant from soybean cell line SB-1 which was irradiated by 1000-R y rays, and cultured for 20 days on y-rayed B5 medium. It is believed that irradiation might trigger chemical changes in culture medium in addition to the direct effects on the treated cells.

10

HU HAN AND SHAO QIQUAN

Thus the mutation rates may be increased by exposing the calli to the direct and indirect irradiations. Ho et al. (1980) were able to select a lysine analog (0L)-resistant tobacco callus mutant at a low frequency (ca. lo-') by first subculturing tobacco (N. tubacum L. cv. Gexin No. 5 ) callus through three successive passages on a medium containing lg/ 1 L-Coxalysine (OL), an inhibitor of callus growth, then culturing the treated tissue through 6 passages on a medium without the selective agent OL, and finally treating the callus cultures with 2% EMS. The resultant mutant is 10 times less sensitive to OL than the parent line, and in general it was stable through 12 passages of subcultures in the absence of the selective agent. It was found that the lysine analog-resistant tobacco callus mutant has accumulated twice the normal level of lysine content, while a different pattern of peroxidase isoenzyme spectrum has been discovered electrophoretically in the mutant as compared with that of parent lines.

VI. MISCELLANEOUS: IN VITRO PROPAGATION THROUGH PLANT TISSUE CULTURE A number of economically important plants, including seaweeds, have been propagated through tissue culture in China. In addition, many drug-producing plants have also been cloned in vitro. A. TISSUE CULTURE OF DRUG-PRODUCING PLANTS

Peking Pharmaceutical Institute was able to cultivate ginseng callus proliferated from young stems and roots. It was found that the total ginseng saponins was the same as that of the garden ginseng. For example, the ginseng callus contains 4% (dry weight) saponins, whereas the 6-year-old garden ginseng contained 3% (dry weight) of this chemical. Zheng and Liang (1976) were able to cultivate the famous herb Panax noroginseng, and chemical analysis showed that the herb contained 10.25% (dry weight) of gross saponin and sapogenin, in contrast to the tuberous roots of the natural plants, which contain 6.06% (dry weight) of the said drugs only. It has been found that callus tissue of Seopolia ocutagula plant produced 0.55% (dry weight) of Hyocyamine and seopolamine, whereas fieldgrown plants had only 0.139% of these drugs (Zheng and Liang, 1977). B. In Vitro PROPAGATION BY PLANTTISSUE CULTURE

Since 1975, Wang et al. (1975a,b; Wang and Chang, 1978a,b) have successfully regenerated seedlings from embryo (diploid 2n = 18) and endosperm

PLANT CELL AND TISSUE CULTURE IN CHINA

11

callus (triploid 3 n = 27) of Citrus sinensis. The diploid plantlets have been easily induced either from bud primordia or through embryoid formation. The triploid seedlings were regenerated through embryogenesis of endosperm calli, which had been induced at the cell stage. A successful system for obtaining “virus-free’’ seed potato has been worked out by a number of institutions. Virus-free potato seed is now successfully produced and planted in about 20 provinces. The yield has increased significantly through the use of ‘‘virus-free’’ seed (unpublished results). Clones of sugarcane have been obtained through tissue culture in all canegrowing provinces (unpublished results). Meristem of stem apex or lateral buds as well as young leaves are used as initial culture materials for the induction of callus from which young seedlings are regenerated. The yield of canes raised through “test-tube” methods is the same as that of the control, but the tissue culture-obtained clones are more variable as some have more tillers and others have smaller stems. I n vitro fertilization of corn ovules was first achieved in 1977 with a simplified medium prepared mainly from natural extracts of potato (Shao et al., 1977). Fourteen seeds were produced in 1978, and the portion of seed set was 0.42%. Hybrid kernels matured in about 20-22 days after pollination and germinated in a test tube. After transferring to pots, only two plants survived. One of them had clear purple markers on leaves and stem. The chromosome number of root cells was 20. These results indicated that plants produced from test-tube fertilization were intervarietal hybrids but not parthenogenetic haploids (Jiang et al., 1979). ACKNOWLEDGMENTS The authors wish to thank H. W.Li for his help in preparing the manuscript and R. B. Tan for her technical assistance.

REFERENCES D’Amato, F. 1978. In “Frontiers of Plant Tissue Culture” (T. A. Thorpe ed.), pp. 287-295. University of Calgary Offset Printing Office, Calgary, Alberta, Canada. Breeding Laboratory, Heilungchiang Sugarbeet Institute. 1973. In “Proceedings of Symposium on Anther Culture,” pp. 304-305. Science Pxess, Peking. Chen, C. H., Chen, F. T., Chien, C. F., Wang, C. H., Chang, S. J., Hsu, H. E., Ou,H. H., Ho, Y.T., and Lu, T. M. 1978. In “Proceedings of Symposium on Plant Tissue Culture,” pp. 11-22, Science Press, Peking. Chen, L., Xu, Z., Yin, G.C. Zhu, Z. Y., and Bi, F. Y. 1979a. Acra Genet. Sin. 6 , 421-426. Chen, Z. H., Qian, C. F., Qin, M., W a g , C. H.,SUO,C. J . , Xiao, Y.L., and Hsu, H. 1979b. In “Annual Report of the Institute of Genetics, Academia Sinica,” pp. 88-90. Chen, S. L., Tian, W. Z . , and Zhang, G.H. 1980a. In “Annual Report of the Institute of Genetics, Academia Sinica,” pp. 100-102. Chen, Y., Wang, R. F., Tian, W. Z., Zuo, Q. X., Zheng, S. W. Lu, D. Y., and Zhang, G.H. 1980b. Acta Genet. Sin. 7. 46-54.

12

HU HAN AND SHAO QIQUAN

Chen, Z. G., Wang, M. Q., and Liao, H. H. 1980c. Acta Genet. Sin. 7, 189-191. Cheng, K . C., and Liang, C. 1978. In “Proceedings of Symposium on Plant Tissue Culture,” pp. 469-479. Science Press, Peking. Chu, C. C., Wang, C. C., Sun, C. S . , Hsii, C., Yin, K. C., Chu, C. Y., and Bi, F. Y. 1975. Sci. Sin. 18, 659-668. Chu, Z. C., Wang, C. C., Sun, C. S . , Chien, N. F . , Yin, K.C., and Hsii, C. 1973. Acra Bot. Sin. 15, 1-11.

Chuang, C. C., Ouyang, T. W.,Chia, H., Chou, S. M., and Ching, C. K. 1978. In “Proceedingsof Symposium on Plant Tissue Culture,” pp. 51-56. Science Press, Peking. Chung, C. H., Jen, Y. Y., and Tai, W. K. 1977. In “Proceedings of Symposium on Anther Culture’’ (H. Hu, ed.), pp. 200-201. Science Press, Peking. Cytology Lab., Dept. Biology, Lanchow Univ. 1977. Acta Genet. Sin. 4, 242-247. 1st Group, 3rd Lab., Institute of Genetics. 1976. Acta Gener. Sin. 3, 25-31. 502 Research Group, Institute of Genetics. 1974. Acta Gener. Sin. 1, 59. Guangdong Institute of Botany. 1978. In “Proceedings of Symposium on Cell Culture and Somatic Hybridization,” p. 155. Lanzhou Univ., Langzhou. He Ze Chinese Medicine Lab., Shangton Province. 1978. Plant Mag. No. 4, p. 8. Ho, C. P., Xu, Z. Y., Xu, S. P., and Loo, S. W. 1980. Acra Phytophysiol. Sin. 6, 213-219. Hsia, C. A. er al. 1979. Int. Protoplasr Symp., 5th. Szeged. (Abstr.). Hsu, C. Y. 1978. In “Proceedings of Symposium on Plant Tissue Culture,” p. 285. Science Press, Peking. Hu, H., Hsi, T. Y., and Chia, S. E. 1978. Acra Genet. Sin. 5 ( I ) , 23-30. Hu, H., Xi, Z. Y., Zhuang, J. J., Ouyang, J. W., Hao, S. He, M. Y., Xu, Z. Y., and Zou, M. Q. 1980. Sci. Sin. 23, 905-912. Hu, Z., Pang, L. P., and Cai, Y. H. 1981. Acta Genet. Sin. (in preparation). Jiang, X. C., Nui, D. S., and Shao, Q. Q. 1979. Acta Gener. Sin. 6, 339-342. Ku, M.K.,Cheng, W. C., Huang, C. F., Huang, C. H., Kuo, L. C., Kuan, Y. L., and An, H.P. 1975. Acra Genet. Sin. 2, 138-143. Li, W. A., Wang, K. Y., and Tang, T. 1978a. In “Proceedings of Symposium on Plant Tissue Culture,” p. 325. Science Press, Peking. Li, W. P., Sun, Y. J., and Li, H. H. 1978b. Acfa Genet. Sin. 5 , 57-60. Li, X. H. 1979. Int. Protoplast Symp., 5th, Szeged pp. 261-267. (Abstr.). Li, X. H., and Schieder, 0. 1981. Plant Sci. Lett. (in preparation). Li, X. H., and Yen, C. S. 1978. In “Proceedings of Symposium on Plant Tissue Culture,” p. 361, Science Press, Peking. Liang, C. C., Chou, Y. H.,and Chen, W. M. 1978. In “Proceedings Symposium on Plant Tissue Culture,’’ pp. 161-166. Science Press, Peking. Ouyang, T. W., Hu, H., Chuang, C. T., and Tseng, C. C. 1973. Sci. Sin. 16, 79-95. Ouyang, T. W., Chuang, C. C., Chia, H.,Chou, S. M., and Ching, C. K. 1977. In “Proceedings of Symposium on Anther Culture” (H. Hu,.ed.), pp. 58-64. Science Press, Peking. Potrykus, I., Harms, C. T., and Thomas, H. 1977. Mol. Gen. Genet. 156, 347. Qin, M., Qian, C. F., Wang, C. H., Chen, Z. H.,and Xiao, Y. L. 1979. In “Annual Report of the Institute of Genetics, Academia Sinica,” pp. 85-87. Shao, C. C., Chiang, H. T., Li, C. K., and Chen. Y. C. 1977. Acta Genet. Sin. 4, 329-332. Shao, M . W. 1979. Zhongguo Tiancai 1, 27-30. Sun, C. S. 1973. Acta BOI. Sin. 15, 163-173. Sun, C. S. 1978. In “Proceedings of Symposium on Plant Tissue Culture,” pp. 117-124. Science Press, Peking. Sun, C. S., Wu, S. C., Wang, C. C., and Chu, C. C. 1978. In “Proceedings of Symposium on Plant Tissue Culture,” p. 248. Science Press, Peking.

PLANT CELL AND TISSUE CULTURE IN CHINA

13

Sun, H. T. 1979. Kexue Tongbao 24, 948-950. Teng, L. P., and Kuo, Y. H. 1977. I n “Proceedings of Symposium on Anther Culture” (H. Hu, ed.), pp. 198-199. Tian, B. e t a / . 1977. Acra Microbiol. Sin. 17, 306-310. Tsai, C. K., Chien, Y. C., Chou, Y. L., and Wu, S. H. 1978. I n “Proceedings of Symposium on Plant Tissue Culture,” p. 317. Science Press, Peking. Tseng, C. C., and Ouyang, 2. W. 1980. Acta Genet. Sin. 7, 165-172. Wang, C. C., Chu, C. C., Sun, C. S . Wu, S. H . , Yin, K. C., and Hsii, C. 1973. Sci. Sin. 16, 2 18-222. Wang, C. C., Chu, C. C., and Sun, C. S. 1975a. Acra Bot. Sin. 17, 56-59. Wang, C. C., Chu, C. C., Sun, C. S . , Hsii, C., Yin, K. C.. and Bi, F. Y. 1975b.Acta Gener. Sin. 2, 72-80. Wang, C. C., Sun, C. S . , Chu. C. C., and Wu, S. C. 1978. I n “Proceedings of Symposium on Plant Tissue Culture,” pp. 149-160. Science Press, Peking. Wang, P. T., Chen, J. Y.,Zhao, S. M.,and Xu, J. X. 1981. Kexue Tongbao 26, 373-375. Wang, T. Y., and Chang, C. J . 1975. Acta Bot. Sin. 17, 149-152. Wang, T. Y., and Chang, C. J . 1978a. Acta Gener. Sin. 5, 133-137. Wang, T. Y., and Chang. C. J. 1978b. Sci. Sin. 16, 823-827. Wang, Y . Y., Sun, C. S . , Wang, C. C., and Chien, N . F. 1973. Sci. Sin. 16, 147-151. Wu, J . L., and Zhong, Q. L. 1980. Acta Phyrophysiol. Sin. 6, 221-224. Wu, J . L., Zhong, Q. L., Nong, F. H., and Zhong, T. M. 1980. Hereditas (Beijing) 2, 23-26. Wu, S. C., Lin, C. P., Ma, C., Wang, Y . S., Chao, Y. C., and Liu, H. C. 1977. Acta Genet. Sin. 4, 462-464. Xu, S. 1979. Plant M a g . No. 6 , 5. Yen, C. S., and Li, X. H. 1979. Heredifas (Beijing) No. 5, p. 25. Yin, G. C., Li, X . Z., Xu, 2. C., Li. Z. Z., and Bi, F. Y. 1980. Kexue Tongbao. 25 ( I t ) , 976. Zhang, 2. H. 1980. I n “Proceedings of Symposium on Celebrating the 20th Anniversary of IRFU” (in preparation). Zhao, W. B. 1978. Acta Genet. Sin. 5, 337-338. Zheng, G. Z., and Liang, Z. 1976. Acra BOI. Sin. 18, 163-169. Zheng, G. Z., and Liang, 2. 1977. Acra Bot. Sin. 19, 210-215.

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY. VOL. 34

HOW MUCH NITROGEN DO LEGUMES FIX? Thomas A. LaRue and Thomas G. Patterson Boyce Thompson Institute for Plant Research, Ithaca, New York

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . ........... A. The Importance of Obtaining Reliable Es ........................... B. The Energy Cost of C. Published Estimates of Nitrogen Fixation by Legume Crops . . . . . . . . . . . . . . . . . . 11. Methods of Estimating Fixation by Crops ........................... A. Nitrogen Accumulation . . . . . . . . . . . . . . . . . . . . . . . ......... B. Difference Methods ............................... C. Isotopic Methods . . ............................... D. Acetylene Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . E. Other Methods of C 111. Estimates for Major Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. The Forages . . . . . . . . . . . ........ B. Seed Legumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Evaluation of Studies.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . for Future Research ..........................

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

15 15

18 19

20 23

27 27 30 34 34 35 36

1. INTRODUCTION A . THEIMPORTANCEOF OBTAINING RELIABLE ESTIMATES

In the coming decades our increasing world population and the growing needs for higher quality food will place severe demands on agriculture. During this time the cost of nitrogen fertilizer will continue to rise with the price of energy. The processes of decreasing fertility, soil erosion and desertification will force agriculture onto land presently marginal and will boost efforts to reclaim land. The role of legumes in agriculture is certain to increase in importance. The pulses are sources of good-quality protein. The forages have a long-documented history of supporting livestock on poor soil. Both these factors are attributed in 15

Copyright 0 1981 by Academic Ress. Inc. All rights of reproduction in any form reserved. ISBN 0-12-000734-7

16

THOMAS A. LARUE AND THOMAS G. PATTERSON

part to the ability of legumes, in symbiosis with rhizobia, to obtain nitrogen from the air. But how much nitrogen is obtained? The direction of research and the future management of legumes in agriculture will require accurate knowledge of the amounts fixed by crops in the field. If fixation is less than we now think, then we will not achieve the expected returns of N. If fixation is as high or higher than some reports say, we must search out the reason why some legume crops can deplete soil N. Recent work in plant physiology indicates that symbiotic fixation is not ‘‘free fertilizer”; the plant must provide energy in the form of photosynthate. Whether a legume crop “pays” for fixation with decreased yield is not yet determined. Legumes require more phosphate fertilizer than cereals, and many are more demanding of water. The cost of inoculant will not be negligible to cash-poor farmers in developing countries. Does the N fixed compensate for these inputs? Ultimately, plant breeders and agronomists will require accurate estimates of the amount of fixation, and of its cost, to determine whether increasing fixation is economically justified. B. THEENERGYCOSTOF SYMBIOTIC FIXATION VERSUS NITRATEUTILIZATION

Despite apparent differences, there is a fundamental similarity between symbiotic nitrogen fixation and the industrial production of nitrogen fertilizer. Energy is required for both methods, for thermodynamics requires this of all possible methods of fixation. Nitrogenase requires energy in the form of ATP and electrons. In addition, there are energy costs associated with nodule formation and maintenance, hydrogen loss, and incorporation and transport of newly fixed N. There must also be an energy cost for using soil nitrate, but comparisons of the two have been difficult to examine experimentally. Silsbury (1977, 1979) estimated the respiratory burden of subterranean clover grown under artificial light with nitrate or NZ.He calculated the growth coefficients (the fraction of net C02uptake in light associated with the synthesis of new dry matter) and found them constant over a 50-day period. For nodulated plants they were significantly higher (0.189) than for plants using nitrate (0.137). Nodulated plants used 810 mg C02 for the synthesis of 1 g dry weight, while nonnodulated plants used only 510 mg C02. Mahon (1977, 1978) calculated the energy cost of fixation after measuring root respiration and estimating N fixation by acetylene reduction. He compared the respiration of plants grown on N, with similar plants treated a few days previously with ammonium nitrate. It was assumed that the respiration due to

HOW MUCH NITROGEN DO LEGUMES FIX?

17

growth and maintenance was the same in both populations, and that the decreased root respiration in the treated plants was associated with decreased nitrogenase activity. Mahon obtained a value of about 6.7 g C/g N fixed with soybeans, cowpeas, Phaseolus, and peas. Ryle and his co-workers (1979) compared growth, photosynthesis, and shoot and root respiration of soybean, clover, and cowpea grown on N2 or nitrate under bright light. The plants provided with nitrate grew larger. Gross photosynthesis did nor differ in the two populations, nor did shoot respiration. The fixing plants had respiration rates about twice those of nitrate-grown plants. Expressed as a percentage of the gross photosynthesis, the root respiration of fixing and nonfixing plants was 22 versus 11 for soybean, 27 versus 14 for cowpea, and 34 versus 21 for clover; that is, for three species, plants fixing their nitrogen respire 11-13% more of their photosynthate. The relationship between root respiration and N fixed varied during plant growth. In all species it was highest (- 15 g C/g N) in young plants. Presumably, energy was used in forming nodules. In soybean and cowpea it dropped to a minimum of 3-5 g C/g N just before the nodules senesced during pod fill. The average cost was 6.3 g C/g N-a figure very close to Mahon’s. It is remarkable that the same respiratory cost, approximately 6.5 g C/g N, was determined by two investigators using five legumes. The close similarity amongst species suggests that it may not be easy to find significant differences in efficiency within a single species. There is good evidence that photosynthate supply to nodules is a major limitation to symbiotic fixation (Hardy and Havelka, 1975). The carbon cost of 6.5 g C/g N estimated by Ryle et al. and Mahon suggests that fixation of 1 kg NH, would cost 15-20 kg dry weight. In evaluating some of the extraordinarily high claims for hundreds of kilograms of N per hectare from symbiotic fixation, one should question whether the crop is capable of fixing the required carbon and translocating it to the root. Would a nodulated legume crop ever yield less than one obtaining all its N from soil? The common observation on soils very low in available N is that effective nodulation increases yield compared to uninoculated controls. In soils of very high fertility, the soil N suppresses nodulation and fixation, and yields are generally equal. The aforementioned experiments, however, indicate that on soils of intermediate fertility, the energy cost of nodule formation and fixation might lower the yield of the crop obtaining some of its N via the symbiosis, compared to an uninoculated control. Such a result was observed in a Brazilian soil in which soil N was not limiting to growth. With Phaseolus cultivars there was a positive correlation of nodule weight with plant N content but a negative correlation with grain production (Pessanha et al., 1972). This is what is expected if fixation requires more

18

THOMAS A . LARUE AND THOMAS G. PATTERSON

photosynthate than nitrate utilization. We did not find similar results elsewhere in the literature. It may be that the difference in energy cost between nitrate and N2 usage is not so great in the field as in lab experiments.

C. PUBLISHED ESTIMATES OF NITROGENFIXATION BY LEGUME CROPS

In reviews and texts there are many published tabulations of the amount of fixation by legume crops. Most are derived from a very few sources. Two favorites are publications by Erdman (1949) and Lyon and Bizzell(l933, 1934). The first was an extension pamphlet promoting inoculation, and containing data that were not substantiated by methodology. Erdman stated that his estimates had been calculated in most cases from controlled pot experiments which when magnified to an acre basis gave results higher than the actual figures. Unfortunately others citing his figures did not include his caveat. The study of Lyon and Bizzell, as we will see, did not approximate the field crop condition and probably overestimated fixation. Bums and Hardy (1975) averaged a great many published estimates to arrive at an average figure of 140 kg N fixed per year per hectare of arable land under legumes. Shortly thereafter, that figure was considered by a group of scientists attending a conference on nitrogen-fixing microbes. They concluded (Burris, 1978) that a realistic figure would be half the Bums-Hardy calculation. This reassessment, however, was apparently based as much on intuition as on new data. In the literature much of the work reported on estimating fixation cannot be extrapolated to field conditions. A very common procedure, especially when isotopic N is used, is to grow legumes in small pots, often in the greenhouse. The substrate is so unlike the soil and the conditions of plant growth so different from the field that the results cannot be used to calculate fixation by crops. Many experimenters, including those using field plots, give results only as milligrams N per plant or percentage N of the yield or percentage plant N derived from fixation. In the absence of information about yield per area or planting density, it is impossible to calculate the fixed N per hectare. The published results of fixation as a percentage of plant N may, however, serve in estimating a realistic figure for fixation in farm crops. It is a common observation that yields in experimental or demonstration plots are much higher than the average crop yields in the area. It is likely true that symbiotic fixation is also less on the farm. Nodule formation and function are depressed by a variety of environmental factors-water stress, flooding, herbicides, improper fertilizer placement, etc. It is unlikely that the percentage N from fixation will be higher on farms than on well-tended research plots. Therefore the percentage N fixed in

HOW MUCH NITROGEN DO LEGUMES FIX?

19

test plots might be used in calculating a realistic upper limit on fixation in agriculture.

II. METHODS OF ESTIMATING FIXATION BY CROPS A. NITROGEN ACCUMULATION

The standard procedure for nitrogen analysis is the Kjeldahl determination (Bremner, 1965). Its major advantages are simplicity and low expense. The simplest estimate of N fixation is by total N accumulation of the crop. This is based on the intuitive assumption that the crop derives all its N via symbiotic fixation. We can find no published evidence that this is ever the case under field conditions. The many estimates based on total plant N certainly overestimate fixation. Growth on low-fertility soils or on soils artificially impoverished in available N is no guarantee that all N is obtained by fixation. Kohl et al. (1980) decreased available N in a soil by the addition of 34 tonnesha of corn cobs. The nodulated soybean neverthelessobtained 47% of its plant N from this soil, and a nomodulated isoline yielded 2106 kg/ha of grain. These results demonstrate the ability of a legume to scavenge soil N. A closer approximation to N fixed may be achieved by analyzing changes in soil N as well as that removed in the crop. The frequently cited study by Lyon and Bizzel (1933, 1934) was such an approach. These workers placed a mixture of 60% silty clay loam and 40% sand in outdoor frames of unstated dimensions. Various crops were alternated or grown together for 10 years and the N in the crops was assayed at each harvest. The N in the top 28 cm of substrate was analyzed at the beginning and end of the trial, and the “apparent fixation of nitrogen” was calculated on an annual basis. There was an accumulation of soil N under clovers, alfalfa, and vetch, and a decrease with soybeans, peas, beans, barley, rye, and oats. The clovers had an apparent fixation averaging 166-200 kg N/ha. Alfalfa accumulated 296-330 kg N/ha. Soybeans, peas, and beans averaged 125, 57, and 70 kg N h a annually, respectively. Nonlegumes accreted 2 1-37 kg Nha. This study is considered a classic for documenting the advantage to topsoil N of proper forage-cereal rotations. However, its limitations are that the change in soil N has based on only two estimates 10 years apart and that only the top layer was analyzed. Legume roots may extend downwards 3 m (Weaver, 1926). The unaccountably high “apparent fixation” by nonlegumes suggests that there was appreciably mobilization of nutrients from below the sampled zone.

20

THOMAS A . LARUE AND THOMAS G . PATTERSON

Experiments on legumes growing on artificially impoverished soils are not uncommon in the literature. The substrate used by Lyon and Bizzell was an artificial one and for that reason the data they obtained should not be extrapolated to estimate fixation by farm crops. Long-Term Nitrogen Balance Studies Using Lysimeters

A lysimeter is an enclosed soil system in which the addition and removal of nutrients, water, and plant material can be controlled and measured. Although many designs have been described (Kohnke et al., 1940), the basic construction consists of a tank or box placed in the ground and filled with soil. Leachate from the soil in the tank is collected from the bottom of the lysimeter. Lysimeters have been used in long-term studies on nitrogen balance in crops under different management systems (Chapman et al., 1949). In addition, lysimeters can be used to compare and calibrate different techniques for estimating N fixation under controlled conditions (Williams et al., 1977). Several disadvantages are inherent in lysimeter studies. They are expensive to install because they require the excavation of large volumes of soil and must be constructed of materials resistant to corrosion. Due to their expense, the size and number of lysimeters, and thus the number of treatments and replications, is limited. In some lysimeter studies, long-term nitrogen balances have shown unaccounted losses of nitrogen (Collison et al., 1933). This is attributed to volatilization of N from the lysimeters (Chapman et al., 1949; Patwary and Raikavich, 1979), although this assumption has been challenged (Craswell and Martin, 1975a,b). B. DIFFERENCE METHODS

An adjusted measure of fixation by the nitrogen accumulation technique is obtained when the contribution of soil N to the total N of legumes is estimated. This correction for the contribution of soil N is obtained by growing a nonfixing plant in comparison with the N-fixing legume. Total N content of the nonfixing crop (derived solely from soil N) is subtracted from the total N content of the N-fixing legume. The difference between the values is assumed to be the quantity of N derived by N fixation. This procedure is often referred to as the “difference” method (Williams et al., 1977). Three versions of the difference method are commonly used. 1 . Comparison of a Legume with a Nonlegume Soil N contribution to a fixing legume is estimated by growing a nonlegume concurrently with the legume. Annual small grains such as wheat and oats have

HOW MUCH NITROGEN DO LEGUMES FIX?

21

been used (Rizk, 1966). In studies of forage legumes, however, perennial grass species are used. This is preferred in studying N fixation in grass-legume mixed pastures. Sprague (1936) used wheat as a control crop to estimate the soil N contribution. The wheat contained 38 kg N/ha and the vetch 149 kg N/ha. By subtraction, he estimated the net fixation of the vetch to be 111 kg N/ha. Using a nonlegume as a control in the difference method requires two major assumptions. First, the nitrogen contained in the nonlegume is assumed to arise solely from the soil N pool. Secondly, the legume and nonlegume will take up soil N in proportion to the amount available, and differences due to growth patterns and root morphology will not be significant. The validity of the second assumption is open to question. No evidence exists that shows the exploitation of soil N is equal between a legume and its nonlegume control. The fixation estimated will obviously depend on the arbitrary choice of nonlegume control. Wagner (1954) calculated that the N fixation of ladino clover was 165 kg/ha when orchard grass was used as a control or 189 kg/ha when tall fescue was the control. 2 . Comparison of a Legume with a Nonnodulating Legume Another approach to the difference method utilizes the existence of nonnodulating legume genotypes. Soybean genotypes with nonnodulating characters have been described (Williams and Lynch, 1954). The nonnodulation character is controlled by a single recessive gene, labeled rj, . This nonnodulating gene has been backcrossed to produce near-isogenic lines. The advantages of using nonnodulating lines as a control species are that the growth pattern, root morphology, and N uptake patterns are assumed to be nearly identical. However, the nodulated isoline may absorb more soil N than the nonnodulated isoline (Ruschel et al., 1979). This suggests a synergistic effect of fixed N on N derived from other sources. Weber (1966) used nonnodulating isolines of soybean to estimate N fixation under different levels of applied fertilizer nitrogen. By subtracting the N content of the nonnodulating isoline from that of the nodulated soybean, N fixation rates of 14.7-84 kg/ha were calculated. In soybeans, isolines are available for only a few cultivars. Unfortunately, near isogenic nonnodulating and nodulating cultivars are not yet available for other legume species. 3 . Comparison of Inoculated and llninoculated Legumes

Another difference method is a comparison of single cultivars grown on inoculated or uninoculated soil. A variation of this procedure uses separate plots inoculated with effective or ineffective strains of rhizobium (Nutman, 1973). In this situation, the identical variety is used for the control and treatment, with the

22

THOMAS A . LARUE AND THOMAS G. PATTERSON

only variable being the presence or absence of nodules (or effective and ineffective nodules). This method requires a soil free of native rhizobium species capable of establishing an effective symbiosis with the legume being studied. Bezdicek et al. (1978) compared inoculated and uninoculated soybeans under field conditions. Uninoculated soybeans accumulated 76 and 105 kg N/ha in the 2 years, while nodulated soybeans yielded 387 and 368 kg N/ha. By difference, the net fixation was calculated as 31 1 and 263 kg N/ha. In areas where legumes are routinely grown, soils lacking the appropriate rhizobium species may not be available. When using this technique, considerable care is required to prevent contamination of the uninoculated soils with rhizobium. For these reasons this procedure has limited application in the routine evaluation of N fixation. Using the difference method with nonnodulating legumes or uninoculated plots requires the assumption that the root structure and function are not altered by the presence of nodules. This assumption has yet to be proven experimentally. 4 . N Fertilizer Equivalence

In some studies, an indirect estimate of nitrogen contribution from a legume to a nonlegume is used to estimate N fixation. Forage legumes and grasses are often grown together in pasture systems. The legumes have been observed to stimulate the productivity of the grass species (Lyon and Bizzell, 1911). If the response of the grass in the mixed stand is compared to the response of the grass in a pure stand with different amounts of applied fertilizer N, then an estimate of the N contributed to the grass by the legume is obtained. This method has been used in the studies of pasture management systems. Cowling (1961) calculated the “indirect effect” of clover on grass as the amount of fertilizer N applied to a grass plot to give a yield equivalent to the grass component of a clover-grass mix. He found that the effect of clover on cocksfoot was equivalent to 117-134 kg fertilizer N/ha. Another variation of the nitrogen contribution technique is to estimate the value of a legume to a following crop. Green manure crops have often been used to increase soil N (Kroontje and Kehr, 1956; Kamprath et a l . , 1958; Sprague, 1936). The N added to the soil by the legume crop gives a yield in the successive crop that is compared to the response to different levels of added N fertilizer. Kroontje and Kehr (1956) estimated that vetch added the equivalent of 107 kg N/ha to the soil. This resulted in a significant increase in barley yields following the vetch crop. Similarly, Sprague (1936) estimated that a winter crop of vetch was equivalent to 671 kg NaNO,/ha. In comparing the response of follow crops, urea can be used as the nitrogen fertilizer (de Souza, 1969) because it contains no other nutritional element that might contribute an effect. A N fertilizer is subject to leaching and volatilization, whereas the decomposing forage crop releases N more slowly and efficiently to

HOW MUCH NITROGEN DO LEGUMES FIX?

23

the nonlegume crop. Therefore the “equivalents N ” is likely an overestimate of N fixation. The beneficial effects of legumes on nonlegumes are not a direct measure of N fixation. However, they do show the utility of green manures in adding N to the soil, and the cost of the fertilizer spared may serve as an indication of the economic value of N fixation. C. ISOTOPIC METHODS

The use of 15N isotopes for studying N uptake by plants was recently reviewed here (Hauck and Bremner, 1976). Very few of the articles reviewed related to the use of I5N for estimating fixation by legume crops. The technique is likely to become more common, however, because of the lowered cost of the isotope and improvements in methodology. These include relatively inexpensive mass spectrometers designed for low-mass compounds (e.g., Micromass) and increased precision by using double introduction and double collection. Because it is a direct method, fixation of I5N2 remains the method of choice for checking the validity of other estimates of fixation (Burris, 1972). The cost of I5N2 and the volumes that would be required to test replicate plant samples over a growing season preclude estimates based on that gas. However, the technique has been used to test assays based on other methods. The methods are essentially those developed by Burris and Wilson (1957; Bums, 1972). Typically, nodulated roots (Saito et al., 1980) or nodules (Hudd et al., 1980) are briefly incubated in a chamber with a gas phase enriched in 15N1. The fixation is stopped by the addition of strong acid, the tissue is analyzed for total N (Kjeldahl) and I5N by mass spectrometry (Smith et al., 1963) or emission spectrometry (LloydJones et al., 1977). The major limitation to mass spectrometer methods is the high cost involved for the instruments and the isotopes. The procedures are fraught with potential errors. Although the proper techniques are well documented (Bremner, 1965), it remains true that the method is demanding and requires skilled operators. When N gas is excited by a high-frequency oscillator, the wavelengths of emitted light depend on the isotope composition. Commercial emission spectrometers, less expensive than mass spectrometers, are available for measuring 15N. The instruments are not as precise as mass spectrometers. Their major advantages are that they require only small samples (10 p g N) and they are not so demanding of operators.

I . Methods Based on Isotope Dilution Nitrogen fixation can be estimated by isotope dilution. In this method the fixing crop and a nonfixing control are grown in soil to which I5N has been added

24

THOMAS A. LARUE AND THOMAS G . PATTERSON

as a small amount of labeled nitrate or ammonium (McAuliffe et al., 1958; Legg and Sloger, 1975). The N in the control plant should have the same 15Ncontent as the available soil N. The plant obtaining part of its N from the atmosphere will have less of the isotope. The percentage of total N from fixation is calculated as %N=( 1-

at. % 15N excess test crop at. % I5N excess control crop

) x 100

This technique involves the assumption that the isotope incorporated into soil N is equally available to both crops. Moreover, if the soil N is low, the nonfixing crop may grow less well than the test crop and thus not be a comparable control. Fried and co-workers (1975, 1977) proposed the “ A N value” modification. Adequate labeled fertilizer N can be added to the soil for the nonfixing control to promote growth. I5N in fertilizer N is added to soil for test plants if the effect of fertilizer N on fixation is required. The 15Ncontent of the plants is measured after harvest. For each crop and treatment, the “AN” value is determined. The A N value (Legg and Stanford, 1967) is an indication of how much of a nutrient is available to a crop and is expressed as AN = B ( l - y ) / y where B is the amount of fertilizer N added (kilograms N per hectare) and y is the proportion of N in plant derived from that added: excess 15N in plant = excess 15N in fertilizer

The fraction of legume N from fixation is calculated as Nitrogen fixed (kg/ha) = (AN test plant - A N nonfixing control) N in crop x excess 15N X N in fertilizer x excess I5N It should be possible to calculate for each crop and treatment the contribution made by soil, fertilizer, and symbiotic fixation. Procedural details and potential sources of error in these methods have been reviewed by Hauck and Bremner (1976) and Rennie et al. (1978).

3 . Methods Based on Natural Isotope Abundance There are very slight isotope effects during chemical or biological processes involving N compounds. Symbiotic fixation seems to favor a very slight increase of 15N. This can be determined by growing a plant on N-free nutrient so that all its N arises by fixation. The isotope discrimination of fixation for soybean has been calculated as 1.0038 (Bardin, 1977) and 1.0014 (Amarger et al., 1979) but also as 0.999 (Kohl and Shearer, 1980).

HOW MUCH NITROGEN DO LEGUMES FIX?

25

Denitrification produces N 2 lowered in 15N and the nitrate remaining in soil slightly enriched. The degree of enrichment is measured as parts per thousand difference from the l5NP4N ratio in a standard (usually atmospheric nitrogen) (Hauck and Bremner, 1976) 615N = (15NP4N) - (l5NP4N) atm (I5NP4N) atm where one 6

15N

unit equals 0.00037 at. %

looo

15N.

A plant obtaining all its N from soil (e.g., nonlegume or nonnodulated control) will have a slightly enriched I5N relative to the atmosphere. A plant obtaining N from symbiosis will have a lower I5N composition. The I5N enrichment is not uniform throughout the plant, although it is claimed that the seed content is representative of the whole soybean plant (Shearer et al., 1980). The percentage of N from fixation is calculated (Domenach et al., 1979): % N fixed = 100 x

(% 15N control - % I5N legume) % I5N control - (% 15N airlb)

where b is the isotope discrimination of fixation. An appealing advantage of this method of estimating fixation is that it does not necessitate the purchase of isotopic N to add to soil. The major disadvantage is that the enrichment may not be uniform in the soil (Rennie et al., 1976). Karamanos and Rennie (1980) found that the P 5 N dropped with depth in welldrained profiles but was constant to a depth of 5 m in upper slopes. Pedogenic processes produced marked differences in the 6I5N profile within short distances. These results extended and confirmed those of previous workers that there is an unpredictable lack of uniformity of enriched N in soils. The possible isotope discrimination due to symbiotic fixation may favor the fixation of I4N. If this is not included in the calculation, the level of fixation will appear to be higher than it is (Rennie et al., 1978). Experienced critics (Hauck and Bremner, 1976) view the technique based on natural abundance as giving only qualitative or semiquantitative information. However the standard errors of the procedure when applied to field-grown nodulating and nonnodulating soybeans compared favorably with the errors of estimation based on N difference (Kohl et al., 1980). D. ACETYLENE REDUCTION

Nitrogenase reduces acetylene to ethylene and, so far, it is the only biological agent reported to do so. Typically, freshly excised roots are incubated in a chamber with 1-20% C2H, for 30-120 min. A sample of the gas phase is then

26

THOMAS A . LARUE A N D THOMAS G . PATTERSON

removed and the ethylene produced is measured by gas chromatography. Innumerable variations of the method have been described (Hardy et al., 1973). The principal assumption in the method involves the ratio of acetylene reduced to nitrogen fixed. The reduction of nitrogen to ammonia uses six electrons, while the production of ethylene requires two. Therefore, the ratio 3 : 1 was originally assumed, i.e., a mole of ethylene was equivalent to % mole of N 2 reduced. It is now realized that the reaction of nitrogenase approximates N2 + 8 H+ + 8e ---t 2 NH,

+ H?

Protons and acetylene compete for electrons, and, with the amounts of GH, usually used, only small amounts of H2 may appear. In the intact nodule the hydrogen may be metabolized by hydrogenase in appropriate strains of rhizobia and therefore not be detected. There is no adequate method now for calibrating ethylene formation with nitrogen fixation, but it seems that the ratio is approximately 4 : 1 for legumes (Hardy et al., 1973). The acetylene reduction method has the advantages of sensitivity, speed, and economy. A detection limit of pmole C,H,/ml gas permits estimation of nitrogenase activity even when only a few nodules are forming. It is possible to measure 40-80 samples/operator-day. The operation of a gas chromatograph can be easily learned, and the analysis may be conducted by semiskilled staff. No expensive reagents are necessary. If sensitivity is not required, there are inexpensive alternatives to measuring ethylene (LaRue and Kurz, 1975). Because the plant is usually destroyed in the assay, acetylene reduction is a one-time estimate of fixation. The incubation must be done immediately after the root is harvested. There is a diurnal variation in fixation by many legumes (Sloger et al., 1975). Much plant-to-plant variability is observed, and the entire nodulated root may not be recovered from the soil. Therefore an estimate of fixation over a growing season by a crop requires a mathematical summation of many frequently obtained assays on replicate plots. E. OTHERMETHODSOF COMPARING FIXATION

Several indirect methods have been used for estimating the nitrogen fixing ability of a legume. These include indices of nodulation-number of nodules, fresh or dry weight of nodules, and leghemoglobin concentration in nodule or amount per plant. Within a single cultivar these may be closely related to nitrogen fixation and will continue to be useful in screening rhizobial strains or scoring the effects of environment on nodulation. We have seen no evidence, however, that these nodule-related characters can be used to calculate the amount

HOW MUCH NITROGEN DO LEGUMES FIX?

27

of fixation by crops. Nodules that are pink or red internally are only an indication, and not a proof, of nitrogen fixation (de Souza, 1969). The concentration of the ureides allantoic acid and allantoin in the shoot is positively correlated with the amount of fixation by soybeans grown in the greenhouse under defined conditions (McClure er al., 1980). This suggests that estimates of fixation might be made by assays, nondestructive to the plant, of N compounds in the shoot parts. Several laboratories are pursuing this topic. While there are indications that it may aid in the qualitative ranking of cultivars, the procedure seems unlikely to measure the amount fixed by a crop.

Ill. ESTIMATES FOR MAJOR CROPS

A. THE FORAGES

Most published estimates of fixation by forages assume that plant N arises only from N fixation. Excluding these, there remain very few estimates, most of which are based on long-term lysimeter studies (Table I). Except where noted, all studies used pure stands of legumes. Chapman et al. (1949) studied the nitrogen balance of vetch (Vicia villosa) and sweetclover (Melilotus alba) used as winter cover crops at Riverside, California, over a 10-year period. Only one lysimeter per treatment was used, due to their large size (3.04-m diameter) and small number (12). The legumes were grown in a winter rotation and incorporated into the soil prior to planting a summer crop of barley or sudan grass. Nitrogen fixation of the legumes was estimated by the increase in soil N due to the incorporation of the green manure into the soil. A 10-year average of N added to the soil gave an annual fixation rate of 140 kg N/ha for sweetclover and 184 kg N/ha for vetch. Estimates for alfalfa (Medicago sativa), white clover (Trifolium repens), red clover (Trifolium pratense), and Korean lespedeza (kspedeza stipulacea) came from an 1 1-year lysimeter study by Karraker et al. (1950) in Lexington, Kentucky. They studied the nitrogen balance in a continuous cropping system with forage legumes and Kentucky bluegrass (Poa prarensis). The lysimeters were 57 cm in diameter, and each treatment was replicated twice. Although the authors presented only the total N accumulation over the 1 1-year experiment, the presence of the Kentucky bluegrass control permits our calculation of the average annual N fixation by the difference method. Averages for each legume, minus the bluegrass control, are 212, 128, 154, and 193 kg N/ha for alfalfa, white clover, red clover, and Korean lespedesa, respectively. Jones et al. (1977) used a lysimeter-difference technique to estimate N fixa-

Table I Estimates of Nitrogen Firation by Forage Crops Contml species

Duration of experiment (years)

Location

~

16 II 6O

Geneva. NY Lexington, KY R m m o u n t , MN Lexington. KY Northern Ireland Beltsville, MD

Collison cr d.,1933 Karralicr cr al.. 1954 Heichel cr d.,I 9 8 1 Karraker cr ol.. 1950 Halliday and Pate, 1976 Wagner. 1954 Karraker er 01.. 1950 S p g u e . 1936 chapman cr al.. I949 Sprague. 1936 Phillip and Bennca. 1978 Joms er al.. 1977 Rizk. 1%2

165- 189

Lysimeter Lysimeter-differeme Isotope dilution-"A" Lysimeter-difference C& reduction Difference

154 I 7b 140 9b 21-183 207 62-235

Lysimter-difference Difference Lysrmeter Difference "A" value Lysimete-difference Difference

Winter wheat Soh chess Soh chess chicory

3 2

Lexington, KY New Brunswick. NJ Riverside, CA New Brunswick, NJ Hopland, CA Davis, CA Giza,E g y p

21'

Difference

Winter wheat

5

New Bmnswick. NJ

Sprague, 1936

641

Difference

Winter wheat

5

New Brunswick. NJ

Sprague. 1936

Difference Lysimeter Lysimeter-differencc

Winter wheat

5 10

Kentucky bluegrass

II

New Brunswick, NJ Riverside, CA Lexington. KY

Sprague. 1936 chapman e r a / . , 1949 K&cr ct d..1950

Alfalfa (Medicago sariva)

229-290 212 I48

White clover (Trifoliurn repms) Ladin0 clover (Trifoliurnrepens var. Ladino) Red clover (Medicago prmense) Sweet clover (Mclibrn a h ) Subclover (Trifoliurn subrerraneurn) Egyptian clover (Trifolium alexandrinurn) Alsike clover (Trifolium hybridurn) Crimson clover (Trifoliurn indica) Vetch (Vicia vilbsa) K o m kspedeza (kspedew sripulacca)

I28 258

aDuration given in months. *Winter crop rotation.

-

I lob

I84 193

Kentucky bluegrms Reed canarygrass Kentucky bluegrass

II

-

1

Orchard grass-call fescue

I

Kentucky bluegrass Winter wheat

-

-

II 5

10 5 7'

HOW MUCH NITROGEN DO LEGUMES FIX?

29

tion in subclover (Trifolium subterraneum L.) at Davis, California. Over a 3-year period, subclover fixed 261, 398, and 207 kg N/ha, as estimated by difference. The soft chess control (Bromus mollis L.) accumulated 37, 32, and 72 kg N/ha. Sprague (1936) evaluated a series of legumes for their usefulness as winter green manure crops in New Jersey. Over a 5-year period, winter wheat (Triticum aestivum L.), hairy vetch (Vicia villosa), crimson clover (Trifolium incarnatum), red clover, alsike clover (Trifolium hybridum), and sweet clover were planted in August following a corn crop. In April of the following year samples of the tops and roots (to a depth of 28 cm) were harvested in N analysis. The remainder of the crop was incorporated into the soil and the corn summer crop was planted. Over the 5-year period, the average fixation by difference was 17, 9, 21, 64, and 110 kg N/ha for red clover, sweet clover, alsike clover, crimson clover, and vetch. The relatively low rates of fixation are most likely due to the winter cropping system used in this study. Hairy vetch was found to be most suited to a winter green manure rotation. Sears et al. (1965) reported an extremely high rate of 406-681 kg N/ha for white clover in New Zealand (difference method-mixed grass control). These figures were obtained, however, in an artificially depleted soil system. The top 15 cm of soil was removed, and the remaining subsoil was mixed with additional subsoil and replaced. Thus the experimental system had an artificially lowered soil N level. Additionally, these rates may exceed estimates from other locations because the growing season was a full 12 months. A white clover perennial ryegrass lye i n Northern Ireland fixed N for 8 months (Halliday and Pate, 1976), but there was essentially no acetylene-reducing activity during November-February when the soil temperature averaged 4-5°C. Integration of the seasonal acetylene reduction profile indicated a fixation of 268 kg N/ha. Wagner ( 1 954) examined N fixation in ladino clover by the difference method at Beltsville, Maryland. Over a 2-year period, the clover averaged 165-189 kg N/ha. A study by Rizk (1962) in Egypt estimated the fixation of two varieties of Egyptian clover (Trifolium alexandrinum) utilizing the difference method with chicory (Chicorium intybus) as a control. A considerable difference between the cultivars was found, with the average nitrogen fixation ranging from 62-235 kg N/ha. Phillips and Bennett (1978) utilized the “AN value” technique and compared it to acetylene reduction activity in rangeland plots of subclover and soft chess in California. The influence of planting density and percentage clover in the plots was also studied. The amount of N fixed (kilograms per hectare) was found to be dependent on both stand density and composition. Low-density stands with 50% clover fixed 21.2 kg N/ha and derived 84.5% of the clover N from fixation. A

30

THOMAS A. LARUE AND THOMAS G . PATTERSON

low-density stand of 100% clover fixed 58.1 kg N/ha but derived only 50.1% of the clover N from fixation. High-density stands fixed 103 and 183 kg N/ha (94.5 and 88.0% clover N from N2) in 50 and 100% clover stands, respectively. Estimates for alfalfa in a northern environment came from Heichel et al. (1981) in Rosemount, Minnesota. They used isotope dilution and A N value to estimate N fixation in two populations of alfalfa selected for high nitrogenase activity. Reed canary grass (Phalaris arundinacea) was used as a control. Average N fixation in the establishment year was 148 kg N/ha, with an average of 43% of the nitrogen derived from fixation. Both the amounts of N fixed and percentage of N derived from fixation varied over the season. Nitrogen fixation was 8-20 kg N/ha at the first and fourth harvests (25-30% of plant N from N2). During the second and third harvest period the legume fixed 47-87 kg N/ha (60% of plant N from N2). Estimates for forage legumes vary widely between studies. White clover, for example, was reported to fix 128 kg N/ha in Kentucky and 408-681 kg N/ha in New Zealand. Differences in climate, management practices, and growing season cause considerable variation in the N fixation rates of forage legumes. Direct comparisons of the values in Table I therefore must be approached with caution. B. SEEDLEGUMES The soybean Glycine max is the most important cash crop among legumes in North America, and its N fixation has received more attention than other crops. Most studies do not lend themselves to estimates of amounts of fixation, though several yield data on the percentage of N derived from fixation (Tables I1 and 111). Rizk (1 966) used the difference method to calculate fixation by soybean in a calcareous sand at Kutch, Egypt. Sesame (Sesamum indicum) was the nonfixing control and soil N was measured before sowing and after harvesting. Rizk reported a fixation of 16.7 kg N/ha and a seed yield of 13.7 kg N/ha. These low figures support his statement that he had not found the most suitable tillage practice for this crop. Weber (1966) compared the above-ground N content of a nodulating and a near-isogenic nonnodulating soybean cv. Lee at Ames, Iowa, for 6 years. In 2 years when moisture was limiting, the difference was 14.7 kg N/ha, representing about 13% of the plant N. In 4 years with good growth conditions, the mean difference was 72.3 kg N/ha in seed or 84 kg N/ha in total dry matter. In both cases this represented 40% of the total N in the nodulated plant. When the soil was amended with 45,000 kg/ha of corn cobs to immobilize N, the average difference in 3 years with good growth conditions was 160 kg N/ha in dry matter, or 74% of the total N . This higher figure may illustrate a potential for fixation by

31

HOW MUCH NITROGEN DO LEGUMES FIX? Table I1 Estimates of Nitrogen Fixation by Soybean

Location

Estimate (kg N/ha) 39.7 14.7-84

Egypt Iowa

Delaware

Method Difference Difference

GH, reduction

33-42 80- 120 103 56- I61

“ A ” value Difference

Minnesota

263 I05 76-152

Difference GH, reduction “ A ” value

Nebraska

43-146

“A” value

Southern Ontario

78-161

C,H, reduction

Illinois Arkansas Washington

Remarks Sesame Nonnodulating isoline, soil low in N Fertile soil

Nonnodulating isoline Uninoculated Soil low in N Nonnodulating isoline Nonnodulating isoline lnoculant trial

References Rizk ( 1 966) Weber (1966)

Hardy et a / . (1973) Johnson et al. (1975) Bhangoo and Albritton (1976) Bezdicek et a / . (1978) Ham and Caldwell (1978) Deibert et al. (1979) Muldoon et al. (1980)

soybean, but the extraordinary soil amendment is not representative of normal practice. Soybeans were grown in soils high in available N near Wilmington, Delaware, and assayed by the acetylene reduction technique. In 3 years with a low planting density, they fixed an average of 38 kg N/ha. In other trials with higher planting densities, rates of 80- 120 kg N/ha were calculated. These approximated to 25-30% of the total plant N (Hardy et a l . , 1973). Table 111 Estimates of Nitrogen Fixation by Pulses

Pulse Phaseolus vulgaris“ Pisum sativum Vicia faba Lupine Chick-pea Lentil Arachis hypogea

“25-120 mg/plant. bCalculated.

Estimate (kg N/ha) 1Ob

17-69 121-171 121-157 67-141 62-103 79 87-222

Method

Reference

GH, reduction GH, reduction Difference (barley) Difference (barley) Difference (barley) Difference (barley) Difference (sesame) Difference (uninoculated check)

Westerman and Kolar (1978) Mahler et al. (1979) b z k (1 966) Rizk (1966) Rizk ( I 966) Rizk ( I 966) Rizk ( 1 966) Ratner et al. (1979)

32

THOMAS A. LARUE AND THOMAS G . PATTERSON

Bhangoo and Albritton ( 1 976) compared nodulating and nonnodulating lines of cv. Lee for 3 years at Pine Bluff, Arkansas. The N differences in seed and dry matter were 130, 161, and 56 kg N/ha, representing 51, 64,and 22% of total plant N. In the first 2 years the trials were at the same plots, whereas in the third year the soybeans were planted on an area that had previously grown corn and had been heavily fertilized with NPK. This latter figure is probably more representative of the way soybean is generally grown. A comparison of inoculated and uninoculated soybeans cv. Merit was made in an irrigated low-N soil at Prosser, Washington (Bezdicek et al., 1978). The total plant N was 368 and 105 kg/ha; the difference of 263 kg N/ha representing 71% of the inoculated plant. An estimate based on acetylene reduction was 105 kg N/ha. The threefold difference in seed yield and plant top N suggests that the roots might not be comparable in the two treatments; perhaps inoculation provided sufficient N to support additional root growth to obtain soil N. In that case, the difference method would overestimate fixation. Ham and Caldwell (1978) tested cv. Clay at Rosemount, Minnesota, with plots amended with I5N and with different P levels. Fixation calculated by the “ A value” method was 76-152 kg N/ha. The authors reported that these results agreed with measurements made by Kjeldahl N and by weekly acetylene reduction experiments made with the same variety on an adjacent plot. In Nebraska, soybeans cv. Ford were grown with sprinkle irrigation (Deibert er al., 1979), with rates of added N of 0-134 kg/ha. At the & (beginning seed) stage, the total plant N was 70 and 178 kg/ha in nonnodulating and unfertilized nodulating lines, respectively, and at the highest levels of fertilizer were 170 and 180 kg N/ha. A , value calculations indicated that fixation had contributed 60 and 33%, respectively, under the two regimes. At harvest, the difference in seed N was 51 and 61 kg N/ha, and the percentage of N in seed from fixation was calculated at 66% in unfertilized and 31% in fertilized plants. Muldoon et al. (1980) compared a seed-applied inoculant and two soil-applied granular inoculants at three sites in southern Ontario, Canada. The fields had not been planted with soybean previously. Acetylene reduction assays indicated a range of 12-17 kg N/ha in uninoculated plots. Fixation with the largest amount of seed-applied inoculant was 78 kg N/ha. At the heaviest applications of two granular soil inoculants, the fixation rates were calculated at 128 and 161 kg N/ha. Differences in total plant N between inoculated and uninoculated plots were usually lower than the N (GH,)calculated. Interestingly, the added yield barely paid for the cost of granular inoculants if they were used at recommended levels. Four varieties of soybeans were grown in the field at Diome, France (Amarger et al., 19791, and fixation was estimated from variations in natural abundance. Yield figures were not presented, but fixation apparently contributed 13-37% of the plant N. The effect of soil fertility and added fertilizer on the degree of symbiotic

HOW MUCH NITROGEN DO LEGUMES FIX?

33

fixation was estimated by isotope dilution at Urbana, Illinois (Johnson et al., 1975). Soybeans in unfertilized fields fixed 48% of their N; this decreased to 10% as fertilizer rates increased to 224 kg N/ha. The authors concluded that the upper figure was a good estimate for their region. Calculations indicated that soybean is a good scavenger for soil N and that soybean removes larger quantities of N from soil than corn does. The natural abundance of I5N was measured in normal and nonnodulating isolines of soybean cv. Harosoy at Urbana, Illinois. The percentage of N contributed by fixation was calculated by the N difference method and calculations from isotopic concentrations (Kohl et al., 1980). In 2 years the N from fixation was 14 and 37% (by difference) or 33 and 43% (by W”). Depleting soil N by adding corn cobs increased the fixation to 5 3 or 56% estimated by the respective methods. The studies listed here are from every soybean-growing area of North America. Yields, when reported, were over 2000 kg/ha. Except when soil N was low, the contribution of fixation rarely exceeded 50%. The average American soybean yield (1975) was 1471 kg/ha. Assuming that the harvested seed represents about two-thirds of plant N, the whole plant N approximates 150 kg/ha on farms. Since it is unlikely that fixation on farms is more than 50%, the upper estimate for average fixation by soybeans in American agriculture is 75 kg N/ha.

I. Viciu faha L . Faba beans use soil or fertilizer N in preference to symbiosis. Pot experiments (Richards and Soper, 1978) show that they are as adept as barley in extracting available N. In pots low in N, 87% of the plant N was derived from fixation ( A N value method). Fertilizer addition decreased fixation without increasing total shoot N. In the field, faba bean yields are not responsive to added fertilizer. From this it was assumed that symbiotic fixation was adequate and that “the nitrogen in the grain is an approximate estimate of the atmospheric nitrogen “fixed” (McEwen, 1970). There was no experimental support for this view. The results of Sprent and Bradford (1977) indicate that during pod fill fixation estimated by acetylene reduction accounts for about one-third of the plant’s N gain. Estimates in soils at Giya and Sids, Egypt, were based on crop and soil accumulation of N, compared to barley, and were in the range 49-171 kg N/ha (Rizk, 1966).

2 . Arachis hypogaea L . Compared to sesame, peanuts at Kutah, Egypt, accumulated 79 kg N/ha in crop and soil (Rizk, 1966). In 3 years at Lakhish, Israel, the shoot N difference between uninoculated and

34

THOMAS A. LARUE AND THOMAS G . PATTERSON

inoculated peanuts was 222, 93, and 87 kg N/ha. These corresponded to 58,40, and 30% of total N in the inoculated plants (Ratner et a l . , 1979). The highest estimate of fixation was associated with a hay yield of 8667 kg/ha and a pod yield of 7042 kg/ha containing 5056 kg/ha kernel. The soil was not low in nitrogen and the crops were frequently imgated. The extraordinary yields suggest that the estimates for total fixation should not be applied to typical peanut crops, which have an average yield of 980 kg/ha (FAO, 1976). 3 . Pisum sativum L .

Mahler e f al. (1979) measured CpH, reduction in pea plots on the ridge top, south slope, and bottomland of a catina near Pullman, Washington. The three sites were within a few hundred meters of each other. The total plant N at the three sites was 75, 99, and 210 kg/ha. The calculated fixed N was 17 (23%), 22 (23%), and 69 (33%) kg/ha. The lower yields on the ridge and slope were associated with water stress. The differences in fixation observed within a small region demonstrate the futility of extrapolating amounts fixed in agriculture from single investigations. 4 . Phaseolus vulgaris L .

Acetylene reduction rates were estimated for 18 dry bean cultivars at Kimberly, Idaho (Westerman and Kolar, 1978). The calculated N fixation was in the range 25-120 mg N/plant. Fixation on an area basis was not reported, but the planting density was about 100,000 plants/ha. This suggests a fixation of about 10 kg N/ha, a small part of the total uptake of 150-400 kg N/ha. Unfortunately there are few estimates of fixation by pulses important to farmers in developing nations. The accumulation of plant and top soil N in low-N Egyptian soils was estimated by Rizk (1966). The rates of fixation by lupine (121-157 kg N/ha), chick-pea (67-141 kg N/ha), and lentil (62-103 kg N/ha) are not necessarily representative of the rates on farms. Cowpeas grown in a greenhouse can fix 88% of their N (Huxley, 1980). Cowpeas are generally grown as intercrops with maize or sorghum. Since the average yield of cowpeas is only 200 kg/ha (FAO, 1976), the amount of N fixed per hectare must be small.

IV. SUMMARY A. EVALUATION OF STUDIES

There is not a single legume crop for which we have valid estimates of the N fixed in agriculture. There are good estimates for soybean grown in representative

HOW MUCH NITROGEN DO LEGUMES FIX?

35

locations in experimental plots. However, extrapolation from this data to fixation in agriculture is speculative. For other pulses the few documented reports do not permit estimation of fixation on farms. The data on forage legumes are sparse and mostly derived from studies of pure stands. Estimates from more locations are required. In practice the forages are often planted with a grass, or over time become admixed by grasses. There is an almost complete lack of data on fixation by mixed stands, or on the amounts of fixation when forages are harvested or browsed. It is regrettable that fixation by legumes, which are important to the developing world, is not better documented. Dry beans are the most important legume crop for human consumption in Latin America. Cowpeas are a staple in much of Africa and Asia, and chick-pea is important in semiarid regions. For none of these are data available. There is no good evidence that any legume crop satisfies all its N requirements by fixation. Soybean fixation has been estimated in several areas of the United States and from consideration of all the data we must conclude that this crop depletes soil N. There are no substantiated reports that 100% N of any plant derives from symbiosis. The highest estimates (-80%) are typical of lowfertility soil or soils artificially made N-poor by admixture or carbon amendment. B. REQUIREMENTS FOR FUTURERESEARCH

Appropriate techniques for examining N fixation in perennial forage legumes must be developed. The interpretation and accuracy of long-term isotope studies needs to be evaluated. Since forage legumes are often grown in association with perennial grasses, techniques for estimating N fixation and N transfer between species need to be developed. A simpler technique, such as the difference method using a nonnodulating control, would find wide application if equivalence to the isotope procedures can be confirmed. The identification and development of more nonfixing forage genotypes should be a priority. Viands et al. (1979) have identified an ineffective nodulation character in alfalfa. Nonfixing strains in other species would be valuable. With the rising cost of N fertilizer, the potential of using legumes to increase soil N becomes increasingly attractive. Green manure systems were once much used to increase soil N (Sprague, 1936). A reexamination of the potential of legumes as a soil N source is called for. Quantitative data on the contribution of fixed N to the soil and to succeeding crops is needed for an adequate economic analysis of green manuring. Adequate methods of estimating fixation in pulses exist, but they are not yet so simple and inexpensive as to be commonly used. The acetylene reduction assay as now used requires too many samples over the growing season to be convenient

36

THOMAS A . LARUE A N D THOMAS G . PATTERSON

for analyzing many cultivar, strain, location, fertilizer, or environmental combinations. Nondestructive in situ measurements may be made with inexpensive apparatus (Mahon and Salminen, 1980), but the calibration of this assay with N fixed has not been accomplished. In developed countries with access to the necessary apparatus, isotopic techniques are likely to become more common. The 615N method of natural enrichment should be tested as a method of estimating fixation in agriculture by sampling seeds and comparing them with cereal grain harvested in the same region. This might serve as a way of estimating, albeit approximately, fixation over large areas. In developing countries there is need for inexpensive methods not requiring expensive imported apparatus. The Kjeldahl method for N accumulation in plant and soil would be useful if a way were found for estimating N mobilized from lower strata. This can be accomplished most easily if in each geographic region nonlegumes are found that utilize soil N in a manner similar to that of the major legume crops. There is then an urgent requirement to test nonlegumes that might serve as suitable controls for the difference method. Indirect methods based on N compounds in legume shoots might be used if they could be calibrated with fixation. Plant physiologists should determine whether symbiotic fixation in the field is more demanding of photosynthate than use of soil N. They should determine why legumes prefer to use soil N, and why even low levels of available N will decrease nodulation and fixation in the field. If this knowledge were available, it might be possible to breed legumes, especially forages, that would supply more of their N needs by fixation and leave soil N for grasses or cereals. REFERENCES We have omitted many articles and reviews stating estimates of symbiotic fixation without citing the source of information. We have also left out estimates, often anonymous, in commercial brochures and Institute or Station annual reports in which the experiment leading to the estimate is not documented. We noted only a few of the many reports that assume, without evidence, that total plant N, shoot N , yield, or some multiple of nodule weight, etc. is the measure of fixation. Amarger, N., Mariotti, A . , Mariot’ti, F . , Durr, J. C., Bourguignon, C., and Lagacherie B. 1979. Plant Soil 52, 269-280. Bardin, R., Domenach, A. M., and Chalamet, A. 1977. Appl. Rev. Ecol. Biol. Sol. 14, 395-402. Bezdicek, D. F . , Evans, D. W., Abede, B., and Witters, R. E. 1978. Agron. J . 70, 865-868. Bhangoo, M. S . , and Albritton, D. J. 1976. Agron. J . 68, 642-645. Bremner, J . M. 1965. In “Methods of Soil Analysis” (C. A . Black, ed.), Part 2, pp, 1149-1 179 and 1256-1286. Amer. SOC.Agron. Madison, Wisconsin. Bums, R. C., and Hardy, R. W. F. 1975. “Nitrogen Fixation in Bacteria and Higher Plants.” Springer-Verlag. New York. Bums, R. H. 1972. I n “Methods in Enzymology” (A. San Pietro, ed.), Vol. 24, pp. 415-431. Academic Press, New York.

HOW MUCH NITROGEN DO LEGUMES FIX?

31

Bums, R. H. 1978. In “Environmental Role of Nitrogen-fixing Blue-green Algae and Symbiotic Bacteria” (U. Granhall, ed.). NFR Editorial Service. Stockholm. Bums, R. H., and Wilson, P. W. 1957. In “Methods in Enzymology” (S. P. Colowick and N. 0. Kaplan, eds.), Vol. IV, pp. 355-366. Academic Press, New York. Chapman, H . D . , Liebig, G . F . , and Raynes, D. S. 1949. Hilgardia 19, 57-128. Collison, R. C.. Beattie, H. G., and Harlan, J. D. 1933. N.Y. Agric. Exp. Sta. Tech. Bull. No. 212. Cowling, D. W . 1961. J. Br. Grassl. Soc. 16, 281-290. Craswell, E. T . , and Martin, A . E. 1975a. Aust. J. SoilRes. 13, 43-52. Craswell, E. T., and Martin, A . E. 1975b. Aust. J. SoilRes. 13, 53-61. Deibert, E. J. Bijeriego, M . , and Olson, R. A. 1979. Agron. J . 71, 717-723. de Souza, D. I. A. 1969. East Afr. Agric. For. J . 34, 299-305. Domenach, A. M . , Chalamet, A., and Pachiaudi, C. 1979. C.R. Acad. Sci. (Paris) D289,291-293. Erdman, L. W. 1949. USDA Farmers Bull. No. 2000, pp. 1-20. F A 0 1976. “1975 Production Yearbook,” Vol. 29. Food and Agricultural Organization of the United Nations, Rome. Fried, M., and Dean, L. A. 1952. Soil Sci. 73, 263-271. Fried, M., and Middelboe, V. 1977. Plant Soil 47, 713-715. Halliday, J . , and Pate, J. S. 1976. J. Br. Grassl. SOC.31, 29-35. Ham, G. E., and Caldwell, A. C. 1978. Agron. J. 70, 779-783. Hardy, R. W. F., and Havelka, U. D. 1975. In “Symbiotic Nitrogen Fixation in Plants” (P. Nutman, ed.), Int. Biol. Prog. Ser. Vol. 7, pp. 421-439. Cambridge University Press, London. Hardy, R. W. F., Bums, R. C., and Holsten, R . D. 1973. Soil Biol. Biochem. 5, 47-81. Hauck, R. D., and Bremner, J. M. 1976. In “Advances in Agronomy” (N. C. Brady, ed.), Vol. 28, pp. 219-266. Academic Press, New York. Heichei, G. H . , Barnes, D. K.,and Vance, C. P . 1981. Crop Sci. 21, 330-335. Henzell, E. F. 1962. Aust. J. Exp. Agric. Anim. Hush. 2, 133-140. Hudd, G . A., Lloyd-Jones, C. P., and Hill-Cottingham, D. C. 1980. Physiol. Plant. 48, 1 I 1-1 15. Huxley, P. A . 1980. Trop. Agric. 57, 193-202. Jones. M. B., Delwiche, C. C . , and Williams, W . A. 1977. Agron. J . 69, 1019-1023. Johnson, J. W . , Welch, L. F., and Kurtz, L. T. 1975. J. Environ. Q u a / . 4, 303-306. Kamprath, E. J . , Chandler, W. V., and Krantz, B . A. 1958. N.C. Agric. Exp. Sta. Tech. Bull. No. 129. Karamanos, R. E., and Rennie, D. A. 1980. Can. J. Soil Sci. 60, 337-344, 365-372. Karraker, P. E . , Bartner, C. E., and Fergus, E. N . 1950. Kentucky Agric. Exp. Sta. Bull. No. 557. Kohl, D. H . , and Shearer, G. 1980. Plant Physiol. 66, 51-56. Kohl, D. H . , Shearer, G . , and Harper, J. E. 1980. Plant Physiol. 66, 61-65. Kohnke, H.. Dreibelhis, F. R . , and Davidson, J. M. 1940. USDA Misc. Puhl. No. 372. Kroontje, W . , and Kehr, W. R. 1956. Agron. J. 48, 127-131. LaRue, T. A , , and Kurz, W . G. 1973. Plant Physiol. 51, 1074-1075. Legg, J . O., and Sloger, C. 1975. In “Proceedings of the Second International Conference on Stable Isotopes” (E. R. Klein and P. D. Klein, eds.), pp. 661-666. Argonne National Laboratory, Argonne, Illinois. Legg, J. 0.. and Stanford, G. 1967. Soil Sci. Soc. Am. Proc. 31, 215-219. Lloyd-Jones, C . P . , Adam, J. S., Hudd, G. A., and Hill-Cottingham, D. G . 1977. Analyst 102, 473-476. Lyon, T . L., and Bizzell, J. A . 1911. Cornell Agric. Exp. Stu. Bull. No. 294. Lyon, T. L., and Bizzell, J . A . 1933. Agron. J. 25, 266-272. Lyon, T . L., and Bizzell, J. A. 1934. Agron. J. 26, 651-656. McAuliffe, C., Chamblee, D. S . , Uribe-Arango, H., and Woodhouse, W . W. 1958. Agron. J. 50, 334-337.

38

THOMAS A. LARUE A N D THOMAS G . PATTERSON

McClure, P. R . , Israel, D. W., and Volk, R . J. 1980. Plant Physiol. 66, 720-725. McEwen, J . 1970. J . Agric. Sci. (Cambridge) 14, 61-66; 67-72. Mahler, R. L., Bezdicek, D. F., and Witters, R. E. 1979. Agron. J. 71, 348-351. Mahon, J. D. 1977. Plant Physiol. 60, 812-816, 817-821. Mahon, J. D. 1978. Plant Physiol. 63, 892-897. Mahon, J . D., and Salminen, S. 0. 1980. Plant Soil 56, 335-340. Muldoon, J. F., H u m , D. J., and Beversdorf, W. D. 1980. C a n . J . Plant Sci. 60, 399-409. Nutman, P. S. 1973. I n “Symbiotic Nitrogen Fixation by Plants’’ (P. S. Nutman, ed.), pp. 211-237. Cambridge Univ. Press, London and New York. Patwary, S. U., and Raikovich, Z. 1979. Plant Soil 52, 209-217. Pessanha, G. G., Franco, A. A . , Dobereiner, J., Groszmann, A., and de S. Britto, D. P. P. 1972. Pesqui. Agropecu. Bras. Ser. Agron. 7, 49-56. Phillips, D. A,, and Bennett, J. P. 1978. Agron. J . 70, 671-674. Ratner. E. I., Lobel, R., Feldhay, H., and Hartzook, A. 1979. Plant Soil 51, 373-386. Rennie, D. A , , Raul, E. A., and Johns, L. E. 1976. C a n . J . Soil Sci. 56, 43-50. Rennie, R. J., Rennie, D. A , , and Fried, M. 1978 In “Isotopes in Biological Dinitrogen Fixation,” IAEA, Vienna. Richards, J. E.. and Soper, R. J. 1978. Agron. J . 71, 807-811. Rizk. S. G. 1962. J . SoilSci. U . A . R . 2, 253-270. Rizk, S. G. 1966. J . Microbiol. U . A . R . 1, 33-45. Ruschel, A. P., Vose, P. B., Victoria, R. L., and Salati, E . 1979. Plant Soil 53, 513-525. Ryle, G. J. A., Powell, C. E . , and Gordon, A . J. 1979. 1. Exp. Bor 30, 135-144, 145-154. Saito, S. M. T., Matsui, E., and Salati, E. 1980. Physiol. Plant. 49, 37-42. Sears, P. B., Goodall, V. C., and Jackman, R. H. 1965. New Zeal. J . Agric. Res. 8, 270-283. Shearer, G . , Kohl, D. H., and Harper, J. E. 1980. Plant Physiol. 66, 57-60. Silsbury, J. H . 1977. Nature (London) 267, 149-150. Silsbury, J. H . 1979. Aust. J . Plant Physiol. 6, 165-176. Sloger, C., Bezdicek, D., Milberg, R., and Boonkerd, N. 1975. In “Nitrogen Fixation by FreeLiving Organisms” (W. D. P. Stewart, ed.), IBP No. 6, pp. 271-284. Cambridge Univ. Press, London and New York. and Carter, J. N. 1963. Soil Sci. 96, 313-318. Smith, J. H., Legg, J. 0.. Sprague, H. B. 1936. N . J . Agric. Exp. Sfa. Bull. No. 609. Sprent, J. I . , and Bradford, A. M. 1977. J . Agric. Sci. (Cambridge) 88, 303-310. Viands, D. R., Vance, C. P., Heichel, G. H., and Barnes, D. K . 1979. Crop Sci. 19, 905-908. Wagner, R. E. 1954. Agron. J . 46, 233-237. Weaver, J. E. 1926. “Root Development of Field Crops.” McGraw-Hill, New York. Weber, C . R. 1966. Agron. J . 58, 46-49. Westerman, D. T . , and Kolar, J. J. 1978. Crop Sci. 18, 986-990. Williams, L. F . , and Lynch, D. L. 1954. Agrori. J . 46, 28-29. Williams, W. A,. Jones, M. B., and Delwiche, C. C. 1977. Agron. J . 69, 1023-1024.

ADVANCES IN AGRONOMY, VOL. 34

PEANUT BREEDING' J. C. Wynne and W. C. Gregory Crop Science Department, North Carolina State University, Raleigh, North Carolina

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Germ Plasm Resources

Ill.

IV.

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

A. Subspecies and Commerci B. Related Wild Species. . . . C . Induced Variability . , , . . . , . , . . . , . . . . . . . . . . . , . , . , . . . , . , . . , , . . . . , . . . . , . , D. Origin and Distribution , . . . . . . . . . . . . . . . . . . , . , . . . , . . . , . . . . . . . . . . . . , . , , . , E. Germ Plasm Collection , . . . , . . . . , . , , . , . , , . , , . , . , . . , . . , . . . . , . . , . . , . , . , . , Economic Importance and Breeding Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Higher Yields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Oil and Protein Content.. . . . . . . . . . . . . . . . . , . . . . . . . . . . ............... ... .. ..., ., ., . , C . Pest Resistance.. . . . . . . . . . . . . . . . . . . . , . . . . . , . . . . . . . . Breeding and Quantitative Genetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Variability, Heritability, and Correlation , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , B . Type of Gene Action.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Effects of Selection . . . . , . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . , , . . . D. Genotype x Environment 1 .................. E. Evaluation of Stability . . . .................. F. Implications for Breeding ..................................... Breeding Methods .............................. A. Introduction and Pure-Line Selection , . , . , , , , , . , . , . . . , , . , . . . . , . . . . . , . , , , . . B. Hybridization . , , , . , , , , , . . , , . , . . , , , , , . . , , . . . , . , . , . , . , . _ _, . . . , . , . , . . . . Interspecific Hybridization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . '

V.

VI. VII.

39 40

42 42 42

44 44 44 45 49 49 54 59 59 61 62 63 63 63 65 68 68

1. INTRODUCTION Peanut (Aruchis hypogaea L.) breeding received little attention in the United States before the early work by Hull, Carver, and Higgins in Florida and Georgia in the 1930s and by Gregory in North Carolina in the 1940s. As a consequence, peanut breeders are now investigating the somewhat dated estimates of quantita'Paper No. 6741 of the journal series of the North Carolina Agriculture Research Service, Raleigh, North Carolina. 39

Copyright 0 1981 by Academic Press. Inc. All rights of reproduction in any form reserved.

ISBN 0-12-ooO734-7

40

J. C. WYNNE A N D W . C. GREGORY

tive genetics and methods of field testing while needing to investigate novel approaches to interspecific hybridization and gene introduction into cultured cells. Ever since the natives of South America collected the ancestors of modern peanuts from the wild, the difficult process of adapting the peanut to new environments has been in progress. Some adjustments of the environment to peanut production have been made and will continue to be made in the future, but the great adaptive contingency lies with modification of the plant. Peanut breeding, like all other plant breeding efforts, is based upon confidence that application of the Darwinian theory of selection alters the hereditary constitution of the plant. Ideally, in a self-pollinated crop such as the peanut, it would be advantageous to insert genetic sequences of known and specific desired effect into the superior cultivars one already has, thus maintaining or enhancing the superiority already gained through the transformation of wild ancestors into cultivated plants. If this insertion were possible, the devastating effects of outcrossing to less effective germ plasm could be avoided and the temptations to narrow the genetic base could be resisted. The major costs of peanut breeding would be confined to the search for, identification of, and the insertion of the desired genetic sequences into the optimum number of sites and the final testing of the product. It has been demonstrated that there exist environments in which ordinary land races of peanuts are equal in productivity to the best products of the modern plant breeder (Hildebrand and Smartt, 1980). As Hildebrand and Smartt point out, the extraordinary response in Zimbabwe of the land races of Bolivia (5 1500 m above sea level at 16-17"s) to the plateau of Africa (15-16"s) and their corresponding low productivity near sea level in the United States is matched by the poor performance in the plateau of Africa of the exceptionally productive cultivars bred for the southeastern coastal plain of the United States. Elevation, photoperiod, and temperature-the environmental factors presumed to be involved in these differential responses-annot be changed. The task is to change the genome. It is the study of how these changes can be recognized and managed amongst the multitude of variables within local environments that is the principal theme of the review presented here.

II. GERM PLASM RESOURCES A . SUBSPECIES A N D COMMERCIAL NAMESOF CULTIVATED PEANUTS

Krapovickas and Rigoni (1960) subdivided A . hypogaea into two subspecies ( A . hypogaea ssp. hypogaea and A . hypogaea ssp. fastigiata Waldron), each

41

PEANUT BREEDING

composed of two botanical varieties. Krapovickas (1968, 1973) further clarified the classification of peanuts and identified five centers of diversity: Guarani, Bolivia, Peru, Amazon, and the region of Goias and Minas Gerais. Gregory and Gregory (1976) extended the number of regions to six to include the northeast of Brazil. The subspecific nomenclature was related to geographical regions as follows (Gregory el al., 1980): A . hypogaea L.

ssp. hypogaea var. hypogaea-Bolivian, Amazonian (Virginia type) var. hirsura-Peruvian ssp. fastigiata Waldron var. fastigiara-Guaranian, Goias and Minas Gerais, Peruvian northeast Brazil (Valencia type except for Peruvian) var. vulgaris Ha.-Guaranian, Goias and Minas Gerais, northeast Brazil (Spanish type) B. RELATEDW I L DSPECIES

The genus Arachis is comprised of 22 described species, including cultivated peanuts, and possibly 40 or more as yet undescribed annual or perennial species all originally collected from South America (Gregory er al., 1980). These species have been assigned to seven sections, several of the sections being subdivided into series (Table I) (Gregory et al., 1973; Krapovickas, 1973; Gregory et al., 1980). Table I Taxonomic Classification of the Genus Arachis L." Section Arachis nom. nud.

Erectoides Krap. et Greg. nom. nud.

Series Annuae Krap. et Greg. nom. nud. Perennes Krap. et Greg. nom. nud. Amphiploides Krap. et Greg. nom. nud. Trifoliolatae Krap. et Greg. nom. nud. Tetrafoliolatae Krap. et Greg. nom. nud. Procumbensae Krap. et Greg. nom. nud.

Caulorhizae Krap. et Greg. nom. nud. Rhizomarosae Krap. et Greg. nom. nud. Prorhizomatosae Krap. et Greg. nom. nud. Eurhizomatosae Krap. et Greg. nom. nud. Exrranervosae Krap. et Greg. nom. nud. Ambinervosae Krap. et Greg. nom. nud. Triseminalae Krap. et Greg. nom. nud.

"After Gregory er a l . (1973, 1980) and Krapovickas (1973).

Ploidy level 2n 2n

4n 2n 2n 2n 2n 2n

4n 2n 2n 2n

42

J . C. WYNNE AND W . C. GREGORY

C . INDUCED VARIABILITY

Several induced mutation studies have bees conducted with peanuts (Ashri and Goldin, 1965; Emery et al., 1965, 1972; Avadhani and Rao, 1968; Gregory, 1968; Emery and Wynne, 1976). Gregory and his associates (Gregory, 1966, 1968) demonstrated that considerable variability controlled by both major and minor genes could be induced in the peanut using a mutagen. Mutants with higher yield (Gregory, 1968), Cercospora leafspot resistance (Cooper and Gregory, 1960), and potato leafhopper resistance (Campbell et al., 1976) were identified in irradiated populations. D. ORIGINA N D DISTRIBUTION

The cultivated peanut is found throughout the tropical and temperate regions of the world; however, wild species of Arachis are found only in South America. The species of Arachis are found in an area bounded by the Amazon River to the north, the Rio de la Plata to the south, the Andes Mountains to the west, and the Atlantic to the east. The Bolivian region was identified as the center of origin of the cultivated peanut, with the regions shown in Fig. 1 as secondary centers of diversity. Gibbons et al. (1972) classified the variability that had originated in or been introduced to Africa during the last five centuries. They emphasized that each of the botanical varieties of Krapovickas’ classification contains a large array of local cultivars. Gibbons et al. (1972) suggest that the variation in the African collection has arisen by hybridization and subsequent selection in Africa, making it an important secondary center of variation of both the Guarani and Bolivian peanuts of South America. Gregory et al. (1973) caution that additional exploration of the Bolivian center of origin is needed to resolve the relationship between South American and African variability. Such exploration has been in progress since 1976. E. GERMPLASMCOLLECTION

Collections of cultivated peanuts are already quite extensive (Banks, 1976). Much of the present collection was obtained from specific expeditions made to South America under the sponsorship of the United States Department of Agriculture with cooperation from state experiment stations and foreign countries. More recently, the International Board for Plant Genetic Resources has funded collection efforts. The most important collections are those of Archer in 1936;

43

PEANUT BREEDING

* I I

SECONDARY CENTERS :

I GUARANI I1 111 IV V VI

G O I A S AND M I N A S G E R A I S WESTERN B R A Z I L BOLIVIA PERU NORTHEASTERN BR A Z I L

FIG. 1. Centers of origin and diversity of the groundnut, Arachis hypogaea (from Gregory and Gregory, 1976).

Stephens and Hartley in 1947-1948; Gregory, Krapovickas, Pietrarelli, and others in 1959, 1961, and 1967; Hammons, Langford, Krapovickas, Pietrarelli, and others in 1968; and Gregory, Banks, Simpson, Krapovickas, Pietrarelli, and others in 1976, 1977, 1979, and 1980. Much of the germ plasm from Africa was introduced to the United States by Smartt in 1959. In addition, material has been received into the United States from other foreign sources (Banks, 1976). Approximately 4000 accessions are maintained by the Southern Regional Plant Introduction Station at Experiment, Georgia. The most extensive collection of cultivated peanut germ plasm is now maintained by the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) at Hyderabad, India. Their collection consists of over 8000 accessions (Rao, 1980).

44

J. C. WYNNE AND W. C. GREGORY

Ill. ECONOMIC IMPORTANCE AND BREEDING OBJECTIVES The peanut is both an oilseed crop and a food grain legume. It is grown on an estimated 18.9 million ha in 82 countries for use as food, oil, and a high-protein meal. Peanut improvement programs have placed major emphasis on the development of cultivars with high yields, high oil content, good flavor, and characteristics desired by the processor. Breeding for pest resistance is receiving much effort as well. The incorporation of resistance to the many peanut diseases into adapted types has become a major goal of breeding programs. Breeding for insect resistance has not received as much emphasis as selection for disease resistance because insect problems are usually more localized and thus command less attention. A. HIGHER YIELDS

Despite the doubling of yields over the past 40 years, increased productivity remains one of the principal objectives of peanut breeding. Duncan et al. (1978) monitored the partitioning of the total assimilates in both new cultivars and old land races and found that the rate and duration of the fruit filling period were the factors that accounted for most of the variation in yield. They concluded that selection for yield has resulted in peanut cultivars that partition more of their daily assimilate to fruit (Duncan et al., 1978). This conclusion suggests that increased yields have resulted primarily from more efficient exploitation of existing fruiting sites rather than from increasing the number of sites available. If this suggestion proves to be correct, breeders may do well to alter their priorities or to achieve a greater insight into possible structural alterations in the peanut plant for increasing fruit production. B . OIL A N D PROTEINCONTENT

Recent oil analyses suggest that much could be done to increase oil production in peanuts. For example, Cherry (1977) determined the quantity of oil from 37 selected wild species and 21 cultivars and found oil content in the seeds was 46.5-63.1% for the wild species and 43.6-55.5% for the cultivars. The variability in oil yield, especially in the wild species, suggests that oil content of the peanut can be increased, perhaps dramatically. Oils from seed of different subspecies of peanuts differ in their tendency to develop oxidative rancidity and undesirable odors and flavors (Worthington and

PEANUT BREEDING

45

Hammons, 1971). Virginia-type peanuts produce oil with a lower linoleic acid content and tend to have greater oil stability than Spanish or Valencia types. Hybrids from subspecific crosses also vary in chemical composition of the oil (Tai, 1972). Khan et al. (1974) examined the refractive indices of oil from hybrids and segregating populations of six peanut crosses in order to assess fatty acid compositions of the groups. The iodine values were under the control of a few additive genes and thus highly heritable. A wide range of genetic variability in iodine value in the F2 populations of these crosses indicated that selection could improve quality and stability of oil. Peanut seeds contain about 26% protein and the resulting peanut meal about twice that percentage (Woodroof, 1966). Although variation in the protein level has been demonstrated for a wide range of genotypes (Holley and Hammons, 1968; Young and Hammons, 1973), little breeding work to exploit this variation has been reported. C. PESTRESISTANCE

In many countries there is a concerted effort to develop cultivars resistant to leafspots and rust, two of the most important diseases of peanuts worldwide. Norden (1973, 1980) reviewed the progress and difficulties encountered in breeding for disease resistance in peanuts. The progress in breeding for resistance to several important pests is summarized in what follows.

I . Disease Resistance a. Rust. Peanut rust, caused by Puccinia arachidis Speg., has become established throughout Asia, Australia, and much of Africa and has occurred with increasing frequency in the United States (Hammons, 1977). Sources of resistance to rust were identified over a decade ago (Bromfield and Cevario, 1970; Bromfield, 1974; Hammons, 1977). According to Hammons (l977), the resistant sources consist of three breeding lines, which are as follows:

(a) Tarapoto (PIS 259747,34 1879,350680,38 1622,405 132) originating from Peru. (b) Israel line 136 (PIS 298115 and 315608), which was selected from the cultivar Virginia Odom (Cook, 1972). Bromfield and Bailey (1972) concluded that the resistance of PI 2981 15 was controlled by two recessive genes. Fourteen progeny from the cross of PI 298 1 15 and an unknown pollen parent identified as FESR 1-14 were released as rust-resistant germ plasm. (c) DHT 200 (PI 314817), which was collected from Peru and is being considered for release in Australia (Shorter, 1978).

46

J . C. WYNNE AND W . C. GREGORY

At ICRISAT, Subrahmanyam et al. (1980) have screened over 4000 lines of cultivated peanuts for their rust reaction in the field and screenhouse. The resistance of Tarapoto and Israel line 136 was confirmed and several additional resistant germ plasms were identified. Two land races, NC Ac 17090 and EC 76446 (292), were more resistant than previously identified resistant sources. Several species of Arachis also have been shown to be resistant to rust, developing no symptoms when tested in India by Subrahmanyam et al. (1980) (Table 11). b. Cercospora Leafspots. Cercospora leafspots, the most common disease of peanuts, are caused by Cercospora arachidicola Hori (early leafspot) and Cercosporidium personatum (Berk. and Curt.) Deighton (late leafspot). Potentially valuable sources of resistance to C. arachidicola were identified in cultivated peanuts in Georgia (Sowell et a l . , 1976). Resistance was confirmed by Foster et al. (1980), who proposed to increase the level of resistance using recurrent selection and a detached leaf technique developed by Melouk and Banks (1978). Monasterios et al. (1978) evaluated the reaction of 180 peanut accessions for resistance to leafspot caused by both C. arachidicola and C . personatum. Of the genotypes screened, 14 had very few lesions caused by C . arachidicola, 10 had very few caused by C. personatum, and 15 had few lesions from either pathogen. Nigam et al. (1980) found that several accessions resistant to rust were also resistant to late leafspot. The wild species of Arachis are also resistant to leafspot (Abdou et al., 1974; Gregory et a l . , 1973; Moss, 1980). Sharief et al. (1978) inoculated three species of Arachis and three of their F, and F2 hybrids with spore suspensions from C. arachidicola and C. personatum. They concluded that variation in disease tolerance resulted from multifactonal genetic differences in the hosts. The authors suggested that introgression of genetic factors from the two resistant Arachis species studied ( A . chacoense Krap. et Greg. nom. nud. and A . cardenasii Krap. et Greg. nom. nud.) into the genomes of A . hypogaea could increase host Table I1 Wild Species of Arachis Immune to Pucciniu armhidis Speg." ~~~

Species A . duranensis Krap. et Greg. nom. nud. A . correntina (Burk.) Krap. et Greg. nom. nud. A . cardenasii Krap. et Greg. nom. nud. A . pusilla Benth. A . sp. 9661 A . sp 10596

"After Subrahmanyam et a l . (1980).

USDA PI No.

Section

Origin

2 19823 33 1 194 262141 338448 262848 216233

Arachis Arachis Arachis Triseminalae Rhizomatosae Rhizomatosae

Argentina Argentina Bolivia Brazil Brazil Paraguay

PEANUT BREEDING

47

resistance in the cultivated peanut. Smartt et al. (1978) advocated maximizing resistance by selection in the A . chacoense x A . cardenasii population before attempting gene transfer to the cultivated species. Hexaploids from crosses of cultivated peanuts with A . chacoense, A . cardenasii, and other diploid species have been exposed in field trials to C. permnatum in India and C. arachidicola in Malawi (Moss, 1977, 1980). Lines from the hexaploids of A . hypogaea x A . cardenasii were found to be resistant to both pathogens (Moss, 1977, 1980). Stalker et al. (1979) also reported resistant selections to both pathogens in 40-chromosome derivatives from an A . hypogaea x A . cardenasii hybrid population. c . Ajlatoxin. Aflatoxin produced by Aspergillus jlavus (Link) Fries is a major problem facing the peanut industry in the United States. Mixon and Rogers (1973) and Mixon (1976) found two Valencia accessions (PI 337394 and 337409) and several breeding lines resistant to seed colonization by two aflatoxin-producing strains of A.j7avus. LaPrade (1974) found that an intact testa is required for resistance, the testa appearing to act as a mechanical barrier to the fungus. Many additional factors seem to influence susceptibility of seeds to invasion by Aspergillus (Bartz et al., 1978). Therefore breeding of resistant cultivars has proven difficult. Seeds of genotypes resistant to invasion by A.flavus under field growth contained aflatoxin concentrations equal to genotypes easily colonized when the two were stored under humid conditions (Wilson et a l . , 1977). Nevertheless, Mixon (1976, 1980) has developed a productive line resistant to colonization. d . Rosette Virus. Rosette virus is transmitted by an aphid (Aphis craccivora Koch.). Cultivars resistant to rosette were developed in Senegal and the Upper Volta (Gillier, 1978). These cultivars are being used as source of resistance to deveiop cultivars adapted to other growing areas of Africa. e . Southern Stern Rot. No presently available lines of peanuts appear to be highly resistant to stem rot caused by Sclerotium rolfsii Sacc. Garren (1964) concluded that the Virginia cultivar NC 2 was the least susceptible to this disease and lines of the fastigiate type were the most susceptible. This observation was confirmed by Muheet et al. (1975). Despite breeding efforts to develop resistant cultivars at several institutions (Lin, 1959; Beute et a l . , 1976), except for NC 2 (Gregory, 1970), no cultivars resistant to stem rot have been released. f. Cylindrocladium Black Rot. Cylindrocladium black rot (CBR) is caused by a soil-borne fungus, Cylindrocladium crotalariae (Loos) Bell & Sobers. Valencia types are most susceptible and Spanish types are most resistant to CBR (Wynne et al., 1975b; Coffelt, 1980b). NC 3033 and Va GPl have been released as resistant germ plasm (Beute et al., 1976; Coffelt, 1980a). Advanced breeding lines resistant to CBR have been developed in North Carolina using an accelerated breeding program (Wynne, 1976b). Backcrossing to an adapted cultivar to

48

J. C. WYNNE A N D W. C. GREGORY

improve quality may be required before cultivar release. Hadley et al. (1979) studied the F, , F2, and parental generations from a four-parent diallel cross (two resistant to black rot, Argentine and NC 3033, and two susceptible, NC 2 and Florigiant) under greenhouse conditions to determine the inheritance of CBR resistance. Because general combining ability was significant for both generations, indicating that CBR resistance is primarily controlled by additive genetic effects, they proposed that genetically distinct sources of resistance be combined to produce genotypes with superior CBR resistance. Such material should buffer potential pathogen adaptation to overcome resistance. g. Nematodes. Lesion nematodes. Smith et al. (1978a) evaluated six peanut introductions and two commercial cultivars (Starr and Spancross) for resistance to lesion nematodes [Pratylenchus brachyurus (Godfrey) Felip Sch. Stek] under field conditions. Three introductions were found to have some resistance. Root-knot nematodes. Minton and Hammons ( 1975) conducted greenhouse tests for resistance to root-knot nematodes, Meloidogyne arenaria (Neal) Chitwood, a major parasite of peanuts in the southeastern United States. None of 479 entries tested showed resistance to the nematode. Banks (1969) tested about 400 accessions of cultivated peanuts and 33 species collections for resistance to the northern root-knot nematode, Meloidogyne hapla Chitwood. Although differences in susceptibility among the cultivated lines were observed, Banks (1969) concluded that resistance was not great enough to be useful in developing resistant cultivars. One wild species from section Rhizomatosae, PI 262286, had a moderately high level of resistance, but this species has not yet been successfully hybridized with cultivated peanuts. h. Other Diseases. Peanut breeders are also attempting to develop cultivars resistant to Verticillium wilt (Verticillium sp.), Sclerotina blight [Sclerotinia sclerotiorum (Lib.) DeBary], Webb blotch (Phoma arachidicola), collar rot (Diplodia gossypina), and pod breakdown (Pythium sp.). Resistant lines have been reported for Verticillium wilt (Smith, 1961; Frank and Krikun, 1969; Khan et al., 1972, 1974), Sclerotina blight (Porter et al., 1975), collar rot (Porter and Hammons, 1975), and Webb blotch (Smith et al., 1978b). 2 . Insect and Mite Resistance a . Southern Corn Rootworm. The southern corn rootworm (Diabrotica undecimpunctata howardi Barber) is the most destructive insect pest of peanuts in the Virginia-Carolina production area. The cultivar NC 6 has a high enough level of resistance to rootworm to eliminate the need for insecticide application in most soils of the region (Campbell et al., 1977). NC 6 is also moderately resistant to the potato leafhopper (Empoascafabae Harr.), although sources with a higher level of resistance have been isolated (Campbell et al., 1976).

PEANUT BREEDING

49

b . Lesser Cornstalk Borer. The lesser cornstalk borer, Elasmopalpus lignosellus (Zeller), is a major insect pest on peanuts in the southeastern and southwestern United States. Leuck and Harvey (1968) screened germ plasm for lesser cornstalk borer resistance but their results were inconclusive. Smith et al. (1980) screened seedlings of 490 genotypes for their reaction to lesser cornstalk borer larvae under greenhouse conditions. Eighty entries gave a significantly better reaction than the check cultivar, Starr. c. Mites. Johnson et al. (1977) have found resistance to the two-spotted mite, Tetranychus urricae Koch, a major pest of peanuts in North Carolina during dry seasons in several species of Arachis. No evidence of highly resistant commercial cultivars was reported. Several of the wild species of Arachis were found to have resistance approaching immunity. d . Thrips. Young et al. (1972) screened 872 peanut entries for resistance to tobacco thrips, Frankliniella fusca (Hinds). Only moderate resistance was found. Immature and adult thrip populations in the buds and flowers of developing peanut plants were studied by Tappan and Gorbet (1979) to evaluate thrip damage. They concluded that thrip control on peanuts is not economical because injurious populations occur too early in the growing season to influence yields. e . Fall Armyworms. Leuck and Hammons (1974) observed that the nutrition of the host plant can influence the expression of resistance to armyworm [Sportoptera frugiperda (J. E. Smith)]. They conducted a study using three growth media and various fertilizer combinations and found that armyworm preferred foliage of peanuts grown in Tifton loamy sand field soil over that of peanuts grown in washed sand or vermiculite. The data also showed that fall armyworm resistance varied with the types of fertilizer and with vigor and appearance of the single cultivar utilized in the study. Resistance to fall armyworm has been reported (Duke, 1980).

IV. BREEDING AND QUANTITATIVE GENETICS A . VARIABILITY, HERITABILITY, A N D CORRELATION

Bernard (1960) estimated genetic and environmental variability for several traits, including seed yield, number of pods, and weight per seed in the F, through F4 generations of 4 crosses between 8 diverse cultivars and 15 crosses between 6 F4 selections from the first group of crosses. Seed size (weight per seed) had a higher estimate of heritability than did seed yield. Several traits were correlated with yield, but a selection index including yield and any or all of the

50

J . C. WYNNE A N D W . C. GREGORY

remaining nine characters was not superior to selection for yield alone (Table 111).

Syakudo and Kawabata (1965) found that genotypic correlations among 15 characters in the 6 possible crosses of Virginia-, Valencia-, and Spanish-type peanuts were higher than phenotypic correlations. Estimates of broad-sense heritability were low for all traits of economic importance. Lin (1966) also found that estimates of heritability for number of pods and seed yield were relatively low. He reported that the major portion of genetic variance among F2 and F3 progenies of a Spanish X Virginia cross was due to dominance effects for number of pods and yield. Estimates of broad-sense heritability for an Fs bulk population were higher for yield and number of pods in high planting densities than in low densities (Lin et al., 1971). Martin (1967) obtained narrow-sense heritability estimates of approximately 70% for oil content, shelling outturn, and yield using F, and backcross progenies between two cultivars. He reported that cultivar differences were due to two pairs of alleles for oil content, one for shelling outtum and five for seed weight. Oil content was not correlated with yield. Coffelt and Hammons ( 1974) reported correlation coefficients and heritability estimates for nine components of yield in an F2 population between Argentine (Spanish type) and Early Runner (Virginia type). The characters measured were number of pods and seeds per plant, pod and seed weight per plant, 100-seed weight, length and breadth of 10 pods, number of seeds per pod, and pod length-to-breadth ratio. They found highly significant positive correlations between number of pods and pod weight, number of seeds and seed weight, pod weight and number of seeds, pod and seed weights, and number of seeds and seed weight. Selection for increases in any of the four characters, number of pods, pod weight, number of seeds, or seed weight, should result in a corresponding increase in the remaining traits. Pod breadth was also significantly correlated with 100-seed weight. Other significant correlations were obtained, but they were small in magnitude. Broad-sense estimates of heritability for 100-seed weight, pod length, pod breadth, and the pod length-to-breadth ratio were high (71-90%). Low heritability estimates were observed for number of pods, pod weight, number of seeds, seed weight, and seeds per pod. Tai and Young (1975) studied the inheritance of protein content and oil content using six cultivars and their F2 populations. They concluded that both protein content and oil content were quantitatively inherited. Correlations between protein content and oil content were negative and varied from nonsignificant to highly significant in the various populations. Holly and Hamrnons (1968) had previously reported a tendency for a reciprocal relationship between oil content and protein content. However, enough exceptions were found for the 26 cultivars tested to invalidate an absolutely negative relationship between oil and protein.

PEANUT BREEDING

51

The inheritance of amino acid and fatty acid composition of three crosses in Fz generation and their parents were also reported by Tai and Young (1975). These traits were also found to be controlled by genes acting in a quantitative manner. Some transgressive segregants were found for some of the amino and fatty acids. Correlations among the 18 amino acids and 8 fatty acids were inconsistent over parental and F2 cross populations. In another study Tai and Young (1977) used nine F2 families from crosses among six peanut cultivars and breeding lines to investigate the inheritance of dry matter accumulation and free arginine (as a measure of maturity). Dry matter accumulation was found to be a quantitative trait, whereas the free arginine level was found to be controlled by two major genes with partial dominance for low arginine. Broad-sense heritabilities were 38-78% for dry matter and 60-93% for arginine level. Mohammed et al. (1978) estimated heritability, phenotypic correlations, and genotypic correlations for yield, fruit size, and maturity using the F, and F, generations of two crosses between one Virginia and two Spanish lines. Broadsense heritability estimates based on intraplot variance for yield ranged from 42 to 82% for four year-location environments. Broad-sense heritability estimates were also high for fruit length, ranging from 79 to 92%. Estimates of heritability for several maturity traits were lower and less consistent over environments. Estimates of heritability computed by parent-offspring regression were much lower for all traits than those estimated by the variance partitioning method. Parent-offspring regression heritability for the two crosses for yield of pods was 21 and 16%, for weight of seeds 10 and 6%, for fruit size 42 and 50%, for fruit length 18 and 27%, for weight per seed 41 and 51%, and for a fruit maturity index 20 and 35%. The discrepancy between the variance and regression estimates of heritability for the F, populations suggests that broad-sense heritabilities based on intraplot variances are poor predictors of genetic advance from selection. The variance estimates of heritability were biased upward, probably from inflated genotypic estimates resulting from competition among plants within plots. The regression estimates of heritability were biased less by nonadditive variance and genotype X environment interaction and thus seem to be more useful as predictors of response to selection. Gibori er al. (1978) used a 9 x 9 diallel cross involving widely divergent cultivars as parents to estimate heritability and correlations for pod size, pod yield per plant, days to first flower, and shoot weight measured in the F2generation. Their estimates of heritability were calculated using the methods of Hayman (1954, 1958) and Jinks (1954, 1956). These authors suggest that the high heritability estimate obtained for pod yield per plant (79%) indicates that visual selection of promising plants in large F, populations followed by careful progeny testing can be used to increase productivity. Pod yield per plant was not highly correlated with the other three traits, suggesting that selection for yield cannot be

Table HI Estimates of Heritability in Peanuts from Populations Derived after Hybridization Yield bl

w

Parental material

Estimated

Reference population

Pod

F2 families in F3

Diverse set of 5 parents Intervarietal crosses among 3 parents Virginia x Spanish

H

Virginia X Virginia Virginia X Virginia Spanish X Virginia

Ill

F2 population

h’

F4, Fs families

Seed 0.54-0.68 0.83-0.85 0.66

F3 families in F, F2 individuals

Weight per seed

Oil

Protein

Other traits

Investigators Bernard (1960)

0.73-0.87 0.92-0.97

Syakudo and Kawabata (1965)

F5 bulks

Lin et al.

1 x 105 plantdha 1.5 x 105 3.0 x 105

0.64, 0.56 0.53, 0.49 0.38, 0.30

(1971)

0.66-0.76

0.70

O.7Oa

Martin (1967)

0.69-0.94b

Gupton and Emely ( 1970)

H

0.43

0.43

0.90

Coffelt and Hammons ( 1974)

Intervarietal sample of 6

Virginia X Spanish

w

"Leafspot resistance.

H

F2

0.72

0.47 0.38-0.78" 0.60-0.93"

F2 and F3 families

"Meat. bMaturity (oil index). "Total dry matter. dAM.

F2 individuals

H 'h

Intervarietal 9 parent diallel Virginia x Virginia Virginia x Virginia Virginia X Spanish VI

H

h'

Variance comp. Po regression F2 families

0.42-0.82 0.16-0.21 0.79

h2

F5 and F6 families

0.54

H

F3 families

0.28, 0.41

0.57, 0.83

0.47, 0.72

0.52, 0.68

0.77, 0.82'

Tai and Young ( 1974) Tai and Young (1977) Mohammed et a1 (1978) Gibori et al. ( 1978) Wynne and Rawlings ( 1978) Sandhu and Khehra ( 1977)

54

J. C . WYNNE A N D W. C. GREGORY

accomplished by indirect selection. They found a positive but low genetic correlation between fruit size and yield, supporting the practice of selection for both large pods and high yields. Layrisse et al. (1980) estimated correlation coefficients based upon F2 cross means and Spearman rank correlations based upon general combining ability effects for nine traits from the F2 generation of a diallel cross involving ten diverse parents. Correlation coefficients based on cross means are phenotypic; those based on general combining ability effects are phenotypic correlations that approximate genetic correlations. Fruit yield and seed yield were significantly correlated with oil content and protein content. Oil content and protein content were positively correlated but only the phenotypic correlation was significant. Wynne and Rawlings (1978) estimated heritability for yield and several fruit traits for the F5 and F6 generations of a cross between two Virginia cultivars. Narrow-sense estimates of heritability over reciprocal crosses and environments ranged from 54% for yield per plot to 89% for fruit length. Progress from selection in late generation should be expected from these crosses. Sandhu and Khehra (1977) determined heritability and predicted genetic advance for the F3 progenies of two peanut crosses for resistance to leafspot, pod yield, 100-kernel weight, oil content, and protein content. Broad-sense estimates of heritability were high for all traits except yield in both crosses. However, the estimated advance from selection was only high for resistance to leafspot. Hadley et al. (1979) estimated heritability for CBR resistance to range from 48 to 65% depending upon the method of calculation. Their estimates were obtained in the greenhouse for the F, and F2 generations of a four-parent diallel. B. TYPEOF GENEACTION

Although methods for characterizing genetic variability in self-fertilizing species are available (Hanson and Weber, 1961; Cockerham, 1963; Stuber, 1970), little information has been obtained on the types of gene action and their relative magnitude for important traits in peanuts. Brim (1973) has emphasized the importance of developing more efficient breeding procedures through a better understanding of the type of gene action governing the inheritance of quantitative traits .

I . Heterosis Heterosis, followed by inbreeding depression, usually indicates that nonadditive gene action is important. Marked heterosis for vegetative traits and pod yield were obtained for several combinations when Higgins (1941) crossed 16 cultivars in all combinations. Individual plant yields were highest for Spanish x Virginia

PEANUT BREEDING

55

crosses. Gregory el al. (1980), in a diallel cross of 10 diverse peanut lines made in 1944, found hybrid vigor for F, hybrids between subspecies. Most F2 hybrid means were equal to midparental values, although some F2 means were exceptionally high or low. Syakudo and Kawabata (1963) found appreciable heterosis for shoot weight in Virginia X Spanish and Valencia X Virginia F, hybrids. Heterosis was not present in crosses between cultivars within each botanical variety nor in Spanish x Valencia crosses. Lin (1966) found significant heterosis for length of main stem and branches for F2 plants grown in Taiwan from the cross of a Spanish type with Florispan Runner (Virginia type). The superiority of the F, hybrids over their better parents for yield, as well as for the number of branches and leaflet length, was shown by Hassan and Srivastava (1966) using crosses among three cultivars differing in maturity and growth habit. Parker et a f . (1970) noted that F, crosses of Valencia X Virginia gave greater heterosis than did crosses of Virginia X Spanish or Valencia X Spanish for several seedling characters measured in a controlled environment. Wynne et al. (1970), using the same parents as Parker et al., reported that F, hybrids from Virginia x Valencia parents gave greater heterosis than other crosses for vegetative plant characters. Crosses of Valencia x Spanish gave greatest heterosis for yield and fruit characters. The highest yielding population, however, resulted from a cross of Virginia x Spanish parents. Hammons (1973a) reported heterotic responses for fruit yield for F, hybrids resulting from crosses between the subspecific peanut groups. Five cultivars representing Virginia and Spanish types and all their possible hybrid combinations were evaluated in Senegal by Garet (1976). Heterosis was found for pod and seed size, pod and seed number per plant, and shelling outturn. In all cases where heterosis was observed, the cross was between Virginia and Spanish parents. Layrisse et al. (1980) found that hybrid vigor for fruit yield, seed yield, and 100-seed weight persisted in F2progenies of a diallel cross among ten lines, two from each of five centers of genetic diversity in South America. The parents of the crosses displaying significant heterosis most often came from different centers. Arunachalam et al. (1980) classified parents of two diallel crosses as high or low based on their general combining ability as computed for 15 characters. High x low crosses produced greater heterosis than high x high or low x low crosses. Isleib and Wynne (1980) crossed 28 diverse peanut lines with an elite Virginia breeding line and grew the F, and F2 generations at two North Carolina locations. Included in the parental sample were genotypes from five South American centers of diversity, Africa, China, and A . monticola. Positive heterosis was observed for pod yield, number, and size. Fastigiate parents generally produced greater heterotic responses than parents from ssp. hypogaea. Maximum responses were noted for fastigiate parents from the Peruvian center of diversity. The evidence is convincing that heterosis in peanuts, like heterosis in other crop species such as wheat (Fonesca and Patterson, 1968; Sun et al., 1972;

56

J . C. WYNNE AND W . C. GREGORY

Widner and Lebsock, 1973), alfalfa (Sriwatanapongse and Wilsie, 1968), cotton (Marani, 1963, 1968), corn (Moll et a l . , 1962), and tobacco (Matzinger and Wernsman, 1968), is related to genetic diversity. Heterosis in peanuts is most often observed in crosses between the subspecific groups. These results suggest that gene action differs in crosses made within and crosses made between botanical varieties. Additive genetic variance appears to be of primary importance in crosses made between parents chosen from a single botanical variety, but both additive and nonadditive genetic variance may be significant in crosses made between parents from different botanical varieties. 2 . Combining Ability Mating designs such as the diallel have been used in partitioning genetic variability into portions due to general combining ability (GCA) and specific combining ability (SCA). GCA indicates additive genetic effects, while SCA indicates nonadditive genetic effects. Gregory et al. (1980) crossed ten of the most diverse peanut lines in his collection in 1944 and estimated combining ability in the F, generation by using vegetative cuttings. He found GCA to be highly significant and several times greater in magnitude than SCA for yield and several yield components. In a series of experiments Parker et al. (1970) and Wynne et al. (1970, 1975a) reported the results from a series of combining ability analyses using six diverse parents. Parker et al. (1970) estimated combining ability for 17 characters of F, hybrid seedlings in a diallel set of crosses of 6 lines, 2 each from 3 centers of diversity in South America. In the controlled environment of a phytotron estimates of GCA were found to be higher than SCA. Wynne et al. (1970), however, reported combining ability estimates for SCA higher than those for GCA for yield and several yield components for the same F, hybrids in the field. However, when a more appropriate analysis of the data was made (Baker, 1978), estimates of GCA were found to be significant for all 17 characters. Furthermore, GCA estimates were larger than estimates for SCA for all except one character. Estimates of combining ability were also obtained for the F2 generation of these 15 crosses in both spaced and drill-planted tests (Wynne et al., 1975a). Estimates of both GCA and SCA were highly significant for yield, fruit length, seeds per kilogram, percentage extra-large kernels, and percentage sound mature kernels. GCA estimates were larger than SCA estimates for all traits except percentage sound mature kernels in the drilled tests. In the space-planted test, GCA and SCA were significant for all traits except for SCA for weight of sound mature kernels. GCA estimates were likewise of greater magnitude than SCA for all traits. Garet (1976) evaluated the F, hybrid progeny from a complete diallel of five cultivars chosen to represent a wide range of variation in Senegal. Estimates of

PEANUT BREEDING

57

GCA were significant for pod and seed yield per plant, the number of pods and seeds per plant, 100-pod weight, 100-seed weight, oil content, and shelling outturn. SCA and reciprocal effects were also significant for all traits except oil content. Since GCA effects were larger than SCA estimates for all traits except shelling outturn, Caret (1976) concluded that the major part of the total genetic variability was additive for all characters except shelling outturn. A graphic analysis of the data for pod yield per plant, 100-pod weight, and shelling outturn using the methods of Hayman (1954) confirmed the conclusions reached through the analysis of variance for combining ability. Pod yield per plant, days to first flower, pod size, and plant weight were studied by Gibori et al. (1978) by analyzing F, data from a 9 x 9 diallel cross utilizing cultivars of Virginia, Valencia, and Spanish types. They reported bidirectional dominance for pod yield per plant and days to first flower, whereas the alleles giving small pods were dominant and the alleles for large plants showed dominance and overdominance. Estimates of genetic components of variance indicated that additive genetic effects were significant for all traits and accounted for more of the variation than nonadditive effects for all traits except plant weight. Layrisse er al. (1980) used ten peanut lines, two from each of five centers of diversity in South America, and the Fz generation of all possible crosses among them to estimate combining ability for yield, fruit and seed traits, protein content, and oil content. Both GCA and SCA were significant for all traits except for SCA for protein percentage. The component of variation for GCA was larger than that for SCA for all traits. Hadley er al. (1979) (using greenhouse-grown plants of the F, and F, generations from a four-parent diallel) determined combining ability for CBR resistance. GCA was significant for both generations, suggesting that resistance was primarily due to additive genetic effects. Kornegay er al. (1980), using fieldgrown F I and F2generations from a six-parent diallel, determined the inheritance of resistance to early and late leafspot in Virginia-type peanuts. GCA was significant in both generations, indicating that variation in resistance to both fungi depended upon additive genetic effects. Crompton et al. (1979) used a complete diallel among four Virginia and two Spanish lines to estimate combining ability for seed calcium concentration and total adenosine phosphates. GCA, SCA, maternal effects, and reciprocal effects were significant for calcium concentration, while only SCA was significant for total adenylates. Reciprocal and SCA components of variation were larger for calcium concentration than the GCA component of variation, although GCA was sufficiently large to also be important. Isleib et al. (1980) crossed 10 South American cultivars in diallel to analyze the gene action for traits indicative of nitrogen fixation. SCA was significant and accounted for more variability than GCA for nodule number per plant, nodule mass, specific nitrogenase activity,

58

J . C. WYNNE A N D W. C. GREGORY

shoot weight, and total nitrogen for greenhouse-grown F, plants, suggesting that nonadditive gene action is important for these traits. 3 . Variance Studies

Mohammed et al. (1978) estimated additive and nonadditive genetic effects for crosses between one Virginia and two Spanish lines using a generation means analysis. Estimates of additive effects were significant for yield, maturity, and fruit size traits. Nonadditive genetic effects were also significant for yield and fruit size. Wynne and Rawlings (1978), using maximum-likelihood procedures from a nested mating design, estimated genetic variances for yield and several fruit traits for the F5 and F6 generations of an intercultivar cross. Estimates of additive and additive by environmental variances were significant for yield and fruit traits. Estimates of additive x additive epistatic variance were essentially zero for all traits; however, estimates of additive x additive x environmental variances were larger than their associated standard deviations for all traits except yield. The available evidence suggests that the principal component of genotypic variance for traits of economic importance in peanuts is additive. The importance of nonadditive effects is not known. The significant heterosis observed in some peanut crosses suggests that dominance deviations occur but hybrids cannot presently be utilized in peanut improvement. In the self-pollinated peanut, epistatic variance of the additive x additive type is potentially useful to breeders because it can be fixed in homozygous genotypes. Hammons (1973a) suggested that many important traits may be governed by epistatic variance. Significant estimates of epistatic variance for quantitative traits would not be surprising since the peanut is an allotetraploid and several qualitative traits have been found to be controlled by duplicate genes (Hammons, 1971, 1973b). A generation means analysis was used by Sandhu and Khehra (1 976) to determine the importance of epistatic variance for two crosses at two locations in India. Nonadditive genetic effects were more important than additive effects for pod yield, number of mature pods, and 100-kernel weight in one cross and for pod yield in the second cross at a single location. These authors concluded that epistasis cannot be ignored in peanut crosses. Isleib el al. (1978) tested for the presence of epistatic effects using progeny from a six-parent half-diallel of diverse peanut cultivars. Significant variability attributable to specific combining ability persisted over generations for yield and seed characters. Epistasis was indicated since dominance deviations could not account for the variance due to SCA in the F5 generation. Although their estimates were obviously biased by linkage disequilibrium, the authors reported that epistatic variance was greater than dominance variance for all traits. This study suggests that considerable epistatic variance may account in part for the heterosis in crosses derived from

59

PEANUT BREEDING

diverse parents. Cahaner et al. (1979) used a diallel in an attempt to detect genic interactions. Six traits, measured in the F2 generation of crosses made among four parents, were analyzed. A duplicate genic type of interaction and complementary interactions were detected using the methods suggested by Mather (1 967). They concluded that duplicate genic interactions were involved in the inheritance of pod yield and mean pod weight. The number of pods per plant, dry weight of plant, and the ratio of reproductive to vegetative branches were found to be controlled by additive-dominant genes. Genetic variance, heterosis, and epistatic variance estimates need to be integrated into the mainstream of selection procedures for the major objectives of peanut breeding programs. Information on the type and magnitude of genetic variance for important traits of both adapted and exotic intersubspecific crosses needs to be part of the ongoing process of cultivar development. C. EFFECTS OF SELECTION

In the absence of selection experiments to report in peanuts, we can only say that while the dramatic improvements in cultivar performance in peanuts attest to the breeders' skills and luck, they do not indicate firm courses of action for the future. The best we can report here are data comparing new cultivars with the old land race standards with which they have been tested (Table IV). D. GENOTYPE X ENVIRONMENT INTERACTIONS

Genotype x environment interactions influence the progress that a breeder makes in his breeding program. Well-buffered cultivars, those with small Table IV Comparison of Yield of Land Races of Peanuts with Improved Cultivars

Improved cultivar

Year of release

NC 2 NC 5 NC 6 Florigiant Florigiant

1952 1964 1976 1960 I960

Mean yield of land race Land race

(%)

NC 4" NC 4 NC 4 Va 56R" 35 Bolivian introductions 53 Guarani introductions 22 Peruvian introductions

103 112 119 120 137 134 145

Years tested

"Single plant selections from land races grown in North Carolina and Virginia.

6 6 4

8 3 3 3

No. of tests 20 19 13 38 3 3 3

60

1. C. WYNNE AND W. C. GREGORY

genotype x environment interactions, are usually desired. Conversely, if cultivars are to be selected for specific environments, then cultivar development may be easier when large genotype x environment interactions are present. Several investigators have reported the presence and magnitude of genotype X environment interactions in peanuts. Chen and Wan (1968) measured the genotype x environment interaction using 13 peanut cultivars grown in Taiwan at 10 locations for 2 years. Both cultivar X year and cultivar x location interactions were small for yield; the cultivar x year X location interaction was highly significant. Ojomo and Adelana (1970) determined cultivar X environment interactions for 16 cultivars grown at 3 locations for 3 years in western Nigeria and found cultivar x location and cultivar x year X location interactions to be significant. The cultivars consisted of introduced lines and a few local standards. In Punjab, India, Sangha and Jaswal (1975), using 12 Virginia peanut cultivars, found significant cultivar x location and cultivar x year x location interactions for pod yield. Tai and Hammons (1978) estimated the magnitude of cultivar x environment interaction for pod yield, percentage sound mature kernels, percentage extralarge kernels, percentage fancy-sized pods, 100-seed weight, and other fruit traits for tests conducted under irrigated and nonirrigated management in Georgia at two locations for two years. The 19 cultivars used represented both early- and late-maturing groups. Significant cultivar x location x year interaction was found for most traits. The cultivar component of variance was larger than the first- and second-order interactions. Wynne and Isleib (1978) estimated cultivar x environment interactions for yield and several fruit traits for two groups of Virginia cultivars. A large cultivar X location x year interaction was observed for yield in both North Carolina studies. Both cultivar x location and cultivar x year interactions were small when compared to variation among cultivars. Yield, percentage sound mature kernels, and percentage extra-large kernels were determined for two years at two locations for nine crosses represented by eight lines per cross in the F4 and F5 generations by Wynne and Coffelt (1980) in North Carolina and Virginia. Cross populations and lines within crosses were significantly different for all traits. Cross populations interacted with the yearlocation environments for all traits, whereas lines within crosses interacted with the environment for all traits except yield. Although genotype x environment interactions vary with the material tested and the site chosen for testing, genotype X environment interactions in peanuts appear to be similar to those in several other autogamous species. Matzinger (1963) concluded that second-order interactions were prevalent in cotton, soybeans, and tobacco. In general, despite the results of Tai and Hammons, the second-order interaction also tends to be most important for peanuts. Thus the

PEANUT BREEDING

61

yield of a peanut cultivar in each individual experiment is unique and the environmental conditions differentiating the tests cannot be grouped according to years or locations. This is not surprising considering the indeterminate nature of the peanut. E. EVALUATION OF STABILITY

Because of limited resources peanut breeders have generally been interested in developing cultivars that are stable, that is, show a minimum of interaction over environments. Regression techniques have been used to characterize responses of genotypes under varying environmental conditions. Although many of the regression analyses used to measure phenotypic stability do not meet rigorous statistical requirements (Moll and Stuber, 1974), they are useful indicators of stability. Joshi et a / . (1972) measured the stability of five bunch genotypes and a local standard at seven environments in Gujarat, India, using the analysis suggested by Eberhart and Russell (1966). Cultivars showed average stability in all environments for yield. The local standard was low yielding in both good and poor environments, whereas one genotype, released for cultivation as “Junagadh, performed consistently well in both poor and good environments. Singh et al. (1975) evaluated eight cultivars for yield and stability at four locations in India during a single growing season. Their data were also analyzed using Eberhart and Russell’s (1966) methods. A significant genotype x environment interaction was found. Cultivar “M 13” (a subline of NC 13) had both high yields and average stability. Wynne and Sullivan (1978) found that eight Virginia cultivars differed in stability over environments for the percentage of seedlings that emerged. They tested at five locations in North Carolina during a 3-year period. Two cultivars, Florigiant and NC-Fla 14, produced a high percentage of emerged seedlings and gave greater stability over environments than the remaining six cultivars. The genotype x environment interaction for pod yield and days to maturity was found to be significant by Yadava and Kumar (1978) for 15 bunch genotypes grown in 4 environments at Hissar, India. The linear component of the genotype x environment interaction was significant for both traits, while the deviations from regression were also significant for days to maturity. The individual cultivars differed in stability for both traits as measured by both the regression coefficients. One cultivar was consistently early-maturing and high-yielding in all environments. Yadava and Kumar (1978) also used 17 genotypes grown in 4 environments to estimate genotype x environment interactions and stability parameter for 100-kernel weight, oil content, and shelling percentage. The linear and nonlinear portions of the genotype X environment interactions were signifi-

”

62

J . C. WYNNE AND W. C. GREGORY

cant for all three traits. One cultivar had consistently high 100-kernel weight and oil content in all environments. Another cultivar had high shelling percentage and was stable over all environments. The stability parameters for the different traits were apparently governed by independent genetic systems. In order to achieve stability of yield over a wide geographical area and over seasons, Norden (1 980) released cultivars that are early generation composites of 4-10 sister lines selected in the F4 through FBgenerations. Two such cultivars, Florigiant and Florunner, have been grown in the southeastern United States with outstanding results. The relative yield advantage and stability of this type of multiline was compared to its pure-line components in North Carolina and Virginia using 2 multilines composed of 4 sibling homozygous lines grown in 16 environments (Schilling et al., 1980). An analysis of stability was conducted to obtain a measure of relative stability among lines and a measure of the adaptation of each line to a range of environments. The two multilines did not yield significantly more than the better pure lines or the average of pure-line components. The stability for seed yield of the multilines was not different from the stability of the pure lines. The regression coefficients indicated that the mixtures were adapted to all environments, whereas the pure lines were different in their adaptation to the environment. These stability studies, although limited in number and scope, suggest that the adaptation and stability of a peanut line, both traits being under genetic control but acknowledged to be difficult to determine (Simmonds, 1979), should be evaluated and considered before a line is released for production. F. IMPLICATIONSFOR BREEDING PROCEDURES

With the evidence that additive genetic effects are the primary components of genotypic variance for traits of economic importance in peanuts, selection should lead to pure-line cultivars. Significant estimates of hybrid vigor in intersubspecific crosses notwithstanding the difficulty of obtaining hybrid seeds makes it unfeasible to develop hybrid varieties with favorable heterozygous gene combinations. Although evidence from soybeans indicates that in most cases homozygous lines surpassing the F, hybrid can be obtained (Brim, 1973), the amphidiploid character of peanuts might permit a breeding procedure that exploits both hybrid vigor and pure-line selection (see Fig. 3). Even if additive X additive epistatic variance is shown to be important, breeders will still have as their primary goal the production of pure-line cultivars. Breeding procedures such as the pedigree, modified pedigree, bulk, and backcrossing methods should continue to predominate in cultivar development.

PEANUT BREEDING

63

V. BREEDING METHODS Few novel breeding and selection methods for improving the peanut and other self-pollinated crops have been developed in recent years. The selection methods applied in peanut breeding are the same as those used in improving other autogamous species. These methods include mass selection, pedigree, bulk population, backcrossing, and other modifications of pure-line breeding methods, traditional procedures that have proved effective in the development of our best modern cultivars. A. INTRODUCTION A N D PURE-LINE SELECTION

Breeding programs at their inception can often fulfil their objectives through introduction and mass selection from, or direct use of, cultivars from other countries (Hildebrand and Smartt, 1980; Norden, 1980). Eventually, however, the breeder is confronted with the problem of combining the desirable traits of existing cultivars if advances are desired. Shorter (1978) found that although some introductions were higher-yielding than the local cultivars grown in Australia, these introductions were not acceptable to the peanut processing industry. This is generally the case in the presence of specialized markets and local pests. It is under these conditions that cultivar improvement through hybridization and selection provides its greatest opportunity (Norden, 1973). B . HYBRIDIZATION

I . Pure-Line Methods

Hybridization followed by selection among and within the offspring is currently the most common method used for improving peanuts. Norden (1973, 1980) suggested that parents be chosen that are adapted to and produce high yields within a given environment and that at least one parent of the cross be capable of reasonably high yields in the area where the new cultivar is to be grown. The pedigree breeding method with or without early testing or some modification is most often used by peanut breeders following initial hybridization. Cultivars developed by the pedigree method in Florida have been released as multiline strains (Norden, 1973). Each cultivar is a composite of 4-10 sister lines selected in the F4 to F8 generations. Each component line of a particular cultivar is so similar in phenotype that the individual lines are virtually indistinguishable.

64

J. C. WYNNE AND W. C. GREGORY

The compositing of sister lines is expected to produce wider adaptability and stability than would result from any of the pure lines taken individually (Norden, 1973; Schilling et al., 1980). The individual lines are maintained separately by the breeder for future compositing. A number of peanut breeders utilize a bulk breeding method. The early generations are usually grown in bulk with single plant selections being made in the F6 or later generation. One of the major limiting factors in making progress in peanut breeding is the time required for cultivar development. After hybridization, it usually takes 12-15 years to develop a new cultivar. A modified pedigree method of selection using single-seed descent, proposed by Goulden (1941) and described by Brim (1966), has been applied to peanuts by Emery (Norden, 1973) and Wynne (1976b) in North Carolina. Using this method, each F2plant in the population is advanced by a single seed into late generation, where selection is practiced among essentially homozygous lines. The main advantages of this procedure are that the genetic variance for characters with low heritability is maintained and more than one generation can be obtained per year. This method has been used successfully at North Carolina to transfer southern corn rootworm resistance to lines with desirable agronomic qualities and to develop a high-yielding cultivar with early maturity. Details of the procedure are given by Wynne (1976b) and Wynne and Isleib (1980). This method combined with the use of a greenhouse, a phytotron unit, and a winter nursery in a tropical environment has been used to develop CBR-resistant peanut lines. Only 24 months were needed from the time of the initial cross to the evaluation of F5 progeny rows for pod characteristics and disease resistance. Early-generation testing for yield and other traits during the inbreeding phase is an alternative to rapid generation advance. Coffelt and Hammons (1974) used early-generation yield trials to test a large number of F2 lines from a single intersubspecific cross. They concluded that acceptable breeding lines can be selected using early-generation trials. Wynne (1976a) conducted a similar study but concluded that early-generation testing of crosses between lines from different botanical varieties was effective for fruit and seed size but was of limited value when selecting for yield. The first “hybrid” variety of peanuts produced in North Carolina, NC 2, was released from an early-generation testing program after only 5 years of yield trials of an F, progeny bulked in the F3 generation. Backcrossing is now being used in several peanut breeding programs (Krapovickas et al., 1968; Gibbons and Mercer, 1972; Norden, 1973). The procedure is mostly used for transferring desired traits from a donor source to an existing cultivar. With an increased emphasis on breeding for disease resistance, the backcrossing procedure may come to be used more frequently than in the past.

PEANUT BREEDING

65

2 . Population Improvement Traditional pure-line breeding methods impose severe restrictions on the amount of recombination among linked genes during the rapid approach to homozygosity. Branch (1 980), Norden (1 980), and Wynne and Isleib (1 980) report the use of multiple-crossing systems, such as the convergent cross, to create genotypic variability before selection. Hanson et al. (1967) suggested that the use of recurrent selection for self-pollinators should overcome the restrictions on progress imposed by selecting among homozygous lines generated from biparental crosses. Selection procedures that maximize recombination have not been used in peanuts because of the difficulties in making sufficient crosses to initiate the recombinational portion of each cycle and the length of time involved in each cycle (Norden, 1973). A modified recurrent selection scheme proposed by Compton (1968) and applied to peanuts by Wynne (1976b) minimizes the number of pollinations required per cycle while providing a large number of crosses between several genotypes. A winter nursery decreases the time required for each cycle to only 2 years even though the crosses are tested in the F4 generation. After only two cycles of selection, several crosses exceeded the yield of the check cultivar (Wynne and Isleib, 1980). A breeding procedure for self-pollinators called ‘‘a diallel selective mating system” (Jensen, 1970) is also being applied to peanuts (Wynne and Isleib, 1980). The method allows the breeder to use a selected group of parents and provides opportunity for multiparental gene recombination. As with other population improvement procedures, it is difficult to obtain the mass-crossing required to initiate each cycle.

VI. INTERSPECIFIC HYBRIDIZATION The utilization of the wild species in improving the cultivated crop is hampered by cross-incompatibility and sterility (Gregory and Gregory, 1979). Gregory and Gregory (1979) obtained 22 interspecific hybrids using cultivated peanuts as one of the parents. All of the hybrids were limited to section Arachis. It appears that the cultigen may be totally isolated from all species except those in section Arachis (Hull and Carver, 1938; Smartt and Gregory, 1967; Pompeu, 1977; Gregory and Gregory, 1979). However, because the diploid species of section Arachis will cross with species from sections Erectoides and Rhizomatosae, the genetic resources of more distantly related sections may eventually become available to peanut breeders through species bridging (Fig. 2 ) . With the introduction of germ plasm from the wild species of the section

66

J. C. WYNNE AND W. C. GREGORY

FIG.2. Diagram of the crossing relationships in the genus Arachis. The symbols designate the sections and series of the genus as follows: A,, A,, AS, series of section Arachis; E l , &, &, series of section Erecroides; R,, &. series of section Rhizomarosae; C , section Caulorhizae; EX, section Extranervosae; T, section Triserninalae; and AM, section Ambinervosae. The arrows show the intraand intersectional hybridizations that have succeeded. The direction of the arrows indicates the direction of the cross, female X male (from Gregory and Gregory, 1979).

Arachis into breeding programs for peanuts, the simultaneous exploitation of additive genetic variance and hybrid vigor might be possible (Fig. 3). A similar scheme for another amphidiploid autogamous species, Nicotiana tabacum, was generated by Wernsman and Matzinger (1966). Smartt et al. ( 1 978) have suggested that A . hypogaea arose from an amphidiploid derived from a completely sterile hybrid similar to A . cardenasii (series Perennes) x A . batizocoi (series Annuae). If true, then adding genes from wild Perennes or Annuae in duplicate to each of the genomes of the amphidiploid A . hypogaea would be most difficult. In the absence of complete “dominance” for resistance across genomes, achievement of high levels of expression of the desired gene would be impossible. In case of a complete “recessive,” the whole effort would be worthless. However, Mroginski and Fernandez (1980) have recently succeeded in culturing plantlets from diploid cells of anthers of two different species of Arachis,

67

PEANUT BREEDING

both from section Erectoides. We speculate that the day of haploid plantlets from anther cells of A . hypoguea will shortly be at hand and that the breeding scheme for the simultaneous exploitation of hybrid vigor and additive genetic variance will be a viable alternative to the present prospect of gene substitution from wild species to cultivated peanuts by conventional methods. Even more remote is the possibility of recovering genes from distant species of Aruchis and by use of gene amplification to improve traits using techniques similar to those described by Schimke (1980) for experiments on the development of drug resistance in cultured mammalian cells (mouse and hamster). PROGRAM EXPLOITING ADDITIVE GENETIC VARIANCEAT Selected lines Line selections of species of Perennes genome haploid A . hypogaea

THE

DIPLOID LEVEL Selected lines species of Annuae genome 1

1

’

2

3 etc. PROGRAM EXPLOITING F, HYBRID VIGORWITHIN Selected 2n F, Selected 2n F, recurrent crosses to recurrent crosses to A . hypogaea haploids A . hypogaea haploids Annuae genome Perennes genome Induce to 4n level

THE

AMPHIDIPLOID

I

1

F:.--.2

2

3

.

1

3

I

Agronomic selection for combinability field testing for hybrid vigor at 4 n level and variety release program

I

Conventioh 4 n level breeding program and variety release program

FIG. 3. Proposed breeding scheme to exploit simultaneouslyadditive genetic variance and hybrid vigor.

68

J. C. WYNNE AND W. C. GREGORY

VII. CONCLUSIONS Peanut yields in the United States have more than doubled during the past 40 years. Breeders have contributed from one-quarter to one-third of this increase through the development of high-yielding cultivars. The progress from breeding has been hampered by several factors including ( a ) the late initiation of peanut breeding programs (1930-194Os), (b) the regional importance of the crop in the United States, ( c ) the relatively few breeders assigned to peanuts, (d) the lack of financial resources, and (e) the difficulty associated with breeding an indeterminate crop whose fruit are formed underground. On a worldwide basis, international breeding efforts were boosted with the adoption of the peanut as a target crop by the International Crops Research Institute for the Semi-Arid Tropics. Recognition of the importance of the peanut, particularly in developing nations, has given impetus to programs of breeding for disease and insect resistance, incorporation of diverse germ plasm into breeding populations, and population improvement. During the last two decades much emphasis has been placed on collection of wild species and exotic cultivars. As emphasis moves from collection to evaluation and exploitation of these genetic resources, progress in improving the productivity, pest resistance, and adaptability of the peanut can be expected. REFERENCES

Abdou, Y. A., Gregory, W. C., and Cooper, W. E. 1974. Peanut Sci. 1, 6-11. Arunachalam, A., Bandyopadhyay, Nigam, S. N., and Gibbons, R. W. 1980. “National Seminaron the Application of Genetics to Improvement of Groundnut,” pp. 1-19. Tamil Nadu Agricultural University. Ashri, A., and Goldin, E. 1965. Radiat. Bot. 5, 431-441. Avadhani, K. K., and Rao, B. V. R. 1968. Indian J . Sci. 2 , 77-82. Baker, R. J . 1978. Crop Sci. 18, 533-536. Banks, D. J . 1969. J. Peanut Res. Educ. Assoc. 1, 23-28. Banks, D. J . 1976. Crop Sci. 16, 499-502. Bartz, J . A , , Norden, A. J., LaPrade, J. C., and DeMuynk, T. J. 1978. Peanut Sci. 5, 53-56. Bernard, R. L. 1960. Ph.D. Thesis, North Carolina State Univ. Diss. Abstr. 21, 1028-1029. Beute, M. K., Wynne, 1. C., and Emery, D. A. 1976. Crop Sci. 16, 887. Branch, W. D. 1980. Proc. Am. Peanut Res. Educ. SOC. p. 48. Brim, C. A. 1966. Crop Sci. 6, 220-221. Brim, C. A. 1973. In “Soybeans: Improvement, Production, and Uses” (B. E. Caldwell, ed.), pp. 155-186. Amer. SOC. Agron., Madison, Wisconsin. Bromfield, K. R. 1974. F A 0 Plant Protect. Bull. 22, 29. Bromfield, K. R., and Bailey, W. K . 1972. Phytopathology 62, 748. Bromfield, K . R . , and Cevano, S. J. 1970. Plant Dis. Rep. 54, 381-838. Cahaner, A., Hillel, J., and Ashri, A. 1979. Theor. Appl. Genet. 55, 161-167. Campbell, W. V . , Emery, D. A., and Wynne, J. C. 1976. Peanut Sci. 3, 40-43. Campbell, W. V . , Wynne, J. C., Emery, D. A., and Mozingo, R. W. 1977. Crop Sci. 17, 346.

PEANUT BREEDING

69

Chen, C. Y., and Wan, H. 1968. J. Agric. Assoc. China 64, 1-12. Cherry, J. P. 1977. J. Agric. Food Chem. 25, 186-193. Cockerham, C. C. 1963. In “Statistical Genetics and Plant Breeding,”pp. 53-94. NAS-NRC Publ., Washington, D. C. Coffelt, T. A. 1980a. Crop Sci. 20, 419. Coffelt, T. A. 1980b. Peanut Sci. 7, 91-94. Coffelt, T. A., and Hammons, R. 0. 1974. Oleagineux 29, 23-27. Compton, W. A. 1968. Crop Sci. 8, 773. Cook, M. 1972. Plant Dis. Rep. 56, 382-386. Cooper, W. E., and Gregory, W. C. 1960. Agron. J. 52, 1-4. Crompton, C. E., Wynne, J. C., and Isleib, T. G. 1979. Oleagineux 34, 71-75. Duke, James A. 1980. In “The Handbook of Legumes (of World Economic Importance).” Plenum, New York. Duncan, W. G., McCloud, D. E., McGraw, R. L., and Boote, K. J. 1978. Crop Sci. 18, 10151020. Eberhart, S. A., and Russell, W. A. 1966. Crop Sci. 6, 36-40. Emery, D. A., and Wynne, J . C. 1976. Environ. Exp. Bot. 16, 1-8. Emery, D. A., Gregory, W. C., and Loesch, P. J., Jr. 1965. Radiar. Bor. 5, 339-353. Emery, D. A., Wynne, J . C., and Rawlings, J. 0. 1972. Radiat. Bor. 12, 7-18. Fonesca, S., and Patterson, F. L. 1968. Crop Sci. 8, 85-88. Foster, D. J., Wynne, J. C., and Beute, M. K. 1980. Peanur Sci. 7, 88-90. Frank, Z. R., and Krikun, J. 1969. Plant Dis. Rep. 53, 744-746. Garet, B. 1976. Oleagineux 31, 435-442. Garren, K. H. 1964. Phytopathology 54, 279-281. Gibbons, R. W., and Mercer, P. C. 1972. J. Am. Peanur Res. Educ. Assoc. 4, 58-66. Gibbons, R. W., Bunting, A. H., and Smartt, J. 1972. Euphyrica 21, 78-85. Gibori, A., Hillel, J . , Cahaner, A., and Ashri, A. 1978. Theor. Appl. Genet. 53, 169-179. Gillier, P. 1978. Oleagineux 33, 25. Goulden, C. H. 1941. Proc. Int. Gent. Congr. 7th, Edinburgh pp. 132-133. Gregory, M. P., and Gregory, W. C. 1979. J. Hered. 70, 185-193. Gregory, W. C. 1966. In “Plant Breeding” (K. J. Frey, ed.), pp. 189-218. Iowa State Univ. Press, Ames. Gregory, W. C. 1968. Radiar. Bor. 8, 81-147. Gregory, W. C. 1970. Crop Sci. 10, 459. Gregory, W. C., and Gregory, M. P. 1976. In “Evolution of Crop Plants” (N. W. Simmonds, ed.), pp. 151-154. Longmans, Green, New York. Gregory, W. C., Gregory, M. P., Krapovickas, A., Smith, B. W., and Yarbrough. 1973. In “Peanuts-Culture and Uses. ” Amer. Peanut Res. Educ. Assoc., Stillwater, Oklahoma. Gregory, W. C., Krapovickas, A., and Gregory, M. P. 1980. In “Advances in Legume Science” (R. J . Summerfield and A. H. Bunting, eds.), pp. 469-481. Royal Botanic Gardens, Kew. Gupton, C. L., and Emery, D. A. 1970. Crop Sci. 10, 127-129. Hadley, B. A., Beute, M. K . , and Wynne, J. C. 1979. Peanut Sci. 6, 51-54. Hammons, R. 0. 1971. Crop Sci. 11, 570-571. Hammons, R. 0. 1973a. “Peanuts-Culture and Uses,’’ pp. 135-173. Amer. Peanut Res. Educ. Assoc., Stillwater, Oklahoma. Hammons, R. 0. 1973b. J . Am. Peanur Res. Educ. Assn. 5 , 193-194. Hammons, R. 0. 1976. Crop Sci. 16, 527-530. Hammons, R. 0. 1977. PANS 23, 300-304. Hanson, W. D., and Weber, C. R. 1961. Genetics 46, 1425-1434. Hanson, W. D., Probst, A. H., and Caldwell, B. E. 1967. Crop. Sci. 7, 99-103.

70

J. C. WYNNE AND W. C. GREGORY

Hassan, M. A., and Srivastava, D. P. 1966. J . Indian Bot. Soc. 45, 293-295. Hayman, 9. 1. 1954. Genetics 39, 789-809. Hayman, 9. I. 1958. Heredity 12, 371-390. Higgins, 9. 9. 1941. G a . Agric. Exp. Sin. Bull. 213, 3-11. Hildebrand, G. L., and Smarit, J . 1980. Zimbabwe J . Agric. Res. 18, 39-47. Holley, K. T., and Hammons, R . 0. 1968. Univ. G a . Coll. Agric. Exp. Sta. Res. Bull. No. 32. Hull, F. H., and Carver, W . A. 1938. Fla. Agric. Sta. Annu. Rep. pp. 39-40, Isleib, T. G., and Wynne, J. C. 1980. Proc. Am. Peanut Res. Educ. Soc. 12, 74. Isleib, T. G., Wynne, J . C., and Rawlings, J. 0. 1978. Peanut Sci. 5, 106-108. Isleib, T. G., Wynne, J. C., Elkan, G. H . , and Schneeweis, T. J. 1980. Peanut Sci. 7, 100-105. Jensen, N. F. 1970. Crop. Sci. 10, 629-635. Jinks, J. L. 1954. Genetics 39, 767-788. Jinks, J . L. 1956. Genetics 10, 1-30. Johnson, D. R., Wynne, J. C., and Campbell, W . V. 1977. Peanut Sci. 4, 9-1 1 . Joshi, S. N., Vaishanani, N. L., and Kabaria, M. M. 1972. Indian J . Agric. Sci. 42, 145-147. Kahn, A. R., Emery, D. A . , and Singleton, J. A. 1974. Crop Sci. 14, 464-468. Khan, 9. M. 1974. Ph.D. Thesis, Oklahoma State Univ. Diss. Abstr. 34, 4788b. Khan, 9. M., Wadsworth, D. F., and Kirby, J. S. 1972. Proc. Am. Peanut Res. Educ. Assoc. 4, 145.

Kornegay, J . L., Beute, M. K., and Wynne, J. C. 1980. Peanut Sci. 7, 4-9. Krapovickas, A . 1968. In “The Domestication and Exploitation of Plants and Animals” (P. J . Ucko and 1. S. Falk, eds.), pp. 427-441. Duckworth, London. Krapovickas, A. 1973. In “Agricultural Genetics, Selected Topics” (R. Moav, ed.), pp. 135-151. Wiley, New York. Krapovickas, A., and Rigoni, V. A. 1960. Rev. Invest. Agric. 14, 197-228. Krapovickas, A., Cenoz, H. M., and Ojeda, H. R. 1968. Bor. Cat. Genet. Fitotec, Corr. 5, 96. LaPrade, J. C. 1974. Ph.D. Thesis, University of Florida. Laynsse, A , , Wynne, J. C., and Isleib, T. G. 1980. Euphytica 29, 561-570. b u c k , D. B., and Hammons, R . 0. 1974. J . Econ. Entomol. 67, 564. b u c k , D. B., and Harvey, J. E. 1968. J . Econ. Enrornol. 61, 583-584. Lin, H. 1959. J. Agric. Assor. China 26, 44-48. Lin, H. 1966. J . Agric. Assoc. China 54, 17-24. Lin, H.,Chen, C. C., and Lin, C. Y . 1971. J . Agric. Assoc. China 47, 27-35. Marani, A. 1963. Crop Sci. 3, 552-555. Marani, A. 1968. Crop Sci. 8, 299-303. Martin, J . P. 1967. Oleagineux 22, 673-676. Mather, K. 1967. “The Elements of Biometry.” Methuen, London. Matzinger, D. F. 1963. In “Statistical Genetics and Plant Breeding” (W. D. Hanson and H. F. Robinson, eds.), pp. 253-279. Natl. Acad. Sci.-Natl. Res. Counc. Publ. No. 982, Washington, D.C. Matzinger, D. F., and Wernsman, E. A . 1968. Crop Sci. 8, 239-243. Melouk, H. A., and Banks, D. J . 1978. Peanut Sci. 5, 112-1 14. Minton, N. A., and Hammons, R. 0. 1975. Plant Dis. Rep. 59, 944. Mixon, A. C. 1976. Proc. Am. Peanut Res. Educ. SOC. 8, 54-58. Mixon, A. C. 1980. Proc. Am. Peanut Res. Educ. Assoc. 12, 5 3 . Mixon, A . C., and Rogers, K. M. 1973. Agron. J . 65, 560-562. Mohammed, J.. Wynne, J. C . , and Rawlings, J. 0. 1978. Oleaginew 33, 81-86. Moll, R. H., and Stuber, C. W. 1974. Adv. Agron. 26, 277-313. Moll, R. H.,Salhuana, W. S., and Robinson, H. F. 1962. Crop Sci. 2, 197-198. Monasterios, T., Jackson, L. F., and Norden, A. J. 1978. J . Am. Peanut Res. Educ. Assoc. 10,154.

PEANUT BREEDING

71

Moss, J. P. 1977. J . Am. Peanut Res. Educ. Assoc. 9, 34. Moss, J . P. 1980. In “Advances in Legume Science” (R. J. Summerfield and A. H. Bunting, eds.), pp. 525-535. Royal Botanic Gardens, Kew. Mroginski, L. A., and Fernandez, A. 1980. Oleagineux 35, 89-92. Muheet, A., Chandran, L. S., and Agrawal, 0. P. 1975. Madras Agric. J . 62, 164. Nigam, S. N., Dwivedi, S. L., and Gibbons, R. W. 1980. In “Proceedings of the International Workshop on Groundnuts.” Patancheru, A. P., India. Norden, A. J. 1973. In “Peanuts-Culture and Uses,” pp. 175-208. Amer. Peanut Res. Educ. Assoc., Stillwater, Oklahoma. Norden, A. J. 1980. In “Advances in Legume Science” (R. J. Summerfield and A. H. Bunting, eds.), pp. 515-523. Royal Botanic Gardens, Kew. Ojomo, D. A., and Adelana, B. 0. 1970. J . Agric. Sci. 75, 419-420. Parker, R. C . , Wynne, J . C., and Emery, D. A. 1970. Crop Sci. 10, 429-432. Pompeu, A. S. 1977. Cienc. Cult. 29, 319-321. Porter, D. M., and Hammons, R. 0. 1975. Peanut Sci. 2, 23-25. Porter, D. M., Beute, M. K., and Wynne, J. C. 1975. Peanut Sci. 2, 78-80. Rao, V . R. 1980. In “Proceedings of the International Workshop on Groundnuts.” Patancheru, A. P., India. Sandhu, B. S., and Khehra, A. S. 1976. Crop Improv. 3, 9-17. Sandhu. B. S . , and Khehra, A. S . 1977. Indian J . Genet. Plant Breed. 37, 22-26. Sangha, A. S . , and Jaswal, S. V. 1975. Oilseeds J . 5, 8-12. Shilling, T. T., Mozingo, R. W., Wynne, J. C., and Steele, J. L. 1980. Proc. Am. Peanut Res. Educ. SOC. 12, 71. Schimke, R. T. 1980. Sci. Am. 243, 60-69. Sharief, Y . , Rawlings, J. 0..and Gregory, W. C. 1978. Euphytica 27, 741-751. Shorter, R. 1978. Peanut Industry Workshop. Cons. Fertilizers, Kingaroy, Queensland. Simmonds, N. W. 1979. “Principles of Crop Improvement.” Longmans, Green, New York. Singh, M., Badwal, S . S., and Jaswal, S . V . 1975. Indian J . Genet. Plant Breed. 35, 26-28. Smartt, J . , and Gregory, W. C. 1967. Oleagineux 22, 1-5. Smartt, J . , Gregory, W. C . , and Gregory, M. P. 1978. Euphytica 27, 677-680. Smith, J . W., Jr., Posada, L., and Smith, 0. D. 1980. The Texas A&M Univ. System Publ. MP-1464, College Station, Texas. Smith, 0. D., Boswell, T. E., andThames, W. H. 1978a. CropSci. 18, 1008-1011. Smith, 0. D., Smith, D. H., and Simpson, C. E. 1978b. J . Am. Peanut Res. Educ. Assoc. 5 , 193. Smith, T. E. 1961. Phytopathology 51, 411-412. Sowell, G., Jr., Smith, D. H., and Hammons, R. 0. 1976. Plant Dis. Rep. 60, 494-498. Sriwatanapongse, S., and Wilsie, C. P. 1968. Crop Sci. 8, 465-466. Stalker, H. T., Wynne, J. C., and Company, M. 1979. Euphytica 28, 675-684. Stuber, C. W. 1970. Crop Sci. 10, 129-135. Subrahmanyam, P . , Gibbons, R. W., Nigam, S. N., and Rao, V. R. 1980. Peanut Sci. 7, 10-12. Sun, P. L. F.. Shands, H. L., and Fornsberg, R. A. 1972. Crop Sci. 12, 1-5. Syakudo, K., and Kawabata, S . 1963. Jpn. J . Breed. B (3). 137-142. Syakudo, K., and Kawabata, S. 1965. Jpn. J . Breed. 15 (3). 167-170. Tai, P. Y. P., and Hammons, R. 0. 1978. Peanut Sci. 5, 72-74. Tai, P. Y. P.,and Young, C. T. 1977. Peanut Sci. 4, 1-6. Tai, Y. P. 1972. Ph.D. Thesis, Oklahoma State Univ. Diss. Abstr. 33, 5698B. Tai, Y. P., and Young, C. T. 1975. J . Am. Oil Chem. SOC. 52, 377-385. Tappan, W. B., and Gorbet, D. W. 1979. J . Econ. Entomol. 72, 772-776. Wernsman, E. A., and Matzinger, D. F. 1966. Crop Sci. 6, 298-300. Widner, J. N., and Lebsock, K. L. 1973. Crop Sci. 13, 164-167.

72

J. C. WYNNE AND W. C. GREGORY

Wilson, D. M., Mixon, A. C., and Troeger, 1. M. 1977. Phytopathology 69, 159-162. Woodroof, J . G. 1966. “Peanuts: Production, Processing, Products.” AVI Publ., Westport, Connecticut. Worthington, R. E., and Hammons, R. 0. 1971. Oleagineux 26, 695-700. Wynne, J. C. 1976a. Peanut Sci. 3, 62-66. Wynne, J . C. 1976b. Proc. Am. Peanut Res. Educ. Assoc. 8, 44-47. Wynne, J . C., and Coffelt, T. A. 1980. Proc. Am. Peanut Res. Educ. Assoc 12, 75. Wynne, J. C., and Isleib, T. G. 1978. Peanut Sci. 5, 102-105. Wynne, J . C., and Isleib, T. G. 1980. Proc. Am. Peanut Res. Educ. Assoc. 2, 48. Wynne, J. C., and Rawlings, J. 0. 1978. Peanut Sci. 5, 23-26. Wynne, J . C., and Sullivan, G. A. 1978. Peanut Sci. 5, 109-111. Wynne, J. C., Emery, D. A,, and Rice, P. W. 1970. Crop Sci. 10, 713-715. Wynne, J . C., Rawlings, J. 0.. and Emery, D. A. 1975a. Peanut Sci. 2, 50-54. Wynne, J . C., Rowe, R. C., and Beute, M. K. 1975b. Peanut Sci. 2, 54-56. Yadava, T. P., and Kumar, P. 1978. Indian J . Agric. Res. 12, 1-4. Young, C. T., and Hammons, R. 0. 1973. Oleagineux 28, 293-297. Young, S., Kinzer, R. E., Walton. R. R., and Matlock, R. S. 1972. J . Econ. Entomol. 65,828-832.

ADVANCES IN AGRONOMY, VOL. 34

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS Umesh C. Gupta* and John Lipsett'f *Research Branch, Agriculture Canada, Charlottetown, Prince Edward Island, Canada and ?Division of Plant Industry, CSIRO, Canberra City, Australia

1. Introduction

73

11. Molybdenum Fertilizers, Their Rates and Methods of Application, and Industrial Uses

of Molybdenum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ill. Physiological Role of Molybdenum in Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Determination of Molybdenum in Soils and Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Total Molybdenum in Soils.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Available Molybdenum in Soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... C. Molybdenum in Plants.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Responses to Molybdenum on Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Factors Affecting the Molybdenum Uptake by Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Parent Rock and Chemistry of Molybdenum in Soils . . . . . . . . . . . . . . . . . . . . . . . . B. Soil pH ...................... C. Plant Species and Varieties . . . . . . D. Effect of Phosphorus and Sulfur . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Stage of Plant Growth and Plant Part Sampled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Organic Matter. . . G. Drainage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Deficiency and Sufficiency Levels of Molybdenum in Plants . . . . . . . . . . . . . . . . . . . . VIII. Molybdenum Deficiency and Toxicity Symptoms in Plants A. Deficiency Symptoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Toxicity Symptoms . . . . . . . . . . , . . , . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . , . . . . . . IX. Molybdenum Toxicity and Molybdenum-Copper-Sulfur Interrelationships in Animals X. Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

75 78 81 82 82 84 85 89 89 94

96 96 97 98 99 99 100 100 104 105 107 109

1. INTRODUCTION Molybdenum is one of the most recently recognized nutrient elements considered to be essential for the growth of plants. The biological importance of Mo was not realized until Bortels (1930) showed that the element was highly beneficial in the fixation of N, by the nitrogen-fixing bacterium Azotobacter chroococcum.The essentiality of Mo for higher plants was first established by Arnon and 73

Copyright @ 1981 by Academic Press. Inc. All rights of reproduction in any form reserved.

ISBN 0-12-000734-7

74

UMESH C. GUPTA AND JOHN LIPSETT

Stout (1939), who found that deficiency symptoms in tomatoes (Lycopersicon esculentum, Mill.) were controlled by the addition of minute quantities of Mo in the nutrient solution. This discovery later led to the detection of Mo deficiency in clover (Trifolium subterraneum L.) by Anderson (1942) and of “whiptail” in cauliflower (Brassica oleracea var. borrytis L.) by Davies (1945) and Mitchell (1945). Responses to Mo have since been reported in a variety of crops in various countries of the world. Of the annual crops, tomatoes, lettuce (Lactuca sativa L.), spinach (Spanica oleracea L.), beets (Beta vulgaris L.), all the Brassica species [cauliflower, broccoli (Brassica oleracea var. italica Plenck. ), and rape (Brassica napus L.)] are very sensitive to restricted Mo supply (Johnson, 1966). Forage legumes, such as clover (Robinson er al., 1957) and alfalfa (Medicago sariva L.), also respond to Mo (Gupta, 1969). Deficiency symptoms for most micronutrients appear on the young leaves at the top of the plant because most micronutrients are not readily translocated. Molybdenum is an exception in that it is readily translocated and its deficiency symptoms generally appear on the whole plant (Bergmann, 1976). Symptoms include a general yellowing and stunting of the plant (Bergeaux, 1976), and interveinal mottling and cupping of the older leaves followed by necrotic spots at leaf tips and margins (Hewitt and Jones, 1948). Root interception and mass flow are considered to be the important mechanisms controlling the movement of Mo to plant roots growing in soil (Barber er a l . , 1966). Plants appear to absorb Mo in the form of the anion MoOiL- (Chesnin, 1972). This anion appears to be mobile as indicated by the experiments of Kannan and Ramani (1978), who found that when it was fed to one of the primary leaves of bean (Phaseolus vulgaris L.), most of it was transported to the stem and root. Tiffin (1972) stated that since H,MoO, is extensively dissociated in the range of pH 5-6, the anion MOO%- would therefore be the predominant form of Mo in plant xylem, assuming it does not associate with other plant constituents. It was further suggested by Tiffin (1972) that the form of Mo translocated in plants is unknown, and the possibility of organic complexing cannot be excluded. In soils Mo occurs in the following forms: (1) water-soluble present in the soil solution, (2) adsorbed by soil colloids, (3) held in the crystal lattice of minerals, and (4) present in organic matter (Bergeaux, 1976). Forms most available for plant use are the soluble forms present in the soil solution and Mo adsorbed by soil colloids. Highly weathered acid soils are apt to be more deficient in Mo (Giddens, 1964b). On the other hand, soils that are derived from granitic rocks, shales, slates, or argillaceous schists tend to be high in Mo (Kubota, 1972; Mitchell, 1974). Many peats, alkaline soils, and poorly drained soils with a high water table tend to produce forages high in Mo (Dye and O’Hara, 1959). The presence of large quantities of Mo in plants generally does not produce harmful effects or yield reductions in crops. Plant tissue Mo concentrations in

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

75

excess of 100 ppm (Allaway, 1977) andof 218 ppm (Gupta et al., 1978) in onion (Allium cepa L.) tops were not found to be associated with any toxicity. However, animals are quite sensitive to a high-Mo forage ration, which can cause Mo-induced Cu deficiency. Dietary levels of Mo of 10-20 ppm are nearly always associated with some evidence of disturbed Cu metabolism in ruminant animals (Allaway, 1977). Nonruminants, especially horses, are less subject to Mo toxicity. The objective of this article is to present an up-to-date report on soil as the source of supply of Mo, having regard to Mo availability to the plants, functions and methods of applying Mo in crops, factors affecting the Mo uptake by crops, and the interrelationship of Mo, Cu, and S in plants as associated with Mo toxicity in animals. The interaction effect of various factors in soils and plants will also be discussed.

II. MOLYBDENUM FERTILIZERS, THEIR RATES AND METHODS OF APPLICATION, AND INDUSTRIALUSES OF MOLYBDENUM Molybdenum can be applied to crops by various methods in the form of different fertilizer sources. In the most commonly used sources, Mo is present in the Mo6+ oxidation state as the ion in Na,MoO4.2HzO and (NH,),MoO, (Lehr, 1972). A list of some of the common sources of Mo fertilizers as outlined by Murphy and Walsh (1972) is given in Table I. Another source of Mo introduced recently is a yellow ammonium phosphomolybdate (NH4)3[P(Mo30,0)~] containing 61% Mo that is a reagent-grade experimental material (Cyanamid, 1970). Martinez et al. (1977) used molybdic acid for treating soybean (Glycine max (L.) Merr.) seeds. Molybdenum frits containing 2-3% Mo (Murphy and Walsh, 1972) have also been used. Henkens Table I Percentage of Molybdenum and Chemical Formulae of Molybdenum Sources

Source

Formula

Sodium molybdate Ammonium molybdate Molybdenum trioxide Molybdenum sulfide

N%MoO,. 2 H 2 0 (NH&Mo,02, .4H20 Moo3 MoS,

Approximate Mo (%) 39 54

66 60

76

UMESH C. GUPTA AND JOHN LIPSETT

and Smilde (1966) found that frits are good sources of Mo and reported that their residual effect is somewhat larger than that of sodium molybdate. Molybdenum can be applied as a seed treatment, soil application, or as a foliar spray. The data summarized by Murphy and Walsh (1972) indicate that 50-100 g Mo/ha are generally needed for soil treatments on most agronomic crops and that as much as 400 g Mo/ha may be needed on vegetable crops, such as cauliflower. Applications of greater than 1 kg Mo/ha (Gupta and MacLeod, 1975) produced forages that could prove toxic if fed to livestock. For acid soils, broadcast applications of Mo are best mixed with limestone to prevent fixation of Mo (Berger, 1962). Molybdenum has also been applied to soils in combination with superphosphates and has been found to be readily available in a few New Zealand soils (Widdowson, 1966). Approximately 40-60% of the Mo applied by this method to beans (Phaseolus spp.) was recovered in the tops and seeds. Lipsett and David (1977) in Australia used molysuper, which is supposed to contain 0.04% Mo in each bag. However, the percentage distribution of Mo varied according to the fraction size, which was not evenly distributed within the bag. The fine material, which was mostly in the lower layer, contained an average of about 1900 ppm Mo compared with 1500 ppm in the medium-sized particles and only 260 ppm in the coarse ones. It was suggested that the addition of Mo be made during the early stages of manufacture, such as with the acid that is poured on the rock phosphate to produce even mixing.' The residual effect of Mo added to the soil varies from one soil to another. McLeod (1976) reported, based on his studies in New Zealand, that Mo applications of 140 g sodium molybdate would last 4-5 years. Gupta (1979) reported that the residual effect of Mo added at 0.4 kg/ha on some podzol soils should last 2-3 years from the crop sufficiency point of view. Foliar-applied Mo for rapid uptake and for overcoming Mo deficiency has been a common practice for many crops (Murphy and Walsh, 1972). Inden (1975) recommends use of wetting agents in the spray when applying Mo in foliar sprays on cauliflower or onions. Results of Gupta (1979) showed that foliar sprays may be more desirable than soil applications under dry conditions. It has been suggested that foliar sprays should allow a reduction in the rates of Mo needed to maintain adequate levels in certified hybrid maize (Zea mays L.) seeds (Weir et al., 1976); this method of application would also avoid the problem of Mo fixation in acid soils (Bergeaux, 1976). Weir etal. (1976) reported that both soil and foliar treatments of Mo raised the Mo concentration in corn (Zea mays L.) grain and leaves, but the foliar sprays were more effective. Spraying when the maize plants were 80 cm tall increased the Mo concentration in the seeds 'The company manufacturing the material now uses an oil carrier to try to improve adhesion.

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

77

more than earlier or later spraying. Likewise Boswell er al. (1967) found that the concentration of Mo in the kernels of peanuts (Arachis hypogaea L.) increased as the time of spraying was delayed for up to 6 weeks after bloom, after which there was a decline in the effect. Because of the extremely low Mo requirement of crops, the most common method of correcting a Mo deficiency is to treat the seeds with a Mo preparation. Seed treatment of Mo has been used to prevent Mo deficiency in Brussels sprouts (Brassica oleracea, var. gemmifera Zenker) (Gupta and Cutcliffe, 1968) and to increase the Mo concentration of soybeans (Golov and Kazakhkhov, 1973). The Mo content of seeds is important; for example, Hagstrom and Berger (1965) observed that large-seeded crops, such as peas (Pisum sativum L.), responded to soil applications of Mo when the seeds contained less than 0.2 pprn Mo, but not when they contained enough Mo (0.5-0.7 ppm) to supply the Mo needs of the crop. Gurley and Giddens (1969) also reported that high Mo content in large seeds may supply enough Mo to plants grown on Mo-deficient soils. In maize, severe Mo deficiency can be expected when seeds containing less than 0.02 ppm Mo are sown on Mo-deficient soils, but not when the seeds contain more than 0.08 ppm (Weir et al., 1966). Application of Mo in the lime of lime-pelleted legume seeds is a practical way of applying Mo in close proximity to the seeds and without detriment to the rhizobia in the inoculum (Date and Hillier, 1968). Results of Martinez et a f . (1977) indicated that when molybdic acid was seed-applied (0.04 ppm on soil basis), soybean growth was reduced. Although seed-applied Mo was at a low level when calculated on a soil basis, the concentration near the seeds would be higher than in the soil application and would account for the greater decrease in growth. Gupta and Kunelius (1980) found that use of moist seeds treated with a commercial Mo preparation at the rate of 14 g Mo/ha resulted in large quantities of Mo in forage crops, which when fed to the livestock could produce molybdenosis (Mo-induced Cu deficiency). In order to avoid excessively high concentrations of Mo in the crops, it would therefore be advisable not to treat germinating seeds with a Mo preparation. The rate of 28 g commercial source of Mo per 27 kg of seed (1 oz/bushel) is sufficient to meet the Mo requirement for soybeans (Bergeaux, 1976). According to the recommendation of Lancaster (personal communication from J. D. Lancaster of Mississippi State University in 1968) about 7-35 g Mo/ha (0.1-0.5 oz Mo/acre) annually is sufficient for seed treatment. Reisenauer (1963) also showed that seed application of sodium molybdate was much more effective than soil application for peas. The Mo supply was considered to be adequate when applied at the rate of 18-36 g/ha. Besides its use in fertilizers, Mo is used industrially as a component of hard, corrosion-resistant alloys with steel, a lubricant, and a pigment or other reagent,

78

UMESH C. GUPTA AND JOHN LIPSE'IT

Table I1 Production of Molybdenum in 1976" Production ( lo3 tons)

Source

50.8

United States Canada Chile U.S.S.R.

13.4

9.8 8.8

Source

Production (lo3 tons)

China PeN Bulgaria Japan

I .8 0.8 0. I 0. I

"From Manheim and Landergren (1978)

notably catalytic. These uses account for the production shown in Table 11. Reserves amounting to a supply of 6- 10 years at this rate of usage appear to have been identified. Molybdenum is not abundant in the Earth's mantle, but it is widespread, as one would expect from its essential role in plants. Large amounts of Mo occur in sedimentary formations, especially marine manganiferous concretions. Concentrations may exceed 0.04%, but these amounts are not yet recoverable since the Mo tends to be dispersed in the sediments (Manheim and Landergren, 1978). The concentrations that are attributed to various materials are given in Table 111, based on Norrish (1975) and Manheim and Landergren (1978).

111. PHYSIOLOGICAL ROLE OF MOLYBDENUM IN PLANTS Molybdenum is a component of at least five distinct enzymes that catalyze diverse and unrelated reactions, namely nitrogenase, nitrate reductase, xanthine oxidase, aldehyde oxidase, and sulfite oxidase (Nicholas, 1975). Three of these enzymes, nitrate reductase, nitrogenase, and sulfite oxidase, are found in plants. The principal functions of Mo in plants are implicated in the electron-transfer system; for instance, nitrate reductase and nitrogenase require Mo in the reducTable 111 Concentration of Molybdenum in Rocks, Soils, Natural Waters, and Coal" Earth's CNSt

Igneous rocks

Sedimentary rocks

Soils

1

2

1-2

2 .O-2.5

"Values given in parts per million.

River, lake, and groundwater

0.5 x

10-3

Ocean

Coal

11.0 x 10-3

3-5

MOLYBDENUM IN SOILS, PLANTS, A N D ANIMALS

79

tion of NO,- and in the fixation of N,, respectively. This section will include a discussion of these two enzymes (molybdoproteins) as they function in the plant system. The first molybdoprotein, nitrate reductase, is known to require Mo and flavin for its activity and in the reduction of NO,- to No,- as follows: Reduced N A D

+ NO,-

+ NAD

+ NO,- + H,O

(1)

where NAD is nicotinamide adenine dinucleotide. The reduction mechanism from NO,- to NOz- has been proposed by Nicholas (1975) as NO,NO,-

+ 2H' + 2 e - + NO,- + H,O + H+ + 2e- -+ NO,- + OH-

Reaction (2) is based on the acidic half reaction, whereas reaction (3) allows for OH- participation at physiological pH. Nitrate reductase is found in most plant species as well as fungi and bacteria (Price et al., 1972). The increased Mo requirement of most plants grown on No,--N compared with NH,+-N can be almost completely accounted for by the Mo in nitrate reductase (Evans, 1956). The other major known molybdoprotein of plants, nitrogenase, fixes elemental nitrogen in the form of NH,, which is then assimilated by the plant (Koch ef al., 1967). The role of Mo in the fixation of N, has been reviewed in detail by Chatt (1974). The unique role of Mo in biological systems is exemplified by nitrogenase, the enzyme that converts N, into NH, at room temperature and normal pressure (Schrauzer, 1976). The nitrogenase is an enzyme complex composed of two distinct components that combine to reduce N, to NH, [reaction (4)] or acetylene to ethylene [reaction (5)] (Nicholas, 1975):

+ 6H+ + 6 e C,H, + 2H+ + 2eN,

-+

2NH3

+

GH,

(4)

(5)

Nitrogenases have been isolated from a variety of different sources, for example, from Azotobacter vinelandii, Rhizobium japonicum, Azotobacter chroococcum, and Klebsiella pneumonianum (Schrauzer, 1976). Recent studies by Agarwala et al. (1978) have shown that in addition to reduced nitrate reductase activity, Mo deficiency in corn resulted in significantly lower activities of catalase, aldolase, and alanine aminotransferase and higher activities of peroxidase, P-glycerophosphatase, and ribonuclease. In addition to the involvement of Mo in the fixation of N, and nitrate reduction, Mo is associated with other processes in plants. However, many of these processes are interrelated with the two main functions. For example, experiments of Malonosova (1 968) showed that addition of Mo to the soil resulted in better development of lupine (Lupinus spp.) and increased weight of its roots and nodules although the N, was not fixed. In this review of Russian work it was

80

UMESH C . GUFTA AND JOHN LIPSETT

reported that on soils containing little Mo, plants will develop many nodules on roots, but N, is not fixed. Merkel et al. (1975) showed that Mo deficiency in tomatoes decreased organic nitrogen content of the leaves to the same degree that it decreased the organic anion content of the leaves. This change was mainly in the contents of malate and citrate. Studies by Anderson and Spencer (1950) showed that deficiency of Mo decreased both the protein nitrogen and the nonprotein nitrogen percentages in the clover. The percentage crude protein in a number of plant species has been found to increase with optimum rates of Mo applied to the soil (Reddy, 1964). Hagstrom and Berger (1965) found that applications of Mo increased the nodulation and nitrogen content of peas and soybeans. The effect of Mo in plants is to increase the content of proteins and to create favorable conditions for the biosynthesis of nucleic acids (Peyve, 1969). Inden (1975) stated that deficiency of Mo can cause plant chlorosis, which is due to the inability of the plant to form chlorophyll. Further, since the deficiency of Mo reduces the rate of NO3- reduction in plants, photosynthesis decreases because the end products are not removed by combination with nitrogenous compounds. Das Gupta and Basuchaudhuri (1977) found that Mo applications enhanced the nitrate reductase activity in the rice (Oryza sativa L.) plant, particularly under the high-nitrogen nutrition. This ultimately led to greater concentration of reduced nitrogen and thereby created a concentration gradient for the uptake and subsequently greater assimilation of nitrogen in the tissues as suggested by Wardlaw (1968). This suggests that there exists a close functional relationship between the nitrate reductase activity and chlorophyll content as observed by Das Gupta and Basuchaudhuri (1977). The chlorophyll content of corn has been found to decrease due to a deficiency of Mo (Agarwala et al., 1978). Molybdenum was also considered to be associated with the metabolism of Fe and phosphoric acid (Inden, 1975). Premature sprouting of maize grain on the cob was shown to be controlled in glasshouse and field experiments by the Mo concentration of grain (Tanner, 1978). It was found that when the Mo concentration of grain fell below 0.05 ppm, sprouting occurred, the severity of which was enhanced by heavy, late sidedress of nitrogenous fertilizer. The explanation offered was that Mo-deficient plants have a decreased ability to reduce NO3- and consequent accumulation of inorganic NO3- promotes sprouting. Under Mo deficiency, plants accumulate low-molecular-weight nitrogen compounds and may have defective vascular bundles in the midrib in young leaves. There is a deposition of a brown substance on tissue surface and in intercellular spaces and cells (Busslar, 1970). Hewitt (195 1 ) pointed out that a low concentration of ascorbic acid in tissues is a characteristic of Mo deficiency in a number of species. Agarwala (1952) also demonstrated that cauliflower plants grown with various nitrogen sources, in-

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

81

cluding urea and ammonium sulfate, developed characteristic Mo deficiency symptoms known as “whiptail” and contained reduced concentrations of ascorbic acid. Subsequent studies by Agarwala and Hewitt (1955) showed that Mo deficiency decreased the total and reducing sugars in the leaves of young cauliflower plants. In nutrient culture studies in flax (Linurn usitatissirnurn L.), Mo has been found to be closely associated with the regulation of the deleterious effect of Mn, Zn, Cu, Ni, or Co on the physiological availability of Fe to the plant (Millikan, 1947).

IV. DETERMINATION OF MOLYBDENUM IN SOILS AND PLANTS Thiocyanate and dithiol are the two most commonly used reagents employed in the colorimetric determination of Mo from biological materials. The dithiol method used by Piper and Beckwith (1948), Clark and Axley (1955), and Bingley (1963) for plants and soils has been found to be more sensitive and precise than the thiocyanate method because the green-colored complex formed between Mo and dithiol was stable for at least 24 hours (Gupta and MacKay, 1965a). Later, Fuge (1970) developed a rapid and simple method in which a Technicon AutoAnalyzer is used for the determination of Mo on the basis of its catalytic action on the potassium iodide-hydrogen peroxide reaction. Fernandez et al. (1978) used 2,2-dihydroxybenzophenone reagent and considered it to be superior to the most commonly used Mo-thiocyanate complex method because an extra step is not necessary and the results are more reproducible. Molybdenum has also been determined spectrochemically after chemical concentration, using the cathode-layer arc technique (Mitchell, 1974) and polarography (Dekhkankhodzhayeva and Kruglova, 1972). Trace quantities of Mo have been determined by atomic absorption spectroscopy (AAS), both flameless (Henning and Jackson, 1973; Jarrel and Dawson, 1978) and flame (Khan et al., 1979). Little and Kemdge (1978) used a carbon rod analyzer for determining very low levels of Mo. The high temperature required to atomize Mo in this procedure makes it easy to remove matrix materials during the ashing phase. The colorimetric method using dithiol and the most recently used AAS are probably the most commonly used techniques for determining Mo in soil and plant materials. The detection limits for determination of Mo by AAS using flame and graphite furnace have been found to be 10 and 2 ng/ml, respectively (Khan et al., 1979). Using the dithiol colorimetric method (Gupta and MacKay, 1965a), the satisfactory detection limit is about 20 nglml. The recovery of Mo added to the plant material as determined by these two methods has been found to range from 92 to 95%.

82

UMESH C. GUPTA AND JOHN LIPSETT

A. TOTAL MOLYBDENUM I N SOILS

The most common method for extracting Mo from soils is by perchioric acid digestion (Reisenauer, 1965). Dry ashing of soil and the extraction of ash using concentrated acids was employed for determining total Mo in soils by Perrin (1946) and Grigg (1953a). Total Mo has also been extracted by Na&O, fusion of soil (Purvis and Peterson, 1956). Unpublished data of the first author of this article showed that such extracts contained large quantities of interfering materials and required purification, which is time consuming. Little and Kerridge (1978) used HF-HC104 digestion for determining total Mo in soils. As for other plant nutrients, total Mo content of soils, except for very low levels, is generally not a good indicator of plant Mo availability (Little and Kemdge, 1978; Williams, 1971). Available Mo content has not been found to be closely related to the total Mo content of soils (Stone and Jencks, 1963). However, soil with a total Mo content of more than 20 ppm may be regarded as potentially “teart” (producing Cu deficiency in animals) in Scotland (Williams, 1971). Soils with low total Mo and neutral to alkaline pH may be depleted by many years of intensive cropping (Davies, 1956). Liming can correct Mo deficiency; therefore an estimate of total Mo content may provide some indication of the Mo supplying power of acid soils. Details of the effect of liming on Mo availability will be dealt with in Section VI,B. Little information exists on the levels of Mo in various soils but, in general, contents of 0.5-5 ppm are normal (Robinson and Alexander, 1953; Williams, 1971) and in agreement with the relative abundance of Mo in the lithosphere (2.3 ppm), whereas figures of 0.5 pprn or less would be considered particularly low (Williams, 1971). The Mo content of a few soils selected from areas of Canada close to industrial plants ranged from 1 .O to 11.3 ppm (Warren, 1973). MacLean and Langille (1973) reported that the Mo content of Nova Scotia (Canada) podzol soils ranged from 0.05 to 12.1 ppm. B. AVAILABLE MOLYBDENUM IN SOILS

The presence of extremely small quantities of Mo in the soil, the influence of chemical characteristics of soils (Karimian and Cox, 1979), the importance of seed reserves (Gurley and Giddens, 1969), and the possibility that seed reserves may mask a deficiency in the soil make the problem of determining Mo availability more difficult than for the other micronutrients. The first report on the available Mo in soils, which related extracted Mo to plant uptake, was by Grigg (1953b) in New Zealand. This involved an acid oxalate extractant buffered at pH 3.3. The responses and lack of responses as related to Mo extracted by Grigg’s reagent for a number of crops have been summarized by Reisenauer

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

83

(1967). However, oxalate-extracted Mo has not been found to correlate with Mo uptake by plants on various soils (Karimian and Cox, 1979; Little and Kerridge, 1978). The extracted-Mo values on some iron-rich soils may, however, be misleading (Little and Kerridge, 1978). Water has been used as an extractant for determining Mo in soils (Gammon et al., 1954; Lavy and Barber, 1964; Gupta and MacKay, 1965b). Difficulties are encountered in the analyses because quantities extracted are very low. Lowe and Massey (1965) used 10-hour leaching of soil with hot water and found the method to be more reliable for extracting available Mo than the Grigg method. Pathak et al. (1969) found hot-water-soluble Mo to be related to plant uptake of the element in a number of soils from India. The hot-water method, however, is time consuming and has not been tested extensively. Recently, anion-exchange resin (Dowex 1-X4) has been used to extract Mo from soils in Oregon (United States) by Bhella and Dawson (1972). The method involved equilibration of an aqueous soil-resin suspension and measurement of the sorbed molybdate using 2 M NaCl to displace it. The method was considered to be satisfactory, theoretically reasonable, and practically acceptable in assaying available soil Mo within the pH range 4.95-7.10. Later Jarrell and Dawson (1978) found that Mo uptake by subterranean clover ( T . subterruneum L., cv. Mt. Barker) grown in field experiments was significantly correlated with Mo extracted by anion-exchange resin. Although the Grigg and resin-soil test procedures may be used to separate Mo-responsive from non-Mo-responsive soils, neither was considered by Karimian and Cox (1 979) to be a good predictor of the size of response to Mo fertilization. They suggested that a measure of the degree of crystallinity of soil iron oxides, the active Fe ratio (amorphous Fe/free Fe), along with the soil pH could be used to reasonably predict response magnitude to Mo fertilization until a more reliable procedure is developed. Ammonium acetate and EDTA have also been used to extract Mo from soils; the extracted Mo together with soil pH and organic matter provided useful information on the potential uptake of Mo by pasture (Williams and Thornton, 1972). The most recent studies suggest extraction of soil after wetting it to field capacity and leaving it at this moisture content overnight (Little and Kemdge, 1978). A concentration of less than 4 p g Mo/liter thus obtained appeared to indicate soils that would respond to Mo application. A 12-hour extraction with 1 M (NH,),CO, at pH 9.0 was proposed by Vlek and Lindsay (1977) to reflect the more labile fraction that could come into solution. The amount of Mo obtained using this extractant on soils, with pH a 7 . 0 , amended with 0-2 ppm Mo as N+MoO, showed a correlation of r = 0.977 with Mo uptake by alfalfa. However, when soils with pH <7 were included, the correlation dropped considerably due to the extraction of more soil

84

UMESH C. GUFTA AND JOHN LIPSETT

Mo by (NH4)&03 than was available to plants. The method was considered to be good for predicting Mo toxicity in alkaline soils. With so many factors affecting the availability of Mo in soils and the uptake of Mo by plants, it is rather obvious that soil tests are often unsuccessful in predicting Mo deficiencies and toxicities. Microbiological assay of available soil Mo using A . niger has been used by Mulder (1948). It was later pointed out that, using this method, the discrepancy between the growth of cauliflower and the microbiological test could be ascribed to the presence of a high concentration of available Mn in the acid soils, which inhibited uptake of Mo by cauliflower but not by A . niger (Mulder, 1954). Soil Mo values obtained with A . niger have been reported to be only moderately correlated with field response (Donald et al., 1952) and not well related to soil solution Mo (Lavy and Barber, 1964). More recently Franco et al. (1978) used Azotobacter paspali nitrogenase activity (measured by acetylene reduction) to detect Mo deficiency in soils. The test is based on the percentage increase of nitrogenase activity of Azotobacter paspali growing in a defined medium with a small amount of soil as the Mo source. Using this test, 36 out of 41 soils tested showed a significant response to added Mo with a tropical forage legume (Centrosema pubescens Beth) and gave better correlation than the oxalate extractable Mo. C . MOLYBDENUM I N PLANTS

Due to the presence of Mo in microquantities in plant materials, its analysis is somewhat more difficult. However, due to advances made in the methodology for Mo as discussed earlier in this section, its analysis is carried out more precisely than before in most laboratories. Because of the unreliability of various methods of extracting Mo from soils from one region to another and from one group of soils to another, the analysis of plant materials for Mo continues to be one of the better indicators of Mo availability in soils as related to plants. The two most common methods of extracting Mo are dry ashing and wet digestion. Dry ashing has been used by Gupta and MacKay (1965a) and Fuge (1970). Recently Khan et al. (1979) used the wet digestion method and found that the results obtained were reasonably similar to those obtained by the dry ashing technique used by Gupta and MacKay (1965a). Wilson (1979) used matrix-compensated Mo standards to determine Mo in plant tissue by flameless AAS. The need for preparing Mo standards in an appropriate matrix solution was emphasized by the fact that an aqueous Mo standard gave an apparent Mo concentration 41% greater than a Mo standard prepared with the matrix solution of 0.55 M HC104. This acid concentration was used because the concentration of HClO., in the diluted plant digests ranged from

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

85

0.2 to 0.7 M. Concentrations of HCIOl ranging from 0.25 to 1.5 M had little effect on the observed Mo values. The recovery of Mo added to the plant digests of orchard leaves and alfalfa tops ranged from 95 to 105% (Wilson, 1979). Using this procedure, tissue Mo concentration of 0.1 ppm could be determined with good precision. However, tissue Mo values of 0.05 ppm and less were substantially less precise using this procedure.

V. RESPONSES TO MOLYBDENUM ON CROPS The criterion for response to an added nutrient in agricultural terms is usually an increase in growth of the plant, either in rate or in final yield. Sometimes a visible symptom of deficiency may be present, and its disappearance or prevention form part of the response. The first reported responses by higher plants to application of Mo were by tomato in solution culture, following scrupulous removal of all traces of Mo from the other nutrients (Amon and Stout, 1939). The purification was necessary because only extremely small amounts of Mo are required for normal plant growth. Some lower organisms-Azotobacter (Bortels, 1930), Clostridium (Bortels, 1936), and Aspergillus (Steinberg, 1936)-had already been shown to require Mo for the fixation of N, and for utilization of NO3-. Mulder (1948) showed that symbiotic Rhizobium in peas also required Mo for nitrogen fixation. It was further reported that “the growth-rate curve and the increasing sporulation of Aspergillus niger with increasing amounts of Mo were used in estimating very small amounts of this element in various materials. ” The first reports of responses by agricultural plants in the field soon appeared, notwithstanding reservations on the part of the first workers that a nutrient needed in such small amounts would be lacking in the field (Stout, 1972). Anderson (1942) obtained responses to Mo (applied as ammonium molybdate at a rate of 1 kg/ha) by the clover component of subterranean clover-perennial ryegrass (Lolium perenne L.)-Phalaris tuberosa (now P . aquatica L . ) pastures in South Australia. Much work followed in Australia and other countries, and several conditions (“diseases”) in plants that had hitherto been unexplained were shown to respond to application of Mo. Probably the best known and most widespread of these was in cauliflower. “Whiptail” of cauliflower was marked by distortion of the leaf lamina. Quite early, cereals and other grasses were shown to respond to Mo (Fricke, 1947; Lobb, 1953; and Mulder, 1954, in Tasmania, New Zealand, and Holland, respectively). Several compilations (which will not be reviewed here) report the accumulated knowledge of Mo deficiency in agricultural plants, including field, horticulture, and pasture species (Wallace, 1951; Hewitt, 1956; Johnson, 1966; Lucas and Knezek, 1972).

86

UMESH C . GUPTA AND JOHN LIPSE'M

The early reports of responses by cereals to Mo application seem not to have affected commercial practice in cereal growing to any marked extent. Particularly in Australia (where deficient soils proved to be widespread), Mo deficiency was regarded as a complaint of Rhizobium, particularly that associated with subterranean clover. Because of this, subterranean clover with fertilizer N omitted became the standard experimental indicator plant. Molybdenum was applied in the field primarily in order to secure symbiotic fixation of N, for permanent pastures or leys. It was not applied unless the clover showed a response; associated grasses, or cereals in the rotation, were not assessed. However, Mo deficiency was reported in maize in a coastal area of New South Wales (Noonan, 1953). Further work of Weir and Hudson (1966) clarified the association of deficiency symptoms with seed reserves: symptoms were unlikely, even on low-Mo soils, for Mo contents in seed >0.08 ppm and likely for <0.02 ppm. Accordingly, application of Mo to crops producing hybrid seed of unsatisfactory content was stipulated as a requirement for certification (Weir et al., 1966). However, it was found that high rates of sidedress were often required. These were expensive and potentially hazardous to grazing ruminants in the rotation, so foliar sprays are now preferred (Weir et al., 1976). Maize appears to be relatively susceptible to Mo deficiency, particularly on acid soils. Tanner (1976) readily produced symptoms in the greenhouse and suggests that the problem is widespread in Rhodesia. Premature sprouting of the grain on the cob in Rhodesia has been linked with Mo-N balance (Tanner, 1978). Most of the corn in the United States seems to be adequately supplied with Mo from the soil. However, Brown and Clark (1974) reported Mo deficiency in a pot study of two inbred lines, grown on acid (pH 4.3) soil. One line developed symptoms (twisted leaves, chlorosis, and necrosis), whereas the other did not. Application of either Mo or lime cured the symptoms. This suggests different genetic abilities in taking up Mo from soil of low Mo availability. A relationship between deficiency symptoms and seed reserves has been found in some cases (Mulder, 1954; Hewitt, 1956), and has been advanced (Anderson, 1956) as a probable reason for the relative absence in early experience of deficiency in large-seeded legumes. Nevertheless, these species are by no means immune (de Mooy, 1970; Sedbeny et a f . , 1973; Parker and Hams, 1977), and the results suggest that deficiency may occur in the period of dependence on seed reserves of Mo before the new seedling of whatever species taps soil sources by means of an expanding root system. On the other hand, it is possible for a seed to contain more Mo than the whole new plant will require (Meagher et al., 1952). Meanwhile, application of Mo specifically to wheat (Triticum aesiivum L.) came to be practiced in Australia on soils that grew subterranean clover satisfactorily without it. In Western Australia, Gartrell (1966) found responses in grain yield by wheat and oats (Avenu sativa L.) on light sandy soils, particularly if ammonium sulfate was added. The untreated plants had a pale color and many

MOLYBDENUM IN SOILS, PLANTS, A N D ANIMALS

87

unfilled heads. In southern New South Wales, Freney and Lipsett (1965) and Lipsett and Simpson (1971, 1973) found that high levels of available N would render wheat seedlings responsive to Mo. The Mo had a protective function in alleviating damage (see Section 111) caused by accumulation of NO3- in the small plant. Of course, the protection breaks down if NO,- reaches high levels. It has long been known that fertilizer N commonly reduces grain yield in Australian wheat crops (Stonier, 1965; Dann, 1969), an effect most often ascribed to early exhaustion of moisture reserves in the profile, with consequent restriction of grain filling and hence of yield. There has been little experimental backing for this explanation, and the results with Mo (which attribute the reduction in yield to damage in the early seedling stage) may better account for the damage in many cases. Whatever the explanation, fertilizer N was avoided by farmers and bare fallowing was relied on to mineralize N for the crop. It appears that this intention may succeed too well at times, particularly in clover-based rotations, by mineralizing undue amounts of N. Aspects of the matter still to be investigated include the following: seasonal effects on mineralization and distribution of NO3- within the root profile; the role of seed reserves of Mo; the full geographical and pedological extent of the deficiency or imbalance; managerial practices in respect of N mineralization and N and Mo fertilizers. Plainly the temperate cereals are not immune to Mo deficiency. Since they are normally grown under cultivation, the appropriate techniques of applying Mo are not necessarily those that were worked out for undisturbed pastures. There is a relatively large gap between the plant requirement and the usual Mo fertilizer dressings, which provides an opportunity for economies by such techniques as seeddress or placement. These would appear mainly suited to crops that are reseeded frequently. It is not clear whether the tropical cereals suffer from Mo deficiency to any extent. Low available N and neutral to alkaline soils with adequate available Mo probably combine to safeguard the dryland crops in general. The nutrition of paddy rice (Oryza sativa L.) with respect to Mo is not clear, but should be considered because of the status of rice as a staple foodstuff. Rice in aerobic culture presumably exhibits at least the conventional requirement for Mo for nitrate reductase. Shukla et al. (1976) reported a field response to Mo in India by a new high-yielding variety. It might be expected that supplying N predominantly as NH4+ would lessen the demand for Mo in the paddy, whereas the likelihood of the presence of Fe'+ and Sz- ions would reduce the availability of Mo in the flooded soil. Although grasses in general have been thought to have low requirements for Mo, Lipsett (1975) found that Phularis tuberosa (now P . aquatica L.) was like wheat in showing a marked response to Mo on a soil on which perennial ryegrass grew strongly. It was not established whether the ryegrass was sustained by seed reserves, recovery of more soil Mo, or some other means. Phalaris aquatica

88

UMESH C. GUPTA AND JOHN LIPSETT

(tuberma) proved to be extremely sensitive to Mo deficiency in the seedling stage, particularly when supplied with NO3-. Johansen (1978a) examined three tropical grasses but found no marked sensitivity. However, there were indications of the interaction between Mo and Nog- on which this response to Mo is based. Since the yield response to Mo actually begins at an early growth stage, attempts have been made to describe and diagnose responsiveness by characterizing the biochemical processes that are affected most directly by the application of Mo. The activity of nitrate reductase in producing NO2-, and hence protein, and in lessening NO,- within the plant is the process primarily involved. Randall (1969) developed a method for diagnosing Mo deficiency in wheat by a bioassay of leaf material. Johnson et al. (1976) have suggested that nitrate reductase activity might serve as a predictive test of crop yield. One might expect a correlation where N is the main limiting factor. Responses by plants to Mo are closely related to soil properties, and consequently there are established geographical patterns of deficiency and of excess. Large areas of North America, Australia, New Zealand, and probably eastern Europe are potentially deficient. In Canada, responses to Mo have been limited to the eastern part of the country. The soils of this area are leached, are acidic, and have given response to Mo in controlling whiptail of cauliflower (Robinson and Campbell, 1956) and in increasing the yield of grass-legume hay and the nodule weight of red clover (Robinson et a!., 1957). Gupta (1969) reported that in greenhouse experiments, crops grown on coarse-textured soils gave response to Mo even when the soil was limed to pH 6.5. Results of field experiments conducted in Prince Edward Island showed that in the case of a suspected Mo deficiency, addition of about 0.2 kg Molha as a foliar spray or 0. 4 kg Molha applied to the soil should alleviate a Mo deficiency problem, and the residual effects at these levels of Mo should last 2-3 years (Gupta, 1979). There are many abstracts of reports dealing with Mo in the Soviet Union and associated countries, frequently with positive responses in yield. Valdek ( 1 974) and Agafonova et al. (1975) indicate that Mo is in regular use. Again, mainly light acid soils appear to be involved. Excessive amounts are of main concern in the western United States and Europe. Information is lacking for most of Africa, South America, and Asia, with the exception of India, where Mo appears to be generally in moderate supply. There is a dearth of reports for tropical areas in particular, yet there are large areas of leached, acid, ferruginous soils that would seem to be highly prone to Mo deficiency. Prasad and Page1 (1976) examined a range of tropical soils for ammonium acetate-extractable Mo, and reported a high incidence of deficiency in ferrallitic soils. Molybdenum deficiencies have been readily found in Queensland, Australia, in investigations for pasture establishment and growth on infertile soils (Jones and Crack, 1970; Bishop, 1974). The preference often shown

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

89

agriculturally for recent volcanic soils (as between Java and Sumatra) may reflect better Mo supply. It seems likely that Mo deficiency may be widespread on yellow earths and similar soils in resettlement areas in Indonesian Borneo (personal communication with L. F. Myers, CSIRO, Australia). The expectation is that any plant species sown on such soils would be potentially at risk of Mo deficiency, particularly if the crops involved are dependent on fixation of N, as their source of N. It is to be expected that the response in yield to Mo will be accompanied by an increase in Mo content, since applied Mo (or soil Mo on liming) is readily taken up by plants. Molybdenum contents can, in fact, reach levels at which the material is toxic to animals, notably ruminants. A figure of 10 ppm Mo in forage is widely assumed to be dangerous (see Section IX). This problem of molybdenosis may reflect soil properties, but the use of lime, the rates and frequency of application of Mo fertilizers, the composition of irrigation water, and the possibility of contamination from mining or from burning coal are all aspects to be considered. There were firm reports (Allaway, 1968), which seem not to have been followed up, that a relatively high Mo content in plant material in the diet favors dental health in humans.

VI. FACTORS AFFECTING THE MOLYBDENUM UPTAKEBYPLANTS A . PARENT ROCKAND CHEMISTRY OF MOLYBDENUM IN SOILS

1 . Parent Rock

Molybdenum is a transition element in the fifth row of group VIB of the periodic table. It is metallic and closely resembles tungsten (W) in chemical properties. General principles of the occurrence of Mo in the igneous rocks of the Earth’s crust are now fairly well established, since molten magmas represent comparable starting points of relatively well-blended materials. However, the concentration and form of Mo in other rocks and soils tend to vary according to particular origins and conditions of formation. Molybdenum is a versatile element insofar as valence is concerned, and it may precipitate under either oxidizing (Mofi+predominant) or reducing (Mo4+) conditions (Manheim and Landergren, 1978). Consequently, there may be local enrichments or depletions, and recent work is largely concerned with elucidating sequences of occurrence, mobilization, and deposition in particular situations. a . Occurrence in Igneous and Metamorphic Rocks. Igneous rocks make up some 95% of the crust of the Earth (Mitchell, 1964), and Mo occurs in both acid

90

UMESH C. GUPTA AND JOHN LIPSETT

and basic igneous rocks. Manheim and Landergren (1978) suggest an overall figure of nearly 2.0 ppm Mo for granitic rocks and somewhat lower for basalts. The Mo is found in feldspar and ferromagnesian minerals such as biotite and olivine, respectively. Although the occurrence of Mo in metamorphic rocks has not been widely studied, metamorphism would be expected to alter the form and site of occurrence rather than the amount of Mo present. New minerals may be formed that must undergo weathering-again, in the case of sedimentary parent materialbefore the Mo becomes available to plants. b. Occurrence in Sedimentary Rocks. The sedimentary rocks that are formed following weathering and transport usually retain some of the Mo of the parent material. Concentrations of Mo may be high if the rocks have formed under conditions favoring accumulation and precipitation of Mo, viz., at depth in oceans or in the presence of carbon (coals, oil shales, some limestones). The weighting given such sediments determines the average value actually quoted for Mo content. Manheim and Landergren (1978) suggest < I ppm Mo overall, but Norrish (1975) suggests 2 ppm. The lowest values are found in sandstones that contain stable minerals and have undergone high drainage losses. The carbonaceous materials are of interest mainly in relation to either contamination of the environment by spoil from mining and industrial uses or, in the case of some limestones, their deliberate addition to the soil for agricultural purposes. c . Weathering and Occurrence in Water and Soil; the Sedimentary Cycle. The Mo is released from rocks by weathering (Mitchell, 1964), which involves one or more cycles of solution, oxidation, and precipitation before the Mo from a given rock either appears in the soil formed from that rock or is transported to ocean sediments as part of the sedimentary cycle. Molybdenum is fairly readily released from primary minerals by weathering and, compared with other metals, it remains relatively mobile as potentially soluble molybdates (Mo")). Consequently, movement by leaching is likely, unless iron, aluminum, or manganese oxides interfere under conditions appropriate for occlusion on these minerals (Davies, 1956). Entry of Mo into surface or groundwaters is normal (Table III), and may be marked near ore bodies (Jackson et al., 1975), where the concentration may reach several ppm. Manheim and Landergren (1978) quote high natural values for rivers in arid regions, up to 10 p g Mo/liter, but suggest that pollution from industry and agriculture regularly leads to much higher values, and has caused a probable doubling in recent time of dissolved runoff of Mo. The amounts of Mo in coal (Table 111) indicate it to be one of the sources of the extra Mo. It appears in the ash and possibly in smoke and fumes. This dissolved Mo may be intercepted by anaerobic layers in lakes, and incorporated as the sulphide in bottom deposits, but the ocean is the ultimate sink for

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

91

the Mo that leaves the land. Currently there is an average concentration in the ocean of 11 p g Mo/liter (Table 111). The oceanic Mo may come out of solution, again into bottom deposits, under either oxidizing or reducing conditions. The former condition probably occurs under most oceans, and the latter occurs notably in the Black and Baltic seas (Manheim and Landergren, 1978). On land, the Mo content of the soil reflects the amounts in the parent rock. This and the effect of mobility are clearly seen near ore bodies where the influence of mineralization may be detected some distance away from the body. Measurements of the Mo content of soil have been used to locate deposits of Mo and associated metals such as Cu (Manheim and Landergren, 1978). Thornton and Webb (1 973) used the Mo content of stream sediment to map likely patterns of high-Mo soils in England and Wales. The soils of highest content are underlain by marine black shales. Soils on sediments in general vary more widely in Mo content than those on igneous rocks because the latter are more uniform. The variations of interest concern the extremes about the average figure of 2-2.5 ppm (Table 111) for soils. Since the requirements of plants for Mo are so low, removal in produce is not great as a rule, and absolute deficiencies were originally thought to be unlikely (Stout, 1972). However, they were duly found, initially on old soils in Australia (Anderson, 1956) and subsequently on the other continents. Soils formed on sandstones, which lack suitable minerals and may have already experienced heavy losses, are most likely to be deficient in total amount. Deficiencies also occur that are probably due to secondary reactions that reduce the availability of Mo. High Mo contents may be found in soils near ore bodies or formed on some shales. Plants may tolerate the high levels of Mo without difficulty (see Section I), but animals grazing on them develop molybdenosis (see Section IX). This was first reported for the teart disease of ruminants in Somerset, England (Ferguson et al., 1943). Further occurrences have been reported, particularly since it was realized that the toxic effects may be insidious, being quite substantial before clinical symptoms are shown. Particularly in the United States it seems that excess Mo is currently of more concern than are deficiencies. This reflects the predominance of young soils and the wide extent of Mo mineralization in the source areas of important imgation waters. 2. Chemistry of Molybdenum in the Soil The three dominant aspects of Mo as a soil species are occurrence as the anion, the effect of pH on solubility, and the possibility of changes in form with loss of availability. a . Occurrence as the Anion. It is now well accepted that Mo, having been released from crystal lattices by weathering, is found in well-aerated agricultural

92

UMESH C. GUPTA AND JOHN LIPSETT

soils (pH in water >5.0) mainly as molybdate ions, MoOL2- (Lindsay, 1972). This predominance of the anionic form makes Mo unique among metallic plant nutrients. It leads to known interactions with the other major anions SOg- and Pod3-, notably the reduction by SO%- of uptake of Mo (see Section VI,D). Thus it is feasible to control Mo content of forage at safe levels by applying SOL2(Gupta and Munro, 1969; Gupta and MacLeod, 1975). Reports disagree to some extent on whether PO-: affects uptake of Mo (Stout et al., 1951; Gupta and Cutcliffe, 1968); displacement from adsorption sites and competition for entry might be expected to have opposing effects. Whether NO3- ions as such interact also is not known, but it would seem possible. High concentrations of NO3appear to increase the requirement for Mo (J. Lipsett, CSIRO, Australia, unpublished results), but assessment of the precise interaction is hindered by the direct participation of Mo in removal of NO3- in the plant through the action of nitrate reductase (see Section 111). It seems that the chemical similarities between Mo0i2- and Po,3-(which were extensively relied on to establish a soil chemistry of Mo by analogy) were overemphasized. The similarities were most persuasive at acid pH, with respect to adsorptive reactions (see Section VI,A,2,c) with hydrous oxides of iron and aluminum (Smith and Leeper, 1969). At higher pH, on the other hand, Mo0L2and Pod- do not behave alike. High pH is marked by the presence of Ca2+ions, and whereas calcium phosphate has low availability to plants, calcium molybdate (CaMoO,) is readily available. Thus CaMoO, is too soluble to form or persist in most soils (Lindsay, 1972). b. The Effect of p H on Solubility of Soil Molybdenum. It is now well-known that availability of soil Mo to plants is, in general, increased by raising the pH and reduced by lowering it (see Section VI,B). However, high organic matter may in some way protect the Mo at low pH (Mitchell, 1964), which may explain the relative infrequency (Anderson, 1956) of Mo dressings applied to many of the Australian soils on which the deficiency was first observed. On these soils, clover growth was made possible by applying Mo with superphosphate, and soil organic matter increased substantially. Evidently, the only Ca added, that in the superphosphate, was not enough to saturate the new exchange capacity, and soil acidity increased (Donald and Williams, 1954). Nevertheless, the applied Mo apparently kept its availability, possibly because of protection by the organic matter. The use of lime has been postponed (Anderson, 1956) in this way. c . Forms of Molybdenum in Soils. The forms in which Mo occurs in soil are not yet completely described because of the difficulties in dealing with small amounts of relatively labile material. Five potential fractions are recognized: 1. Primary crystalline material. 2. Water-soluble molybdates in the soil solution. 3. Organically complexed Mo.

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

93

4. Molybdate adsorbed on positively charged surfaces. 5. Discrete, secondary compounds, either crystalline or amorphous. A full description of soil Mo would include extents of occurrence in these five categories, the transformations between them, and the availability of Mo in each to plants. It seems that high amounts of available Mo are probably the result of high values in fractions 2 and 3; thus wet soils high in organic matter commonly yield plants high in Mo (Davies, 1956; Kubota et al., 1963). Low amounts may reflect low content in the soil overall, but may also arise from the occurrence or formation of the fractions 4 and 5. The relative roles of these fractions are the main concern in the chemistry of deficient soils and in the correction of the deficiency. Smith and Leeper (1969) considered that in many of their soils residual availability of Mo applications was lost by slow reactions with the soil rather than by leaching. The implication is a transition into fraction 5. Recent work suggests that fractions 4 and 5 may be rather vaguely differentiated, particularly at certain pH values. The principal positively charged sites in the soil for anionic adsorption are provided by hydrous oxides of Fe and A1 (Reisenauer et al., 1962; Childs and Leslie, 1977). The importance of Fe in the reaction is shown by consistent correlations between it and Mo (Norrish, 1975). Molybdenum so held would presumably be well placed to form definite compounds with Fe, but Vlek and Lindsay (1974) believe that Fe,(MoO,), as such is too soluble to persist in soils as a major inert form. They favor PV+ as the most drastic combining cation, if it is present. Even PbMoO, would become more soluble if pH were to rise. Occlusion of Mo in fermginous concretions would seem to impose positional unavailability, and further work is needed on the association between Mo and Fe, which begins with adsorption of MOO,'- ions. For a time it was suggested that Mn played a part in immobilizing Mo, but its oxides do not develop a positive charge even under alkaline conditions (Childs and Leslie, 1977). Furthermore, the marine nodules in which there are conspicuous amounts of Mo and Mn are formed from alkaline solutions relatively high in both (Manheim and Landergren, 1978), so extrapolation to soil conditions does not seem to be valid. There seems to be little direct evidence for the mechanism that puts Mo into Mn marine nodules at high pH. d . Effect of Iron and Aluminum Oxides. The effect of adsorption in restricting the rate of supply of to plants is shown by the significant negative interaction term in the regression of extracted on plant response (Barrow and Spencer, 1971). Iron oxides found in acid soils carry positive charges and can react with MoOil-. Jones (1957) showed that a soil high in free iron oxide sorbed the largest quantities of MOO,'- from aqueous solution. Likewise, Karimian and Cox ( I 978) showed that Mo adsorption on mineral soils increased with free iron oxide content. Aluminum oxides are also capable of removing Mo from

94

UMESH C. GUPTA AND JOHN LIPSETT

aqueous solutions, but their effectiveness is less than that of iron oxide under the same conditions (Jones, 1957). Adsorption of Mo by Fe,O, is pH-dependent, as reported by Jones (1956), who showed that adsorption of Mo by Fe,O, i n a solution containing 100 p g Mo decreased from 98 p g at pH 7 to 22 p g at pH 9 after shaking for 15 hours with 100 mg of amorphous Fe,O,. It was suggested that Mo, in fermginous soils, is held on the surface of the colloidal femc oxides as the anion, which is replaceable by OH- ions. Since H+ and OH- are considered potential-determining ions, they cause a change in surface potentials of minerals with pH changes (Vlek and Lindsay, 1977). The adsorption maximum of MOO,'- on hematite was found to be reduced by 80% if the pH was changed from 4 to 7.75 (Reyes and Jurinak, 1967). The zero point of the charge of the hematite occurred at pH 8.0. A method for quantifying the amount of reactive hydroxides present on the surfaces of soil colloids has been suggested that involves reaction of soil with 0.85 M NaF solution at pH 6.8 (Perrott et al., 1976). The measurement of the released OH- in the first 25 min correlated well with extracted A1 and Fe. Fluoride ions reportedly bring structural AI and Fe"+ ions into solution, releasing OH- as ligands on the central Al or Fe ions (Huang and Jackson, 1965). However, Jarrell and Dawson (1978) reported that the Mo content of plants correlated poorly with Mo sorbed by the soil and NaF-released hydroxyls. Their work showed that Mo adsorption was related to the (NH,),C,O,-extractable Fe in the soil but not to other fractions of extractable Fe or to extractable Al forms. The (NH,),GO,-extractable Mo has also been found to be correlated with amorphous and free iron oxide contents, and this was probably a result of strong dissolution of iron oxide by oxalate (Karimian and Cox, 1979). B . SOILpH

Soil pH is one of the most important single factors affecting the uptake of Mo in plants. The beneficial effects of liming acid soils to increase Mo solubility are obvious. The concentration increases 100-fold for each unit increase in pH (Lindsay, 1972). With increase in pH, the soluble Moo4,- species in equilibrium with soil Mo is much greater than for HMoO,- and H,MoO,. At a pH of 5 or 6, the ion HMoO,- becomes dominant and at very low pH values the unionized acid H,MoO, and the cation MOO,'+ are the principal species present (Krauskopf, 1972). The anion exists in an exchangeable form in the soil. Thus, the fact that Mo availability to plants increases with increasing pH may possibly be explained by an anion exchange of the type 2 OH- Ft (Berger and Pratt, 1965). Even wulfenite (PbMoO,), the least soluble of possible soil compounds, becomes more soluble as pH increases (Vlek and Lindsay, 1974), so that the mode of action of liming in increasing the availability of Mo seems straightforward.

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

95

Earlier, Amin and Joham (1958) proposed a Mo cycle within the soil in which it is possible that oxidation may contribute to Mo availability in the long term and that immediate availability of the element may be governed by the equilibrium between MOO, and the salt form. Under acidic conditions the balance would be in favor of MOO,, thus reducing the amounts of molybdate salts present. This would account for the fact that Mo availability is lowered by increasing acidity. The highest amounts of available Mo are associated with soils derived from limestone and shale parent material (Stone and Jencks, 1963). In general, the need for Mo fertilization of most soils can be met by adequate liming of the soil. Ahlnchs et a1. (1963) stated that for Minnesota soils, the recommended liming practices substantially increase the Mo content of vernal alfalfa and thus direct Mo fertilization for those soils was not essential. Likewise, Rolt (1968) stated that with 2 tons of lime plus Mo or 4 tons of lime alone the maximum yield of white clover could be obtained. In a number of crops, a significant interaction effect between applied Mo and lime has been noted, whereby the increases in Mo concentration with added lime are of much greater magnitude with applied Mo than without added Mo (James el al., 1968; Gupta, 1969; Gupta et al., 1978). Unpublished results of U. C. Gupta and J . A. MacLeod on podzol soils of eastern Canada showed a continuous increase in the Mo concentration of barley (Hordeum vulgare L.) plants with increases in the rates of lime application from 0 to 20 g/kg soil (Table IV). The lime rates were associated with a pH range of 4.9-8.0 and very small pH increases were noted beyond the 10-g lime rate. The soil pH increases from 4.9-5.6 to 7.0 increased the plant tissue Mo concentrations from 0.13-0.19 to 0.66 ppm. However, further pH increases to pH 7.8-8.0 resulted in much higher tissue Mo concentrations of 1.69-3.42 ppm. The principle of Mo release upon liming is dependent upon the total Mo content of soils. For example, lime applications to pH 6.5 on two soils of Prince Edward Island did not produce maximum yields of alfalfa and cauliflower in a greenhouse experiment (Gupta, 1969). Some addition of Mo was necessary to achieve maximum yields. Likewise on a sphagnum peat soil some addition of Mo, besides liming, was necessary to obtain maximum yield of some vegetable crops (Gupta et a / . , 1978). The observation of Anderson (1956) that plants responded to wood ashes but not to lime indicated the need for something in the ashes that was simply not provided by raising soil pH. Well-limed or naturally neutral soils may also be depleted of available Mo by cropping or by leaching (Johnson ef al., 1952). Rubins (1956) noted that some Mo deficiencies have occurred on well-limed or naturally neutral soils and on acid soils derived from calcareous parent materials. On highly oxidized soils of the southeastern United States, a high initial rate of 1 1,200 kg lime/ha or the equivalent amount in annual applications did not eliminate the need for Mo application to maintain alfalfa stands on these soils (Giddens and Perkins, 1972).

96

UMESH C. GUPTA AND JOHN LIPSETT

Table IV Effect of Increasing Lime Levels on the Mo Concentration of Barley Boot Stage Tissue (BST) on Three Soils“ Albeny sl

Charlottetown fsl Lime rate (g/kg soil)

Soil pH

Mo in BST (ppm)

0 0.8 1.7 2.5 3.5 5.0 10.0 15.0 20.0

5.6 6.0 6.4 6.9 7. I 7.6 7.8 7.9 7.8

0.19 0.22 0.40 0.66 1.26 1.93 2.79 3.21 3.42

Soil pH 4.9 5.5

6.2 7.0 7.1 7.5 7.7 7.8 7.9

Culloden sl

Mo in BST (ppm)

Soil pH

0.14 0.13 0.26 0.60 0.68 0.94 1.09 I .34 1.69

5.4 5.6 6.2 6.7 7.1 7.6 7.8 1.9 8.0

Mo in BST (ppm) 0.13 0.15

0.18 0.45 0.58 1.16 I .62 2.03 2.70

“Unpublished results of U. C. Gupta and J . A. MacLeod, Research Station, Charlottetown, Prince Edward Island, Canada.

C . PLANTSPECIES A N D VARIETIES

Plant species have been found to have some influence on the plant tissue Mo concentration. Findings of Kubota and Allaway (1972) showed that different forage legumes grown on the same soil have nearly the same amounts of Mo. They further stated that grasses generally contain less Mo than legumes. Studies by Brogan et al. (1973) also showed that the Mo concentration is higher in clover than in grass growing on soils high in Mo. Gupta (1979), however, working on podzol soils of eastern Canada could not find consistent differences in the Mo concentration of red clover and alfalfa when compared to that of timothy. The cultivars of plant species have been found to differ in their ability to absorb Mo. For example, Chipman et al. (1970) found that among cauliflower cultivars “Pioneer” was not as efficient as “Snowball 84” in the uptake of Mo, especially at low levels of supply. D. EFFECTOF PHOSPHORUS A N D SULFUR

Plant uptake of Mo is usually enhanced by soluble P and decreased by available s. Earlier studies by Stout et af. (1951) showed that higher P levels in culture solutions increased Ma uptake as much as tenfold and therefore soil applications of Mo with P fertilizers may be effective. Since then a number of researchers‘have observed a similar effect of P on Mo uptake by plants.

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

97

The effect of P in increasing the concentration of Mo in plants has been reported to be associated with the stimulating effect of Po43ions on the uptake of Mo due to the formation of a complex phosphomolybdate anion, which is absorbed more readily by the plants (Barshad, 1951). Addition of S has been found to decrease the Mo content of crops, such as Brussels sprouts (Gupta and Munro, 1969), raya (Brassica juncea L.) (Pasricha and Randhawa, 1972), Berseem (Trifolium Alexandrinum L.) (Sisodia et al., 1975), soybeans (Singh and Kumar, 1979), tobacco (Nicotiana tabacum L.) (Sims et al., 1979), and perennial ryegrass (Lolium perenne L.) and clover (Williams and Thomton, 1972). Significant Mo rate X rate interactions have been reported for Mo concentrations whereby adding fertilizer had a detrimental effect on the growth of burley tobacco at low soil Mo levels but positive effects at high soil Mo levels (Sims et al., 1979). A recent study by Singh and Kumar (1979) showed that applications of 40 ppm S significantly increased the total Mo uptake in 45- and 110-day-old soybean plants. This increase was suggested to be governed by yield and plant tissue Mo concentration. The inhibitory antagonistic effects of SO,’- on Mo content have been suggested to occur primarily during the absorption process, with some antagonistic mechanism involved during translocation from roots to leaves. McLachlan (1955) reported that deficiencies of P and S can limit the response to Mo unless P and S are applied. In both instances there is a deficiency of P and S rather than an indirect effect on Mo uptake. It was further stated that a deficiency of either P, S , or Mo is no indication of deficiency or sufficiency of the other two. E. STAGEOF PLANTGROWTHA N D PLANTPARTSAMPLED

The Mo content of tobacco leaves was found to be greater than other plant parts (Pal et al., 1976). Studies on tomato by Stout et al. (1951), on alfalfa by Reisenauer (1956), and on soybeans by Singh and Kumar (1979) showed that leaf tissue Mo was considerably higher than that of the stems. The interveinal areas of leaves have been found to preferentially accumulate Mo (Stout and Meagher, 1948). In the case of cereals, the Mo concentration of grain is generally lower than that of the boot stage tissue or straw in low-Mo soils. However when Mo was applied at 0.5 ppm or higher, the Mo concentration of grain was considerably lower than that of the boot stage tissue (Gupta, 1971a). With maturity, the Mo content of leaves and stems was found to decrease in soybeans (Singh and Kumar, 1979). It was also noted that grain contained higher quantities of Mo than either leaves, stems, or pod husks. In solution culture studies on Phaseolus vulgaris L. roots were found to contain higher quantities of Mo than stems and leaves (Wallace and Romney, 1977).

98

UMESH C. GUF'TA AND JOHN LIPSEIT

Unpublished data of U. C. Gupta (Research Station, Charlottetown, Prince Edward Island) also showed that leaves of a number of plant species contained more Mo than petioles or stems with the exception of red clover (Table V). The above-ground vegetative portion of the plant is generally lower in Mo than the seeds. The Mo content of seeds varies from one plant species to another. In a study conducted on Prince Edward Island, the unpublished results of the first author of this chapter revealed that samples collected from adjacent sites showed the Mo content of seeds of a few crops to be in the following order: peas > barley > wheat. F. ORGANIC MATTER

The level of Mo adsorbed has been found to be closely related to soil organic matter (Karimian and Cox, 1978). Since Mo acts like an anion in the soil, it is difficult to explain its adsorption by organic matter. However, certain soils high in organic matter have been found to be deficient in Mo (Mulder, 1954; Davies, 1956). Mulder (1954) found a pronounced response to added Mo in several crops grown on these soils. The presence of organic matter may also promote the availability of certain elements, presumably by supplying soluble complexing agents that interfere with the fixation of elements. For example, soils rich in organic matter contain a readily exchangeable Table V Range of Mo Concentration of Plant Parts of Various Species from Five Locations Plant species Brussels sprouts (Brassica oleracea var. gernmifera Zenker) Broccoli (Brassica oleracea var. iraka Plenck.) Cauliflower (Brassica oleracea var. borryyris L.) Rutabaga (Brassica napohrassica Mill.) Alfalfa (Medicago snriva L.) Red clover (Trifolium pratense L.) "Mean ppm Mo given in parentheses.

Leaves

Petioles

0.11-2.98 (0.90)"

0.12-0.62 (0.36)

2.16- I I . 16 (3.76)

0.37-4.28 (1.76)

0.54-3.72 ( I .65)

0.37-1.10 (0.98)

0.20-1.59 (0.65) 0.10-0.72 (0.28) 0.06-0.23 (0.12)

0.17-0.78 (0.32)

Stems

0.08-0.28 (0.15) 0.08-0.35 (0.15)

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

99

mobile Mo0i2- ion (Koval’skiy and Yarovaya, 1966). Laboratory experiments involving additions of various organic materials, such as compost, farmyard manure, and peat, to the soil increased the extracted Mo after 4-12 weeks of incubation (Gupta, 1971b). However, in the presence of added Mo the extracted Mo was consistently lower where organic materials were added. Bloomfield and Kelso (1973) reported that the anionic form of Mo persisted in solution even after a 3-week incubation period with anaerobically decomposing elant material. G . DRAINAGE

Generally, there have been few studies on the chemistry of Mo in the wet soils that produce high-Mo plants (Allaway, 1977). This may be due to the difficulties in transfemng the wet, poorly aerated conditions found in the field to conditions suited to precise laboratory studies of Mo solubility. However, soil wetness is one of the chief factors affecting the availability of Mo. Poorly drained soils accumulate so much available Mo0i2- that the plants grown on them produce toxicity when fed to animals (Davies, 1956; Kubota et al., 1961). Peats and mucks are products of a wet environment and are associated with Mo toxicity in the United States in the soils of the California delta, the Klamath area i n Oregon, and the Everglades in Florida (Kubota, 1972). Some Scottish soils with poor pedological drainage produced plants containing large quantities of Mo (Mitchell, 1974). The early term “peat scours” for Mo toxicity in cattle and sheep strongly suggests that the problem occurs frequently on wet soils. Well-drained soils, e.g., podzols and krasnozems, are likely to be low in Mo, and there is little enrichment in any layer (Mitchell, 1974). Wet soils, such as those in swamps and lakes that are not leached, tend to be high in organic matter and to have large amounts of Mo that may be highly available, low pH notwithstanding (Kubota et a l . , 1961). It might be expected that Australian soils, being in areas that are now quite arid, would be well-supplied with Mo. However, the deficient soils invariably show signs of having formed in earlier geological periods under heavy leaching. Thus the low present contents of Mo reflect substantial earlier losses that were never replenished.

VII. DEFICIENCY AND SUFFICIENCY LEVELS OF MOLYBDENUM IN PLANTS Among the micronutrients that are essential for plant growth, Mo is required in the smallest amounts. In the majority of soils, as reviewed in Section VI,B, the Mo requirements of plants can be met by liming the soil. Because of the low

100

UMESH C. GUPTA AND JOHN LIPSETT

requirements, the deficiency and sufficiency levels in most plants are extremely small. Rarely, if ever, has Mo been found to be toxic for plants under field conditions. Molybdenum toxicity in plants could be induced only under extreme experimental conditions (Johnson, 1966). Therefore this section will not deal with the toxicity levels of Mo in plants. The plants, however, could under certain conditions accumulate large concentrations of Mo and induce molybdenosis in ruminants consuming such material. This will be dealt with in Section IX. The precision of modem analytical methods is such that even microquantities of Mo in plants can be detected accurately, and considerable data have been obtained in the last 20 years on the Mo levels in a number of plant species. The purpose of this section is to report the sufficiency and deficiency levels of Mo in a number of cultivated crop species as found by workers around the world. Often when one talks about the deficiency and sufficiency levels of nutrients in crops, there is a range in values rather than one definite figure that could be considered as critical. A value considered critical by workers in certain areas may not be critical under conditions in other areas. Likewise the term “optimum” levels of nutrients, as used in the literature by some researchers to express a relationship to maximum crop yield, is sometimes not clear. Theoretically, such a level for a given nutrient should be sufficient to produce the best possible growth of a crop. Often a single value is published on the optimum level. In practice there can be no single number to describe this relationship adequately, and it would be more appropriate to describe the nutrient status of the crop by a range of values. Therefore, for the purposes of this chapter, wherever possible the term sufficiency will be used rather than critical or optimum. As is evident from Table VI, data on the deficiency levels of Mo for many crops are still missing. Because of the very small requirements for Mo, there appear to be much greater variations in the deficiency and sufficiency levels of Mo as shown by various workers than have been reported for other micronutrients. Some of the differences are likely due to differing procedures used to determine Mo. The percentage variations seem to be of a much larger magnitude, but the actual differences seem to be of the order of about 0.5 ppm or less.

VIII. MOLYBDENUM DEFICIENCY AND TOXICITY SYMPTOMS IN PLANTS

A . DEFICIENCY SYMPTOMS

The symptoms associated with deficiency of Mo are closely related to metabolism of N. They are conveniently grouped according to the actual function

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

101

of the Mo that applies. If the Mo is needed for nitrogenase activity, then its lack prevents the fixation of N2.This sequence involves higher plants through the symbiotic organisms such as Rhizobium, and so symptoms associated with deficiency of N are displayed by the higher plant. This syndrome was typical of subterranean clover in the first reported instance of Mo deficiency in the field (Anderson, 1942, 1956). If the Mo is needed directly by the plant for nitrate reductase, then symptoms peculiar to Mo occur, although the plants can be considered to be suffering essentially from a shortage of protein due to the failure of the initial processes of NO3- reduction. These specific symptoms commonly involve leaves, as in “whiptail” of cauliflower (Davies, 1945). The first set of symptoms can be relieved by supplying fixed nitrogen. Even NO,- may serve the purpose since the Mo requirement of nitrate reductase is lower than for nitrogenase. The second type of symptom, however, indicates severe deficiency, and addition of NO3- in that case is likely to make things worse, since the plant is unable to process it and undue accumulation of NO3may occur as a symptom of Mo deficiency (Anderson, 1956; Freney and Lipsett, 1965). An accumulation of NO3- is commonly observed in plants, the NO3- being derived from soil processes or from fertilizer (Hanway and Englehorn, 1958; Moore and Hutchings, 1967). It is often associated with toxicity of the plant to consumers (Everist, 1974) and may impair commercial use in other ways, as in fruit of papaya (Carica spp.) (Menary and Jones, 1972). The possibility exists that attention to the Mo-N balance in such plants might overcome at least part of the accumulation. The visual symptoms shown by various plants are described by Wallace (1951), Hewitt (1956), Anderson (1956), Johnson (1966), and others. They are pictured in color in the March 1956 issue of Soil Science. The symptoms of deficiency and/or imbalance shown by wheat in the Australian reports and in a recent report from India will be described, since the condition appears to be widespread. Doyle et al. (1965) reported paler color and presence of empty heads at maturity. Agarwala and Sharma (1979), working under sand culture in Lucknow, India, reported that Mo deficiency in wheat resulted in golden yellow coloration of older leaves along the apex and the apical leaf margins. Plants had short internodes and reduced foliage. Lipsett and Simpson (1971) described symptoms in young plants of white, necrotic areas extending back along the leaves from the tips. Normal tillering did not occur on affected plants, which were paler in color. Severely affected plants died. Healthy individual plants commonly occurred near affected plants. This suggests variation in soil, in seed reserves, or in ability to extract Mo. Particularly in pots, deficient plants tended to appear gray and limp. However, the characteristic visual symptoms were of terminal scorch on the leaves (suggestive of excessive salt accumulation), paleness, and stunting at an early stage.

Table VI Deficient and Sufficient Levels of Molybdenum in Plants

Plant Alfalfa ( Medicago

sativa L.)

Barley (Hordeum vulgare L . )

Beans

Part of plant tissue sampled

Mo in dry matter (ppm) Deficient

Sufficient

Leaves at 10% bloom Whole plants at harvest Top 15.2 cm of plant prior to bloom Upper stem cutting at early flowering stage Shortly before flowering (top % of plants) Whole tops at 10% bloom Blades 8 weeks old Whole tops at boot stage Grain Tops 8 weeks old

0.26-0.28 0.55-1.15 <0.4

0.34

Tops 8 weeks old

0.05

Tops 8 weeks old

0.04

Whole plants when sprouts began to form Whole plants when sprouts began to form

c0.08

Leaves Leaves Above ground portion of plants at the first appearance of a curd

Whole plants before the appearance of curd Young leaves showing whiptail Leaves Roots Stems At tassel middle of the first leaf opposite and below the lower ear

<0.2

References

1-5

Reisenauer ( I 956) Evans and Purvis (1951) Jones ( I 967)

0.5"

Melsted ef a/. (1969)

0.5-5.0

Neubert er a / . (1970)

0.12-1.29 0.03-0.07 0.09-0.18 0.26-0.32 0.4

Gupta (1970) Johnson er a / . (1952) Gupta (1971a) Gupta (1971a) Johnson et a l . (1952)

0.62

Johnson et a / . (1952)

(Phaseolus vulgaris)

Beets ( Bera vulgaris

L.) Broccoli

Johnson et al. (1952)

(Brussica oleraceu var. iralicu Plenck.)

Brussels sprouts (Brassica oleracea var. gemmifera

0. I6

Gupta and Munro ( 1 969)

0.1 1-0.69

Gupta (1970)

0.09 0.09 C0.26

0.61 0.42 0.68- I .49

Plant (1952) Plant (1952) Chipman et a / . (1970)

10.11

0.56

Gupta ( I 969)

0.19-0.25 2.8-11.9 I .4-7.0 >0.2 0.2"

Peterson and Purvis (1961) Plant (1951a) Dios and Broyer ( 1 965) Dios and Broyer (1965) Neubert er a / . (1970) Melsted ef al. (1969)

Zenker) Cabbage (Brassica oleracea var. cupitatu L.)

Cauliflower (Brassica oleracea var. borryis L.)

Corn (Zea mays L . )

0.07 0.02-0.07 0.023-0.3 0.013-0.1 I


(continued)

I02

Table VI-Continued

Plant Lettuce (Lactuca sativa L.) Pasture grass (Graminae) Red clover ( Trifolium pratense L.)

Soybeans Spinach (Spinacea oleracea L . ) Sugar beets (Beta vulgaris L.)

Part of plant tissue sampled Leaves

Mo in dry matter (pprn) Deficient

Sufficient

0.06

0.08-0.14

Plant ( I 95 I b)

0.2-0.7

Neubert et ul. (1970)

0.3-1.59

Neubert et al. (1970)

0.26

Hawes et a / . (1976) Gupta and MacLeod (1975)

0.45 0.46-1.08

Hagstrom and Berger (1965) Gupta (1970)

1.61 0.15-1.09

Peterson and Purvis (1961) Johnson et al. (1952) Gupta ( 1 970)

0.01-0.15

0.2-20.0

Ulrich and Hills (1973)


0.2-2.0 >20b

Neubert et al. (1970)

>o. I

Johansen (1978b)

First cut at first bloom Total above ground plants at bloom First cut at flowering Plants at 10% bloom Whole plants at the bud stage Whole plants at the bud stage Plants when 26-28 cm high Leaves 8 weeks old Whole tops at normal maturity Blades shortly after symptoms appear Fully developed leaf without stem (taken end June or early July) Plant shoots

Temperate pasture legumes Whole tops at prebloom, Timothy (Phleum head fully emerged from the panicle pratense L.) Tobacco Leaves 8 weeks old (Nicotiana tabacum L.) Tomatoes Leaves 8 weeks old (Lycopersicon esculentum Mill) Plant shoots Tropical pasture legumes in mixture with (Panicum muximum cv. Gatton) Whole tops at boot stage Wheat (Triticum Grain aestivum L.) Winter wheat Above ground plants at (Triticum ear emergence, when 40 aestivum L.) cm high

(0.15

<0.22 0.1-0.2

0.19 0. I

Gupta and MacKay (1968)

0.1 I

0. I3

"Considered critical. "Considered toxic. I03

References

1.08

Johnson et al. (1952)

0.68

Johnson et a / . ( I 952)

>0.02

Johansen (1978b)

0.09-0.18 0.16-0.20

Gupta (1971a) Gupta (1971a)

20.3

Neubert et al. (1970)

104

UMESH C. GUPTA AND JOHN LIPSETT

These symptoms, though definite, are not highly specific as is, for example, the whiptail symptom in cauliflower. In the field in Australia, where wheat is grown through winter, they resemble the effects of other possible agencies, such as frost and sap-sucking insects or mites. It seems possible that these other agencies may have been blamed in the past for what was really the result of Mo deficiency or NO,--Mo imbalance. Similarly, the widespread belief that nitrogenous fertilizers tend to reduce grain yield by promoting the depletion of water reserves needed for grain filling, may be better explained in terms of the Mo status of the young plant. B. TOXICITY SYMPTOMS

Molybdenum is not readily toxic to plants, and marked toxicity is not known in the field. Warington (1937, 1954) and Millikan (1949) made pioneering observations on toxicity under extreme experimental conditions; their evidence suggests that toxicity, marked by chlorosis and yellowing, involves interference with metabolism of Fe in the plant, and that species differ in sensitivity. Molybdenum is freely taken up and apparently normal plants may exhibit a considerable range of contents. As little as 0.1 ppm Mo in the dry matter may suffice, but several hundred parts per million may be found without obvious damage (Johnson, 1966; Jones, 1972). Such high values are unlikely in the field under ordinary conditions, but even 10 ppm is undesirable for ruminants. This level might conceivably occur following the application of fertilizers containing Mo or of lime to high-Mo acid soils. Soils naturally high in available Mo or contaminated by appropriate residues may also produce unacceptable forage. Gupta (1971a) has reported a reduction in yield of grain from wheat and barley following application of Mo. Application of Mo to leaf material of wheat in bioassays for nitrate reductase activity tends to reduce the activity if the material is from a plant that had already been well-supplied with Mo (J. Lipsett, CSIRO, Canberra, Australia, unpublished results). Apart from these instances of effects of Mo tending to reduce growth, the most likely impairment of function appears to be in seed germination. Molybdenum is often applied by coating or soaking seeds in media containing Mo compounds. This ensures economy and even distribution. However, Davies (1945) suggested in his original report on “whiptail” that seed-applied Mo reduces germination and wheat is similarly affected (Lipsett, unpublished results). When the seeds involved are leguminous and inocula are being applied simultaneously, the possibility arises of damaging the rhizobia. Some reports (Date and Hillier, 1968; Kerridge el al., 1973) approve the practice, but others as clearly condemn it (Giddens, 1964a; Gartrell, 1969), particularly if the Mo compounds and inocula remain in contact, as in dry soil.

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

105

IX. MOLYBDENUM TOXICITY AND MOLYBDENUM- COPPER-SULFUR INTERRELATIONSHIPS IN ANIMALS

Although an extensive volume of literature is available on the toxicity of Mo in livestock, data on the Mo requirements for livestock are very limited. Molybdenum is recognized as an essential element for animals, but a deficiency has never been developed or observed in cattle, indicating that the minimum requirement is very low. High levels of Mo in forages can induce Cu deficiencies in animals (Kubota, 1975; Vlek and Lindsay, 1977). This disorder, referred to as molybdenosis, occasionally results in death. Severe molybdenosis in cattle occurs under natural grazing conditions in many countries (Ferguson et al., 1943; Britton and Goss, 1946; Cunningham, 1950; Neenan et al., 1956; Hornick el al., 1977; Underwood, 1976). In some parts of Colorado it was found that using the irrigation water from areas with Mo-producing mines caused molybdenosis (Vlek and Lindsay, 1977). The disease is known by different names in various parts of the world; for example, in England it is known as “teart” and in New Zealand as “peat scours.” Species differences in tolerance to Mo are substantial. All cattle are susceptible to molybdenosis with milking cows and young stock suffering the most, with sheep next in susceptibility, whereas horses and pigs are the most tolerant farm livestock (Underwood, 1976). In general, nonruminants are less subject to Mo toxicity (Allaway, 1968). Scouring of cattle and sheep may occur in restricted areas where the pastures are subnormal in Cu and higher than normal in Mo, even though the Mo levels are well below those typical of teart herbage (Underwood, 1976). An amount of 20-100 ppm Mo in the herbage on a dry matter basis was related to scouring of cattle and sheep (Ferguson et al., 1943). The onset of scouring is delayed until the tissue Cu stores are depleted; thus, control of the scouring can be achieved merely by raising the Cu content of the pastures or the Cu intake of animals to normal levels (Dick, 1956). The toxicity of any level of dietary Mo is affected by the ratio of the dietary Cu to dietary Mo. The critical Cu : Mo ratio, with respect to the incidence of hypocuprosis, in animal feeds from western Canada was found to be 2 (Miltimore and Mason, 1971), whereas on some English pastures the ratio was reported to be closer to 4 (Alloway , 1973). Lower ratios than these can be expected to result in Mo-induced Cu deficiency. In some studies on beef cattle, corn silage rations containing 7 ppm Cu with a Cu : Mo ratio of 6.6 did not warrant supplementation, and this ratio was considered favorable (Pringle et al., 1973). In general, Cu concentrations <5 ppm in feeds and forages have been found to be deficient for most animals, whereas concentrations >20-30 ppm have been

106

UMESH C. GUPTA AND JOHN LIPSE'M

reported to be toxic to sheep but not to poultry, swine, or laboratory animals (Corbett and Leach, 1976). The Cu levels in the blood in the mature cattle suffering from molybdenosis were in the range 0.28-0.38 ppm compared with normal levels of 0.75-1.5 (Hornick et al., 1977). Exactly how Mo and Cu affect each other is not known, but it is certain that an imbalance in these two elements means trouble (Kubota, 1975). Whereas too much Mo induces Cu deficiency, too little Mo induces Cu toxicity, if Cu levels are high. Under naturally occumng conditions, a true Mo deficiency has never been reported in man or farm animals, and nutritional interest in Mo for animals is overwhelmingly concerned with its toxic effects and its interactions with Cu and S (Underwood, 1976). The dietary concentration of Mo required to produce Mo toxicity varies according not only to the Cu but also to the SO*'- concentrations in the diet. Swayback in lambs born of ewes that are Cu deficient during pregnancy was found on pastures not deficient in Cu and the incidence was due to an imbalance of Cu, Mo, and S (Todd, 1976). Various methods have been used to alleviate the scouring or Mo toxicity problem in cattle. Scouring has been found to be cured by feeding or drenching with CuSO, (Ferguson et al., 1943), by the acidification of soil by applying acidic fertilizers or by regular applications of S (Lewis, 1943), and by the addition of CuSO, to the feed or a subcutaneous injection of copper glycinate to the calves and adult cattle (Hornick et al., 1977). In Pennsylvania, the current practice includes supplemental feeding of ground corn mixed with CuS04 5 H20 (at 2.27 kg/ton) and fed at a rate of 0.9-1.35 kg/cow/day (Hornick et al., 1977). The oxidation in vivo of methionine- and cystine-S to SO;2- allows these S-containing compounds to be as effective as Cu in alleviating Mo toxicity; the resulting action is probably like that of inorganic SO;'-, namely reduced retention of Mo in the tissues and increased excretion in the urine (Underwood, 1976). Mills and Fell (1960) found that ewes receiving high levels of both SO;2- and Mo in the diet retained rather more Cu in their livers than the ewes of the group receiving SO;2- only. However, their results showed that the transfer of Cu from the ewe to the lamb was diminished by feeding Mo. An increment of 4 mg Mo/kg feed was sufficient to reduce the availability of dietary Cu by 50%, whereas an increment of 1 g S/kg feed had a similar effect (Suttle, 1975). Ovine hypocuprosis has been found in New South Wales, Australia, on pastures containing 3-10 pprn Cu, 3-9 ppm Mo, and 0.1-0.7% (Wynne and McClymont, 1955). These workers have suggested that it is not so much the absolute intake of Cu, Mo, and So,'- that is important, but rather the ratios in the diet. Studies by Scott (1972) indicated that neither Mo nor SO;;?alone interfered with Cu retention and that the effectiveness of either one increased to a maximum as the intake of the other was increased. Inorganic Soil- and organic S used as sources of S to inhibit Cu depletion and decrease plasma Mo concentrations were found to be similar (Suttle, 1975).

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

107

Suttle (1975) proposed that the Cu X Mo X S antagonism involved a lowering of the availability of both Cu and Mo in the rumen and that inorganic and organic S could potentiate the process. Ruminants differ from nonruminants in having the ability to reduce Soil- to S2-, which may either be absorbed and detoxified in the liver or incorporated into S amino acids (Huisingh and Matrone, 1976). High levels of Mo and Soil- restrict Cu utilization in animals by depressing Cu solubility in the digestive tract through the precipitation of insoluble CuS. Molybdenum has been found to interact with organic S, as well as inorganic S, in limiting the utilization of dietary Cu by sheep (Suttle, 1974). The effect of dietary Mo is dependent on dietary levels of Soil-,particularly in ruminants (Huisingh and Matrone, 1976). It should be noted that Mo plus S04z- can either increase or decrease the Cu status of an animal, depending on their intakes relative to that of Cu (Underwood, 1976). Chronic Cu poisoning can occur in sheep with moderate Cu intakes and very low levels of Mo and SOi2-. In Northern Ireland, the incidence of Cu poisoning during the 1960s increased sometimes in calves, but mostly in sheep, and was found to be due to the use of CuS04 as a growth promoter in pig meal-this meal on occasions was being fed to sheep or calves (Todd, 1976). Dick (1954), on the other hand, discovered that Mo limits Cu retention only in the presence of adequate dietary or endogenous SOP-. Depletion of the animal’s Cu reserves to the extent of clinical symptoms of Cu deficiency can arise on normal Cu and high Mo plus SOL2- intakes (Wynne and McClymont, 1955).

X. SUMMARY AND CONCLUSIONS This review discusses the place now occupied by Mo in the agronomy that services the agriculture on which the world depends for food. The detailed results from a number of disciplines are shown to be integrated into a mature body of knowledge concerning Mo in agricultural practice. In addition, it indicates some of the likely biological effects in the biosphere, and hence broader applications of scientific effort. The earliest studies, some 50 years ago, established Mo as an essential nutrient, first for bacteria in fixing N, and later for higher plants in reducing NO3-. Deficiencies were soon found in many species and in several countries, and symptoms became well known. The research that followed has established detailed information about sources, supply, and uptake of Mo, on the one hand, and function in the plant, on the other. Sodium and ammonium molybdates are common Mo fertilizers, but other compounds are being investigated to counter costs and shortages. Application to the seed is cheaper and more uniform than soil application but may unduly raise

108

UMESH C. GUPTA AND JOHN LIPSE'IT

levels of Mo in the plant. Foliar application is effective and fast-acting. The practice of adding solid Mo compounds to another fertilizer as carrier for application to the soil can lead to uneven distribution. The residual value of Mo applications to the soil is not clear, and this topic will become more important as the cost of Mo rises. The main physiological role of Mo in plants appears to be its participation in two important enzyme systems, namely nitrogenase and nitrate reductase, but others may also be involved. (Both systems affect higher plants since the nitrogenase of Rhizobium is the basis of the nitrogen economy of nodulated legumes.) Hence protein synthesis depends on Mo. Surprisingly, there is little recent work on the role of Mo. The Mo concentration of plants is an important index of the supply from the soil, since some of the total Mo usually is not available. Chemical measures of availability are of varying quality, so attempts to improve such methods continue. Resin extraction and a microbiological assay based on nitrogenase (which also catalyzes the reduction of acetylene) have recently been suggested. The dithiol method is the best of the wet methods of analysis, and atomic absorption coupled with a graphite furnace is the best of the physical methods. Any method must be extremely sensitive to cope with the low amounts of Mo in deficient plant material. The reported instances of responses to Mo in agricultural practice, beginning with clover, for a number of species were consolidated within the first 20 years. Since then the main development has concerned cereals, particularly in Australia, and Cole crops in Great Britain and Canada. Maize evidently is quite susceptible to Mo deficiency, particularly on acid soils and if seed reserves of Mo are low. The response to Mo may take the form of an alleviation of damage to seedlings due to excess NO,-. Changes in nitrate reductase activity and overall content of Mo are associated with yield responses. Geographically, most of the responses have been reported from lightly acid soils in temperate countries. However, there are reports of deficiencies from the wetter parts of tropical Australia, and leached soils in other parts of the tropics are fairly clearly at risk. The geochemistry and factors that influence supply and uptake of Mo are discussed. Molybdenum is a metallic transition element and is widespread in the Earth's crust, usually in low concentrations (<3 ppm). The predominant primary mineral is the disulfide molybdenite (MoS,), which forms readily and is highly insoluble but also weathers readily under oxidizing conditions, typically to molybdate (Mo0i2-). Molybdenum is relatively mobile in the soil and follows a sedimentary cycle. Increasing industrial use-in fertilizers, alloys, catalysts, and lubricants-has led to increasing contamination. The anionic form and direct correlation of solubility with pH distinguish Mo from the other trace metals in plant nutrition. Plants take up Mo0i2- readily, except at low pH and in some highly organic soils. Other organic soils may

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

109

supply excessive amounts, particularly if wet. Sulfate reduces uptake of Mo and PO,3- usually increases it. Different species and cultivars within species may differ in their uptake of Mo from the same soil. Most of the Mo is probably to be found in vegetative parts, but seeds carry highly variable amounts as reserves. Deficiency of Mo may result from low overall amounts in the soil or from unavailability, usually associated with low pH. The latter is related primarily to adsorption reactions with active oxides of Fe and possibly Al, which are positively charged at the levels of pH involved. Manganese, which was formerly implicated in the loss of availability because Mo is commonly found occluded in deposits or nodules along with Mn, is probably involved only incidentally in soils. Whatever the precise mechanism, liming to shift the pH is well established as a corrective measure. Deficiency levels and symptoms are of more concern than Mo toxicity, which is rare and virtually confined to germinating seeds. However, even moderate amounts of Mo in forage can harm ruminants. Sufficiency and deficiency levels of Mo are reported for a number of species. The visual symptoms of deficiency are related to nitrogen status. In cereals and grasses the symptoms may be essentially those of NO,- accumulation, which can be reduced by supplying Mo. Many crop plants are consumed by animals, and the Mo status of the plant can affect the animal that eats it. Deficiencies of Mo in livestock have not been reported, but toxicity (molybdenosis) does occur and is associated with disturbance of Cu metabolism. Ruminants are mainly at risk. Toxicity of any level of dietary Mo is affected by the ratio of the dietary Mo to dietary Cu and by the S status. High Mo relative to Cu can induce symptoms of Cu deficiency, and Mo can protect against Cu toxicity. Sulfate affects the balance, both indirectly through reduction in the uptake of Mo and directly through being present in the diet. In the latter case, S2- is probably formed in the rumen from SO,‘- or organic S, and Mo and Cu then coprecipitate. The topic shows a consistency and breadth of approach that have effectively consolidated knowledge of Mo as an essential element. The recent advances have furthered this synthesis and established the position of agricultural uses relative to other industrial uses. Future research on Mo in agriculture should be about interaction effects with a range of other elements rather than gross deficiency. In some parts of the world where Mo has been substituted for lime the soils have become more acidic, thus making it difficult to crop on such soils. Liming in these areas will become a major topic and it might be suggested that the “safe” soils and crops may be less safe than hitherto believed. REFERENCES Agafonova, A. F., Danilova, T. A , , and Kayurova, L. V. 1975. Khim. Selsk. Khoz. 13, 826-833 (in Chem. Abstr. 84, 42383s).

110

UMESH C. GUPTA AND JOHN LIPSETT

Agarwala, S. C. 1952. Nature (London) 169, 1099. Agarwala, S . C., and Hewitt, E. J . 1955. J. Hortic. Sci. 30, 163-180. Agarwala, S. C., and Sharma, C. P. 1979. “Recognizing Micronutrient Disorders of Crop Plants on the Basis of Visible Symptoms and Plant Analysis,” p. 72. Prem, Lucknow, India. Aganvala, S. C., Sharma, C. P., Farooq, S., andchatterjee, C. 1978. C a n . J. Bot. 56, 1905-1908. Ahlrichs, L. E., Hanson, R . G . , and MacGregor, J . 54. 1963. Agron. J. 55, 484-486. Allaway, W. H. 1968. A h . Agron. 20, 235-274. Allaway, W. H. 1977. /ti “Molybdenum in the Environment” (W. R. Chappell and K . K . Petersen, eds.), Vol. 2, pp. 317-339. Dekker, New York. Alloway, B. J . 1973. J. Agric. Sci. 80, 521-524. Amin, J . V., and Joham, H. E. 1958. Soil Sci. 85, 156-160. Anderson, A . J . 1942. J. Aust. Inst. Agric. Sci. 8 , 73-75. Anderson, A. J. 1956. Adv. Agroiz. 8, 163-202. Anderson, A. J . . and Spencer, D. 1950. Ausr. J. Sci. Res. 3, 414-430. Arnon, D. I., and Stout, P. R. 1939. Plant Physiol. 14, 599-602. Barber, S. A,, Halstead, E. H . , and Follett, R. F. 1966. Trans. Comm. I I , IV, Inr. Soc. Soil Sci.. Aherrleen pp. 299-304. Barrow, N . J . , and Spencer, K . 1971. Aust. J. Exp. Agric. Anim. Hush. 11, 670-676. Barshad, I . 1951. Soil Sci. 71, 387-398. Bergeaux, P. J . 1976. “The Need for Micronutrients in Georgia Agriculture.” Bull. No. 784, pp. 1-18. Univ. Georgia, Coil. Agric. Coop. Ext. Service. Berger, K . C. 1962. Agric. Food Chern. 10, 178-181. Berger, K . C., and Pratt, P. F. 1965. I n “Fertilizer Technology and Usage” (M. H. McVickar, G. L. Bridger. and L. B. Nelson, eds.), pp. 313-317. Soil Sci. SOC. Am., Madison, Wisconsin. Bergmann, W. 1976. “Emahrungsstorungen bei Kulturpflanzen in Farbhildem,” pp. 164-172. Fischer, Jena. Bhella, H. S . , and Dawson, M. D. 1972. Proc. Soil Sci. Soc. Am. 36, 177-179. Bingley, J. B. 1963. Agric. Food Cheni. 11, 130-131. Bishop, H. G. 1974. Queens/. J. Agric. Anim. Sci. 31, 329-335. Bloomfield, C., and Kelso, W .I . 1973. J. Soil Sci. 24, 368-379. Bortels, H. 1930. Arch. Mikrobiol. 1, 333-342. Bortels, H. 1936. Zenrralhl. Bakteriol. Parasitenkd. 95, 193-218. Boswell, F. C., Anderson, 0. E., and Welch, L. F. 1967. “Molybdenum Studies with Peanuts in Georgia.” Bull. No. 9, pp. 5-23. Univ. Georgia, College Agric. Exp. Sta. Res. Britton, J . W., and Goss, H. 1946. J. Am. V r f . M e d . Assoc. 108, 176-178. Brogan, J . C., Fleming, G. A,, and Byrne, J . E. 1973. I r . J. Agric. Res. 12, 71-81. Brown, J . C., and Clark, R. B. 1974. Proc. Soil Sci. Soc. A m . 38, 331-333. Bussler, W. 1970. Z . Pjlanzenernaehr. Boderzkd. 125, 36-50. Chatt, J. 1974. J. Less Common Merals 36, 429-435. Chesnin, L. 1972. I n “The Fertilizer Handbook” (W. C. White and D. N . Collins, eds.), pp. 65-84. Fertilizer Institute, Washington, D.C. Childs. C. W., and Leslie, D. M. 1977. Soil Sci. 123, 369-376. Chipman, E. W., MacKay, D. C., Gupta, U . C., and Cannon, H. B. 1970. C a n . J. Planr Sci. 50, I 63- 167. Clark, L. J., and Axley, J . H. 1955. A d . Chem. 27, 2000-2003. Corbett, W. S., and Leach, R. M., Jr. 1976. Sci. Agric. (Penn. Stare U n i v . ) 23(3), 2. Cunningham, I. J. 1950. I n “A Symposium on Copper Metabolism” (W. D. McElroy and B. Goss, eds.), pp. 246-273. Johns Hopkins Press, Baltimore, Maryland. Cyanamid of Canada Limited. 1970. Tech. Bull. I, 41. Dann, P. D. 1969. Aust. J. Exp. Agric. Anim. Hush. 9, 625-629. Das Gupta, D. K . , and Basuchaudhuri, P. 1977. Plant Soil 46, 681-685.

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

111

Date, R . A., and Hillier, G . R. 1968. J . Ausr. Inst. Agric. Sci. 34, 171-172. Davies, E. B. 1945. Nature (London) 156, 392-393. Davies, E. B. 1956. Soil Sci. 81, 209-221. Dekhkankhodzhayeva, S. Kh., and Kruglova, YE. K. 1972. Agrukhimiyu 4, 109-1 13. (Soviet Soil Sci. 1972, 247-251.)

de Mooy, C . J. 1970. Agron. J. 62, 195-197. Dick, A . T. 1954. Aust. J . Agric. Res. 5, 51 1-544. Dick, A . T. 1956. In “A Symposium on Organic Nitrogen Metabolism” (W. D. McElroy and B. Glass, eds.), pp. 445-473. Johns Hopkins Press, Baltimore, Maryland. Dios, R. V . , and Broyer. T . V. 1965. Agrochimicu 9, 273-284. Donald, C . , Passey, B . I., and Swaby, R. J. 1952. Aust. J. Agric. Res. 3, 305-325. Donald, C . M., and Williams, C. H. 1954. Aust. 1. Agric. Res. 5, 661-687. Doyle, R. J., Parkin, R. J . , Smith, J. A. C . , and Gartrell, J. W . 1965. J. D e p . Agric. West. Aust. 6, 699-703.

Dye, W . B., and O’Harra, J. L. 1959. “Molybdenosis.” U n i v . Nevada Agric. Exp. Sta. Bull. No. 208, p. 32. Evans, H. J . 1956. Soil Sci. 81, 199-208. Evans, H. J . , and Purvis, E. R. 1951. Agron. J. 43, 70-71. Everist, S. L. 1974. “Poisonous Plants of Australia,” p. 684. Angus & Robertson, Sydney. Ferguson, W. S . , Lewis, A . H . , and Watson, S. J. 1943. J. Agric. Sci. 33, 44-51. Fernandez, J. M. L . , Perez-Bendito, D., and Valcarcel, M. 1978. Anul.yst 103, 1210-1214. Franco, A. A , , Peres, J. R. R., and Nery, M. 1978. Plant Soil 50, 1-1 1 . Freney, J. R., and Lipsett, J. 1965. Nurure (Londun) 205, 616-617. Fricke, E. F. 1947. J. Aust. Inst. Agric. Sci. 13, 75-76. Fuge, R. 1970. Analyst 95, 171-176. Gammon, N . , Jr., Volk, G. M., McCubbin, E. N., and Eddins, A. H. 1954. Proc. Soil Sci. Soc. A m . 18, 302-305. Gartrell, J. W . 1966. Nature (Londun) 209, 1050. Gartrell, J. W. 1969. Ausr. J. Ex,. Agric. Anim. Husb. 9, 432-436. Giddens, J. 1964a. Agron. J . 56, 362-363. Giddens, J. 1964b. F e r t . Sulutions 7(5), 18-19, 21. Giddens, J . , and Perkins, H. F. 1972. Agron. J. 64, 819-820. Golov, V. I . , and Kazakhkov, YU. N. 1973. Agrokhimiyu 10, 103-109. (Soviet Soil Sci. 1973, 551-558).

Grigg, J . L. 1953a. Analyst 78, 470-473. Grigg, J . L. 1953b. N. Z . J. Sci. Techno/. See. A . 34, 405-414. Gupta, U. C . 1969. Proc. Soil Sci. Soc. Am. 33, 929-932. Gupta, U. C. 1970. Soil Sci. 110, 280-282. Gupta, U. C. 1971a. C a n . J. Soil Sci. 51, 415-422. Gupta, U. C. 1971b. Plant Soil 34, 249-253. Gupta, U. C . 1979. C a n . J. Soil Sci. 59, 183-189. Gupta, U. C., and Cutcliffe. J. A. 1968. C a n . J. Soil Sci. 48, 117-123. Gupta, U. C., and Kunelius, H. T. 1980. C a n . J. Plant Sci. 59, 113-120. Gupta, U. C . , and MacKay, D. C. 1965a. Soil Sci. 99, 414-415. Gupta, U. C., and MacKay, D. C. 1965b. Proc. Soil Sci. Soc. Am. 29, 323. Gupta, U. C . , and MacKay, D. C. 1968. Can. J. SoilSci. 48, 235-242. Gupta, U. C . , and MacLeod, L. B. 1975. SuilSci. 119, 441-447. Gupta, U. C . , and Munro, D . C . 1969. SoilSci. 107, 114-118. Gupta, U . C . , Chipman, E. W . , and MacKay, D. C . 1978. Can. J. Plant Sci. 58, 983-992. Gurley, W. H., and Giddens, J. 1969. Agron. J. 61, 7-9. Hagstrom, G. R., and Berger, K. C. 1965. Soil Sci. 100, 52-56.

112

UMESH C. GUPTA AND JOHN LIPSETT

Hanway, J . J., and Englehorn, A. J. 1958. Agron. J . 50, 331-334. Hawes, R. L., Sims, J . L., and Wells, K . L. 1976. Agron. J . 68, 217-218. Henkens, C. H., and Smilde, K. W. 1966. Neth. J . Agric. Sci. 14, 165-177. Henning, S . , and Jackson, T. L. 1973. Atom. Absorpr. Newsl. 12, 100-101. Hewitt, E. J. 1951. Annu. Rev. Plant Physiol. 2, 25-52. Hewitt, E. J . 1956. Soil Sci. 81, 159-171. Hewitt, E. J., and Jones, E. W. 1948. Annu. Rep. Long Ashton, Brisrol Res. Sta. pp. 81-90, Hornick, S. B., Baker, D. E., and Guss, S. B. 1977. In “Molybdenum in the Environment”(W. R. Chappell and K . K . Petersen, eds.), Vol. 2, pp. 665-684. Dekker, New York. Huang, P. M., and Jackson, M. L. 1965. Proc. Soil Sci. SOC. Am. 29, 661-665. Huisingh, J., and Matrone, G. 1976. In “Molybdenum in the Environment” (W. R. Chappell and K . K . Petersen, eds.), Vol. I , pp. 125-148. Dekker, New York. Inden, T. 1975. Ext. Bull. No. 5 5 , ASPAC Food Fert. Technol. Cent. (Asian Pac. Council), pp. 1-20. Jackson, D. R., Lindsay, W. L., and Heil, R. D. 1975. J . Environ. Quai. 4, 223-229. James, D. W., Jackson, T. L., and Harward, M. E. 1968. Soil Sci. 105, 397-402. Jarrell, W. M., and Dawson, M. D. 1978. Soil Sci. SOC. Am. J . 42, 412-415. Johansen, C. 1978a. Aust. J . Exp. Agric. Anim. Hush. 18, 732-736. Johansen, C. 1978b. Commun. Soil Sci. Planr Anal. 9, 1009-1017. Johnson, C. B., Whittington, W. J., and Blackwood, G. C. 1976. Nature (London) 262, 133-134. Johnson, C. M. 1966. In “Diagnostic Criteria for Plants and Soils” (H. D. Chapman, ed.), pp. 286-301. Univ. of Calif. Press, Riverside. Johnson, C. M., Pearson, G. A,, and Stout, P. R. 1952. Planr Soil 4, 178-196. Jones, J. B., Jr. 1967. “Soil Testing and Plant Analyses,” Spec. Publ. No. 2, pp. 49-58. Plant Analyses, Part 11. SSSA. Jones, J . B . , Jr. 1972. In “Micronutrients in Agriculture” (J. J. Mortvedt, P. M. Giordano, and W. L. Lindsay, ‘eds.), pp. 319-346. Soil Sci. SOC.Am., Madison, Wisconsin. Jones, L. H. P. 1956. Science 123, I 1 16. Jones, L. H. P. 1957. 1.SoilSci. 8, 313-327. Jones, R. K., and Crack, B. J. 1970. Aust. J . Exp. Agric. Anirn. Hush. 10, 342-349. Kannan, S., and Ramani, S. 1978. Plant Physiol. 62, 179-181. Karimian, N . , and Cox, F. R. 1978. Soil Sci. SOC. Am. J. 42, 757-761. Karimian, N . , and Cox, F. R . 1979. Agron. J . 71, 63-65. Kemdge, P. C., Cook, B. G . , and Everett, M. L. 1973. Trop. Grassl. 7, 229-232. Khan, S . U . , Cloutier, R. O., and Hidiroglou, M. 1979. J . Assoc. Off. Anal. Chem. 62, 1062-1064. Koch, B . , Evans, H. J., and Russell, S. 1967. Plant Physiol. 42, 466-468. Koval’skiy, V. V . , and Yarovaya, G.A. 1966. Agrokhimiya 8, 68-91. Krauskopf, K. B. 1972. I n “Micronutrients in Agriculture” (J. J. Mortvedt, P. M. Giordano, and W. L. Lindsay, eds.), pp. 7-40. Soil Sci. SOC.Am., Madison, Wisconsin. Kubota, J. 1972. In “Geochemical Environment in Relation to Health and Disease” (H. C. Hopps and H. L. Cannon. eds.), pp. 105-1 15. Geological SOC.Am., Boulder, Colorado. Kubota, J. 1975. Soil Conservation 40(9), 18-21. Kubota, J., and Allaway, W. H. 1972. In “Micronutrients in Agriculture” (J. J . Mortvedt, P. M. Giordano, and W. L. Lindsay, eds.), pp. 525-554. Soil Sci. SOC.Am., Madison, Wisconsin. Kubota, J., Lazar, V. A., Langan, L. N., and Beeson, K. C. 1961. Proc. Soil Sci. SOC. A m . 25, 227-232. Kubota, J., Lemon, E. R., and Allaway, W. H. 1963. Proc. Soil Sci. SOC. Am. 27, 679-683. Lavy, T. L., and Barber, S. A. 1964. Proc. Soil Sci. SOC. Am. 28, 93-97. Lehr, J . R. 1972. In “Micronutrients in Agriculture” (J. J. Mortvedt, P. M. Giordano, and W. L. Lindsay, eds.), pp. 459-503. Soil Sci. SOC.Am., Madison, Wisconsin.

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

I13

Lewis, A. H. 1943. J . Agric. Sci. 33, 58-63. Lindsay, W. L. 1972. I n “Micronutrients in Agriculture” (J. J. Mortvedt, P. M . Giordano, and W. L. Lindsay, eds.), pp. 41-57. Soil Sci. SOC.Am., Madison, Wisconsin. Lipsett, J. 1975. Ausf. J . Exp. Agric. Anim. Husb. 15, 227-230. Lipsett, J., and David, D. J . 1977. J . Aust. Inst. Agric. Sci. 43, 149-151. Lipsett, J., and Simpson, J. R. 1971. J . Aust. Inst. Agric. Sci. 37, 348-351. Lipsett, J. and Simpson, J . R. 1973. Ausi. J. Exp. Agric. Anbn. Husb. 13, 563-566. Little, I. P., and Kenidge, P. C. 1978, Soil Sci. 125, 102-106. Lobb, W. R. 1953. N . Z . J . Agric. 87, 3-11. Lowe, R. H., and Massey, H. F. 1965. Soil Sci. 100, 238-243. Lucas. R. E., and Knezek, B. D. 1972. I n “Micronutrients in Agriculture” (J. J. Mortvedt, P. M. Giordano, and W. L. Lindsay, eds.), pp. 265-288. Soil Sci. SOC.Am., Madison, Wisconsin. McLachlan, K . D. 1955. Aust. J . Agric. Res. 6 , 673-684. MacLean, K . S., and Langille, W. M. 1973. Commun. Soil Sci. Planr Anal. 4, 495-505. McLeod, C.C. 1976. N . Z . J . Agric. 133(2), 65-69. Malonosova, 1. A. 1968. Agrokhimiya 10, 144-147. Manheirn, F. T.. and Landergren. S. 1978. I n “Handbook of Geochemistry” (K. H. Wedepohl, ed.), Vol. 11-4, pp. 42-B-1-42-0-2. Springer-Verlag, Berlin and New York. Martinez, J . F., Robertson, W. K., and Martin, F. G. 1977. Proc. Soil Crop Sci. Soc. Fla. 36, 58-60. Meagher, W. R., Johnson, C. M., and Stout, P. R. 1952. Plani Physiol. 27, 223-230. Melsted, S. W.. Motto, H. L., and Peck, T. R. 1969. Agron. J . 61, 17-20. Menary, R. C., and Jones, R. H. 1972. Aust. J . Biol. Sci. 25, 531-542. Merkel, D., Witt, H. H., and Jungk, A. 1975. Plant Soil 42, 131-143. Millikan, C. R. 1947. J . Ausr. Inst. Agric. Sci. 13, 180-186. Millikan, C. R. 1949. Proc. R. Soc. Victoria 61, 25-42. Mills, C. F., and Fell, B. F. 1960. Nature (London) 185, 20-22. Miltimore. J. E., and Mason, J. L. 1971. Can. J . Anim. Sci. 51, 193-200. Mitchell, K . J. 1945. N.Z. J . Sci. Technol. A-27, 287-293. Mitchell, R. L. 1964. In ”Chemistry of the Soil” (F. E. Bear, ed.), 2nd Ed., pp. 320-368. Van Nostrand-Reinhold, Princeton, New Jersey. Mitchell, R. L. 1974. Nerh. J . Agric. Sci. 22, 295-304. Moore, R. M., and Hutchings, R. J. 1967. Aust. J . Exp. Agric. Anim. Husb. 7, 17-21. Mulder, E. G. 1948. Plant Soil 1, 94-1 19. Mulder, E. G. 1954. Plan1 Soil 5, 368-415. Murphy, L. S., and Walsh, L. M. 1972. In “Micronutrients in Agriculture” (J. J. Mortvedt, P. M. Giordano, and W. L. Lindsay, eds.), pp. 347-387. Soil Sci. SOC.Am., Madison, Wisconsin. Neenan, M., Walsh, T., and O‘Moore, L. B. 1956. Proc. Inf. Crassl. Congr. 7rh. N . N . Z . Paper 31A.

Neubert, P., Wrazidlo, W., Vielemeyer, H. P., Hundt, I . , Gollmick, Fr., and Bergmann, W. 1970. Tabellen zur Pflanzenanalyse-Erste orientierende Ubersicht,” pp. 1-40. Inst. fur Pflanzenernahrung, Jena, der Deutschen Akademie der Landwirtschaftswissenschaften zu Berlin. Jena, Naumburger Strasse 98. Nicholas, D. J. D. 1975. In “Trace Elements in Soil-Plant-Animal System’’ (D. J. D. Nicholas and A. R. Egan, eds.), pp. 181-198. Academic Press, New York. Noonan, J. B. 1953. Agric. C a z . N.S.W. 64, 422-424. Nonish, K . 1975. In “Trace Elements in Soil-Plant-Animal Systems.” Proc. Jubilee Symp. Waite Agric. Res. Inst. (D. 1. D. Nicholas and A. R. Egan, eds.), pp. 55-81. Academic Press, New York. Pal, U . R., Gossett, D. R., Sims, J. L., and Leggett, J. E. 1976. Can. J . Bor. 54, 2014-2022.

114

UMESH C. GUPTA AND JOHN LIPSETT

Parker, M. B.. and Harris, H. B. 1977. Agron. J. 69, 551-554. Pasricha, N. S., and Randhawa, N . S. 1972. Plant Soil 37, 215-220. Pathak, A. N., Shankar, H., and Misra, R. V. 1969. J. Inclinn Soc. Soil Sci. 17, 151-153. Perrin, D. D. 1946. N . Z. J. Sci. Technol. 28, 183-187. Perrott, K. W., Smith, B. F. L., and Inkson, R . H. E. 1976. J. Soil Sci. 27, 58-67. Peterson, N . K., and Purvis, E. R. 1961. Proc. SoilSci. Soc. Am. 25, 111-117. Peyve, YA. V . 1969. Agrokhitniva 1, 61-74. Piper, C. S., and Beckwith, R. S. 1948. J. Sac. Chern. Ind. (London) 67, 374-379. Plant, W. 1951a. J. Horr. Sci. 26, 109-1 15. Plant, W. 1951b. Long Ashton Res. Sra. Annu. Rep. pp. 113-1 15. Plant, W. 1952. Trans. Int. Soc. Soil Sci., Comm. 11. I V 2 , 176-177. Prasad, N. R., and Pagel, H. 1976. Beitr. Trop. Landwirtsch. Veterinarmed. 14, 79-87. Price, C. A,, Clark, H. E., and Funkhouser, E. A. 1972. In “Micronutrients in Agriculture” ( J . J . Mortvedt, P. M. Giordano, and W. L. Lindsay, eds.), pp. 231-242. Soil Sci. SOC.Am., Madison, Wisconsin. Pringle, W. L., Dawley, W. K., and Miltimore, J. E. 1973. Can. J . Anim. Sci. 53, 497-502. Purvis, E. R., and Peterson, N. K . 1956. Soil Sci. 81, 223-228. Randall, P. J . 1969. Aust. J. Agric. Res. 20, 635-642. Reddy, G. R. 1964. Indian J. Agric. Sci. 34, 219-233. Reisenauer, H. M. 1956. Soil Sci. 81, 237-242. Reisenauer, H. M. 1963. Agron. J. 55, 459-460. Reisenauer, H. M. 1965. I n “Methods of Soil Analyses” (C. A. Black ed.), Part 2, Agronomy 9, pp. 1050-1058. Soil Sci. SOC.Am., Madison, Wisconsin. Reisenauer, H. M. 1967. “Soil Testing and Plant Analysis,” Part I , SSSA Spec. Publ. No. 2, pp. 7 1 - 102. Reisenauer, H. M., Tabikh. A. A. , and Stout, P. R. 1962. Proc. Soil Sci. Soc. A m . 26, 23-27. Reyes, E. D., and Jurinak, J . J. 1967. Proc. Soil Sci. Soc. Am. 31, 637-641. Robinson, D. B., and Campbell, J . E. 1956. C a n . J. Agric. Sci. 36, 241-244. Robinson, D. B., LeLacheur, K . E.. and Brossard, G . A. 1957. C a n . J. Plant Sci. 37, 193-195. Robinson, W. 0.. and Alexander, L. T. 1953. Soil Sci. 75, 287-291. Rolt, W. F. 1968. N . Z . J. Agric. Res. 11, 193-205. Rubins, E. J. 1956. SoilSci. 81, 191-197. Schrauzer, G. N . 1976. In “Molybdenum in the Environment” (W. R. Chappell and K . K. Petersen, eds.), Vol. I , pp. 243-265. Dekker, New York. Scott, M. L. 1972. I n “Micronutrients in Agriculture” (J. J. Mortvedt, P. M. Giordano, and W. L. Lindsay, eds.), pp. 555-591. Soil Sci. SOC.Am., Madison, Wisconsin. Sedberry, 1. E., Charmaputra, T. S., Brupacher. R. H.,Phillips, S. A,, Marshall, J . G., Sloane, L. W., Melville, D. R., Rabb, J . L., and Davis, J. H. 1973. “Molybdenum Investigations with Soybeans in Louisiana.” Bull. No. 670, La. State Univ. and A and M Coll. Shukla, P., Singh, M. P., and Pathak, A. N . 1976. Indian J. Agron. 21, 177-179. Sims, J. L., Leggett, J. E., and Pal, U. R. 1979. Agron. J. 71, 75-78. Singh. M., and Kumar, V . 1979. Soil Sci. 127, 307-312. Sisodia, A. K., Sawarkar, N. J., and Rai, M. M. 1975. J. Indian SOC. Soil Sci. 23, 96-102. Smith, B. H., and Leeper, G. W. 1969. J. Soil Sci. 20, 246-254. Steinberg, R. A. 1936. J. Agric. Res. 52, 439-448. Stone, K. L., and Jencks, E. M. 1963. W. Virginia Univ.. Agric. Exp. Stn. Bull. No. 484, p. 10. Stonier, R. 1965. Aust. J. Exp. Agric. Anim. Husb. 5 , 317-322. Stout, P. R. 1972. I n “Micronutrients in Agriculture” (J. J. Mortvedt, P. M. Giordano, and W. L. Lindsay, eds.), pp. 1-5. Soil Sci. SOC.Am., Madison, Wisconsin. Stout, P. R., and Meagher, W. R. 1948. Science 108, 471-473.

MOLYBDENUM IN SOILS, PLANTS, AND ANIMALS

115

Stout, P. R., Meagher, W. R.. Pearson, G . A., and Johnson, C. M. 1951. Plant Soil 3, 51-87. Suttle, N. F. 1974. In “International Symposium on Trace Element Metabolism in Animals’’ (W. G. Hoekstra, J. W. Suttle, H. E. Ganthier, and W. Mertz, eds.), 2nd Ed., pp. 612-614. Univ. Park Press, Baltimore. Maryland. Suttle, N . F. 1975. In “Trace Elements in Soils Plant-Animal Systems” (D. J. D. Nicholas and A. R. Egan, eds.), pp. 271-289. Academic Press, New York. Tanner, P. D. 1976. Rhod. J. Agric. Res. 14, 95-99. Tanner, P. D. 1978. Plant Soil 49, 427-432. Thornton, I . , and Webb, J. S. 1973. In “Environmental Geochemistry: Some Recent Studies in the United Kingdom,” pp. 89-98. Proc. 7th Conf. on Trace Substances in Environmental Health. Univ. of Missouri, Columbia. Tiffin, L. 0. 1972. I n “Micronutrients in Agriculture” (J. J. Mortvedt, P. M. Giordano, and W. L. Lindsay, eds.), pp. 199-229. Soil Sci. Soc. Am., Madison, Wisconsin. Todd, J. R. 1976. I n “Molybdenum in the Environment” (W. R. Chappell and K . K. Petersen, eds.), Vol. 1, pp. 33-49. Dekker, New York. Ulrich, A., and Hills, F. J. 1973. I n “Soil Testing and Plant Analysis’’ (L. M. Walsh and J. D. Beaton, eds.), 2nd ed., pp. 271-288. Soil Sci. SOC.Am., Madison, Wisconsin. Underwood, E. J. 1976. I n “Molybdenum in the Environment” (W. R. Chappell and K. K. Petersen, eds.), Vol. 1, pp. 9-31. Dekker, New York. Valdek, E. 1974. T e d . Toode Kogumik Eesti Maaviljeluse Maaparanduse Tead. 32, 12-28 (in Chem. Abstr. 84, 163495). Vlek, P. L. G., and Lindsay, W. L. 1974. Agron. Abstr., p. 126. Am. Soc. Agron., Madison, Wisconsin. Vlek, P. L. G., and Lindsay, W. L. 1977. I n “Molybdenum in the Environment” (W. R. Chappell and K . K. Petersen, eds.), Vol. 2, pp. 619-650. Dekker, New York. Wallace, A,, and Romney, E. M. 1977. I n “Biological Implications of Metals in the Environment” (A. Wallace and E. M. Romney, eds.), pp. 370-379. ERDA Symp. Series 42. Wallace, T. 1951. “The Diagnosis of Mineral Deficiencies in Plants by Visual Symptoms,” p. 125, plus plates. H.M.S.O., London. Wardlaw, I. F. 1968. Bot. Rev. 34, 79-105. Warington, K. 1937. Ann. Appl. Biol. 24, 475-493. Warington, K. 1954. Ann. Appl. Biol. 41, 1-22. Warren, H. V. 1973. I n “Environmental Medicine” (G. M. Howe and J. A. Loraine, eds.), Ch. 2, pp. 9-24. Heinemann, London. Weir, R. G., and Hudson, A., 1966. Aust. J. Exp. Agric. Anim. Husb. 6, 35-41. Weir, R. G., Noonan, J. B., and Boyle, J. W. 1966. Agric. Gaz. N.S.W. 77, 578-582. Weir, R. G., Nagle, R. K., Noonan, J. B., andTowner, A. G. W. 1976. Ausr. J. Exp. Agric. Anim. Husb. 16, 761-764. Widdowson, J. P. 1966. N . Z . J . Agric. Res. 9, 59-67. Williams, C., and Thornton, I. 1972. Plant Soil 36, 395-406. Williams, J. H.1971. “Trace Elementsin Soilsand Crops,”pp. 119-136. Minist. Agric. Fish. Food (U.K.),Tech. Bull. No. 21. Wilson, D. 0. 1979. Commun. Soil Sci. Plant Anal. 10, 1319-1330. Wynne, K. N., and McClymont, G. L. 1955. Nature (London) 175, 471-472.

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY, VOL. 34

INTERGENOMIC INTERACTION, HETEROSIS, AND IMPROVEMENT OF CROP YIELD H. K. Srivastaval Department of Biology, Universlty of Valle, Cali, Colombia

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Genetics of Mitochondria and Chloroplasts

111.

IV.

V.

VI.

V11.

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

A. Organelle Genomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Coding Capacity of Organelles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Transmission of Organelles.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Origins of Organelles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Organelle Involvement in Genetic Phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Mitochondrial Heterosis . . . . . . . . . B. Mitochondrial Complementation., . C. Chloroplast Heterosis and Complem ............................... D. Cytoplasmic Male Sterility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Nutritional Quality in Wheat.. . . . . . . . . . . . . . . Genetic Implications of Intergenomic Interactions . A. Chloroplast Thylakoid Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Chloroplast Mutants . . . . . . . . . . . ................ C. Ribulose-l,5-Biphosphate Carboxylase/Oxygenase . . . . . . . . . . . . . . . . . . . . . . . . . . D. Protoplast Fusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Mechanisms of Organelle DNA Transmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Mitochondrial Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G . Polymorphism of Mitochondria and Chloroplast . . . . . . . . . . . . . . . . . . . . . . Molecular-Genetic Aspects of Heterosis ........................... A . Theories of Heterosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . Mechanisms of Heterosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Multimeric Hybrid Molecules ............ D. DNA of Hybrids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Current Models of Complementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Improvement of Crop Yield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Role of Plasmon Genes . . . . . . . . . . . . . B. Mitochondrial Efficiency and Prediction of Yield C. Photosynthetic Efficiency and Yield. . . D. Conversion of C, Plants to C, Pathway .......... Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

120 124 126 130

139 142

148 149

151 154 157 160 161 164 164 165 170 171 173

180 182

185

'Present address: National Agriculture Research Project, Gujarat Agricultural University, Anand 3881 10 (Gujarat), India. I17

Copyright 0 1981 by Academic Press. Inc. All nghts of reproduction in any form reserved.

ISBN 0-IZWQ734-7

118

H.K. SRIVASTAVA

1. INTRODUCTION The genome of eukaryotic organisms is localized not only in the nucleus, which is undoubtedly the principal source of hereditary information, but is also dispersed throughout the cell in subcellular organelles: mitochondria, chloroplasts, kinetosomes, and centrioles. The genetic hierarchy in the multigenomic cell as a complex of all cell genetic determinants is distributed into chromosomal or nuclear genome and extrachromosomal plasmon, comprising the cytoplasmic organelles-mitochondria (chondriome) and chloroplasts (plastome). Interaction of nuclear and organellular genomes with one another and with the environment determines both the course and the dynamics of cell differentiation, plant growth, development, and maturation. The nucleus as an integrating center of heredity is responsible for the formation of constant organism-specific properties, while organelle genomes presumably determine the cell life tactics through interaction with the nuclear genome, providing greater flexibility and adaptation of the plant as a whole to environmental fluctuations. The paucity of information documenting organellular control of agronomic traits has generally led plant geneticists to believe that such traits are primarily under the control of nuclear genes. Plant breeders, while recognizing the dangers of genetic vulnerability resulting from a uniform nuclear genome base, fail to extend similar dogma for maintaining diversity of plasmon in their improved cultivars. The advent of cytoplasmic vulnerability in the case of the Texas maize cytoplasm, which was recently found to possess a unique mitochondrial genome (Pring ef al., 1977), emphasizes the need for basic understanding of the potential genetic capacity of organelles, and of their subsequent involvement in genetic phenomena like homeostasis, polymorphism, and heterosis. Superior mitochondrial activity has been cited as one of the possible physiological mechanisms of heterosis resulting from organelle genome heterogeneity and polymorphism in several economically important crops (Srivastava, 1972; Berville, 1977). Direct evidence for an association of polymorphic organelles of the hybrids to distinct organelle genomes is lacking mainly because the traditional methods of analysis of nuclear genes are generally not available for studying mitochondrial or chloroplast traits in higher plants. The combined efforts of investigators from fields ranging from conventional genetics to molecular biology have resulted during the last decade in the intelligible demonstration that organelles possess autonomous genetic apparatus to the extent of (1) a distinct mitochondrial DNA (mt-DNA) and chloroplast DNA (ct-DNA) species capable of functioning as stable carriers of genetic information; (2) a mechanism for self-replication of these carrier DNAs within the organelles; (3) an independent transcription-translation system inside mitochondria and chloroplasts; (4) existence of all the components of the protein synthesizing

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

119

system, i.e., 70 S ribosomes, RNA-polymerase, mRNA, tRNA, aminoacyltRNA synthetases, and other translation factors; ( 5 ) ability to synthesize nucleic acids and structural and catalytic proteins; and (6) confirmation that mt-DNA and ct-DNA are essential as master molecules for organelle formation, development, and organization. Further evidence documenting nuclear DNA replication and transcription to be dependent on organelle gene products has also been accumulated (Gillham, 1978). At present, conceptual studies and developmental insights provide only a sketchy framework for understanding the tightly coordinated interactions between organelle and nuclear genomes that are essential for regulation of cellular metabolism, growth, development, and differentiation. Some aspects of the biosynthesis of key enzymes of the organelle system maneuvered by both nuclear and organelle genomes provide convincing evidence that intergenomic interaction is responsible for efficient cellular metabolism and plant growth and development (Srivastava, 1981). It is the purpose of this article to review and synthesize the available evidence that suggests that superior mitochondria1 and chloroplast functions resulting from genomic and intergenomic complementations may indeed be essential components of the heterosis phenomenon. Five basic issues will be discussed. First, what potential coding capacities do organelles possess? Second, how does intergenomic interaction operate to effectively regulate organellular function leading to heterosis? Third, how is heterosis operationally manifested in crop plants? Fourth, to what extent can the results of organelle heterosis and complementation be utilized in the prognosis of combining ability of potential progenitors for breeding? And, fifth, what possibilities exist for extrapolation of biochemical heterosis, in terms of organelle efficiencies, to crop yield improvement? It is not intended to provide a detailed coverage of ‘cytoplasmic inheritance in crop improvement” in this article, and the interested reader is referred to several excellent reviews (Harvey and Levings, 1972; Gillham, 1974; Kung, 1977; and Birky, 1978) and books (Grun, 1976; Gillham, 1978; and Kirk and TilneyBassett, 1978). The transmission genetics of genes in mitochondria and chloroplasts is undoubtedly the hot topic in plant genetics these days and could easily occupy an entire review by itself.

11. GENETICS OF MITOCHONDRIA AND CHLOROPLASTS Cytoplasmic genes have long been recognized in higher plants. Correns (1909) discovered the maternal inheritance of a spontaneous variegated mutation in the four o’clock flower, Mirubilis julupu, and Baur (1909) described the biparental but non-Mendelian transmission of a plastome mutation in Pelurgonium zonule. The possible involvement of the chondriome in maintaining a specific

120

H. K . SRIVASTAVA

phosphorus-to-nitrogen ratio was demonstrated by Caspari (1 956) in mouse mitochondria. The involvement of plasmon in heterosis (Ruebenbauer, 1967) and other hereditary phenomena (Wilkie, 1964; Gillham, 1978) has also been documented. Whaley (1952) observed greater efficiency of plasmon in maize hybrids than in inbreds as judged by the rapidity with which the cytoplasm could duplicate itself. He also commented that the genetic and physiological studies concerned with the early phases of growth and development are most likely to lead to the better understanding of heterosis. The faster germination and growth exhibited by hybrids as well as the heterotic expression of respiration (Srivastava, 1972) and photosynthesis (Sinha and Khanna, 1975) in many crops suggested the involvement of organelles in the phenomenon of heterosis. In yeast Saccharomyces cerevisiae several respiratory-deficient “petite ” mutants have also been identified with chondriome by a number of criteria (Mounolou el al., 1966). These observations spurred the rapid development of investigations into genetics of organelles and their relevance to cell heredity. A. ORGANELLE GENOMES

Chloroplasts and mitochondria contain unique genomes located in the unique DNAs of these cytoplasmic organelles. At the beginning of the 1960s highmolecular-weight double-stranded DNAs were identified in chloroplasts (Sager and Ishida, 1963) and in mitochondria (Luck and Reich, 1964). The organelle DNA molecule differed significantly from that of nuclear DNA in size, nucleotide content, and other physicochemical properties. Subsequently, the mt-DNA was shown to be the camer of the mitochondria1 genome (Michaelis et al., 1973) and the chloroplast genome of the unicellular green alga Chlamydomonas reinhardii was associated with ct-DNA (Schlanger and Sager, 1974). Results of many elegantly designed experiments with the yeast and the green alga by these investigators provide convincing evidence in support of the following conceptions: ( i ) genes are localized in the chloroplast and mitochondria; ( i i ) organelle genomes can be identified by their non-Mendelian behavior in reciprocal crosses and also by the high frequency of segregation and recombination in clonal growth of hybrids; (iii) genetic recombination occurs between the two parental organelle genomes in the progeny of biparental zygotes: biparental zygotes transmit organelle genes from maternal and paternal parents to the meiotic products; in the progeny of such zygotes organelle genes, unlike Mendelian genes, continue to segregate during the postmeiotic mitotic division as well as during the initial meiotic division; ( i v ) the recombinations of nuclear and organelle genes are separated both in time and space: recombination occurs mostly in vegetative growth of progeny clones, rarely in zygotes where nuclear gene recombination occurs during meiosis; and ( v ) two types of recombination events

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

121

occur, reciprocal and nonreciprocal: alleles segregate 1 : 1 in reciprocal recombination, and 1 : I on the average in nonreciprocal recombination. These findings have been generalized to other organisms, in part on the basis of experiments and in part by inference. The recently discovered technique of restriction endonuclease fragment analysis of organelle genomes (Pring and Levings, 1978) from organisms belonging to geographically distinct areas offers promise in the exploration for plasmon heterogeneity and its relevance to crop improvement. The results of genetic mapping by two methods, gene-attachment point distances and coconversion frequencies, reveal that the organelle DNA is genetically circular (Sager, 1977). The genomes of mitochondria and chloroplasts reside in covalently closed circular DNA molecules. Exceptions are the linear mt-DNA molecules of the ciliated protists Tetruhymena and Paramecium. Each species has its own unique mt-DNA and ct-DNA molecule; that molecule carries a complete set of organelle genes. Several important properties of chloroplast and mitochondrial DNAs are summarized in Table I. Comparative data on bacterial and blue-green algal DNAs together with eukaryotic nuclear DNA, giving an idea of the phylogenetic relationship between genetic systems of organelles and prokaryotic organisms, are also presented in Table I. Both the genomes of organelles and prokaryotic organisms are endowed with a circular fibrous DNA organization and an absence of histone. Exceptions are the highly condensed mt-DNA molecules of the true slime molds, which appear to be complexed with a basic protein (Kuroiwa et al., 1976), and the mt-DNA of Xenopus oocytes, which can be isolated in association with protein in the form of structures reminiscent of the nucleosomes of chromatin (Pinon ef al., 1978). The association of circularity with the non-histone-clad DNA is viewed as a primitive one, while the histone-clad DNA of eukaryotic organisms is thought to be a derived state (Uzzel and Spolsky, 1974). The similarities in physicochemical properties shared by chloroplast, mitochondrial, blue-green algal, and bacterial DNAs provide evidence that organelles may have originated as endosymbionts within a eukaryotic cell. A highly significant feature of organelle genome, however, is the occurrence of the eukaryotic properties of intercistronic spacer sequences and polyploidy. Many, but not all, genes from nucleated cells carry within themselves nucleotide sequences (called intervening or spacer sequences) that are not transcribed in the messengers (mRNA) corresponding to these genes. In contrast, bacterial messengers are direct copies of the genes, without any missing segments (Marx, 1978). Thus, the nucleated cells and organelles seem to possess a common mechanism, not found in prokaryotic organisms, for producing mRNAs from which some gene sequences are omitted or deleted. The total amount of ct-DNA per chloroplast in many crops is (2-10) x lo-'" g. This corresponds to a molecular weight of (1.2-6.0) X lo9. Using the renaturation kinetic concepts, the genome size of the ct-DNA in higher plants has been estimated to be ( 1 -2) x lo* MW (Kolodner and Tewari, 1975). Hence the average content per chloro-

H. K. SRIVASTAVA

122

Table I Some Properties of Nuclear, Organelle, and Bacterial and Blue-Green Algal DNA"

Parameters g) DNA content ( X Buoyant density (g/cm3) Dry weight (8) Guanine-cytosine content (8) Histones Repeated DNA sequence Polyploidy (multiple copies) Conformation DNA organization Satellite band 5-Methylcytosine Renaturation rate Recombination (meiosis and mitosis) DNA condensation (cell division) Cytoplasmic ribosome (80 S) Organelle ribosome (70 S ) Intercistronic spacers Ribosomal RNA ( a ) Large subunit (6) Small subunit 5 S RNA (most primitive) Transcription-translation system (streptomycidchloramphenicaYrifampicin) Protein synthesis system (cyclohexamide or a-aminitin) Initiation of protein synthesis

Eukaryotic nuclear DNA

ct-DNA

mt-DNA

BacteriaVbluegreen algal DNA

4-7 1.725 0.6-1.7 40-65 Present Present Present Linear 30-40 di fibrils Present Present Slowhapid Yes Yes Present Absent Present

0.3-0.9 1.697 0.05-0. I5 36-42 Absent Absent Present Circular 25 di fibrils Absent Absent Rapid Yes No Absent Present Present

0.1-0.3 1.706 0.02-0.08 36-45 Absent Absent Present Circular 25 di fibrils Absent Absent Rapid Yes No Absent Present Present

2-3 1.715 0.4-0.6 38-60 Absent Present Absent Circular 25 di fibrils Absent Present Slow

1.3 X lo6 0.70 x loR Present Insensitive

1.28 x 10' 0.75 x lo6 Absent Sensitive

1.23 X 10' 0.65 x lo6 Absent Sensitive

1.1 x 106 0.55 X lo6 Present Sensitive

Sensitive

Resistant

Resistant

Resistant

tRN Arne'

tRNA;""'

tRNA(""

tRNApeL

No No Absent Present Absent

aAdapted with modifications from Lehninger (1976) and Kung (1977). which is considered as bProtein synthesis system in organelle begins with N-formylmethionine (tRNAfmeL), evidence for a prokaryotic origin of organelles.

plast is about 10-30 times that of genome size, suggesting that the genetic material is present in about 10-30 copies (like many chromosomes in a nucleus) per chloroplast. The sequence of organelle genome is therefore amplified manyfold, representing organelle gene amplification. Organelle genes appear to be present in multiple copies in every cell. Organelles are thus analogous to highly polyploid nuclei in an organism with a haploid chromosome number of one. The number of organelle genes is limited by the size of the DNA molecule. Molecules of mt-DNA range from about 5 p m in length (about 10 X lo6 MW or 15 x lo3 base pairs) in animals and Chlamydomonas to about 30 pm(60 x lo6 MW or 90 x lo3 base pairs) in pea plants. Most plants

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

123

have ct-DNA molecules close to 45 p m in length, while the liverwort Sphaerocarpus and the protist Chlamydomonas set the known lower and upper limits with ct-DNA molecules 37 and 62 p m in length, respectively (Birky, 1978). A single cell may contain many molecules of ct-DNA and/or mt-DNA. These organelle genomes are packaged into one or many organelles, each of which may contain many DNA molecules. Finally, there is a third order of packaging; within an organelle, the DNA is generally observed to be localized in discrete regions that are most appropriately called “nucleoids” by analogy with similar regions in prokaryotes (Birky, 1978). The number and packaging of mt-DNA molecules are even more variable in Saccharomyces. The amount of mt-DNA in these cells varies with genotype, ploidy, cell volume, and in at least some strains with the physiological state of the cell (Williamson et al., 1978). These authors conclude that the average haploid cell has about 50 molecules of mt-DNA, with a range of 8-123. Haploid yeast cells have 10-32 “nucleoids,” and each “nucleoid” is estimated to contain 4-5 mt-DNA molecules. A distinction, however, must be made between the repeated DNA sequences and multiple copies of organelle DNA. The organelle genomes do not seem to possess repeated sequences but are amplified in multiple (nonredundant) copies distributed at a number of independent DNA-containing areas within the organelle (Gibbs and Prole, 1973), representing organelle genome polyploidy . Gene amplification in terms of ribosomal RNA genes (rDNA) is of common occurrence in eukaryotes (Tartof, 1975), and the present observation of multiple gene clusters in organelles suggests that gene amplification is not restricted to rDNA only, but DNA amplification may perhaps be found in other extrachromosomal locations. The occurrence and functional significance of gene clusters in prokaryotes, such as yeast and Escherichia coli, have become well documented since the classical studies of Demrec (1964) and Jacob and Monod (1961). In eukaryotes, gene clusters appear to be relatively less frequent than in prokaryotes, although numerous such clusters have been detected in recent years in mitochondria and chloroplasts (Giles, 1978). The plausibility of the presence of a polyploid genome model and its heterogeneity in a heteroplasmon within a cell can be considered supportiveto the view of eukaryotic resemblance of organelles. The eukaryotic nature of modem organelles is further substantiated by the absence of a 5 S primitive ribosomal RNA (120 nucleotides) in mitochondria and chloroplast, while such an RNA is of common occurrence in both bacterial and eukaryotic ribosomes (Schwartz and Dayhoff, 1978). It is associated with the larger ribosomal subunit and is thought to function in the nonspecific binding of transfer RNA to the ribosome during protein synthesis. In most eukaryotic species the 5 S RNA genes (5 S DNA) are of a homogeneous sequence (Tartof, 1975). The undeniable fact remains that organelles contain their own genomes and protein synthetic apparatus: both differ substantially from the corresponding nuclear-cytoplasmic system. The transcription-translation systems of organelles have a strong resemblance to those of prokaryotes.

124

H. K . SRIVASTAVA

B. CODING CAPACITY OF ORGANELLES

Both mitochondria and chloroplasts contain distinct genetic systems based on a covalently closed circular DNA molecule and a 70 S ribosomal proteinsynthesizing machinery. The main question that remains to be resolved is regarding the total content of genetic material present per organelle, and to what extent the organelle genome is involved in the overall genetic functioning of the cell. It is intrinsically difficult to make counts of the number of organelles present in a cell. Rough approximations based on light and electron microscopic views indicate a range from 700-1000 up to an extreme of 200,000 mitochondria in a large egg of higher plant and animal cells (Grun, 1976). Leaf parenchyma cells of Beta vulgaris have been studied intensively by Henmann et a f . (1974), who find that these cells contain 40-50 chloroplasts. Each chloroplast contains 4- 18 nucleoids (discrete regions inside the organelle packed with DNA molecules), and each nucleoid is estimated to have 10-100 chloroplast DNA molecules, and an entire cell possesses a total of about 500-1500 chloroplast genomes. Striking motion pictures of mitochondria of living plant cells suggest that mitochondria could fuse with one another and then split, causing changes in the frequency of mitochondria per cell over time (Honda et a f . , 1971). The numbers of chloroplasts per cell in both palisade and mesophyll tissue in spinach vary from 600 to 6000 depending on the developmental stage and length of leaves (Possingham and Smith, 1972). Nass (1976) estimates that a mouse L cell in culture contains about 250 mitochondria each with about 5 mt-DNA molecules, for a total of 1250 molecules per cell. In contrast, Chfamydomonas haploid cells have about 46 mt-DNA molecules (Ryan et a f . , 1978). The actual count of organelles per cell may not be very important from a genetic point of view. It is imperative to know, rather, how many organelle DNA gene centers there are per cell, how many DNA molecules there are per center, and whether mitochondria differ in their potential genetic capacity within a cell. Electron microscopic studies in yeast reveal that there are on the average 10 mitochondria per haploid cell (Grimes et a f . , 1974). Knowing the total DNA per cell (about 13% of the total DNA is mt-DNA), the authors calculated that there were 2.2 x lo9 MW of mt-DNA per cell; this corresponds to about 44 nucleoids of the yeast mt-DNA, giving about 4 in each mitochondria. Autoradiographic observations coupled with the results of electron microscopic studies in higher plants indicate that each chloroplast possesses a number of independent ct-DNA-containing centers (Gibbs et al., 1974). The number of such centers seems to increase with chloroplast size, varying from one area in small chloroplasts to a maximum of 40 centers in very large chloroplasts. Since organelles are known to reproduce by division (Ridley and Leech, 1970), the ones that contain the greater number of ct-DNA gene centers may be interpreted

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

125

as being ready for division. The chloroplast genome from most higher plants has DNA in the range of (2-10) X g and the unique genetic message is 40-150 p m long (Tewari and Wildman, 1970). A molecule 40- 150 p m long would have a molecular weight of (80-300) X 10" since 1 p m has a molecular weight of 2 x 10". If all the assumptions and estimates involved in this evaluation are correct, then there is room for between 15 and 80 DNA copies of the 40-pm ct-DNA and between 4 and 20 copies of the 150-pm ct-DNA in each chloroplast. The definite account of organelle DNA in light of its potential coding capacity is not yet feasible for want of direct experimental evidence. The mt-DNA may replicate to produce a mass of intramitochondrial DNA with or without mitochondria1 division (Grun, 1976). The organelle DNA replication evidently takes place independently at least on occasion, but the genetic significance of this replicated mass of DNA is unknown. The hypothesis that organelles possessing superior DNA replication machinery may be effective in providing superior organelle activities during plant growth and development is highly speculative, but is nevertheless worthy of consideration in designing future experiments with higher plants. Considering the size of mt-DNA in higher plants to be of the order of 150 X l(r' MW, it could be reasonably assumed that the mitochondrial genome is large enough to code for both large and small ribosomal RNA species, a complete set of transfer RNA species, and polypeptide components of enzyme systems that play key roles in respiratory electron transport and phosphorylation, e.g., cytochrome oxidase and the F,-F, complex in mitochondria. As regards chloroplasts, the total genetic capacity of average length 45-pm circular ct-DNA, assuming asymmetric transcription and allowing for an inverted repeat sequence, is about 6 X l(r' of polypeptides. A subsequent estimate by Smillie et af. (1973) reveals chloroplast genome to contain about 2 x 10' nucleotide pairs. If about 1500 base pairs form one functional gene, the chloroplast genome must contain about (1.3-1.5) x lo3 genes. Taking into account the existence of frequently repeating homologous chains, intercistron spacers, and regulatory genes, one can suppose that the chloroplast genome codes for about 150-300 specific proteins of 50 x 10" MW. It is tempting to conclude that this genetic system synthesizes most of the entire spectrum of chloroplast polypeptides. Nevertheless, genetic studies in both algae and higher plants have located the structural genes for several chloroplast polypeptides in the nuclear genome (Gillham, 1978), and the results of studies of the biosynthesis of many key enzymes of mitochondria and chloroplast activities, which are discussed in detail in later sections, indicate that some of their subunits are synthesized under transcriptional control of organelle genome, while the other partner polypeptides are made by the nuclear DNA. A possible interpretation of such observations would be that probably mitochondria and chloroplasts require the integrated action of both organelle and nuclear genomes for their structural and catalytic functions.

126

H . K . SRIVASTAVA

c. TRANSMISSION OF O R G A N E L L E S In the absence of detailed information about the fate of cytoplasmic organelles before, during, and after syngamy in the sexual offspring, one can only extrapolate from results regarding organelle heredity in terms of the specific source of the organelle genes. Several studies have shown that the organelles of the male gametophyte angiosperms are incorporated into the vegetative cell, and the generative cell and resulting male germ cells receive very few plastids, although, as in most plants, they contain mitochondria (Mascarenhas, 1975). It is of interest to note that in those cases in which plastids were not found in the generative cells (e.g., maize-Larson, 1965; petunia-Sassen, 1964; and impatiens-van Went, 1974) plastid variegation is transmitted only maternally (“status albomaculatus”), whereas in cases in which plastid variegation was found to be transmitted biparentally (“status paralbomaculatus”), numerous proplastids have been identified in the male germ cells (e.g., Pefargonium-Lombard0 and Gerola, 1968; Oneorheru-Walles, 1971 ; Hypericum-Hageman, 1964). Paternal mitochondria and plastids may or may not be excluded during syngamy; they may even prevail and exclude maternal mitochondria and plastids as indicated by the exclusive paternal transmission of organelles in some gymnosperms (Giarnordoli, 1974). In most instances the pattern of non-Mendelian inheritance of organelle traits follows uniparental transmission. Sager (1954) first reported the occurrence of streptomycin-resistant chloroplast mutations in Chlamydomonas showing uniparental transmission. In the literature these mutations are referred to by a variety of terms, including “cytoplasmic,” “chloroplast,” “non-Mendelian, ” “nonchromosomal, ” and “uniparental. ” The occurrence of biparental inheritance, foreshadowed by the findings of Baur (1909), has been documented in few higher plants, of which the most thoroughly studied are Oenothera and Pelargonium (Tilney-Bassett, 1973). Tilney-Bassett (1975) cataloged the various genera and species in which non-Mendelian inheritance of chlorophyll-variegated phenotypes were reported. At present, maternal inheritance for organelle genes appears to be more common than biparental, but both types are found in the two major groups of angiosperms: the dicotyledons, with 24 genera maternal and 14 genera with biparental inheritance, and the monocotyledons, with 8 genera maternal and 2 genera with at least a trace of biparental inheritance. Subsequently, biparental plastid inheritance was reported in Coix lacrymujobi, C. aquatica, and C . gigantea (Rao, 1975), in Browallia (Semeniuk, 1976), and in Pennisetum americunum (Rao and Koduru, 1978). In view of the extranuclear location of mitochondria and chloroplasts, it is not surprising that transmission of the organelle genomes is not governed by the same rules that apply to chromosomal genes. For fungi (Roodyn and Wilkie, 1968), amphibians (David and Blackler, 1972), and mammals (Hutchinson et al., 1974), there is evidence that the

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

127

mitochondria1 genome of an individual is derived solely from the maternal parent. Based on the restriction endonuclease fragmentation analyses of mt-DNA from normal and Texas maize cytoplasms, Pring and Levings (1 978) have made a strong claim for an absence of transmission of the paternal mitochondrial genome; paternal mt-DNA is not expressed at detectable levels in their experiments. However, recent study on ct-DNA distribution in parasexual maize hybrids, as shown by polypeptide composition of ribulose biphosphatase carboxylase-oxygenase, indicates that chloroplasts from both parents stay together and are distributed biparentally to daughter cells, giving a preponderance of one type or the other (Chen et al., 1977). Mammalian eggs are known to be very rich in mitochondria, and if the concentration of mt-DNA is similar to that found in the cytoplasm of echinoderm eggs, an equine egg would contain about 1W molecules of mt-DNA (Hutchinson et al., 1974). This would imply that if paternal and maternal mitochondria replicate at equal rates, then only about of the mt-DNA in a mammal would be paternally derived. This small quantity would presumably be below the limit of experimental detection. The possibility therefore exists that paternal organelle genomes are present in very small numbers in the tissues of higher organisms, and that sufficiently sensitive experiments in the future could detect their presence. It is of interest to consider the possible evolutionary implications of organelle genome transmission. An albomaculatus plant that does not have pollen transmission will inherit each generation only organelles of the maternal parent, and they will all probably be genetically the same. Evolutionary change will occur only when mutants are produced, and the selection will be restricted to one mutant at a time. In paralbomaculatus plants a mixture of the different types of organelles from both parents occurs in the fertilized egg. The heterogenous types present may compete, mitochondria with mitochondria and chloroplasts with chloroplasts. In addition, the paralbomaculate system offers possibilities of fusion between unlike types and recombination of their genes to produce new organelle genotypes. The paralbomaculate hybrid would be more flexible in the evolution of its organelle genotypes than the albomaculate. The hypothesis that the fusion of the two chloroplasts, just like nuclear fusion, brings together the ct-DNA from parental types resulting in an ‘‘interorganellular gene recombination” by physical exchange of pieces of DNA is very attractive (Bastia et al., 1969; Cavalier-Smith, 1970); however, rigorous experimental testing is required to fully understand the possible exchange of genetic material during “interorganellular recombination. ” It would be of further advantage to an organism if the number of organelles in the cytoplasm were reduced before or during fertilization, so that doubling of such particles did not occur each generation. The genetic control mechanism to maintain the specific number of organelles in each species after biparental contribution of contrasting organelles to the zygote is not known. Considering the fact that meiosis serves this purpose for nuclear

128

H . K . SRIVASTAVA

genomes, Jinks (1964) has presented the hypothesis that size differences between male and female gametes in most species serve the purpose of organelle genome transmission in the zygote. There is, however, no correlation between egg size and the number of plastids in Quercus gambelii where the egg and zygote cells contain negligible or very scarce number of plastids (Morgensen, 1972). Although the information is still minimal, the present examples are a warning that one should be extremely wary of making assumptions about the paternal and maternal organelle genome transmission to the zygote. D. ORIGINS OF ORGANELLES

It is now established that the replicating chloroplasts of young spinach leaves actively synthesize DNA, and during their division cycle segregate it to daughter chloroplasts (Rose et a f . , 1975; Possingham and Rose, 1976). Organelle division is not necessarily the same as organelle DNA replication, for the DNA might replicate to produce a mass of intraorganellular DNA but still not produce new organelles therefrom. Direct evidence that organelles divide following one or the other mechanism has been elusive for a number of reasons. The organelles are so small that simply watching them divide in the ordinary microscope is not practical. There are complexities in the interpretations of observations of fixed tissue in an electron micrograph, but some studies of spinach leaves suggest the possibility that both constriction division by cross-wall or baffle formation occur (Gran and Possingham, 1972). In some ways, the constriction division of chloroplasts bears a similarity to cytokinesis in mammalian cells (Robins and Gonata, 1964). In spinach the daughter chloroplasts formed from a single dividing chloroplast are usually of equal size, and in this respect chloroplast binary fission differs from the asymmetric budding that occurs during yeast cell multiplication (Mitcheson, 1971). The constriction process of chloroplasts also differs from division in bacteria, where distinct cross-wall formation from both sides of the cell usually precedes daughter cell separation (Higgins and Shockman, 1971; Slater and Schaechter, 1974). There is evidence that isolated chloroplasts separated from the environment of the cell can divide using mechanisms that exist within the plastid itself (Ridley and Leech, 1970; Kameya and Takana, 1971). A doubling of chloroplast numbers was associated with a doubling of chloroplast silver grains, which is consistent with a duplication of ct-DNA during the chloroplast division cycle (Possingham and Rose, 1976; Possingham, 1976). There have been at least two electron microscopic studies (Marton, 1962; Andre, 1962), on the marine flagellate alga Chromulina pusilla, and an autoradiographic study involving growing Neurospora crassa, that seem to show division of mitochondria by binary fission similar to the constriction division of chloroplasts. The Chlamydomonas chloroplast is the only case where strong data favor

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

129

nonrandom segregation of organelle genomes (Birky, 1978). Because this organism possesses a single large-sized plastid that divides as the cell divides, the ct-DNA segregation can easily be followed during organelle division, whereas in yeast, Paramecium, and higher plants the data are compatible with random segregation of organelles during cell division. It is likely that organelle segregation during cell division has a strong random element, while organelles themselves have mechanisms to assure nonrandom DNA segregation when they divide (Birky, 1978). There is in fact substantial ultrastructural and autoradiographic evidence for numerically equal division of mt-DNA (Kuroiwa et a l . , 1977) and ct-DNA (Rose et al., 1974) during organelle division. It must be noted that the term segregation is applied here to the physical movement of organelles or organelle genomes during cell division. The vegetative segregation of organelle genes is commonly attributed to a random physical segregation and sorting out of the organelles containing them during vegetative divisions, since mitochondria and chloroplasts are usually not closely associated with the mitotic spindle or other visible apparatus that might assure nonrandom segregation. DNA replication in isolated chloroplasts was first reported in spinach and Euglena (Spencer and Whitfield, 1969). Evidence for the semiconservative replication of ct-DNA comes first from the work of Chiang and Sueoka (1967). These results confirm that ct-DNA is essential as master molecule for chloroplast division. According to Kolodner and Tewari (1975), ct-DNA replicates by both the “cairns” and “rolling circle” mechanisms. The “cairns” round of replication is used to initiate the “rolling circle” round of replication of ct-DNA. If this is true, why are two different modes of DNA replication required by higher plants? These authors propose that it is probably related to a developmental aspect of higher plant chloroplast biogenesis. During chloroplast development, a rather rapid synthesis of a number of copies of ct-DNA molecules is required. Therefore, the “rolling circle” mechanism might be used for this rapid synthesis, while the “cairns” mechanism may be used for normal replication. The subject of the evolutionary origin of organelles has aroused much interest and speculation, but the existing evidence is still far from conclusive. Three main theories, among many others, explaining the origin of organelles have been discussed frequently. The first, an endosymbiotic theory, proposes that the cellular organelles of eukaryotes, even including flagella, centrioles, and spindle fibers, originated from symbiotic prokaryotes like blue-green algae or photosynthetic bacteria that originally existed in the host cells and later became established after some modifications as endosymbionts within eukaryotic cells (Stainer, 1974). Margulius (1970) has marshaled a great wealth of morphological, biochemical, and paleontological facts in support of this theory. It has further been supported and somewhat amplified by Raven (1970), as well as by Schnepf and Brown (1971) and Taylor (1974), but has been severely attacked by others, particularly Allsopp (1969), Raff and Mahler (1972), and

130

H. K . SRIVASTAVA

Bogorad (1975). The second theory, the origin of organelles by progressive evolution of prokaryotes, suggests that mitochondria and chloroplasts originated by progressive transformations of the thylakoid membrane systems of blue-green algae (Allsopp, 1969). Raff and Mahler (1972) postulate that mitochondria originated from invaginations of the inner cellular membrane, which first formed respiratory vesicles bound by membranes. These vesicles converted into mitochondria by the addition of a protein-synthesizing DNA plasmid, derived from another prokaryote. They believe that the difference in modem eukaryotes between nuclear-directed and mitochondria-directed protein synthesis and other metabolic activities is due to divergent selective pressures that have acted since the acquisition of mitochondria. The many similarities between organelles and prokaryotic cells (illustrated in Table I) can not be taken as direct support for the endosymbiont hypothesis because the eukaryotes may have evolved by gradual transformation of prokaryotes. Bogorad (1975) offers a third theory, “cluster-clone,’’ for the origin of the eukaryote nucleus and the DNA contained in chloroplasts and mitochondria. He believes that all of the DNA found in eukaryote cells originally existed in a nuclear area, perhaps surrounded by a common nuclear membrane. Later, clusters of genes derived from the protonucleus became separated and surrounded by their own membranes, thus forming the plastids and mitochondria. The earlier considerations that some of the proteins of which these organelles consist are coded by nuclear genes, while other are coded by genes located in the organelle themselves, could be regarded as evidence in favor of the “cluster-clone’’ hypothesis or a nucleoplasmic origin of organelles. It has been reported that there is considerable base sequence homology between ct-DNA and nuclear DNA in tobacco (Siegel, 1975), broad bean (Kung and Williams, 1969), and Euglena (Rawson and Haselkorn, 1973). Furthermore, chloroplast rRNA hybridizes with nuclear DNA in tobacco (Tewari and Wildman, 1968). These studies appear to favor a direct origin of organelles rather than an indirect one. Evidence in support of another hypothesis that organelles originated from an invagination of the nuclear membrane has been presented (Bell, 1970). Whichever theory is finally accepted, evolutionary modification of both nuclear and organelle DNA during early eukaryote evolution must be postulated.

111. ORGANELLE INVOLVEMENT IN GENETIC PHENOMENA When one considers the diverse effects of cytoplasmic genes on plant growth and development of the few species of higher plants studied (Stubbe, 1964), the possibilities of recombination resulting from the biparental transmission of organelle genes (Sager, 1977), and the opportunities for inducing mutations of

INTERGENOMIC INTERACTION, HETEROSIS. AND CROP YIELD

131

organelle genes (Hanna er al., 1978), the need for extending these notions to breeding aspects of crop plants becomes imperative. Jinks (1964) has stated “if differences between reciprocal crosses are to be used as a criterion of extrachromosomal heredity, we must specify that the difference persists through the successive generations that may be derived from them. Using such a criterion, Ressler and Emery (1978) found reciprocal differences in growth habit expression of the F, , F2,F3, and BCIs, generations from reciprocal crosses of peanut lines with reported differences in plasmons and genomes, respectively. Plastome mutants producing chlorophyll-deficient seedlings are generally used as markers for studying gene action, mapping chromosomes, determining the effects of mutagens, and other fundamental studies on higher plants. Plant breeders are beginning to pay attention to the “genome-plasmon” models for plant development and the potential use of plasmon in the screening procedure (Ashry, 1976). The marked variation in mt-DNA and ct-DNA among maize cytoplasms (Pring and Levings, 1978) provides additional evidence that organelles may be involved in many genetic phenomena, including male sterility and disease susceptibility. ”

A. MITOCHONDRIAL HETEROSIS

The involvement of mitochondria in heterosis as judged by increased respiratory function and higher enzyme activities in the hybrids has been demonstrated (Srivastava, 1972). An operational basis for heterosis with regard to mitochondrial metabolism was first advanced in maize involving comparative studies of oxidative metabolism and oxidative phosphorylation from hybrid and inbred lines (McDaniel and Sarkissian, 1966; Sarkissian and Srivastava, 1967). Further studies in hybrids of wheat have provided evidence for the hypothesis that enhanced mitochondria1 activity could be a general physicobiochemical mechanism for the expression of heterosis (Srivastava, 1974). Mitochondria in wheat hybrids were found to possess more efficient systems for channeling electrons to their final acceptor cytochrome oxidase for the production of ATP as judged by relatively greater specific activities of the enzymes NADHcytochrome reductase and cytochrome c oxidase. The oligomycin-sensitive adenosine triphosphatase (ATPase) system of mitochondria from a wheat hybrid and its parents has also been studied and the results show correlation of ATPase activity with heterosis (Srivastava, 1975). The hybrid therefore seems to possess both a higher rate of synthesis of ATP (ADP : 0 ratios) and higher rate of release of ATP energy (ATPase efficiency) than its parents, and this coupled mechanism may be operating in eliciting the heterotic expressions of the hybrids. Apart from the efficient systems for the conservation of ATP energy in hybrid mitochondria, several distinct lines of evidence suggest that the quality and quantity of mitochondria present seem to play an important role in determining

132

H . K . SRIVASTAVA

the degree of vigor exhibited by a given genotype. The increased quantity of mitochondrial protein of seedlings present in heterotic (heavy) seeds has been shown to be indicative of a higher respiratory rate and greater amount of ATP energy conservation (Kittock and Low, 1968; McDaniel, 1969). The hybrid mitochondria possess relatively greater amounts of lipid and phospholipid, and there is some marked difference in fatty acid composition of two important phospholipid fractions-lecithin and cephalin-from the mitochondria of the hybrid and its parents (Srivastava and Sarkissian, 1972). Evidence showing hybrid mitochondria to possess abundance of linoleic acid in their fatty acid fraction (Lahib and Kader, 1977) and a greater amount of internal “bound” water (Khokhlova et a l . , 1975) is indicative of the physicobiochemical changes in the biochemical systems of the hybrid during heterotic expressions. Some significant quantitative differences in the amount of cytochromes a, b, and c between a wheat hybrid mitochondria and its parental mitochondria have also been observed (Sarkissian and Srivastava, 1971). The protein content of seed has been clearly shown to be related to seedling vigor, as measured by seedling dry weight (Ayers, et a f . , 1976), and to grain yield (Lopez and Grabe, 1973). Significant variation in protein content, size, and seedling vigor with position of seed in heads of wheat plant has been observed (Ries et a l . , 1976), suggesting that larger seeds may contain more protein than smaller seeds as a result of size or concentration. Recent results further indicate that mitochondrial ADP : 0 ratios in wheat seedling are correlated to seed metabolism (Flavell and Barratt, 1977), which is in turn determined by the metabolic activity of the plant during formation and maturation of the seeds. Positive correlation between seed and oil yield and mitochondrial activities in oil palm (Elaeis quineensis) has recently been observed (Kouame, 1978). A relationship between mitochondrial metabolism and heterosis in Asparagus has also been demonstrated (A. Berville, personal communication). The more vigorous varieties of maize and soybean have more tightly coupled mitochondria than the less vigorous varieties (Hanson et a l . , 1975). All these observations are in concordance with the view that mitochondria play a significant role in the manifestation of heterosis. The basic mechanism underlying growth and development is protein synthesis (Chen, 1971), and any changes in morphology must ultimately result from changes in transcription or translation of the DNA and RNA of the multiple genomes. An association between heterosis and contents of nucleic acids and of amino acids in generative organs of maize has been established (Scarascia, 1977). Significant differences in terms of ribosomal RNA cistron amplification between total and nuclear DNA of maize hybrids and their inbred lines during the phase of cell elongation have been recorded (Gilyazetdinov et al., 1977). Mitochondria, by way of ATP energy supply, considerably affect the nuclear genome transcription (Wolf and Rempan, 1977) as well as the crucial stages of chloroplast biosynthesis (Schiff, 1975) in higher plants.

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

133

The functions of mitochondria are important not only in ATP energy production but also in cellular metabolic regulation, including nuclear DNA and RNA syntheses. A correlation between mitochondrial heterosis and grain yield in cereals was interpreted by McDaniel ( 1 973) and Sage and Hobson ( 1 973) to indicate that mitochondrial activity is rate-limiting for yield and increased mitochondrial efficiency is the physicobiochemical basis of heterosis. Another interpretation that mitochondrial efficiency may not be a limiting factor for grain yield in view of the marked increased vigor and yield following additional nitrogen application in cereals (Ries and Emerson, 1973; Bingham, 1972) has been put forward by Barratt and Flavell (1977). The basis of their hypothesis is that many parts of cell metabolism, including mitochondrial and chloroplast activities, are coordinately regulated in many biochemical systems. It predicts (i) that heterosis due to a wide range of causes is associated with greater organellular activities; ( i i ) that the genetic control of heterosis is not confined directly to genes concerned with mitochondria and chloroplasts; and (iii) that mitochondria and yield are not under common gene control, but are under the control of a common general regulation of metabolism. There are no direct experimental findings to establish these hypotheses, however, a properly designed genetical experiment should be camed out in which the biometrical evidence from segregating lines would be crucial. It is safe to consider that mitochondrial genomes in higher plants are present in many copies per cell and that they are packaged in one to many mitochondria, and within each mitochondrion, in one to many nucleoids. The segregating units must therefore be groups of mt-DNA molecules, perhaps corresponding to the nucleoids observed by Williamson et al. (1978) or to whole mitochondria or segments thereof. Some evidence that the input ratio of parental mt-DNA molecules in a yeast zygote affects the output ratio of mitochondrial genes in the progeny has been shown (Boker et a/., 1976). Gametes in Chlamydomonas possess one large and sometimes one to four small mitochondria, but these break up into many small mitochondria in the zygote (Grobe and Arnold, 1975). These changes in the number of mitochondria per cell suggest fusion and fission of mitochondria, and can be inferred as equal to the occurrence of genetic recombination. Molecular evidence for mt-DNA recombination in animal cells has also been obtained (Horak et al., 1974; Wallace et a / . , 1976). These findings support the notion that heterotic hybrids contain a population of genetically heterogenous mitochondria and that the recombination between mt-DNA molecules at the time of fusion and fission of biparental mitochondria leads to mitochondrial heterosis. The possibility that recombination between mt-DNA molecules even within the same mitochondrion occurs in those organisms where the real fusion between two mitochondria cannot be detected in crosses has been mentioned (Birky, 1978). Also, there is no sign of mitochondrial gene recombination or of sharing of mt-DNA molecules between mitochondria in an exceptional organism

134

H. K . SRIVASTAVA

Paramecium aurelia (Adoutte, 1977), and one would therefore expect the or-

ganelle as a whole to be the unit of segregation in this case. B . MITOCHONDRIAL COMPLEMENTATION

The biochemical basis for heterosis is fundamentally one of complementation, usually either between proteins or between protein subunits. Genetic complementation involves interaction of gene products to produce a normal phenotype, whereas wild-type recombinants produce a normal phenotype because a normal sequence of nucleotides is actually created along a genome via genetic exchange (Demrec and Hartman, 1959). The ability to generate a normal phenotype is genetically transmitted from parent to offspring when chromosomes arise by recombination, but progeny produced as a result of complementation (interallelic interaction) remain individually defective in genotype. While recombination can occur within a gene, complementation occurs only between one gene and another belonging to either one genome or multiple genomes in the cell. The concept of complementation at organellular level to explain the operational mechanism of heterosis is attractive and could interpret various events at the cellular and subcelMar levels leading to heterotic expressions in higher plants. Recent observations on the potential genetic capacity of mitochondrial genome (Pring and Levings, 1978), chloroplast genome (Coen et a l . , 1977), and intergenomic interactions between organelles and nucleus (Srivastava, 1981) provide insights for the existence of complementation at different levels of structural and functional organization of a plant cell. Mitochondria1 complementation (the enhanced oxidative phosphorylation efficiency of artificially mixed mitochondria of certain inbreds) has been reported as correlated with seedling heterosis in maize (McDaniel and Sarkissian, 1966, 1968; Sarkissian and Srivastava, 1967), wheat (Sarkissian and Srivastava, 1969, 1971; Srivastava, 1974), barley (McDaniel, 1969, 1972), and alfalfa (Schneiter er a l . , 1976). Hobson (1971) observed some complementation effects for grain yield heterosis in wheat cultivars. A small but significant complementation effect for sugarbeet (Beta vulgaris, L) root mitochondria, which seemed associated with root yield heterosis, has been demonstrated (Doney et al., 1972). From these findings, the general conclusion was drawn that measurement of mitochondrial complementation, especially for ADP : 0 ratio, is a useful tool to predict combining ability for yield in higher plants (Sarkissian, 1972; McDaniel, 1973; Srivastava, 1972, 1974). McDaniel (1971) and Sage and Hobson (1973) made a rather strong claim, based on the results of mitochondrial activities coupled with field data, that grain yield heterosis in cereal hybrids was positively correlated with a higher activity of mitochondrial complementing mixtures of the parents. The findings on the

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

135

relationship between mitochondrial complementation to yield heterosis in field crops and their practical utilities in evaluating combining ability for yield potential of individual genotype in breeding programs, however, have not met with all expectations. Zoble et al. (1972) were not able to show mitochondrial complementation i n winter wheat, and Ellis er al. (1973) could not duplicate McDaniel’s (1971) results using the same varieties of barley. The small complementation effects, as observed in sugarbeet by Doney et al. (1976), made it difficult to detect differences in complementation, and thus difficult to predict root yield heterosis. Recent extensive studies of mitochondrial complementation and grain yield in hybrid wheat have given new clues to the detectable correlation between F, yield heterosis and mitochondrial efficiency (Lupton, 1976; Sagi et al., 1976; Barratt and Flavell, 1977). These workers favor using mitochondrial complementation as a biochemical tool over agronomic selection criteria covering a number of different traits in plant breeding programs for a rapid first screening of desirable varieties. It is too early, therefore, to set aside the results of mitochondrial complementation and its prospective use as a selection criterion during young seedling stages of crop plants for breeding purposes. The results of “marginal” complementation of a weak correlation between mitochondrial complementation and yield heterosis as observed by some workers could be a reflection of either mitochondrial preparations per se or mixing of I : 1 parental mitochondria in virro to form complementing mixture of mitochondria for comparative studies on mitochondrial complementation and yield heterosis. The results on heterosis and complementation with regard to some key components of mitochondrial functions in many crops are summarized in Table 11. In addition to the efficient enzyme systems of hybrid mitochondria and 1 : 1 complementing mixture of parental mitochondria, respectively, electron microscopic studies of mitochondria isolated from heterotic cotton hybrids (Arslanova, 1973) show highly developed cristae and respiratory assemblies reflecting the structural superiority of these mitochondria. The mitochondria from the parental lines, on the contrary, exhibit relatively less developed inner-membrane cristae and smaller intramitochondrial respiratory assemblies. The hybrid mitochondria were also found to be dividing at a faster rate than those of the parents. Some inter- and intraspecific cotton hybrids, differing in their degree of yield heterosis, have been shown to synthesize and accumulate bound fractions of DNA and RNA more rapidly than their parents (Rakhmankulov er al., 1976). Complementation in terms of ribosomal RNA cistron amplification during the cell elongation phase using both total and nuclear DNA of the cell in the parents of heterotic maize hybrids has been observed (Gilyazetdinov et af., 1977). A relationship between complementation in the artificially made mixtures of parental mitochondria using three enzyme assays (cytochrome c oxidase, succinic dehydrogenase, and glycerine-1-phosphate dehydrogenase) and morphological heterosis in terms of weight gains of the F, animals has also been demonstrated (Dzapo et al., 1974).

136

H. K. SRIVASTAVA

TABLE I1 Heterosis and Complementation with Regard to Mitochondria1 Activities Organisms

Parameters

Maize (Zea mays)

Amos and Scholl ( 1 977)

Maize (Zea mays) Wheat (Triricum aesfivum)

Berville e f al. (1976, 1977) Srivastava and Sarkissian (1969); Srivastava (1972, 1974); Sage (1973); Lupton (1976); Barratt and Flavell (1977); Flavell and Barratt (1977) Sarkissian and Srivastava (1967); McDaniel and Sarkissian (1970); Sage (1973); Berville et al. (1976, 1977) McDaniel (1975) Hanson e f al. (1975) Israelstam and Fukumato (1977) Schneiter er al. (1976) Grimwood (1972) Sarkissian and Srivastava (1971) Srivastava and Sarkissian (1972) Lahib and Kader (1977) Kouame ( 1978) Knyaseva and Romanova (1977) Arlslanova (1973)

Maize (Zea mays) Beans (Phaseolus vulgaris) Triticum-Agropyron hybrid Wheat (Triricum aesrivum) Maize (Zeu mays)

Adenosine triphosphatase

Malate dehydrogenase Succinic dehydrogenase Isocitrate dehydrogenase Pyruvate dehydrogenase Fumarate dehydrogenase Isocitrate lyase

Rye (Secale cereale) Cotton (Gossypium hirsufum) Wheat (Trificumaesfivum) Maize (Zea mays) Wheat (Triricum aesfivum) Maize (Zea mays) Wheat (Trificumaesfivum) Barley (Hordeum vulgare) Maize (Zea mays) Maize (Zea mays) Cotton (Gossypium hirsutum)

Srivastava and Sarkissian (1969); Srivastava (1974); Lupton (1976); Sagi e f al. (1976); Barratt and Flavell ( 1977) Srivastava (1972) Muresan et al. (1976) Karamanenko (1976) Srivastava (1975) Srivastava (1972); Barratt and Flavell ( 1977) Vecher et al. (1975) Rakhmankulov (1975) Srivastava (1972) Srivastava (1972) Srivastava (1972) McDaniel and Sarkissian (1968) Srivastava (1972) McDaniel (1975) McDaniel and Sarkissian (1968) Roos and Sarkissian (1968) Scholl (1974)

Wheat (Trificumaesfivum)

Cytochrome oxidase

f

Nitrate reductase Glutamine synthetase NADH-glutamate dehydrogenase Oxidative phosphorylation (ADP:O ratio)

t

Respiratory control index (state 3 :state 4 oxidation ratio)

Cytochrome b, c, and a contents Lipid-phospholipid content Linoleic acid content Oil yield (total) Protein content Ultrastructure of hybrid mitochondria

References

Maize (Zea mays)

Barley (Hordeum vulgare) Soybean (Glycine max) Pea (Pisum safivum) Alfalfa (Medicago saliva) Cucurbits (Cucurbifa maxima) Wheat (Triricum aestivum) Wheat (Triricum aestivum) Wheat (Triricum aesfivum) Oil palm (Elaeis quineensis) Potato (Solanum tuberosum) Cotton (Gossypium hirsutum)

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

137

It is of interest to note that all these studies on mitochondrial complementation have been confined to the hybrids and their respective progenitors that were already known for a number of heterotic or nonheterotic phenotypic measurements including yield. There is no doubt that the heterotic hybrids are endowed with a more balanced metabolism than the purebred parents. Since growth and economic yield are the results of a series of metabolic reactions involving multiple-enzyme complexes, it is reasonable to suppose that heterosis as a genetic phenomenon may be dependent on mitochondrial complementation. A working hypothesis based on the conception of nucleus-mitochondria-chloroplast cooperative interaction in the development of the whole plant leading to heterotic phenotypic expression is therefore suggested. The detailed discussion on the implication of intergenomic interaction is presented elsewhere in this article. How should the data on mitochondrial complementation be interpreted in genetic terms? Considering that the origin of mitochondria from the preexisting one is under nuclear genome control and that the differential contribution of mitochondria to the hybrid individual is derived through the biparental transmission, the occurrence of polymorphic mitochondria possessing well-organized ultrastructures as well as efficient enzyme systems in the hybrids, resulting from both organellular recombination and mutation, is conceivable. A similar interpretation of the results on mitochondrial complementation was also given by Wagner (1972). The fact that part of the mitochondrial protein is encoded in the nuclear DNA can not be denied. Therefore, the protein or enzyme of the artificially mixed parental mitochondria could have been highly efficient as a result of the heterozygosity of their chromosomal genes. A nuclear genome heterogeneity could have produced the complementation results in terms of ADP : 0 ratio, R.C. ratio, and oxygen uptake in 1 : 1 mixture of mitochondria from the two inbred parents. However, whether the heterogeneity of mitochondria in 1 : 1 parental mixture in vitro or in the hybrids in vivo is a reflection of heterozygosity of chromosomal genes or extrachromosomal genes still remains to be resolved. There was also a demonstration (Srivastava and Sarkissian, 1970) that the mitochondria of allohexaploid wheats were more vigorous than those of allotetraploid wheats, which in turn exceeded those of diploid wheats. The data were interpreted as evidence that the allohexaploids contained the most polymorphic types and the diploids relatively uniform or unmixed mitochondrial types. Again, the data could reflect the more heterogenous types of proteins supplied to the polymorphic mitochondria by the chromosomal genes of the allohexaploids, as compared to those supplied by the allotetraploids and diploids in wheat polyploid series. Although mitochondrial heterosis appears to be the result of complementation between parent mitochondria that may be present in the hybrid (Srivastava and Sarkissian, 1972), little is known about the precise mechanism by which complementation is accomplished. It was suggested that particle-particle contact between mitochondria is a physical basis of complementation (McDaniel and

138

H. K . SRIVASTAVA

Sarkissian, 1970). The present evidence indicates that the modification of sulfhydryl groups (SH) when p-chloromercuribenzoate reacts with mitochondria results in a complete loss of the kinetic expression of cooperative interaction of the enzymes in the mixture of parental mitochondria (Sarkissian and Srivastava, 1973). Since SH groups are intrinsic to almost all mitochondrial functions (Sanadi et af., 1968), it is a reasonable inference that SH groups of key mitochondrial enzymes may be involved in internal hydrogen bonding to maintain a proper conformation of the active site(s) of the enzymes in the complementing mixture of mitochondria. Mitochondria1complementation and heterosis may therefore be regulated through the conformation of membrane-bound enzymes. Such desired enzyme conformation for efficient metabolic function in the complementing mixture of mitochondria could be produced by exchange of intermediates between contrasting mitochondria from the parents. Some speculation that the control of ADP:O ratios in vitro is mediated via small molecules either bound to or surrounding mitochondria has been made (Barratt and Flavell, 1977). The effective concentration of these would presumably alter the rate of metabolism and tissue growth. If they were in different concentrations in mitochondrial preparations from different varieties, then it is possible that a mixture of the mitochondria would have a different ADP:O ratio from the average of the two “parental” preparations, owing to a new effective concentration of regulatory molecules. The correlation between mitochondrial complementation-heterosis and hybrid vigor would then be a consequence of a similarity between the concentration of regulatory molecules derived from polymorphic mitochondria in the parental mixture and in the hybrid. The mitochondrial efficiency in terms of ADP:O ratios of 5-day alfalfa seedlings has been positively correlated to forage yields of similar lines grown in the field (Schneiter et al., 1974, 1976). The technique of the measurement of mitochondrial efficiency used for selection of genotypes is limited because alfalfa is a very heterogenous crop and an evaluation of individual genotypes can not be made using the large number of seedlings required for adequate sample. A question one is tempted to ask is what relationship exists between mitochondrial efficiency (ADP : 0 ratio) and plant productivity? Calculation of substrate utilization revealed that in higher plants with a relative growth rate of 0.1 g-’ day-’ (which is approximately the growth rate of young maize plants), an increase of ADP :0 from 2 to 3 causes an increase in the efficiency of dry matter production of 8% (Penning De Vries et al., 1974). The substantial differences in ADP: 0 ratio in testing plant material would have significant influence on dry-matter production and yield. Growth analysis of maize hybrids and inbred lines demonstrated that hybrid vigor correlates with the rate of embryo development and the utilization of reserve material during germination (Donaldson and Blackman, 1974). After emergence, hybrid vigor undoubtedly is related to net assimilation rate-to-leaf-area ratio, but interactions between stage of development, environ-

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

139

mental conditions, and, above all, the genotype (together with plasmon) finally determine the path towards heterosis. Recent results indicate that mitochondrial ADP : 0 ratios in germinating seeds are correlated to seed metabolism, which in turn is determined by the metabolic activity of the plant during formation and maturation of the seeds (Barratt and Flavell, 1977). This provides an explanation for the correlation between mitochondrial efficiency in germinating seedlings and grain yield heterosis in the crops so far tested. In view of the results discussed thus far, the employment of the mitochondrial complementation technique in some limited crops as a selection criterion for performing preliminary screening of potential inbred parents that are likely to produce heterotic hybrids in crossbreeding experiments under field conditions is recommended. The findings further establish the relative importance of mitochondrial efficiency as a means of initial increased physicobiochemical superiority in eliciting heterosis. It is suggested that to a certain extent heterosis is the result of mitochondrial complementation that occurs between polymorphic mitochondria of the hybrid. C. CHLOROPLAST HETEROSIS A N D COMPLEMENTATION

It has already been emphasized that heterosis in plants often results in efficient conservation of energy. Therefore, the utility of this phenomenon if observed in both mitochondria and chloroplasts might be of greater importance for increasing crop yield through the manipulation of organelle genomes. Chloroplast heterosis and complementation with regard to several parameters in many economically important crops have been demonstrated (Table 111). The term “chloroplast complementation” is generally used to indicate the greater activity of 1 : 1 parental mixture of isolated chloroplasts when compared to the midparental values. Higher photosynthetic rates in isolated chloroplasts of hybrids of many crop species have been observed in the seedling stages (Sarkissian and Huffaker, 1962; Nagy et al., 1972; Heichel and Musgrave, 1969), and there is a consensus for the occurrence of chloroplast heterosis in crop plants (Sinha and Khanna, 1975). Recent electron microscopic studies have provided evidence that hybrids possess more highly developed chloroplast structure than their respective parents (Hraska, 1978), and the increase in the size of the lamellae and thylakoid membrane structure in the chloroplasts of the hybrids was directly correlated with their chlorophyll contents (Rakhmankulov et al., 1976). In addition to the enhanced activities of the key enzymes of the Calvin cycle in hybrid chloroplasts, heterosis in chlorophyll content in maize and sorghum has also been reported (Nosberger, 1970; Khanna, 1974; Fleming and Palmer, 1975; Planchon, 1976). Results of chloroplast complementation based on Hill’s reaction and cyclic phosphorylation in maize showed 25-60% increased activity, and the enhanced activities due to chloroplast complementation were found to be closely associated with

140

H. K. SRIVASTAVA

Table I11 Heterosis and Complementation with Regard to Chloroplast Activities Parameters Ribulose biphosphate carboxylase/oxygenase Photophosphorylation Hill reaction Chlorophyll content

Chlorophyll a / b ratio Protein content

DNA and RNA contents

Dry matter of chloroplasts Ultrastructure of hybrid chloroplasts

Organisms Maize (Zea mays); sorghum (Sorghum vulgare); and barley (Hordeurn vulgare) Maize (Zen mays) Cotton (Gossypium hirsurum) Maize (Zea mays) Cotton (Gossypium hirsutum) Soybean (Glycine max) Wheat (Triticum aesrivum) Maize (Zea mays) Cotton (Gossypiurn hirsutum) Pea (Pisum sarivum) Maize (Zea mays); cotton (Gossypium hirsurum) Maize (Zea mays) Rye (Secale rereale) Barley (Hordeurn vulgare) Cotton (Gossypium hirsutum) Tomato (Lycopersicum esculentum)

References Sinha and Khanna (1975)

Berville ( I 977) lmamaliev et a / . (1975) Smirnova et al. (1975) Rakhamankulov et al. (1976) Starnes and Hadley (1965) Planchon (1976) Giles (1974); Fleming and Palmer (1 975) Rakhmankulov et al. (1976) Vershinin et al. (1976) Andregeva (1976) Rakhmankulov et al. (1976); Konarev (1976) Vecher er a / . (1977) Horak and Zalik (1975) Rakhmankulov et al. (1976) Kazim (1974)

the degree of grain yield heterosis (Ovchinnikova and Yakovlev, 1978). Some evidence for a developmental complementation in amylase activity when kernels of parental types were germinated together in a 1 : 1 ratio has also been recorded in sorghum (Ghose et al., 1974). These observations on chloroplast heterosis and complementation, like mitochondria1 complementation, could be interpreted to mean that hybrids are endowed with more efficient energy processing systems in their organelles. Photosynthesis is limited by photorespiration in major crop species, including wheat, rice, soybean, potato, peanut, barley, sugarbeet, cassava, and banana. These are called C3 plants .because their primary photosynthetic products are three-carbon compounds (3-phosphoglyceraldehyde)and they have only the reductive pentose phosphate (FU'P) cycle for COP fixation and reduction. The net rates of COz fixation in C, plants are usually about half those of C4plants such as maize, sorghum, pearl millet, sugarcane, and other tropical grasses. The more efficient C4 plants characteristically have, in addition to the RPP cycle, another Cot fixation cycle (Hatch and Slack, 1966; Hatch, 1976). In this cycle C 0 2 is fixed by carboxylation of phosphoenolpyruvate (PEP) to give a four-carbon acid, oxaloacetate or aspartate. Plants with this cycle are called C4 plants because the

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

141

first compounds formed after COz incorporation are four-carbon acids. In all higher plants, irrespective of the type of primary fixation, CO, reduction to carbohydrates occurs in the RPP cycle. The most important feature of C4 plants is the presence in their leaves of two types of chloroplasts which differ in both their ultrastructure and functions. In palisade mesophyll cells the chloroplasts have a pronounced granal structure, while in bundlesheath parenchyma cells they are usually paucigranal (Laetsch, 1974). These chloroplasts differ greatly in their photochemical properties: photosystem I1 is active in mesophyll cells, while photosystem I is active in bundlesheath chloroplasts (Bazzaz and Govindjee, 1973; Bishop et ul., 1977; K u et ul., 1974). Although the precise genetic nature of dimorphic chloroplasts in C4 plants remains to be worked out, it is likely to presume that each type results from the expression of distinct genes. Complementation of dimorphic chloroplasts in C4 plants is thus expected to lead to a concentration of COz in the bundlesheath cells, where it decreases to a certain extent the limitation of photosynthesis by the diffusive COPresistance and therefore aids in increasing the efficiency of the process. This is evidenced by the data on photosynthetic efficiency of C4 plants in which photorespiration is absent (Lorimer and Andrews, 1973). This situation could serve as one of the arguments proving that photorespiration as a process of C02 release or loss in the light has no great physiological significance. Another important component of photosynthesis is photophosphorylation, which is responsible for generating energy in terms of ATP and NADPHz for productive processes in the chloroplasts. Significant variation in cyclic photophosphorylation and Hill reaction activities in isolated chloroplasts from hybrids and their parents was demonstrated by Miflin and Hageman (1966). In these studies the hybrids were either intermediate or followed the activity of one of the parents, indicating either partial or full dominance. Khanna and Sinha (1975) assayed chloroplast activity for cyclic and noncyclic photophosphorylation of maize leaves in I-month-old seedlings. In the hybrids CM 400 X CM 300 and CM 103 x CM 104, the noncyclic photophosphorylation activity was intermediate between that of their parents. In cyclic photophosphorylation the hybrid CM 400 x CM 300 followed the more active parent, but CM 103 x CM 104 followed the less active one. These results suggest a biparental transmission of genetically diverse chloroplasts in these hybrids. Both complementation between parental chloroplasts and intergenomic interactions in heterotic hybrids might lead to such results for traits like cyclic and noncyclic phosphorylations, where the ultimate efficiency of the process is dependent on the formation of effective hybrid enzyme molecules. There are also several reports showing heterotic activities of ribulose- 1,5-biphosphate carboxylase/oxygenase (RBPCase) and PEP carboxylase in barley and sorghum hybrids (Khanna and Sinha, 1974, 1975). Systematic studies have also shown that the activity of RBPCase is differentially influenced during leaf growth and senescence in plants having the C4 pathway

142

H. K . SNVASTAVA

(Kennedy and Laetch, 1973). Direct evidence indicating that RBPCase is formed due to intergenomic interaction in the chloroplast and its implications in the phenomena of heterosis and complementation will be discussed in detail in later sections. In keeping with the results presented thus far, chloroplast heterosis could be viewed in terms of greater photosynthetic capacity of the organelle by means of possessing efficient enzymes of the Calvin cycle and photosynthetic electron transport chain. It is likely that heterotic hybrids are endowed with superior systems of chloroplasts and mitochondria, and such superiority is provided by genomic and intergenomic interactions. D. CYTOPLASMIC MALESTERILITY

It is generally accepted that the causes of pollen abortion in a male sterile plant are complex and are controlled by cytoplasm and male sterile genes of the nucleus. A common denominator of the phenomenon is that as a result of interaction between nuclear genes and plasmon factor certain plants fail to produce functional pollen. The nature of the expressions toward microspore abortion can be attributed to one or a combination of effects (Mascarenhas, 1975): abnormal meiosis during microsporogenesis; abnormal mitosis during microgametogenesis; coherence of pollen mother cells with each other owing to the degeneration of the primary and secondary cell walls; destruction of normal biochemical activities of enzymes-callase and glucanase-in the microsporocytes at all stages of meiosis; degeneration of the cytological and physiological functions of tapetum and anther locule; degradation of the callose matrix of the microspore quartet; and metabolic blockage of amino acid, protein, and nucleic acid syntheses during microsporogenesis and microgametogenesis within the anther of the male sterile plant. A number of alterations of the normal structure or function of organelles in cytoplasmic male sterile plants are also considered to be indications of associated effects leading to disturbed metabolism of sterile anthers. Differences in the level of a number of substances and their accumulation or disappearance during development characterize normal versus sterile conditions of angiosperm pollen development. Many studies have revealed a characteristic deficiency of proline and cystine and accumulation of asparagine in sterile anthers of higher plants (Kern and Atkins, 1972; Izhar and Frankel, 1973). Other relevant studies of sterile anthers indicate some marked alterations in the free amino acid pool and phase-specific changes in the amount of a few free amino acids (McKee, 1962; Fukusawa, 1967). Since the balance of free amino acids in anthers may be subject to the action of several factors, like protein and amino acid breakdown and synthesis as well as their translocation in and out of the anther, alteration in the amino acid profile in sterile anthers could reflect disturbed anabolic or catabolic processes. In addition, Fukusawa et al. (1957),

INTERGENOMIC INTERACTION, HETEROSIS, A N D CROP YIELD

143

Erickson (1967), and Savchenko et a / . (1968) have all reported anthers of malesterile wheat plants to be deficient in sugars relative to normal fertile anthers during pollen maturation. Differences in the enzymes produced by male-fertile and male-sterile anthers have been shown by both Alan and Sandal (1969) in sorghum and Erickson (1967) in wheat, who have reported electrophoretic zymogram band difference for the enzyme cytochrome oxidase. The nature of the expressions of cytoplasmic male sterility and their inheritance depends on cytoplasmic factors like organelle genomes or episomes of certain microorganisms that were unknown until recently. Rhoades (1950) first suggested that mutated mitochondria might induce cytoplasmic male sterility (CMS) in maize. Biochemical analysis of mitochondrial function (oxidative phosphorylation, ATPase activity, and cytochrome oxidase) in several fertile and CMS lines of wheat clearly suggested the involvement of mitochondria in male sterility (Srivastava et al., 1969). Considering the presence of heterogeneous cytoplasm as well as polymorphic mitochondria in CMS lines, these authors suggested (i) cytoplasmic heterogeneity is a persistent phenomenon and is transmissible either uniparentally or biparentally to subsequent generations, and ( i i ) there is a definite interaction between mitochondrial genomes of the two parents involved in the production of a CMS line. Upon crossing of sterile lines with fertility restoring lines, “restoration ” of mitochondrial structure was clearly correlated with “restoration” of fertility in the hybrids (Turbin et al., 1968). A later study of microsporogenesis in pollen of sterile and fertile plants showed abnormal mitochondrial structures in the steriles (Turbin et a/., 1970). Electron microscopic analysis of CMS wheat pollen showed a small number of organelles “subject to rather rapid degeneration.” Fertile pollen was full of cell organelles and therefore well equipped for a complete development (De Vries and Ie, 1970). Slight differences have also been detected in mitochondrial membrane components isolated from normal (fertile) maize cytoplasm and Texas male sterile cytoplasm (Watrud et a l . , 1974). Other studies have shown a higher content of cytochrome oxidase in fertile rather than CMS maize and wheat anthers (Palilova et al., 1966; Erickson, 1967). CMS Sudan grass had fewer isozymes of cytochrome oxidase than did fertile counterparts (Alan and Sandal, 1969). Several distinct lines of evidence further suggest that mitochondria are camers of genetic determinants conditioning CMS. Maize plants with the cytoplasmically inherited “Texas” male-sterile (T) genome, compared with plants having the same nuclear genome and the normal (N) cytoplasmic genome, possess altered mitochondria as judged by their sensitivity to the pathotoxin prepared from Helminthosporium maydis, race T (Miller and Koeppe, 1971; Barratt and Flavell, 1975). Recent studies have shown a banding-pattern difference between N and T cytoplasm mitochondria when the carboxymethylated acidic chloroform/methanol extract of the submitochondrial particles and of the partially purified mitochondrial ATPase complex are analyzed by isoelectric focus-

144

H. K . SRIVASTAVA

ing (Barratt and Peterson, 1977). Present evidences from work on Neurospora crassa and Saccharomyces cerevisiae suggests that mitochondria synthesize probably no more than a dozen hydrophobic polypeptides, which are mainly associated with the mitochondrial inner membrane and four with the oligomycin-sensitive ATPase complex (Schatz and Mason, 1974). The results on mitochondrial banding pattern difference (Barratt and Peterson, 1977) therefore provide additional evidence that cytoplasmic DNA of maize specifies components of the inner mitochondrial membrane. Electron microscopic studies of maize root cap cells exposed to toxins of race T of B . maydis indicate that mitochondria are affected very early after treatment (Aldrich et a l . , 1977). Studies on microsporogenesis showed that mitochondria, but not plastids, degenerate in tapetal cells during meiosis of T, but not N lines, resulting in vastly deranged mitochondria in T lines by the free microspore stage (Lee, 1976; Warmke and Lee, 1977). The mechanism of CMS involving mitochondria in maize has also been studied by isolating mitochondria from seedlings and various anther stages (premeiosis, meiosis, and mature) and analyzing cytochrome oxidase and succinic dehydrogenase biochemically and electrophoretically (Watson et a l . , 1977). CMS anthers exhibited a lack of biochemical activity and fewer isozymatic bands for cytochrome oxidase, while no apparent differences were detected between N and T anthers for succinic dehydrogenase. Considering the facts that cytochrome oxidase is a partially mitochondrially coded enzyme (Ebner et a l . , 1973) and succinic dehydrogenase is controlled completely by the nuclear genome (Vary et al., 1970), the results showing striking differences with cytochrome oxidase on a developmental level provide indirect evidence that mt-DNA may be responsible for CMS. The possibility that mt-DNA may be the site of the genetic factors responsible for CMS has recently been investigated (Levings and Pring, 1976, 1977). When the mt-DNAs prepared from maize with N and T cytoplasm were subjected to restriction enzyme digestion and fragment analysis, the banding patterns were readily distinguishable. The T mt-DNA was characterized by the presence of two faster pieces of molecular weight (3.42-3.48) x 106 and (4.01-4.10) x l(r' (Pring et a l . , 1977). Electron microscopic studies of isolated DNA fragments further showed circular molecules corresponding in size to the two unique DNAs, suggesting that these DNAs may exist in a circular configuration in mitochondria. These authors have hypothesized that the unique DNAs are associated with mitochondria of the CMS cytoplasm group and, in fact, may represent the episomes or fertility elements already postulated by Laughnan and Gabay (1975). Their hypothesis requires that the cytoplasmic element be capable of transposition to, and stabilization in, a chromosomal site. This stabilization would confer fertility, while a cytoplasm mitochondrial location would be associated with male sterility. A possible precedent has been shown in yeast (Grif-

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

145

fiths et al., 1975), where “omicron” DNA was hypothesized to be a mt-DNA episome. Whether ct-DNA is also involved in carrying the traits of male sterility is not clear and requires further investigation. Recently in an extensive effort to study the molecular mechanism of male sterility in maize, Pring and Levings (1978) analyzed both mt-DNA and ct-DNA using restriction endonuclease fragment analyses and provided evidences that the genetic information for CMS is contained in mt-DNA. Restriction fragment analysis of maize ct-DNA failed to demonstrate differences between N and T cytoplasm, with the exception of the CMS-s cytoplasm when digested by HindIII. In the case of Nicotiana, however, ct-DNA seems to be involved. It was reported that fraction 1 protein from a CMS variety of N. tabacum has a pattern of two small subunit polypeptides identical to those of normal N. tabucum (Chen et a/., 1975). However, the three subunit polypeptides were identical to those of an Australian species. Thus it was clear from these results and subsequently confirmed from the breeding records that the CMS plant had been derived from an original cross between the Australian species 0 x N. tabucum 8 . The F, hybrid is male-sterile, and it is maintained by continuous use of N. tabucum as the male. Chen et al. (1976) further provided evidence from an extensive study on seven CMS tobacco cultivars that the genetic factors for male sterility are contained in ct-DNA. Their most convincing evidence is that a male-sterile cultivar was discovered as a spontaneous malesterile mutant from N. tabucum, and that mutation was subsequently shown to be accompanied by a mutation in the ct-DNA genes coding for fraction 1 protein (Berbec, 1974). While much more attention is required in the future to understand precisely the molecular mechanism of male sterility, the present results do suggest that both mt-DNA and ct-DNA are deeply involved in the phenomenon of CMS in higher plants. It is proposed as a working hypothesis that at the operational level, organelle DNA produces some special protein that acts as a repressor molecule to the nuclear genes coding for key enzymes of organelle function. In case of restorer lines the nuclear restorer genes of male sterility could produce another protein to inactivate the repressor molecule. This perhaps explains fertility, male sterility, and its restoration. Some observations in support of the view that the process of mitochondria1 metabolism in male-sterile plants is disturbed due to abnormal mt-DNA have been made (Srivastava et al., 1969; Berville and Demarly, 1970; Berville et al., 1977). E. NUTRITIONAL QUALITY I N WHEAI

Increased plant proteins can be achieved from improved crop varieties and from better management practices. An understanding of the important cellular

146

H. K. SRIVASTAVA

components and mechanisms involved in the synthesis of amino acids and proteins may prove to be useful in devising new strategies to the nutritional improvement of protein. The cytoplasm, as the major component of the cellular environment, has been shown to affect pigment synthesis and chlorophyll mutants in barley (Nilan, 1964; Robertson, 1971; von Wettstein er a l . , 1974). In wheat, the genome of the cultivar “Salmon” was found to perform better in Aegilops caudatu cytoplasm than its own (Kihara and Tsunewaki, 1964). Similarly, field trials of several alloplasmic lines derived by substituting the genomes of two hard red spring Triticum aestivurn cultivars “Chris” and “selkirk” into cytoplasms from T . macha, T. dicoccoides, or Aegilops squarrosa, showed some alloplasmic lines to be significantly earlier-heading and higher-yielding than the original cultivars (Busch and Maan, 1974; Maan, 1976, 1977). The effect of wheat cytoplasm on meiosis of wheat-rye addition lines and hexaploid triticales has been observed (Roupakias and Kaltsikes, 1977; Gour and Singh, 1977). Other earlier reports suggested that T . aestivum cytoplasm has considerable influence on the morphology of triticale ( x Triticosecale Wittmack) (Logodivona, 1962; Berova and Logodinova, 1972; Soler and Jouve, 1971; Shulyndin and Maksimov, 1972; Shulyndin, 1972; and Hsam and Larter, 1974). Based on the cytoplasmic-nuclear ratio in hexaploid triticale, Sisodia and McGinnis ( 1970) reported that hexaploid wheat cytoplasm is more compatible than tetraploid cytoplasm. These studies in a sense are related to the influence of T . aestivum cytoplasm from different sources and ploidy levels in the synthesis of triticale. The genetic control of the content, amino acid composition, and processing properties of proteins by means of cytoplasmic organelle-nuclear gene interaction in wheat has been thoroughly analyzed (Konzak, 1977). Nitrate reductase is the first, and possibly the main, limiting enzyme involved in the synthesis of protein from nitrate, the primary form of N utilized by plants. Cereals can also utilize some ammonium N, but the ammonium form of N is rarely available in soils (Maynard et al., 1976). The nitrate reductase function (NO, --+ NH,) seems to be located within chloroplasts (Dalling et al., 1972; Magalhaes et al., 1974). There is indication that the genes for the synthesis of nitrate reductase, so far found to be localized in chloroplast ribosomes, may be under the direct control of ct-DNA (Neyra and Hageman, 1974). Ferredoxin, also located in the chloroplasts, is required for nitrate reductase activity (Wildman et a l . , 1975; Klepper, 1975). Several restorer lines in T. timopheevi cytoplasm are known to possess greater protein content (Johnson and Mattern, 1975). Differences between the electrophoretic patterns of acid phosphatase proteins produced by triticales with cytoplasm from wheat versus cytoplasm from rye have also been observed (Torres and Hart, 1976). Marked differences in the effects of different cytoplasm sources on gene expression in hexaploid and tetraploid wheat, including effects on spike fertility and characteristics as well as on

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

147

plant development and vigor, have been reported (Srivastava and Sarkissian, 1970; Maan, 1975; Suemoto, 1973; Tsunewaki and Endo, 1973; Mukai and Tsunewaki, 1975). Two malate dehydrogenase (MDH) isozymes of rye were absent when single chromosome of rye was added to wheat, suggesting that rye cytoplasm plus rye chromosome may be required for their expression, or that the rye MDH genes were epistatic to those of wheat (Bergman and Maan, 1973). These results provide indirect evidence that the control of certain proteins in the Gramineae involves intergenomic interaction. Close cooperation between nuclear and cytoplasm genes may be considered an important operational mechanism for an overall functioning of the organism, for most of the enzyme systems present in mitochondria and chloroplasts are coded under the influence of intergenomic interaction.

IV. GENETIC IMPLICATIONS OF INTERGENOMIC INTERACTIONS Sirks (1938) said, “Geneticists forget that genes as such can show their effects only through the instrumentality of the plasm. It is fitting, therefore, that attention be drawn to a number of facts which point to a more or less independent action of the ‘plasmagene’ as a determiner or a contributor to inheritance.” The existence of multiple genomes in a plant cell raises questions pertaining to (1) what are the contributions of plasmon and nuclear genomes toward the structural organization of chloroplasts and mitochondria, (2) where is the genetic code for formation of structural and functional proteins of these organelles actually localized, and (3) whether intergenomic interaction (complementation or cooperative interactions between nuclear and organelle genomes) operates through transcription-translation systems, protein-synthesizing systems, or both. The answers to these questions in higher plants are only partial, but efforts are presently being made in various laboratories to gain insights into the operational mechanisms of plasmon-genome interaction, self-organization, replication, and genetic transmission of organelles. Nuclear-cytoplasmic interactions leading to alterations at gross phenotypic levels are generally accepted to occur by most plant geneticists. While the principal role of the nuclear genome for transcription and translation of most structural or functional components of organelles can not be denied, sufficient data on the formation of some key enzymes of both chloroplasts and mitochondria due to complementation of two distinct genomes are presently available to support a substantial involvement of organelle genomes in the regulation of their own catalytic activities during cell metabolism. A hypothesis based on intergenomic interaction to interpret the operational mechanism of photosynthetic and respiratory efficiencies and the resulting

148

H. K. SRlVASTAVA

heterosis as commonly observed in crop plants has recently been proposed (Srivastava, 1981). A. CHLOROPLAST THYLAKOID PROTEINS

Division of chloroplasts in vitro, found by Ridley and Leech (1970), speaks in favor of their genetic and biochemical autonomy and suggests that ct-DNA is essential as a master molecule for chloroplast formation and development. Differentiation of plastids, however, depends on interaction of both ct-DNA and nuclear genome (Leech, 1976). Photosynthesis is only one of the many biological functions of chloroplasts; others include protein synthesis, reproduction, and nucleic acid replication. In higher plants chloroplast proteins may account for more than 60% of the total cellular proteins. Approximately 50% of the proteins in the chloroplast are soluble in aqueous solutions and are localized in the stroma. In green leaves, the major soluble chloroplast protein is fraction 1 protein. Since this leaf protein (referred to RBPCase, ribulose- 1,5-biphosphate carboxylase/ oxygenase) is found in all photosynthetic organisms, it is the most abundant single protein species on earth. The genetic code for the transcription of the functional molecule of RBPCase, as it is elaborated in greater detail in the subsequent subsections, is contributed by both ct-DNA and nuclear DNA. The remaining chloroplast protein is made up of insoluble or structural proteins that constitute the thylakoid membranes. Photosystem I (PSI), photosystem I1 (PsII), and light-harvesting chlorophyll proteins are found in this fraction. Two major components in the chloroplast membrane of higher plants contain the bulk (75%) of the energy-trapping chlorophyll molecules. These components are the PSI chlorophyll protein and the light-harvesting chlorophyll a h protein (Thornberger, 1975). Genetic studies indicate that PSI chlorophyll protein is coded for by ct-DNA (Hermann, 1971). This conclusion is based primarily on the genetic analysis of a mutant en :alba-1 of Antirrhinum majus. This mutant does not synthesize the PSI chlorophyll protein. Since this mutant is reported to be in the plastome (Hermann, 19711, the site of coding information was construed to be in ct-DNA. Machold and Aurich (1972) have provided evidence that this chlorophyll protein is synthesized on chloroplast ribosomes. This supports the notion that ct-DNA may contain the genetic information for this chlorophyll protein. In contrast nuclear DNA has the coding information for the light-harvesting chlorophyl a h . This inference is deduced from the interspecific hybridization studies of the Nicotiana species and supported by the antibiotic studies that demonstrate that this protein is synthesized on the cytoplasmic ribosomes (Kung et al., 1972). Many reports provide further evidence that mutation of both nuclear and chloroplast genes causes defects in PSI and PsII. Heber and Gottschalk (1964)

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

149

have described a nuclear mutant of Viciafubu with a block in PSI. Chloroplasts of this mutant are incapable of NADP photoreduction. Bishop (1 972) induced a nuclear mutation in Scenedesrnus obliquus with a block in the synthesis of cytochrome f-552, which caused a suppression of cyclic phosphorylation. A defect in PsII caused by mutation on one nuclear gene in the pea lethal mutant and a block of electron transport chain (ETC) components between PSI and PsII in the cotton nuclear mutant have been observed (Nasyrov, 1978). All these results provide substantial evidence for the existence of an interaction between nuclear-plasmon genetic systems ensuring the complete development of photosynthetically active thylakoid membranes. Mutations in genome-plastome genes leading to defects in light-harvesting electron transport, photophosphorylation, or NADP reduction will all result in impaired C 0 2 assimilation. The assemblage of the photosynthetic unit of the thylakoid membrane therefore requires the complementation of both plastome and nuclear genomes and of both chloroplast and cytoplasmic ribosomes. In order to understand the precise genetic control of chloroplast organization, Alberts er ul. (1973) examined the time of appearance of the two major chlorophyll proteins in young green beans (Phaseolus vulgaris). The lightharvesting chlorophyll a l b protein appeared first, followed by the PSI chlorophyll protein. Thus the major chlorophyll a l b protein was first assembled in the thylakoid membrane and the PSI chlorophyll protein was inserted later. These results could be interpreted to mean that the biosynthesis of the lightharvesting chlorophyll a l b proteins serves as an initial step in the assemblage of thylakoid membrane and that the nuclear genome probably regulates the development of the thylakoid membrane. By using the nuclear transplantation technique in Acetabularia, such regulation was also demonstrated at the genomic level (Kloppstech and Schweiger, 1974). Marked differences in the electrophoretic gel pattern of chloroplast membrane proteins from A . calyculus and A . rnedirerrunea have been found. After isolated nuclei of A . rnediterrunea were transplanted into basal parts of enucleated A . calyculus, the gel pattern of some of the chloroplast membrane proteins resembled that of A . rnediterrunea, the source of the nucleus. The converse results were also obtained from a reciprocal transplantation experiment. This indicates that genes for some chloroplast membrane proteins and some ribosomal proteins in Acetabularia are in the nuclear genome and others may be located in the plastome. B. CHLOROPLAST MUTANTS

The genetic backgrounds of many chloroplast mutations with defects in photosynthetic CO, fixation have been described (von Wettstein and Kristiansen, 1973). Most chloroplast mutants investigated so far in higher plants do not

150

H. K . SRIVASTAVA

survive when grown in the field; the stocks are therefore maintained in the heterozygous condition, and in all cases the progeny arising from a heterozygous plant exhibit a Mendelian segregation of three wild-type to one mutant seedling. The homozygous mutants are easily identified by color; they are all yellow or paler green than wild-type plants. Using cytogenetic analysis of pea mutants of chlorina type, Blixt (1968) identified 15 nuclear genes controlling the formation of plastid pigments. An extensive genetic analysis of 198 induced recessive lethal chloroplast mutations in barley (Hordeurn vulgare) revealed that there are 86 nuclear genes responsible for the regulation of chloroplast development (von Wettstein et al., 1971). The 12 mutants represent mutations in 12 different nuclear genes. Five of them (xantha-f, g, h, 1, n ) were classified as structural genes responsible for the conversion of uroporphyrinogen to protochlorophyllide; three genes (tigrina 12, “infrared” 5 and 6 ) were identified as possessing regulatory functions. Unlike struatural genes, two regulatory genes exhibited a similar dominant effect in wild-type and mutant alleles in heterozygotes. The positive correlation between chlorophyll contents and photosynthetic rates on a gross chlorophyll basis among five barley mutants (viridis-b, 1, d, c, and rn) suggests that the reduced absorption of light by the lower amounts of chlorophyll in these mutants is the major factor responsible for their reduced photosynthetic rates on a gram fresh weight of leaves (Carlsen, 1977). On the contrary, if a mutant is characterized by a higher photosynthetic rate in bright light than the wild type, as has been observed in the case of an exceptional mutant viridis-k (Carlsen, 1977), such a result could possibly be interpreted on the basis of intergenomic interaction leading to superior biochemical quality of light-harvesting chlorophyll in the photosynthetically efficient mutant chloroplast. Recognizable mutant plastids have been found in the subgenus Oenothera (Stubbe, 1957), and they could be distinctly identified by the pattern of variegation. Oenothera suaveolens had white defective plastids in some leaf areas, yellow-green defective plastids in other leaf areas, and normal green ones in still others. In some complex Oenothera hybrids there were plants that had all three types of the suaveolens plastids along with normal plastids from Oenothera lamarckiana. Renner ( 1919) first demonstrated that variegation, seedling lethality, or embryo abortion of certain species hybrids were due to incompatibilities between the plastome and nuclear genomes. On the basis of these incompatibilities and other general cytogenetical observations, Kutzelnigg and Stubbe (1974) have recognized five chloroplast genomes (plastome I-V) in subgenus Oenothera, each of which is adapted to distinct nuclear genome class. The close complementation of the nuclear and plastome genes in Oenothera has aroused interest among investigators to devise model experiments at the molecular level to make a distinction between the five plastomes of subgenus Oenothera in respect of their genetic codes. The hypothesis that intergenomic interaction between different allelic forms of nuclear and plastome genes plays a major role in

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

151

the evolution of various Oenothera species is attractive but nevertheless demands specific experimental evidence.

c. RIBULOSE-I.5-BIPHOSPHATE c A R B O X Y L A S E / o X Y C E N A S E The use of ribulose- 1,5-biphosphate carboxylase/oxygenase (RBPCase, EC 4.1.1.39) as a marker enzyme of both nuclear and chloroplast genomes in genetical experiments at the subcellular level in higher plants has recently been advocated. This carboxylase comprises up to 50% of the total soluble protein in leaves, and thus is the most abundant protein in nature. RBPCase in the chloroplast of higher plants participates both in the photosynthetic carbon reduction cycle and in the photorespiratory carbon oxidation cycle (Jensen and Bahr, 1977). In photosynthesis the enzyme catalyzes the only reaction known to give a net increase in the amount of carbon compounds by fixing a molecule of CO, to generate 2 mol of glycerate 3-phosphate, while during photorespiration it functions as an oxygenase, converting ribulose-l,5-biphosphate into glycerate 3-phosphate and phosphoglycolate. The two molecules of glycolate thus yielded are converted in the peroxisomes and mitochondria (Lorimer et al., 1977) into glycerate 3-phosphate with the release of a molecule of CO,. The glycerate 3-phosphate reenters the carbon reduction cycle, making more C 0 2 available for fixation by the carboxylase reaction. The dual function of RBPCase seems to protect chloroplasts against temporary COPdeprivation in the light: it permits the internal generation of C02, the dissipation of harmful photochemical energy through the production of glycolate by oxygenation, and the partial recovery of the diverted carbon. Wildman and co-workers in their pioneer work on the physical chemistry of leaf proteins first discovered this large molecular fraction 1 protein (MW 550,000). The RBPCase molecule is an oligomer of 16 subunits; 8 are termed large (MW 55,000) and carry the catalytic site (Nishimura and Akazawa, 1973), while the other 8 are termed small subunits (MW 12,000-15,000). In addition to the two major products, about 90 other minor products of the enzyme have been separated from isolated pea chloroplasts by electrophoresis on one- and twodimensional polyacrylamide gels (Highfield and Ellis, 1978). One major product is tightly bound to the chloroplast lamellae, and is associated with the ATPsynthase complex. The other major product is soluble, and is identified as the large subunit of RBPCase, the key enzyme in photosynthesis and photorespiration. Peptide mapping and characterization by isoelectric focusing (Kung, 1976; Sakano et al., 1974) of the subunits from Nicotiana species and their respective F, hybrids have shown that the large subunit is maternally inherited and that the small subunit is inherited in a Mendelian fashion. The large subunit is encoded in ct-DNA in maize (Coen et al., 1977), and Chlamydomonas (Gelvin et al.,

152

H. K. SRIVASTAVA

1977), and is synthesized from a messenger RNA lacking poly(A) (Highfield and Ellis, 1978). A ct-DNA fragment from maize generated by restriction endonuclease was cloned and directed the synthesis of the large subunit in an in vitro linked transcription-translation experiment. Messenger RNA for the large subunit has been isolated from chloroplast polyribosomes in Chlarnydornonas with the aid of antibodies and was shown to hybridize with ct-DNA. By contrast, the small subunit is encoded in nuclear DNA; this subunit is not labeled when isolated chloroplasts incorporate labeled amino acids into protein, but it has been identified as an in vitro translation product of cytoplasmic ribosomes (Roy et al., 1976). The small subunit is synthesized on cytoplasmic polyribosomes only in a precursor form (Dobberstein et al., 1977). The precursor form contains an additional amino acid sequence of 44 amino acids at the N-terminus that is cleaved off by an endoprotease in connection with its transfer across the chloroplast membrane. The involvement of intergenomic interaction therefore is envisaged in the formation of a complete and functional molecule of RBPCase having catalytic roles for both photosynthesis and photorespiration. The recent data on the primary structure of RBPCase from a number of plant species confer interpretations in terms of evolution, adaptation, and function of this key enzyme. The marked variation in the N-terminal sequence of the 110-120 amino acids of the small subunit found in different plant species using the automatic Edman degradation technique reveals the existence of allelic polymorphisms of the nuclear genome (Gibbons et al., 1975; Poulsen et al., 1976). Nuclear gene polymorphism has also been reported for the C-terminal end sequences of the small subunit as determined by carboxypeptidase digestion (Sugiyama and Alcazawa, 1970; Strobaek et al., 1976; Poulsen, 1977) in spinach (-Phe-Leu-Thr-Tyr-COOH), tobacco (-Thr-Val-Leu-Tyr-COOH), and barley (-Leu-Tyr-Phe-Val-Asn-AlaCOOH). Sequence information on the large subunit is so far only available for barley and spinach (Poulsen, 1978; Stringer and Hartman, 1978). The large subunit of barley contains 9 methionine residues and yields up to 10 fragments after cyanogen bromide cleavage. Six fragments accounting for about 350 of the 490 residues have been successfully separated and partially sequenced. Comparison of the corresponding sequences from both barley and spinach has revealed 4 amino acid replacements among 35 analyzed residues (Stringer and Hartman, 1978). The results indirectly suggest species differences in the nucleotide sequence of the plastome genes exclusively responsible for coding the large subunit of RBPCase. The intergenomic complementation as evidently involved in the biosynthesis of RBPCase and the occurrence of polymorphic forms of this enzyme in many plant species have led investigators to uncover the specific nature of “genomeplastome” interaction in the subgenus Oenothera (von Wettstein et a l . , 1978). An isolation procedure for Oenothera RBPCase and a peptide mapping proce-

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

153

dure to characterize the subunit polypeptides have been worked out (Holder, 1976, 1978). In order to demonstrate whether the five classes of chloroplast genomes of Oenothera subspecies in the classification scheme suggested by Kutzelnigg and Stubbe (1974) carry different allelic forms (nucleotide sequences) of the genes for the large subunit of RBPCase, von Wettstein et al. (1978) intensified their efforts to precisely define the peptide sequences of the isolated S-carboxymethylated large subunit of the enzyme digested with chymotrypsin. The comparative data on the large subunit chymotryptic peptide maps of nine Oenothera subspecies are summarized in Table IV. Some marked variation in terms of the presence or absence of four peptides designated A, B, C, and D in the five plastome groups of Oenothera has been observed. Although the detectable differences in the peptide maps of the large subunit of RBPCase from the five plastome groups of Oenothera remains small, the results reveal the occurrence of single amino acid replacements among the plastome groups, reflecting nucleotide sequence variation and thus evolutionary divergence of Oenothera subspecies based on plastome genes. The differences in plastome genes among Oenothera subspecies could have been produced due to nucleotide substitutions and recombination leading to allelic polymorphism of chloroplast genomes. The large RBPCase subunit proteins are similar or identical for an Aegilops speltoides accession, the tetraploid wheats, T. dicoccum, T. turgidum, T . timopheevi, and the hexaploid wheat T. aestivum, but are distinctly different in electrophoretic mobility from those of T. boeticum, T. monococcum, T . urartu, and Ae. squarrosa (Chen et al . , 1975). Their results (albeit based on single accession of each Table IV Peptide Maps of the Large Subunit of Ribulose-l,5-biphosphate Carboxylase/Oxygenase from Nine Oenotheru Subspecies Species

Chloroplast genome"

Peptide mapsb

0 . hookeri 0 . strigosu 0 . elata 0 . biennis (Miinchen) 0 . biennis (Citronelle) 0 . lamarckiana 0. parviflora (ammophila) 0 . purvifloru (atrovirens) 0 . urgillicolu

Plastome I Plastome I Plastome I Plastome I1 Plastome 111 Plastome I11 Plastome IV Plastome IV Plastome V

A- B- C+ D+ A- B- C- D+ A- B+ C+ D+ A- B+ C+ D+ A- B+ C+ D+ A- B- C- D+ A+ B+ C+ DAt B+ C+ DA+ B- C+ D-

"The classification of chloroplast genomes is according to Kutzelnigg and Stubbe (1974). bAdapted from von Wettstein er ul. (1978). The superscript + or - after each letter indicates the presence or absence of four chymotryptic peptides in the large subunit of the enzyme.

154

H . K . SRIVASTAVA

species) suggest that the plastome genes are similar for Ae. speltoides and cultivated polyploid wheats, supporting other evidence that an Aegilops species similar to Ae. speltoides may be the B genome progenitor of modem hexaploid wheat.

D. PROTOPLAST FUSION Modem advances in plant cell culture demonstrate that protoplasts from leaf mesophyll cells can be induced to regenerate into entire plants, and that they can be stimulated to fuse by defined experimental manipulations. A combination of these two techniques permits the fusion of protoplasts isolated from two different species and the regeneration of a somatically produced hybrid plant. The two major objectives of producing parasexual hybrids through protoplast fusion in higher plants are ( i ) the creation of new hybrid plants having mixed populations of mitochondria and chloroplasts in their cells, which cannot be achieved sexually because of the unequal or negligible contribution of cytoplasm from the paternal gametes, and ( i i ) the creation of new plant species by somatic fusion of protoplasts that cannot be sexually hybridized-a creation that could not have been visualized until very recently. The first parasexual hybrid derived from the fusion of protoplasts of Nicotiana glauca and Nicotiana langsdofii was obtained by Carlsen et al. ( 1 972). Analysis of the polypeptide composition of RBPCase from this hybrid shows that the polypeptide of the large subunit is similar to that of N . glauca (Kung et al., 1975; Kung, 1976). The large subunits of N . langsdofii in the hybrid were not detected, suggesting that the plastome coding for the large subunit of RBPCase from N . langsdofii is not expressed. However, the hybrid contains the small subunit of both species, indicating that both nuclear genomes are equally expressed in the hybrid. The explanation as to why only one chloroplast genome was transmitted to the hybrid for expression of the large subunit of RBPCase is not yet fully available. Part of the answer comes from the work of Smith et al. (1976), who produced 23 mature hybrid plants by the fusion of protoplasts of N . glauca and N . langsdofii. In a total of 20 hybrid plants analyzed, 12 of them have the plastome of N . langsdofii expressed, whereas 7 have only the plastome of N. glauca expressed. There was, however, one hybrid out of a total of 20 plants in which chloroplast genomes of both species were equally expressed. When the plastome of either species is expressed, the hybrid exhibits the characteristic tissue morphology similar to that derived from sexual hybridization. In the case where plastomes of both species are expressed, the hybrid plants appear abnormal. This is probably caused by incompatibility of interaction between chloroplast and nuclear genomes. The result of such incompatible interaction between the combined plastomes and either one of the two nuclear genomes may probably lead to some disturbance of normal metabolic activity and hence abnormal plant development. More study

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

155

from a large plant population is, however, required to resolve the mystery of uniparental or biparental transmission of organelle genomes in higher plants. Somatic hybrid plants of potato (Solanum tuberosum) and tomato (Lycopersicum esculentum) from fused protoplasts have been successfully produced (Melchers et d.,1978), and ingenious biochemical analyses of RBPCase as phenotypic markers of chloroplast and nuclear genomes from the resulting intergeneric hybrids offer for the first time some clues regarding allelic differences between plastome genes for the large subunit as well as between nuclear genes for the small subunit. The variation patterns based on isoelectric focusing of small subunit bands from four tomato-potato hybrids (6a, 6b, lb, 7a) together with those of tomato and potato for comparative purposes are presented in Fig. 1. It can be seen that all four hybrids contain the three prominent tomato small subunit bands. In addition, they all contain the prominent potato small subunit band and two of the minor bands. One of the minor potato small subunit bands is absent in all the hybrids, while the other minor band is present in only two of the four intergeneric hybrids-6b and 7a. These results reveal that the native RBPCase in the hybrids contains the small subunit products resulting from the expression of both tomato and potato nuclear genes, as could be expected in parasexual hybrids between potato and tomato following Mendelian inheritance. The relative discrepancy for the presence or absence of the four minor potato small

FIG. 1. Isoelectric focusing of S-carboxymethylated RBPCase small subunit from tomato, potato, and their hybrids. (T)tomato, (P) potato, (6a, 6b, Ib, and 7a) interspecific hybrids. Redrawn from Melchers er a / . (1978).

156

H. K . SRIVASTAVA

subunit bands as observed in the case of the hybrids, however, remains unknown. The RBPCase large subunit polypeptides from the hybrid plants are also compared with the tomato and potato large subunit polypeptides in Fig. 2. For each species, the large subunit exhibits two distinct major bands together with at least three less identifiable minor bands in the original photograph (Melchers et al., 1978). In three of the four plants (6a, 6b, and lb), the large subunit polypeptides follow the banding pattern of tomato, indicating uniparental transmission of chloroplast genes in these somatic hybrids. The position of the densely stained major bands of the large subunit from the fourth hybrid (7a) is, however, identical to those of the large subunit from potato. These results were interpreted by the authors to mean that in three out of four hybrids the functional ct-DNA was transmitted from tomato, whereas in the fourth the ct-DNA was contributed from potato. Although the genetic interpretation of these findings remains restricted for lack of more data from a significant number of tomato-potato mature hybrids, the results of Melchers et al. (1978) nevertheless indicate that the products of both tomato and potato nuclear genomes are present in the RBPCase oligomer, whereas the large subunit products are evidently derived from either tomato or potato, i.e., uniparentally. Why should there be expression of plastome genes from only one of the parents during crossing or protoplast fusion? If by sexual hybridization or parasexual protoplast fusion two chloroplast genomes are brought into one cell, why then does only one or the other organelle genome win out and functionally establish itself during the development of the individual? It is not known if a particular type of plastome or ct-DNA produces a repressor molecule against the other or whether plastome competition and its functional preference for a specific nuclear genome is responsible for the perpetuation of

large Subunit hybrids

T

--

P

6a

6b

lb

7a

@mB

-o-@?m 0.0

FIG. 2. Isoelectric focusing of S-carboxymethylated RBPCase large subunit from tomato, potato, and their hybrids. (T)tomato, (P) potato, (6a, 6b, Ib, and 7a) interspecific hybrids. Redrawn from Melchers er al. (1978).

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

157

ct-DNA from only one of the parents in the hybrid. It is likely that the four parasexual hybrid plants so far analyzed for RBPCase were aneuploids rather than true amphidiploids. The other cause of the predominant representation of tomato ct-DNA in three of the hybrids may have been due to the differentiated state of the tomato chloroplast at the time of fusion; the potato protoplasts, on the contrary, contained undifferentiated protoplastids during protoplast fusion (Melchers et d . , 1978). Chiang (1968) has reported the results of experiments in Chlamydomonas using parents pregrown with 3H/'4C-labeled adenine to distinguish the parental ct-DNAs in zygotes, and later (1971) using both CsCl density and radioisotope labeling simultaneously. He reported equal contribution of label from both parents in DNAs extracted several days after mating and from zoospores after meiosis and zygote germination, and concluded that ct-DNAs from both parents were conserved in the zygote. Human-rodent hybrid somatic cell lines were examined for the presence of human and rodent mt-DNA, separable by buoyant density in gradients, and identified by hybridization to complementary RNA (Horak et al., 1974). Of the lines examined, 13 out of 18 showed the presence of both human and rodent mt-DNA sequences covalently linked and hence probably combined by recombination. In other studies by Wallace et al. (1976) hybrids were formed between human chloramphenicol-resistant (mitochondrial capr) and mouse chloramphenicol-sensitive (caps) cell lines. Some of the hybrid lines exhibited growth in the presence of chlorarnphenicol and were classified as cap' like the human parent cells, but contained only or predominantly mouse mt-DNA sequences as judged by hybridization to complementary RNA. Such cell lines could have been produced by recombination, but they might also have resulted from selection of capr mutations in the mouse mt-DNA molecules. E. MECHANISMS OF ORGANELLE DNA TRANSMISSION

There seem to be no set rules at the cellular level for organelle DNA transmission. In oogamous species (gametes from male and female parents differ morphologically) showing purely maternal inheritance, the paternal organelles may be excluded at any step in the reproductive processes (Paolillo, 1974; Hageman, 1976; Birky et al., 1978). Paternal organelles are visibly destroyed in the zygotes of some mammals and algae (Bandlow er al., 1977; Kirk and Tilney-Bassett, 1978). Plants showing biparental inheritance of chloroplast genes almost always have a strong maternal bias; the single exception is the Japanese sugi, Cryptomeria japonica. This tree, the only gymnosperm studied in much detail, exhibits a great preponderance of paternal zygote, a few maternal zygote and biparental zygote, and an overall bias in favor of paternal chloroplasts (Ohba et al., 1971). In the four thoroughly studied species that show biparental inher-

158

H . K . SRIVASTAVA

itance of organelle genes, Chlamydomonas, yeast, Pelargonium, and Oenothera, there is good reason to believe that most or all zygotes receive intact organelles from both parents, and degeneration of entire organelles has not been seen in the zygotes of Chlamydomonas (Cavalier-Smith, 1970), Oenothera (Meyer and Stubbe, 1974), and yeast (Osumi et a l . , 1974). Two pertinent questions could be asked in respect of the precise molecular mechanism of organelle DNA transmission: ( a ) What happens to the mitochondrial or chloroplast genomes that do not appear in the progeny of uniparental transmission? ( b ) Why does the same crossing or mating combination produce both maternal and biparental transmission, and sometimes paternal transmission as well? The validity of the biparental transmission of ct-DNA as demonstrated by Chiang (1968, 1971) has, however, been questioned by Sager (1977), who is of the view that some nuclear factor produced by the maternal parent is required for effective degradation of the paternal cytoplasmic genome, i.e., for maternal inheritance. Sager’s model is based on an analogy with bacterial host modification restriction systems. Ct-DNA from the mt+ (“maternal”) and mt- (“paternal”) parents is supposed to be differently marked, possibly by methylation of the mt+ ct-DNA, prior to fusion of the parental chloroplast in the zygote. A gene closely linked to the mt+ allele produces a restriction enzyme that specifically destroys the unmodified mt- ct-DNA. Destruction or degradation is complete in most zygotes, which consequently transmit only chloroplast genes from the mt+ parent to their progeny and are therefore referred to as “maternal” zygotes even though the Chlamydomonas is an isogamous species (gametes from both parents are alike morphologically). Degradation of mt- ct-DNA fails to occur or is incomplete in rare zygotes, which can then transmit both mt+ and mt- chloroplast genes and are biparental zygotes. Paternal zygotes, which transmit only mtchloroplast genes, are found in significant numbers after treatment of the mt+ gametes with inhibitors or in the presence of the mat-I gene mutation. In an attempt to obtain physical evidence for degradation of the ct-DNA from the mtparent in zygotes in support of the model, Sager and Lane (1972) labeled parental ct-DNAs differentially with [lsNP4N]- or [3H/’4C]adenine, and were unable to detect the paternal ct-DNA on zygotes; they also found a density shift in the maternal ct-DNA, such as might result from methylation. Schlanger and Sager (1974) report that when the mt+ gametes are irradiated with UV the mt- parental ct-DNA is conserved and mt+ ct-DNA is partially or completely degraded, corresponding to the enhanced transmission of mt- markers and reduced frequency of maternal zygotes. Chiang (1976) has verified his own results (1968; 1971) by a set of newly designed experiments in Chlamydomonas, and the recent results suggest that up to 95% of both parental ct-DNAs is degraded in zygotes, but the paternal ct-DNA is degraded more rapidly. This degradation occurred simultaneously with replication and recombination of mt+ and mt- ct-DNAs. Birky

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

159

(1978), however, considers it too early to draw any firm conclusions about the destruction of paternal or maternal ct-DNA in Chlamydomonas zygotes. Another hypothesis for uniparental inheritance in Chlamydomonas that does not depend directly upon differential destruction of ct-DNA was devised by Adams el af. (1976). They proposed that zygotes replicate only a small sample of ct-DNA molecules due to competition for a limited number of replicating “sites.” This is a stochastic process: if the sample by chance contains only molecules from one parent, the result is a uniparental zygote; if it contains molecules from both parents, a biparental zygote results. Adams et al. (1976) further explains the great preponderance of maternal zygotes by assuming that maternal ct-DNA molecules possibly occupy preferential replication sites. Another hypothesis based on a “multicopy model” with a reductional rather than equational segregation of organelle DNA has been proposed (van Winkle-Swift, 1976, 1978). In this model the ct-DNA molecules contributed by both parents tend to remain in separate nucleoids. At the first meiotic division, these nucleoids segregate from each other; gene loci will also segregate unless there has been recombination or transfer of whole molecules between the nucleoids. In subsequent generations, it is assumed that segregation of ct-DNA continues to be mainly reductional due to the maintenance of a fairly rigid spatial arrangement of molecules in dividing nucleoids. Failure of gene segregation is again due to recombination or mixing of molecules. With such a model, high segregation rates are achieved even with large numbers of molecules, provided that the rate of mixing and recombination of ct-DNA molecules from the two parents is sufficiently low or that homogeneity is maintained within nucleoids. Some marked effects of paternal cytoplasm derived from male gametes on the inheritance of yield component traits and grain quality characters in hybrid maize have been observed (Fleming, 1975; Rao and Fleming, 1978). These workers, however, consider that the expression of paternal cytoplasm (ct-DNA and mt-DNA) in the biparental zygote (offspring) may be influenced by the maternal organelle genomes. Thus the real fate of organelle DNA during maternal : biparental :paternal zygotes in Chlamydomonas as well as in higher plants remains an open question. While the practical usefulness of new combinations of “genomeplasmon” genes leading to the creation of new germ plasm derived from protoplast fusion of wild, primitive, and cultivated forms of tomato and potato still remains to be seen, combinations of different allelic forms of the nuclear and chloroplast genes for RBPCase in a single cell and the subsequent formation of hybrid enzyme molecule in vitro will help solve some intriguing problems relating to the precise mode of transmission of organelle genes in higher plants. This would require the detailed analyses of individual hybrid molecules of organelle origin and their in vitro reconstitution in order to leam about the nature of interaction between organelle and nuclear genomes.

160

H. K . SRIVASTAVA

F. MITOCHONDRIAL ENZYMES

It has been established for each of at least five mitochondrial proteins that their subunits are made in different cellular compartments (Table V); for example, three of the seven subunits of cytochrome c oxidase are synthesized by mitochondrial ribosomes, while the other four are made in the cytoplasm (Schatz and Mason, 1974). The cytoplasmic subunits have been reported to exert a specific and positive effect on the synthesis of the other three subunits by isolated mitochondria in the absence of cytoplasmic protein synthesis (Poyton and Kavanagh, 1976). Most of the information on the syntheses of mitochondrial enzymes under complementary effects of both nuclear and mitochondrial genes have so far come from studies on microorganisms. The principles exemplified by these findings may be of general application to the synthesis of the many proteins found in mitochondria. The simplest interpretation, in concordance with present data (Table V; Fig. 2), would be that mitochondria and chloroplasts require the integrated cooperation of both organelle and nuclear genomes for their structural and catalytic activities. A close nuclear-plastid genetic complementation through the transcription-translation system of chloroplast, nucleus, and cytoplasm in the formation of the pigment-protein complexes responsible for primary photochemical processes of light energy conservation during photosynthesis has been demonstrated on nuclear-plastome mutants (Smillie et al., 1977) and by experiments with specific inhibitors (Nasyrov, 1978). Results supporting the existence of intergenomic complementation in maize protoplast heterokaryons using wildtype and mutant strains have also been presented (Giles, 1974). All these cases of intergenomic interaction provide insights into two conceptions: (1) there exists a positive and effective cooperation and coordination between organelle and nuclear genomes for balanced metabolic function and (2) some intergenomic communication operates to regulate growth, development, and maturation in higher plants. The present evidence favors the operation of intergenomic interactions for biosynthetic functions of the cell at messenger RNA level. Messenger RNA has been found to have a higher level in heterotic hybrids (McDaniel, 1973). DifTable V Biosynthesis of Major Mitochondria1 Enzyme Complexes

Enzymes

Total subunits

Cytochrorne c oxidase Adenosine triphosphatase Cytochrorne bc, Ribosomal RNA (large subunits) Ribosomal RNA (small subunits)

7 9 7 30 22

Cytoribosomes

Mitoribosomes

References

4

3 4

Schatz and Mason (1974) Schatz and Mason (1 974) Tzagoloff er al. (1976) Saccone and Kroon (1976) Saccone and Kroon (1976)

5 6 30 21

I 0 1

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

161

ferences in mitochondrial activity of reciprocal hybrids have been observed (Sarkissian and Srivastava, 1967; McDaniel and Grimwood, 1971), thus indicating possibilities for an unequal parental contribution of mitochondria to the hybrid. New types of mitochondria do apparently arise by some process of hybridization (Sarkissian and McDaniel, 1967). The actual transfer of nuclear messenger RNA into the mitochondria where its translation occurred has been demonstrated (Nasyrov, 1978). As for the mechanism of polypeptide partner transport into the membrane system of organelles, two hypotheses have thus far been proposed. The “signal hypothesis” provides an attractive mechanism for explaining how polypeptides that are secreted by cells are synthesized by membrane-bound ribosomes and transported into the lumen of the endoplasmic reticulum (Campbell and Blobel, 1976). A key feature of this mechanism is that the transport through the membrane depends upon concomitant protein synthesis by membrane-bound ribosomes, since the polypeptide chain moves through the tunnel in an extended form as it is being lengthened. The second hypothesis of “envelope carrier” presumes that subunit polypeptides enter the organelle after synthesis by a mechanism involving combination with specific sites in the organelle envelope. This hypothesis requires the presence of a class of proteins in the organelle envelope that recognize sites common to all those proteins that are made on the cytoplasmic ribosomes, but that are destined to function in the chloroplast (Highfield and Ellis, 1978). Figure 3 summarizes the current model for the synthesis of RBPCase in eukaryotic plant cells. The principles involved in this model are applicable to the synthesis of the many proteins found in both mitochondria and chloroplasts. It can be seen from the figure that cytoplasmically synthesized polypeptides enter the organelle by an “envelope carrier” type of mechanism, and control the synthesis of the polypeptides with which they ultimately associate. The precise mechanism of the linking of the two subunits separately encoded by nuclear and organelle genomes in order to form the fully functional enzyme molecule is not known. There are some results suggesting that the products of ribosomes under the genetic control of ct-DNA interact with the products of ribosomes under the control of nuclear genes to synthesize the enzyme RBPCase (Wildman et al., 1975). G . POLYMORPHISM OF MITOCHONDRIA A N D CHLOROPLAST

The heterotic hybrids in many plant species exhibit polymorphic mitochondria (Srivastava, 1972). Electron microscopic studies have further potentiated this claim (Hraska, 1978). The significance of mitochondrial polymorphism or heterogeneity in the manifestation of hybrid vigor and intracellular homeostasis has also been elucidated (Srivastava, 1975). Even in the case of chloroplasts “mixed” heteroplastid cells with normal and mutated chloroplasts are found at

162

H. K . SRlVASTAVA

C hliroplist mslaps

1

k-

00s R i b o m s

i

t

S d l

- subunit

prscurttr prscurttt

rtw

FIG.3. A postulated model for RBPCase synthesis in higher plants. nDNA and ct-DNA stand for nuclear and chloroplast DNA, respectively. The dashed line indicates possible control sites at which small subunit affects synthesis of large subunit. Adapted with modification from Highfield and Ellis (1978).

the boundaries of green and white parts of variegated leaves (Knoth and Hagemann, 1977). The photosynthetically efficient C4 plants like maize, sorghum, millet, and sugarcane have been shown to possess two types of chloroplasts (chloroplast dimorphism) that differ distinctly in terms of their ultrastructure and photochemical functions. The occurrence of intramitochondrial complementation in the heteroplasmon of Neurospora crassa giving rise to vigorous growth provides additional support that the heterozygotes can be better protected or buffered so that environmental disturbances in development do not eliminate these hybrids (Bertrand and Pittenger, 1969). The correlation between organellular complementation-heterosis and hybrid vigor would then be a consequence of interaction between polymorphic organelles in the hybrid. Genetic polymorphism in any form has been considered adaptive by providing an increased diversity of genotypes, thus leading to reduced intraspecific competition and enhanced buffering against environmental conditions. How can organelle polymorphism be explained from a genetic point of view? Is this a result

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

163

of recombination of heterogeneous mitochondria originated from parental types, or are they formed under the control of the nuclear genome? It is difficult to be absolutely sure about one or the other possibility in the absence of any conclusive evidence. One possible hypothesis to interpret the occurrence of polymorphic hybrid mitochondria as a result of the heterozygosity of chromosomal genes has been put forward (Wagner, 1972; Srivastava, 1981). This assumption, at least in the case of maize, which perhaps belongs to the group with the maternal type of mitochondrial inheritance, is more likely, but a hypothesis regarding polymorphism to be the product of organellular recombination caused by biparental transmission of mitochondria from the parents requires rigorous experimentation. Recent study on ct-DNA distribution in parasexual hybrids as shown by polypeptide composition of RBPCase indicates that chloroplasts from both parents stay together and are distributed biparentally to daughter cells, giving a preponderance of one type or the other (Chen et al., 1977; Melchers et al., 1978). The current genetic manipulation of plant genome by the protoplast and organelle transfer techniques (Scrowcroft, 1977) would appear much more important than nuclear transfer, since under normal sexual crossing in higher plants only the female parent presumably contributes organelle genes. Thus it seems there is no potential for organelle genome recombination in sexual crosses, and chloroplasts or mitochondrial transfer by protoplast fusion could open new horizons in plant breeding. Recent results from yeast suggest the possibility of restoring respiratory competent strains due to recombination between different mitochondrial genomes (Oakley and Clark-Walker, 1978). The findings by Coyne (1976) and Singh et al. (1976) that the xanthine dehydrogenase locus is segregating for 23 alleles out of 60 genomes tested (Drosophila persimilis) and for 37 alleles out of 146 genomes tested (Drosophila pseudoobscura)reveal that genic variation at individual loci may be much greater than had been previously imagined. There are three basic paths through which allelic polymorphism may arise and be maintained in a population. These paths produce genetic patterns that are called transient polymorphism, balanced polymorphism, and random fixation of neutral mutations. Transient polymorphism represents a temporary situation, while balanced polymorphism represents a relatively permanent kind of equilibrium in which two alleles a,/+ are present in the population at some steady-state frequencies (frequency-dependent fitness). The relative effectiveness of balanced allelic polymorphism resulting due to the existence of heterogeneity in the multiple genomes (nuclear and organelle genomes) of the eukaryotic cell is also worth considering. Since the existence of balanced polymorphism entails the production of a large number of heterozygotes, the consideration that polymorphic mitochondria or chloroplasts maintain a high frequency of organelle genome heterogeneity extends indefinitely the time during which selection may operate for spreading of advantageous allele combination for heterosis. Direct evidence towards elucidating the principles of

164

H. K. SRIVASTAVA

genomic interaction, intergenomic cooperation, and biparental transmission of mitochondria and chloroplasts in higher plants is required to substantiate the proposed positive relationship between intergenomic interaction and manifestation of heterosis. This is a very exciting research field in genetics and undoubtedly will attract much more attention in the future.

V. MOLECULAR-GENETICASPECTS OF HETEROSIS Heterosis was originally defined as “hybrid vigor.” Shull (191 I ) , who coined the term heterosis, referred it to the superiority of heterozygous genotypes with respect to growth, survival, and fertility in comparison with the corresponding homozygotes. Heterosis is the phenotypic result of gene interaction in heterozygotes and is thus confined, at least in its maximal amount, to increased heterozygosity . A more meaningful and limited definition of heterosis, which involved a measure of the relative fitnesses of genotypes, was proposed by Dobzhansky (1952). Fitness here simply refers to the relative reproductive successes of competing genotypes or, more specifically, the relative contribution of offspring from parental genotypes. The question of whether heterosis is a manifestation of the superiority of the heterozygotes or whether it arises from other causes, has long been debated. The heterozygote may also be superior, or inferior, in Darwinian fitness to both homozygotes. Inferior heterozygotes betoken some degree of hybrid inviability or hybrid sterility. The term “negative heterosis” has also been proposed by Stem (1948) to define examples of “hybrid disvigor,” but it is rather misleading and hence reference to the terms positive and negative heterosis should be avoided. A. THEOFUES OF HETEROSIS

Part of the difficulty in understanding the genetic basis of heterosis has arisen from the variety and vagueness of different theories used to explain heterosis. Unequivocal evidence usually requires a molecular description of the mechanism by which heterosis occurs. The crux of the problem today is in estimating the relative proportion of single-locus polymorphism that is maintained in natural populations by heterosis. The task here is simply to identify the level(s) of biological organization at which heterozygote superiority is expressed and then describe the relevant molecular feature of the genetic difference. Early theories of heterosis fall into two classes. The “dominance” theory (Shull, 1952) argued that hybrid vigor simply resulted from the masking of deleterious recessive mutations fixed in one parental line by nondeleterious alleles found in the other parental line. Since each individual line could be ex-

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

165

pected to be homozygous for a different array of deleterious alleles, the hybrid would contain the fewest number of homozygous recessives. The “overdominance” theory (Lerner, 1958) claimed that heterozygosity per se is adaptive, for, as Lerner argued, heterozygosity reduces phenotypic variance and thus must provide “buffering” against either genetic or environmental modifications of phenotypic characters. This buffering, which in developmental terms Waddington calls “canalization, ” provides the optimum fitness in varying genetic and physical environments. Overdominance requires a heterozygote at a single locus to be superior to both homozygotes; overdominances at different loci are then added up to form heterosis (Srivastava and Balyan, 1977; Srivastava, 1980). Fincham (1972) later subdivided the overdominance model by suggesting that heterosis may occur either ( I ) because the heterozygote properties are unique (outside the range of homozygote properties) and unconditionally superior or (2) because the heterozygote contains two and sometimes three different kinds of gene product, which in total provide an intermediate phenotype, but bestow a biochemical diversity that becomes adaptive in a regularly varying environment. One basic distinction between alternatives (1) and (2) is that, under constant environmental conditions, polymorphism would be lost in the second case, but not the first. Some alternative theories partly in keeping with genomic interaction to explain the genetic basis of heterosis have also been proposed. These include (1) epistasis or nonallelic interaction, where one locus influences the expression of another towards heterozygote-adaptive superiority (Gowen, 1952); (2) physiological stimulus and initial capital, better “buffering” of heterozygote against environmental disturbances (Ashby, 1937; Steward and Krikorian, 1971); and (3) complementation of balanced metabolism, metabolic advantage of the gene products produced by two different kinds of alleles rather than by only one kindheterozygote advantage may produce a unique “hybrid substance” completely different from either of the homozygous products (Schwartz, 1964; Hageman et al., 1967; Srivastava, 1972). All these theories unfortunately are circular, stating in essence that a given heterozygote organism exhibits heterosis because it is adaptively superior over one or both types of parental homozygotes. B . MECHANISMS OF HETEROSIS

In order to overcome the difficulties as to how, in mechanistic terms, heterosis can operate, one may simply examine the criteria that must be fulfilled to define heterosis. First, from a large panmictic population one must be able to demonstrate a reproducible, statistically significant heterozygote superiority at a single locus either in terms of egg-to-adult survival, fecundity, fertility, or some combination of these. Second, the superiority must be shown to involve a selective

166

H. K . SRIVASTAVA

advantage associated with heterozygosity at the locus under examination, or more ideally, the nature of the selective superiority should be defined in molecular terms. The true nature of heterosis has not yet been understood, though there is no doubt that its causes are enclosed in the sphere of genetic processes. It was clear from the preceding discussions that heterosis is not a phenomenon regulated by the nuclear genome only; the plasmon is also involved, particularly through an interaction between nuclear genes and organelle genomes. The genomic and plasmonic regulatory systems of heterosis, along with their genetic causes, are enumerated in Table VI. Apart from genomic and cytoplasmic regulatory systems, manifestations of heterosis at different functional and organizational levels are also illustrated in Table VI. All the genetic causes of heterosis as listed in the table are important. It is most likely that they are all involved in heterosis, but their appearance and the magnitude of manifestation may be different. Heterosis includes purely recombinative elements (Williams, 1959). Heterosis with this genetic background is generally referred to as nonallelic heterosis and includes both genomic and intergenomic complementation. Epistatic heterosis is frequently used as a synonym but should have a more restricted meaning, if the term epistasis is taken in its modem, revised connotation of nonallelic phenotypic suppression (Mackey, 1976). Epistatic heterosis must be considered a part only of nonallelic heterosis, which also includes transgressive (additive, cumulative) and recombinative (complementary) heterosis. In general, these three types of nonallelic heterosis are distinctly classified apart from allelic heterosis, which has its regulatory system rather within than between individual loci. Population geneticists, however, ought to be more disturbed than the practical breeder that such a strict grouping between nonallelic and allelic interaction is impossible in experimental work because of linkage. Fortunately, however, a distinction set by the degree of linkage is just as rational, since the interactions both at the nonallelic and allelic levels are virtually of the same kind. In both cases they can be described as sheltering, inhibiting, stimulating, complementing, or dosageregulating. The sheltering and inhibiting effects of allelic interaction lie behind the dominance or dominance of linked genes hypothesis of heterosis (Bruce, 1910; Jones, 1917), where the negative effects of deleterious recessive genes are masked by the presence of their respective dominant alleles. Alleles cooperating in a stimulatory , complementary, or dosage-adjusting manner fit into the overdominance hypothesis (Shull, 1908, 191 1; East, 1908; Hull, 1945; Berger, 1976), according to which none of the alternative homozygotes is better than the heterozygote ( A A < Aa > ua) at a single gene locus. The mechanisms to explain heterosis based on either nonallelic, dominance or overdominance interaction have had their different advocates. Cytoplasmic contributions were also recognized (Wagner, 1969) and evidence presented to show that organelles are deeply involved in heterosis (Srivastava, 1972). More re-

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

167

cently, a new hypothesis based on intergenomic interaction to interpret the operational mechanism of heterosis has been advanced (Srivastava, 1981). In this respect heterosis must be more effective in those genetic systems which code key positions of metabolism and morphogenesis. In particular such systems may include efficient production of enzymes, hormones, chloroplast and mitochondrial apparatus in the hybrid individual. Enough experimental evidence is available at present to show that all three genetic sources (genomic, plasmonic, and intergenomic interaction) and not just one of them are at work during the manifestation of heterosis. How could the operation of intergenomic complementation in molecular terms be explained‘?Also, it is unknown to what extent observations on mitochondria1 and chloroplast heterosis reflect intergenomic interaction. The existence of heterosis and complementation with regard to key particulate and soluble enzymes appears to be of common occurrence in higher plants. There is no doubt that the heterotic hybrids are endowed with a more balanced metabolism than the inbred parents. The metabolic advantage of the gene products at the cellular or organellular level could be viewed as produced by nonallelic complementation rather than by intergenic complementation pertaining to only one gene locus. Demonstration of multimeric “hybrid enzymes” or “isozyme spectrum” due to nonallelic or intergenomic interaction (Hart, 1978) renders evidence for a buffering mechanism at the subcellular level in the hybrids. Some current research on genetically diverse sugarbeet lines indicates that heterosis is a result of increased cell division rate rather than an increase in cell size (D. L. Doney, personal communication). Genetic studies suggest that genetic differences in cell size are due to additive-type genes, whereas genetic differences in cell division rate are due to nonadditive genes (epistatic interaction). A possible operational mechanism of heterosis would involve complementation of proteins or polypeptide subunits encoded by “multigenomes” instead of unique nuclear genes. The resulting products due to intergenomic interactions in organelles of hybrid organisms would then be expected to ensure for enhanced structural and catalytic functions of these organelles. On the contrary, the inbred parents, by virtue of maintaining higher degree of genomic uniformity and therefore less opportunity for effective intergenomic interaction, might be deprived of such physicobiochemical advantages. It is postulated that heterosis results to a greater extent from intergenomic complementations rather than being confined solely to nuclear genome. The efficiency of mitochondria and chloroplasts caused by both genomic and intergenomic interactions in hybrid organisms must therefore be considered an important component of heterosis. Although the functional efficiency of organelles and seed yield of the plant may not be linked under common gene control, it is likely that they are governed under the control of a common general regulation of metabolism by means of intergenomic interactions. A model situation for such intergenomic complementations could be visualized in

Table VI Regulatory Systems of Heterosis

sources A. Gnomic heterosis e

Genetic causes A. Allelic heterosis i. Dominant heterosis

B. Plasmonic heterosis i. Plastome genes

ii. Chondriome genes

ii. Overdominant heterosis (single-locus heterosis or interallelic complementation)

Manifestation A. Cell and organisms levels i. Cell division rate ii. Growth iii. Differentiation iv. Yield B. Subcellular level i. Mitochondria1 heterosis

B . Nonallelic heterosis

C. Intergenomic interaction heterosis

i. Recombinative heterosis

ii. Epistatic heterosis (nonallelic complementation)

ii. Chloroplast heterosis iii. Other organelles efficiency C. Metabolic and functional levels i . Hybrid enzymes

i. Nucleus-mitochondria

iii. Transgressive heterosis (additive or cumulative genes)

ii. Isozyme spectrum

ii. Nucleus-chloroplast

iii. Mitochondria-chloroplast

C. Intergenomic complementation heterosis

iii. Formation of hybrid-specific molecules with superior function and adaptation iv. Enzyme polymorphism v. Greater production of hormones and other biologically active substances

i. Plasmon-genome ii. Plastome-genome

iv. Mitochondria-chloroplast-nucleus

iii. Chondriome-genome iv. Plastome-chondriome v. Other interactions

D. Molecular-genetic level i . Superior newly formed rnessenger RNA ii. Superior repeated ribosomal RNA loci (rDNA) iii. Superior structural state of DNNchromosome iv. Increased ratio of DNA replication, transcription, and translation v. Repeated DNA sequences in genomes active sites

170

H . K . SRIVASTAVA

interactions between (1) nucleus and chloroplast, (2) nucleus and mitochondria, (3) chloroplast and mitochondria, (4) nucleus, chloroplast, and mitochondria together, ( 5 ) different mitochondria1 types in a cell, and (6) different chloroplast types in a cell. C. MULTIMERIC H Y B R I DMOLECULES

It has been argued that enzymes with multiple substrates (Cillespie, 1976) and enzymes regulating the flux through a pathway (Johnson, 1974) are more variable than enzymes with single substrates or nonregulatory enzymes. The class of nonregulatory enzymes contains more multimers than monomers, so it can be argued that the difference in amounts of variation is ultimately related to function rather than to presence or absence of hybrid bands. It is generally considered that the biochemical properties of heterozygotes are not always intermediate, either for multimeric enzymes or, surprisingly, for monomeric enzymes. The suggestion that heterozygote enzyme ensembles do have unique and superior properties, outside the range of homozygotes (Fincham, 1966), does require a molecular explanation. For enzymes that are multimeric, the model can be quite simple. Heterozygotes contain two distinct types of gene product that interact at random to form the quaternary structure of functional enzymes. If the properties of heteromultimers were unique and in some way superior, this interaction could be expressed as an increased fitness in the organism. Any mutation producing such an effect would apparently be selective. The first real indication that different allelic subunits could interact, leading to interallelic heterosis, comes from the experiments on the dimeric enzyme alkaline phosphatase from Escherichia coli (Schlessinger and Levinthal, 1963). Essentially, the authors found that certain pairs of mutant alleles, which individually produce inactive enzyme, when combined in merodiploids produced nearly wild-type phenotypic properties. When subunits of both inactive enzymes were prepared and allowed to reassociate at random, in v i m , significant, although not wild-type, levels of enzyme activity were restored by virtue of enzymatically active heterodimers. In an informative study of alcohol dehydrogenase variants in maize, Schwartz and Laughner (1969) presented data showing that, in individuals heterozygous for the alleles Cm, which alone produces enzyme with low activity but high stability at pH 10, and F, an allele that alone produces enzyme with high activity but low stability, the heterodimer molecule (FCm) is uniquely stable and active. Studies done on esterase-5 of D . pseudoohscura (Berger, 1974), utilizing the technique of in vitro subunit hybridization (Hubby and Narise, 1967), revealed that for certain allele combinations the formation of heterodimeric enzyme was accompanied by increased catalytic activity or reduced susceptibility to heat inactivation. The heterosis for

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

171

heat stability observed by Singh et al. (1974) for octanol dehydrogenase and tetrazolium oxidase can also be explained only by the presence of uniquely stable heteromultimers. Other cases of enzyme heterosis (Koehn et al., 1971; Miller et al., 1975) also appear to involve uniquely functioning heteromultimers of naturally occurring enzyme variants. Although detailed explanations of the mechanism(s) behind these unique interactions lie in the realm of physical biochemistry, the occurrence of biochemical heterosis based on multimeric forms of hybrid enzyme appears to be a fact of life. One of the most important processes that has affected the evolution of higher plants is allopolyploidy. An alloploid is in fact a “permanent heterozygote” in which positive and negative heterotic interactions between homeoalleles are effectively fixed. Fincham (1969) and Barber (1970) have proposed that new and advantageous oligomeric enzymes may be produced in polyploids by association of polypeptides encoded by differing members of duplicate structural gene sets. Intergenomic variation has been demonstrated for esterases (May er al., 1973; Cubada et al., 1975), malate dehydrogenase (Bergman, 1972), alcohol dehydrogenase (Torres and Hart, 1976), and amino peptidases (Hart and Langston, 1977). Most of the enzymes in maize hybrids exhibit an isozyme spectrum: seven allelic esterase genes, four forms of leucine-aminopeptidase, three isozymes of alcohol dehydrogenase, three forms of catalase, and some isozymes of peroxidase and malic dehydrogenase (Holson, 1967). The results presented so far indicate that most enzymes possess a quaternary structure and are capable of producing hybrid molecules with a higher enzyme activity in the hybrids. D. DNA

OF

HYBRIDS

The presence of quantitative differences of enzymes together with their distinct allosteric properties in the case of heterotic hybrid organisms is widely accepted. The genetic sources of heterosis do involve both genome and plasmone structures. Heterosis can therefore be projected at the level of DNA. In an attempt to evaluate the hybrids and their parents in terms of “active” genome sites and their functional competence, several workers have employed the competitive hybridization technique between newly formed messenger RNA and DNA immobilized by filters (Konarov et al., 1971). The results obtained by these workers on young and mature maize plants indicate that the nucleotide composition of newly formed messenger RNA of hybrids differs from that of their parents to a lesser extent than the parents differ between themselves. The detailed analyses of the nucleotide composition of the messenger RNA molecule from several heterotic hybrids and their respective inbred parents further indicated significant differences in the total content of messenger RNA population and active genome sites. The results were interpreted by the authors to signify that interaction

172

H. K. SFUVASTAVA

between a genome’s active sites creates favorable conditions for cumulative, epistatic, and complementary genes pertaining to major functional loci in heterozygotes. It has also been demonstrated that DNA of hybrids possesses high replicative activity, repetitions of ribosomal RNA loci (rDNA), and relatively repeated nucleotide sequences in the genome’s active sites (Ali-Zade and Aliev, 1973). The basic repeating unit of an rDNA gene is composed of a transcribed segment followed by a nontranscribed or spacer segment (Tartof, 1975). The transcribed segment codes for an rRNA precursor, while the extranucleotides of the spacer are being degraded in the maturation process. There is evidence to suggest that in yeast the rDNA is organized into sets of 10-32 closely spaced rDNA genes that are separated from neighboring sets by very long stretches of DNA (Hartwell, 1970). Since the size of one rDNA gene is between 13,000 and 16,000 base pairs (Tartof, 1975), there must be at least one initiation site for replication with each rRNA gene. It would be of interest to determine the location and sequence of the replication origin for rDNA gene in hybrid and nonhybrid organisms. This might be one class of control sites contained in the spacer region of rDNA to regulate its transcriptional function. Although similar information on the organelle genomes from hybrid and nonhybrid organisms would be highly desirable, there are indications that ct-DNA and mt-DNA may be composed of multiple copies of nucleotide sequences. The relevant findings on redundancy of organelle genome have already been discussed in Section II,B on organelle genomes. Gene duplication by way of repeated DNA sequences produces redundant copy of a locus. Callan (1967) has proposed a very ingenious mechanism by which the organism might escape the hazard-containing multiple (repeated) copies of the genome. He postulates that the one at the end is the master, while all others are slaves: the master-slave theory. When chromosomes duplicate before each cell division, not the slaves, but only the master serves as the template for DNA replication. In any case repeated sequence has been considered a major force of evolution (Ohno, 1970), and the heterotic hybrids seem to carry them for evolutionary advantages. The active state of genome structure can be detected by experiments measuring two biochemical parameters: the nonhistone contents and the nature of histone modifications by phosphojlation or acetylation. These parameters directly reflect on structural transition and availability of active sites of DNA. In terms of the rate and intensity of DNA synthesis, the heterotic maize hybrids excel parental types, while in RNA synthesis they stand at the level of the better parent or sometimes excel the parental values (Konarev, 1976). Besides, the hybrids and their parents differ in the degree and nature of DNA activation. This has actually been observed in experiments designed to measure the rate of DNA and RNA syntheses, in the correlation between the amount of diffused and compact chromatin (euchromatin and heterochromatin) fractions, and in the contents of functionally active sites of nuclear DNA (Konarev, 1973; Konarev and Tuterev,

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

173

1971). It is therefore likely to trace heterotic manifestation at almost all levels of the hybrid plant orgariization. On the DNA level heterosis is manifested by increased DNA replication, transcription, and translation, in the formation of surplus genetic information, in the form of repeated rDNA sequences of favorable gene combinations, and in increased efficiency of the enzymes and many other regulatory elements within the cell. Plant geneticists interested in crop improvement may some day be able to produce heterotic plants by exogenous “active” DNA transformation (Lurquin, 1977; Flavell ef al., 1978) without the necessity of hybridization. E. CURRENT MODELSOF COMPLEMENTATION

Two models of heterosis based on intergenic and nonallelic complementation, respectively, have generally been recognized. A third model based on intergenomic complementation (genome-plasrnon interaction) to explain heterosis has been put forward (Srivastava, 1981). It is, however, considered that the three types of models are not mutually exclusive and that all three types of phenomena are probably involved in the observed association of heterosis and heterozygosity . The physico-biochemical basis of heterosis based on “polypeptide chain complementation” as proposed by Fincham (1966) is widely accepted over an alternative interpretation by Schwartz and Laughner (1969), which supposes the formation of a specific hybrid enzyme or allozyme (in addition to parental types) bearing new catalytic properties to be responsible for heterozygote superiority. The intergenomic interaction, as discussed earlier, is regarded as an essential component of the heterosis phenomenon. Many multimeric enzymes of the mitochondria1 and chloroplast membrane systems sharing genomes of organelles as well as of the nucleus could obtain catalytic superiority by intergenomic complementation. It is suggested, however, that the active principle of intergenomic interactions (cooperative interactions between nuclear and organelle genomes) for vigorous organelle functions lies either with a messenger RNA molecule or the subunit polypeptide encoded by the nuclear genes. Most organelle proteins are multimeric structures, and their functional activity is based on the proper configuration of polypeptide chains; the advantage of a heterozygote in a heteroplasrnon possessing the active principle of intergenomic interaction would therefore be expected to lead to positive polypeptide chain cornplementation and heterosis. The argument presented here in favor of an association between intergenomic interaction and heterosis is selective and speculative, but the implication is worthy of consideration and additional research. Since the first examples of interallelic or intergenic complementation phenomena were identified (Fincham, 1966), several molecular models have been proposed. Interallelic complementation has been demonstrated only in mul-

174

H . K . SRIVASTAVA

timeric proteins, and most models embrace the essential interaction of defective subunits. Kapuler and Berstein (1963) and Crick and Orgel (1964) suggested homologous correction, i.e., for some mutants the nonfunctionality of an enzyme subunit is correctable through an association with an unaltered adjacent wild-type (or complementary) region of another promoter. Heterologous correction has also been hypothesized (McGavin, 1968). In this scheme a mutant with an altered active site may complement a mutant with an altered allosteric site, or two different allosterically defective mutants may complement. Recently, an alternative interallelic complementation model has been proposed (Schwartz, 1975). Rather than having a correctly folded heteropolymer formed by two or more misfolded monomers, Schwartz postulates that protomers may have normal configuration, and nonfunctionality of the homopolymer is due to misfolding during maturation. A novel interallelic complementation of a single allele and a double mutant carrying the original isoallelic single site at the his1 locus of yeast has also been observed (Lax and Fogel, 1978). This study reports that there are 17 his1 -7-c alleles having double-site mutants interallelically complementary to his1 -7. Minimally, they delineate eight distinguishable classes of mutant effects. Each allele may be assigned to one or the other of two mutually exclusive complementation groups. The authors call special attention to the problem of how genetic information affects subunit-subunit interactions. The important considerations for his1 -7/hisl-7-c complementation response from their genetic data are as follows: (1) A single site mutant and various double-site mutants sharing a common mutation exhibit interallelic complementation. (2) The second sites are missense mutations that specify nonfunctional enzymes in an otherwise standard HIS1 gene, though such mutants may complement the original mutant from which they were derived. (3) There are several different and recombinatorially separable second-site alterations, although their distribution is probably limited to specific gene segments. Therefore, (4) there are several amino acid substitutions at various but limited positions within the polypeptide subunit that do not generate an internal correction of the defect due to another amino acid substitution, but that will at least partially correct or stabilize this amino acid substitution in an adjacent subunit to yield a functional phosphoribosyl transferase molecule.

VI. IMPROVEMENT OF CROP YIELD A. ROLE OF PLASMON GENES

Basic physiological and genetic research on crop productivity reveals that the yield potential is rather high and not yet fully exploited. A rich gene pool with

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

175

regard to both genomic and plasmonic factors has not been fully utilized for the purposeful and efficient enhancement of yield-limiting processes that at present limit crop productivity. Identified mitochondria1 and chloroplast genes specifically code subunit polypeptide components of enzyme systems that play key roles in respiratory and photosynthetic electron transport and phosphorylation. In the most thoroughly studied species, like Chlamydomonas reinhardii, Saccharomyces cerevisiae, the geranium Pelargonium, and Oenothera, organelle genes are known to mutate, segregate, and recombine just like nuclear genes. In these cases there is good reason to believe that most or all zygotes receive intact organelles from both parents. Should we not expect, therefore, that if technical difficulties were surmounted and a higher plant system were to afford the opportunity, then selection for recombinant cytoplasmic genomes and the development of cytoplasmic hybrids (cybrids) would lead to corresponding advantages in yield and other desirable agronomic attributes? Treat (1978) has recently shown that, although most yeast diploids are homoplasmic after about 10-20 generations of growth from zygote, some remain heteroplasmic for much longer periods. Thus crop yield improvement, taking advantage of organelle genes, seems an achievable goal in the near future. It has been customary to think of cytoplasmic inheritance as uniparental inheritance, that is, as involving transmission of organelle genes through the female but not the pollen parent. If, in fact, there were no exceptions to the rule that plasmon genes are transmitted through the female parent, it would mean that there is no sexual mechanism in higher plants to bring together in a common cytoplasm the heteroplasmon of different individuals. Many instances of biparental transmission, as illustrated earlier, raise the hope that ways may be found to induce it in economically important crop species. On the face of it, evolutionary advantages should result from biparental transmission in terms of both interorganelle or intraorganelle recombination. If mt-DNA or ct-DNA from both gametes is included in the zygote formed from them, their genetic heterogeneity could result in heterotic vigor by complementation in zygote organelle functions, especially if there is recombination between parental organelle genomes. By contrast, the limitation of transmission to the egg results in restriction of plasmon heterogeneity to the factors that happen to be in the egg. The only genetic changes that can take place are from random mutations occurring one at a time. The evolutionary coadaptation between the cytoplasmic genes of a population of plants and their nuclear genes is a specific product of the natural selection pressures to which both are subjected. For species that show significant proportions of biparental zygotes, it seems that the most appropriate and fruitful course is to consider organelle genetics as a problem in intracellular population genetics. Intraorganism competition between cellular organelles and between organelles and their host cell nucleus must therefore be considered relatively stable phenomena upon which natural selection would operate for

176

H. K . SRIVASTAVA

improved reproduction. Persistent heterozygotes have been found among antibiotic-sensitive revertants from antibiotic-dependent plastome mutants in Chlamydomonas (Sager, 1977). Mitochondria1 gene recombination has very recently been demonstrated in two fungi, Aspergillus nidulans (Mason and Turner, 1975; Lazarus and Turner, 1977) and Podospora anserina (Belcour and Begel, 1977). Recombination of chondriome and plastome genes cannot be studied in many systems because suitable markers are lacking or because paternal organelle DNA is too diluted in the zygote to be experimentally detected. Unipaternal inheritance of organelle genes is phylogenetically ubiquitous and ancient (Birky, 1978). It was earlier thought that there would be a single mechanism of organelle genome transmission, equally ubiquitous and primitive, but this is not the case. The mechanism(s) used by different eukaryotic organisms to bring about preponderance of only one type of organelles (or their genomes) during maternal or uniparental inheritance may well involve any one of the following possibilities: (1) one type of organelle could cause the destruction of genetically different organelles that are in the minority due to interorganelle competition; (2) one type of organelle could have strong input bias and favor its own inclusion in the gametes; (3) one type could reproduce more rapidly after inhibiting the reproduction of other types within a variety of cells very early in development so that its chances of inclusion within gametes were enhanced; and (4) one type possesses more favorable replication sites in its DNA than the other, and thus could multiply faster for its predominance in the progeny. However, if there are very low levels of paternal gene transmission and recombination, these must be measured, for they may become very important over evolutionary time scales. When crosses occur between two distinct species that have different plasmon genes, there may be abnormalities in the hybrids that result from a lack of genotypic incompatibilities between the nuclear genes of one parent and the plasmon genes of the other, i.e., lack of a proper intergenomic interaction. The evolutionary significance of biparental inheritance would seem to be primarily in the production of heterogeneity through recombination of organelle genes. Recent evidence has shown that some subunits of key organelle enzymes, including RBPCase and the F,-F,, of chloroplasts, cytochrome oxidase, cytochromes a and b, and the F,-ATPase of mitochondria are coded by organelle DNAs, and other subunits of the same enzyme are coded by nuclear DNAs. Like viruses, eukaryotic organelles can have only a small number of genes and must rely on the “host,” i.e., the nuclear genome, to code for most organelle components. Intergenomic interaction between heterogeneous nuclear and cytoplasmic genes as a result of biparental transmission for maintaining heterotic advantages of organelle enzymes would be expected to have greater evolutionary significance over the maintenance of genetic uniformity by blocking recombination and allelic polymorphism in the case of uniparental or maternal inheritance of organelle DNAs. Hybrid vigor thus results from an initial superiority of the heterozygous

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

177

zygote produced under the influence of genome-plasmone interaction. Some evidence in support of this hypothesis comes from the experiment of Demarley (1976), where it was demonstrated that a higher rate of cell division during the first three mitosis cycles is sufficient to induce the superiority of a heterozygous individual throughout its life. 9. MITOCHONDRIAL EFFICIENCY A N D PREDICTION OF YIELD

The advance prediction of the degree of yield heterosis of F, combinations from parental observations at young seedling stages could be of immense advantage in crop breeding experiments. Heterosis is generally manifested in quantitative characters and is expressed in vigor of vegetative organs, in increased grain yield, and in resistance to diseases, climatic factor fluctuations, etc. Substantial evidence is available to suggest that hybrids are endowed with greater homeostasis by offering alternative genetic pathways (Lewis, 1954; Mackey, 1970). This reaction will give them an advantage under conditions of stress such as dense plant population, adverse weather, and pathogen attacks. In contrast, other expressions of heterosis might not be fully developed unless the plant is given optimal conditions to exhibit its genetic capacity. The principal components of yield in wheat are heads per unit area, spikelets per head, and seeds per spikelet, and in maize are cobs per plant, kernel rows per cob, kernels per row, and weight per grain. In some cases the main components may be further divided, retaining the multiplicative principle, into tillers per unit area times the percentage of fertile tillers (Thomas and Grafius, 1976). The prediction methods based on parental yield components can be more accurate and meaningful than those based on yield alone. Most of these methods, however, rely on a geometric approach involving the biometrical interrelationship of the components and crop yield. The fairly recent information on the importance of different physiological (photosynthesis, respiration, ferment activity, mitochondria1 activity, chloroplast activity, isozyme spectrum, composition of a number of endogenic regulators) and molecular-genetic (synthetic intensity of DNA, RNA, and genome active sites) processes in the manifestation of yield heterosis leads us to use such contributory parameters in the choice of potential parents, evaluation of heterosis effects, and its prognosis. Some of the physiological parameters commonly used today for the rapid screening of potential inbred lines include effective photosynthetic area, flag leaf size and duration, nitrate reductase activity, photorespiration rate, estimates of photosynthesis and carbohydrate translocation under field conditions, chlorophyll a content, cytochrome c oxidase activity, and adenosine triphosphatase activity (Srivastava, 1975; Lupton, 1976; Planchon, 1976). A new method for the prognosis of heterosis under field conditions using the immunochemical technique of antigen and serum analyses of inbred lines, their

178

H. K . SRIVASTAVA

hybrids, and back crosses in maize (Zea mays) has also been worked out (Dimitrov et al., 1974). In the inbred lines the authors found three protein fractions common to all of them, and a fourth antigen, contained only in those inbred lines that produced heterotic hybrids upon crossing. The gene coding for the synthesis of this individual antigen resides in the nucleus. The heterotic maize hybrids, however, were found to possess an additional fifth protein fraction with increased number of subfractions. Many wheat hybrids exhibit heterosis under field conditions, but the exploitation of this vigor in the form of economic yield has proved very difficult and requires the identification of parental combinations exhibiting greater combining ability (Lupton, 1976). Heterosis may be expressed at any stage in crop development, and the purpose of the breeder is to identify those stages at which heterosis may most conveniently be assessed, and when correlation with yielding capacity is greatest. On the other hand, it is known in wheat that the photosynthetic activity of the flag leaf blade during the blossoming-maturity period determines the yielding capacity to a large extent (Lupton, 1968; Planchon, 1968). In order to obtain better understanding of the combining ability of various wheat cultivars, Planchon (1976) analyzed the photosynthetic activity of the flag leaf blade in the parents and their F, hybrids. The activity in this study was measured using two parameters: chlorophyll a content in the flag leaf and net CO, exchange rate per unit surface area-a trait strictly associated with photorespiration. The results showed a positive correlation between chlorophyll a content and grain yield. It was suggested that parents possessing both a low photorespiration rate and high chlorophyll a content must be crossed to produce highly heterotic hybrids. Several studies which have considered the effects of heterosis prior to maturity have all reported that hybrids exhibiting a high level of grain yield heterosis also show vigor for growth and development at early seedling stages (Sagi et a / . , 1976; Gibson and Schertz, 1977). Organelle heterosis and complementation have been shown to be significantly correlated with grain yield heterosis in many crops. These studies have led us to believe that mitochondrial and chloroplast activity may be rate-limiting in plants; thus organelle activity and efficiency are basic to the control of growth and yield. Organelle complementation is a challenge to conventional biometrical methods and estimations, since apparently both plasmon and nuclear genes are involved. The use of mitochondrial complementation as a tool to preselect well-combining parents under laboratory conditions in crop plants is more amenable, and by this method plant breeders can save considerable time, work, and expense. A relatively good positive correlation r = 0.61) between grain yield per plot of the natural hybrids at low plant density and cytochrome c oxidase activity of the “mitochondrial” hybrids (1 : 1 parental “complementing” mixture) made in v i m has been observed (Sagi et al., 1976). This correlation is in concordance with that found by McDaniel (1972) between

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

179

mitochondrial heterosis and grain yield of several heterotic barley hybrids (r = 0.69). Lupton (1976) also reports a significant positive correlation (r = 0.845) between yield heterosis and mitochondrial complementation based on ADP : 0 ratio in wheat. Both mitochondrial ADP : 0 and respiratory control ratios have further been found to be highly correlated ( r = 0.94) with average grain yield (Flavell and Barratt, 1977). Because of their great significance in plant breeding, our original findings on mitochondrial heterosis and complementation, together with the currently verified results in many other laboratories, demonstrate the possibility of using estimates of ADP : 0 ratios or cytochrome c oxidase activity as an initial screen, independent of such complications as genetic heterogeneity (segregating seed material) in the seeds used, regional adaptation, or disease susceptibility, in assessing the potential of varieties as parents for hybrid breeding.

c.

PHOTOSYNTHETIC

EFFICIENCY AND

YIELD

Low photosynthetic efficiency of our modem crop varieties has become a “bottleneck” for yield improvement. Efficiency of solar energy utilization for many crops does not exceed 1%. Improved varieties of such crops as wheat, cotton, and others have been found to possess lower rates of photosynthesis than their ancestral types (Evans, 1975; Evans and Wardlaw, 1976). However, some improvement i n yield has been made by leaf area increase, by changes in the biomass of reproductive organs to that of vegetative ones, and by other morphological factors. Selection for desirable morphological properties and traits has been practiced, with greater emphasis given to increasing the size of sink and storage organs-the number and size of ears, bolls, tubers, etc. Enhancement of the plant sink and storage capacity could result in an increase of the agronomic yield as long as the photosynthetic potential permitted. However, the imbalance between the accumulation and partition of photosynthetic products in the newly selected varieties and hybrids leads to puniness and immaturity of seeds, falling of ovaries and bolls, reduction in sugar content, and many other disorders (Nasyrov, 1978). Besides, an excessive increase in the sink power can provoke, by the negative feedback mechanism, a decrease in the activity of the photosynthetic apparatus due to the withdrawal of nitrogen from leaves into other organs, thereby causing their premature senescence (McArthur et al., 1975). The improvement of the chloroplast efficiency on physiological and genetic bases is a brand new field in plant breeding research. Breeding of improved high yielding cultivars in cereals and other crops based on chloroplast and mitochondrial efficiency will probably continue to arouse greater interest in modem breeding technology in future, just as dwarf crop varieties presently dominates the scene in plant breeding. Two genetic approaches could be applied to maximize yield potential through chloroplast breeding: (1) germ plasm evaluation for genet-

180

H. K. SRIVASTAVA

ic diversity of chloroplast function and photosynthetic efficiency, and (2) induction of mutations to produce considerable shifts in the structural and biochemical organization of chloroplasts. Ecological and geographical populations of crop species have been found to exhibit a large genetic variability of chloroplast photochemical and enzymatic activities, which are significantly correlated with photosynthetic efficiency and grain yield potential (Wilson and Cooper, 1969; Treharne, 1972). Photosynthetically “plus” (efficient) mutants have been obtained in Chlorella (Mukhamadiev and Zalensky, 1972), pea (Highkin et al., 1969), and cotton (Usmanov et al., 1975). An economically promising mutant variety “duplex” in cotton has been released by gamma irradiation (Nasyrov, 1977). It has a compact type of branching, high frequency of double sympodial bolls, early maturing, and high agronomic yield. Its harvest index amounts to 50%, which points to efficient partitioning and utilization of photosynthetic products. The economic yield of this mutant variety is at the rate of 1 ton/ha of seed cotton higher than that of the parent variety 108-F. Physiological and biochemical analyses have demonstrated that the mutant “duplex” is characterized by a higher photochemical and enzymic activity of the chloroplast during the reproductive period. In this mutant the endogenous delay of leaf senescence is observed, whereas in the parent variety, as well as in many other crops, a decrease in the specific activity of key Calvin cycle enzymes occurs on the 20th-30th day after the emergence of leaves, when there are no visible alterations of chloroplast ultrastructure (Nasyrov, 1975). Physiologic-genetic prolongation of the functionally active state of chloroplasts and prevention of premature leaf senescence are therefore important factors in high photosynthetic productivity. D. CONVERSION OF C3 PLANTS TO C4 PATHWAY

The C4 pathway is observed not only in tropical grasses, but it is also widespread among dicotyledons (Dowton, 1975). Even within the same genus both C3 and C4 species together with their intermediate forms have been reported (Kestler et al., 1975). Zelitch (1973, 1975) and Bassham (1977) have produced data showing that in C, plants (wheat, rice, soybean, potato, peanut, barley, sugarbeet, cassava, banana, alfalfa, chlorella, eucalyptus, and algae) photosynthetic product losses upon photorespiration can reach 50% of the net assimilation. This is further evidenced by the data on the photosynthetic efficiency of C4 plants (maize, sorghum, millet, sugarcane, and napier grass) in which photorespiration is practically absent (Lorimer and Andrews, 1973; Zelitch, 1975). The main source of CO, release in the process of photorespiration is decarboxylation of intermediates of glycolate metabolism (Bishop and Reed, 1976). The glycolate pathway is closely linked with the Calvin cycle, particularly with oxygenation of RBPCase resulting in the formation of phosphoglycolic acid (Tolbert,

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

181

1971). It is generally accepted that the key enzyme RBPCase plays a bifunctional role, catalyzing carboxylation and oxygenation reactions. Oxygen and CO, compete for the reaction substrate ribulose 1,5-diphosphate (Laisk, 1977). In the enzyme oxygenase reaction, the sulfhydryl (SH) groups take part; their total amount is 96, but the number of easily modified, “accessible” ones is 12 (Akhmedov et al., 1976). It has been found that the photorespiration inhibitor glycidate irreversibly inactivates oxygenation, the carboxylase activity of the enzyme not being affected (Wildner and Henkel, 1976). Evidently oxygenationspecific inhibition is connected with blocking of enzyme SH groups located in its active site. In its native state the conformation of this oligomeric protein complex and its functional activity depend on conformation and dimensions of large and small subunits (von Wettstein et al., 1978). Analyzing the peptide composition of RBPCase in 60 tobacco species and 10 plants belonging to variable phylogenetic groups-from blue-green algae to ginkgo-Kung (1976) has identified that in all species studied the large subunit contains only three polypeptides, while the number of polypeptides in the small subunit varies from one to four. How could the C3plants be converted to the C4 pathway? The more efficient C4 plants characteristically have two kinds of photosynthetic tissue (dimorphic chloroplasts). The mesophyll cells fix CO, by way of phosphoenolpymvate (PEP) carboxylase to form oxaloacetate and thence the four-carbon compounds malate or aspartate. A current hypothesis is that one of these acids, depending on the plant species, is transferred to the cells sheathing the vascular bundles of the leaves where decarboxylation occurs (Day, 1977). The peculiar anatomy of C4 species involving complementation between dimorphic chloroplasts is perhaps responsible for some part of their greater yield. In the genus Atriplex, conventional sexual hybridizations between C, and C4 species have been made. Among F, and F3 progeny of hybrids between A . hastata and A . rosea, individuals with the anatomy of the C4parent A . rmea were present, but none had photosynthetic rates that approached it (Bjorkman, 1976). These approaches for selecting plants with C4 anatomy and chloroplast dimorphisms in segregating lines after hybridization are being camed out in many other laboratories (Nasyrov, 1978). Other genetical screening experiments have distinguished a tobacco population from the parent variety “Havana seed” in which 25% of the progeny had low photorespiration (Zelitch, 1975). Some promising varieties with low photorespiration have been obtained through exhaustive genetic screening within populations of C3 species (Wilson, 1972). The C4 pathway is connected with the leaf kranz anatomy, and therefore it seems rather impossible to impart properties of cooperative photosynthesis to C3 plants. Several other experimental approaches to improve photosynthetic efficiency of our crop species include (1) mutant isolation and selection of C3 cultivars with low rates of photorespiration and high rates of net photosynthesis; (2) enhancement of the auxiliary COP fixation through PEP carboxylase; (3)

I82

H . K . SRIVASTAVA

induction of desirable C3 plant mutants by mutagen treatments, and (4)incorporation of C4 dimorphic chloroplasts into protoplasts by somatic cell hybridization. Although plants with low rates of photorespiration and high rates of net photosynthesis were observed in a tobacco cultivar, attempts to establish this phenotype in several generations of pedigree selection were unsuccessful (Zelitch and Day, 1973). Several mutagen-treated tobacco haploid cell lines possessing slower than normal rates of photorespiration and vigorous autotrophic growth in low concentration of C 0 2 have been identified (Zelitch et al., 1977). The possibility of the enhancement of the auxiliary C02 fixation through PEP carboxylase among C , plant mutants produced by genetic manipulations also deserves mentioning. The formation of PEP carboxylase is controlled by nuclear genes, but the enzyme operates in the cytoplasm. It has a higher affinity for HCO, than for C O z (Filmer and Cooper, 1970); therefore it can bind the hydrated Cot more efficiently and transfer it to the chloroplasts, thus decreasing the limitation of photosynthesis by carboxylation resistance. The activity of PEP carboxylase in C , plants is 60 times lower than that in C4 plants (Slack and Hatch, 1967). Somatic cell genetics has already found widespread use in mutant isolation and selection. It may be feasible in future to transfer C4 dimorphic chloroplasts to C3 species through protoplast fusion and thus improve the photosynthetic efficiency and crop yield. The assumption that the C4 pathway originated in the course of evolution as a consequence of plant adaptation to arid and hot climates favors the experimental possibility of any such conversion of C , plants to the C4 pathway.

VII. SUMMARY AND CONCLUSIONS Both mitochondria and chloroplasts are self-replicating organelles, arising only by growth and division of preexisting mitochondria and chloroplasts. The organelle genetic systems are thus genetically autonomous, obeying their own unique laws of uniparental or biparental (non-Mendelian) inheritance. It is apparent that organelle genomes are present in multiple (polyploidy) copies per cell, and that they are packed in one to many organelles and, within each organelle, in one to many nucleoids (DNA gene centers). The population of organelle genomes within a cell may be divided into subpopulations in individual organelles and in nucleoids. These “populations within populations” complicate the analysis of organelle transmission genetics. The application and interpretation of some generalized laws of organelle genetics are made more difficult by the lack of good quantitative data on the number of nucleoids and of DNA molecules per nucleoid, or on the behavior of nucleoids during sexual reproduction in higher plants. It is likely that the segregating units in organelles are groups of mt-DNA or ct-DNA molecules, perhaps corresponding to the nucleoids or

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

183

even to whole mitochondria or segments thereof in some organisms. The transmission of organelles in higher plants appears to be predominantly maternal or uniparental, but the possibility of biparental zygotes resulting due to equal or unequal contributions of organelles from maternal and paternal gametes cannot be ruled out. Paternal organelles are visibly destroyed in the zygotes of some mammals, algae, and higher plants. However, there is evidence suggesting that paternal chloroplast or mitochondrial genes, though in minority, do survive in the biparental zygote and may have a significant role in maintaining plasmon heterogeneity. Heterosis and complementation with regard to mitochondrial and chloroplast activities are observed in several economically important crops. The present evidence indicates that mitochondrial heterosis is the result of complementation between polymorphic mitochondria that may be present in the hybrid as a result of inter- or intraorganelle recombinations. The significance of mitochondrial polymorphism or heterogeneity in the manifestation of heterosis and homeostasis is also illustrated. The genomes of organelles are insufficient to code for all the functional and structural proteins of mitochondria and chloroplasts. Close cooperation between nuclear and organelle genes seems to dominate the scene and determine cell life strategies during plant growth, development, and maturation, for most organelle enzyme systems also contain polypeptides coded by nuclear genes. The results on biosynthesis of key enzymes (RBPCase in chloroplasts, and cytochrome oxidase and ATPase complex in mitochondria) of organelle systems influenced by both nuclear and organelle genomes provide insights into the physicobiochemical and genetical mechanisms responsible for heterotic expressions. A hypothesis based on intergenomic interaction to interpret the operational mechanism of heterosis is discussed with many examples. Superior mitochondrial and chloroplast functions resulting due to genomic and intergenomic complementations are considered as essential components of heterosis. Arguments in favor of organelle polymorphism, which may be transmitted biparentally, and its genetic association with heterosis and Complementation are presented. Recent genetic studies indicate that genetic differences in cell size are due to additive-type genes, whereas genetic differences in cell division rate are due to nonadditive genes (epistatic interaction). Since heterosis is the phenomenon of superior growth and development, a possible molecular mechanism of heterosis would involve complementation of proteins or polypeptide subunits encoded by “multigenomes” instead of unique nuclear genes. The resulting products due to intergenomic interactions in organelles of hybrid organisms would then be expected to ensure enhanced structural and catalytic functions of these organelles and also superior cell division rate. Several distinct lines of evidence ranging from biochemical, physiological, ultrastructural, and restriction endonuclease organelle DNA fragment analyses are available at present to show that all three genetic sources (genomic, plasmonic, and inter-

184

H. K . SRIVASTAVA

genomic interaction) and not just one of them are at work during the manifestation of heterosis. There is evidence to suggest that DNA of hybrids possesses high replicative activity, repetitions of ribosomal RNA loci (rDNA), and relatively greater repeated nucleotide sequences in the genome’s active sites. The plausibility that interaction between the active sites of multigenomes creates favorable conditions for epistatic genes pertaining to major functional loci in heterozygotes and its subsequent role during heterotic expression is also considered. It is suggested, however, that the active principle of intergenomic interactions (cooperative interactions between nuclear and organelle genomes) for vigorous organelle function and enhanced cell division rate lies either with a messenger RNA molecule or the subunit polypeptide encoded by the nuclear genes. The field of transmission genetics of organelles in relation to crop yield improvement appears to be coming of age. Recent results suggest that both mt-DNA and ct-DNA are deeply involved in various genetic phenomena including cytoplasmic male sterility in higher plants. The genetic control of the content, amino acid composition, and quality characters of proteins by means of cytoplasmic organelle-nuclear gene interaction in cereal crops has also been reported. Many studies from a number of different laboratories reveal that hybrids exhibiting a high level of grain yield heterosis also show heterosis for early growth and development at young seedling stages. Organelle heterosis and complementation are also significantly correlated with grain yield heterosis in many crops. The use of mitochondrial complementation or chloroplast complementation as a tool to preselect well-combining parents under laboratory conditions for prediction of grain yield heterosis in crop plants is recommended. Both mitochondrial ADP : 0 and respiratory control ratios have been found to be highly correlated ( r = 0.94) with average grain yield in wheat. Several experimental approaches to enhance grain yield potential of crop plants based on mitochondrial and chloroplast efficiencies have been discussed. The improvement of crop yield in terms of mitochondrial and chloroplast breeding, especially in cereal crops, is a real challenge to modem plant geneticists, and the maturity of the field is suggestive of the fact that complementation data of organelles in 1 : 1 parental mixture should be preferred over time-consuming and expensive biometrical methods and estimations, because apparently both plasmon and nuclear genes are involved. For major innovations in crop yield improvement, plant breeders first need to bridge the technology gap between molecular genetics and plant breeding and to utilize new methods by which they can follow rates of photorespiration, net photosynthesis, and mitochondrial and chloroplast activities. The future task would be to explain the phenomena of heterosis and complementation by mitochondria and chloroplasts at the molecular level, in terms of the behavior of parental organelles and their DNA molecules in a predominantly maternal andor biparental zygote. Another major question that remains to be examined relates to

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

185

the precise operational mechanism of intergenomic interaction (complementation between nuclear and organelle genomes): does it operate through transcriptiontranslation systems, a protein-synthesizing system, or both? It is my belief that recent developments with protoplast fusion of crop plants will lead us to develop new germplasm resources as a result of recombination of both nuclear and organelle genomes for use in orthodox breeding and eventually to ways of selecting improved high yielding varieties. ACKNOWLEDGMENTS Much of the planning and writing of this review was done while the author was asked to deliver a series of lectures in a joint UNESCO/UNDP sponsored international course on “Genetic Hierarchy in the Multigenomic Eukaryotic Cells” for a group of professors belonging to several South American universities at the University of Valle. I wish to thank many colleagues from all over the world who generously shared reprints, manuscripts, and unpublished observations. I only regret that space did not permit a discussion of all their work. My own research is supported by Colciencias and Semilla Valle, S. A. My gratitude also goes to Miss Alba M. Dominguez for patiently typing this manuscript.

REFERENCES Adams, G . M. W., van Winkle-Swift, K. P., Gillham, N. W., and Boyton, J. E. 1976. In “Genetics of Algae” (R. A. Lewin ed.), pp. 69-1 18. Univ. Calif. Press, Berkeley. Adoutte, A. 1977. La Genetique des Mitochondries chez la Paramecie. Ph.D. Thesis. Univ. ParisSud, Centre-d’Orsay. Akhmedov, Y. D., Ulmasov, K . A., and Akhrnedova, L. R. 1976. In “Mater. Resp. Konf. Melodyh. Uchenyh,” pp. 5-11. Dushan be: Donish. Alan, S., and Sandal, P. C. 1969. Crop Sci. 9, 157-159. Alberts, R. S., Thornber, J. P., and Naylor, A. W. 1973. Proc. Narl. Acad. Sci. U.S.A. 70, 134- 137. Aldrich. H. C., Gracen, V. E., York, D., Earl, E. D., and Yoder, 0. C. 1977. Tissue Cell 9, 167- 178. Ali-Zade, M. A,. and Aliev, R. T. 1973. Kokl. Akad. Nauk. USSR 21, 72-74. Allsopp, A. 1969. New Phyrol. 68, 591-612. Amos, J. A,, and Scholl, R. L. 1977. Crop Sci. 17, 445-448. Andre, J. 1962. J . Ultrustruct. Res. (Suppl.)3, 1-185. Andregava, T. V . 1976. Plunr Breed. Abstr. 46, 6839. Arslanova, S. V . 1973. Vzb. Biol. Zhur. 3, 6-9. Ashby, E. 1937. Ann. Bor. 1, 11-41. Ashry, A. 1976. Theor. Appl. Genet. 48, 17-21. Ayers, G. S., Went, V. F., and Ries, S. K. 1976. Ann. Bot. 40, 563-570. Bandlow, W., Schweyen, R. J., Wolf, K., and Kaudewitz, F. 1977. “Mitochondria 1977.” De Gruyter, Berlin. Barber, H. N. 1970. Taxon 19, 154-160. Barratt, D. H. P., and Flavell, R. B. 1975. Theor. Appl. Genet. 45, 315-321. Barratt, D. H. P., and Flavell. R. B. 1977. Ann. Bor. 41, 1333-1343. Barratt, D. H. P., and Peterson, P. A. 1977. Maydica 22, 1-8. Bassham, J. A. 1977. Science 197, 630-638.

186

H. K . SIUVASTAVA

Bastia, D., Chiang, K. S . , and Swift, H. 1969. J. Cell Biol. 43, 1 la. Bauer, E. 1909. Z. Vererbungsl. 1, 330-351. Bazzaz, M. B., and Govindjee. 1973. Plant Physiol. 52, 257-262. Belcour, L., and Begel, 0.1977. Mol. Gen. Genet. 153, 11-21, Bell, P. R. 1970. I n “Control of Organelle Development” (P. L. Miller, ed.), pp. 109-128. Academic Press, New York. Berbec, J . 1974. Z. Pfanzenzvecht. 73, 204-216. Berger, E. 1974. Genetics 78, 1157-1 172. Berger, E. 1976. A m . Nut. 110, 823-839. Bergman, J. W. 1972. Ph.D. Dissertation, North Dakota State University. Bergman, J . W., and Maan, S . S. 1973. Proc. Int. Wheat Genet. Svmp.. 4th. pp. 329-335. Berova, S., and Logodinova, E. 1972. Rastenievud. Nauki 1, 73-79. Bertrand, H., and Pittenger, T. H. 1969. Genetics 61, 643-659. Berville, A. 1977. Agron. Trop. 32, 141-147. Berville, A., and Demarly, Y . 1970. Rep. Eucarpia Meet., Versaifles, June 10-12. Berville, A,, Labib, A., Thiellement, H., and Kouame, B. 1976. Ann. Amefior. Plant. 26, 607-622. Berville, A., Gamier, H., Aranda, G., Fetizon, M., Cassini, R., Poret, B., Comu, A., and Vuillaume, E. 1977. Proc. Congr. Eucarpia, 81h, Madrid, pp. 221-225. Bingham, J. 1972. Proc. Congr. Eucarpia, 6th pp. 15-30. Birkey, C. W. 1978. Annu. Rev. Genet. 12, 471-512. Birky, C. W., Demko, C. A,, Perlman, P. S., and Strausberg, R. 1978. Genetics 89, 615-51. Bishop, D. G., and Reed, M. L. 1976. Photochem. Photobiol. Rev. 1, 1-69. Bishop, D. G., Anderson, K . S . , and Smillie, R . M. 1977. Plant Physiol. 50, 774-77. Bishop, N. I. 1972. In!. Congr. Photosynth. Res.. 2nd. Hague, pp. 459-468. Bjorkman, 0. 1976. I n ‘TO, Metabolism and Plant Productivity” (R. H. Bums and C. C. Black, eds.), p. 343. Univ. Park Press, Baltimore, Maryland. Blixt. S. 1968. Agric. H o r t . Genet. 24, 87-91. Bogorad, L. 1975. Science 188, 891-898. Boker, E., Kaudewitz, F., Richmond, V., Schweyen, R., and Thomas, D. Y. 1976. In “Genetics, Biogenesis, and Bioenergetics of Mitochondria” (W. Bandlow, R. J. Schweyen, D. Y. Thomas, K. Wolf, and F. Kaudewitz, eds.), pp. 99-1 15. De Gruyter, Berlin. Bruce. A. B. 1910. Science 32, 627-628. Busch, R. H., and Maan, S. S . 1974. Wheat News Lett. 20, 163-166. Callan, H. G. 1967. J . Cell Sci. 2, 1-7. Campbell, P. N., and Blobel, G. 1976. FEBS Lett. 72, 215-226. Carlsen, B. 1977. Carlsberg Res. Commun. 42, 199-202. Carlsen, P. S . , Smith, H. H., and Dearing, R. D. 1972. Proc. N a t l . Acad. Sci. U . S . A . 69, 2292-2294. Caspari, E. 1956. Genetics 41, 107-117. Cavalier-Smith, T. 1970. Nature (London) 228, 333-335. Chen, K.. Gray, J . C., and Wildman. S. G. 1975. Science 190, 1304-1306. Chen, K., Johal, S . , and Wildman, S. G. 1976. In “Nucleic Acids and Protein Synthesis in Plants” (J. H. Weil and L. Bogorad, eds.). Plenum, New York. Chen, K., Wildman, S . G., andsmith, H. H. 1977. Proc. N a t l . Acad. Sci. U . S . A . 74, 5109-5112. Chen, P. S. 1971. “Monograph in Developmental Biology,” Vol. 3. Karger, Basel. Chiang, K. S . 1968. Proc. Natl. Acad. Sci. U . S . A . 60, 194-200. Chiang, K . S. 1971. In “Autonomy and Biogenesis of Mitochondria and Chloroplast” (N. K . Boardman, A. W. Linnane, and R. N. Smillie, eds.), pp. 235-249. North-Holland Pub]., Amsterdam. Chiang, K . S. 1976. In “Genetics and Biogenesis of Chloroplasts and Mitochondria,” pp. 305-312. North Holland Publ.. Amsterdam.

INTERGENOMIC INTERACTION, HETEROSIS, A N D CROP YIELD

187

Chiang, K . S., and Sueoka, N. 1967. Proc. N a t l . Acad. Sci. U . S . A . 57, 1506-1513. Coen, D. M . , Bedbrook, J . R., Bogorad, L., and Rich, A. 1977. Proc. Natl. Acad. Sci. U.S.A. 74, 5487-5491. Correns, C . 1909. Z . Vererhungsl. 1, 291-329. Coyne, J. 1976. Genetics 84, 593-607. Crick, F. H . C . , and Orgel, L. E. 1964. J. Mol. Biol. 8, 161-165. Cubada, R., Bozzini, A., and Quattrucci, E . 1975. Theor. Appl. Genet. 45, 290-293. Dalling, M. J . , Tolbert, N. E., and Hagemann, R. H. 1972. Biochim. Eiophys. Acra 283, 505-515. David, J. B., and Blackler, A . W. 1972. Dev. Biol. 29, 152-162. Day, P. R. 1977. Science 197, 1334-1339. Demarley, Y. 1976. In “Heterosis in Plant Breeding” (A. Janossy, & F. G . H. Lupton, eds.), pp. 183-196. Elsevier, Amsterdam. Demrec, M. 1964. Proc. Narl. Acad. Sci. U.S.A. 51, 1057-1060. Demrec. M., and Hartman, P . E. 1959. Annu. Rev. Microbiol. 13, 377-385. De Vries, A . , and le, T. S. 1970. Eupharica 19, 103-108. Dimitrov, P., Nashkova, O., Petkova, S., Nashokov, D., and Marinkov, E. 1974. Theor. Appl. Genet. 45, 91-95. Dobberstein, B., Blobel, G . , and Chua, N. H . 1977. Proc. Natl. Acad. Sci U . S . A . 74, 1082-1085. Dobzhansky, T . 1952. In “Heterosis” (J. W . Gowen, ed.), pp. 218-233. Iowa State College Press, Ames. Donaldson, C . , and Blackman, G. F. 1974. Ann. Bor. 38, 515-527. Doney, D. L., Theurer, J . C . , and Wyse, R. E. 1972. Crop Sci. 12, 493-494. Doney, D. L., Theurer, J. C., and Wyse, R. E. 1976. Eupharica 24, 387-392. Dowton, W. J. S. 1975. Photosynrheticu 9, 96-105. Dzapo, V . , Reuter, H.. and Wassmuth, R . 1974. Z . Zuechtungsbiol. 90, 169-179. East, E. M. 1908. Rep. C t . Agric. Exp. Sta. 1907 pp. 419-428. Ebner, E., Mennucci, L., and Shatz, G . 1973. J. Biol. Chem. 248, 5360-5372. Ellis, J. R. S., Brunion, C . J . , and Palmer, J. M. 1973. Nature (London) 241, 45-47. Erickson, J . R. 1967. Ph.D. Thesis, North Dakota State University, Fargo Diss. Abstr. 29, 2254B (Order No. 69-528). Evans, L. T. 1975. “Crop Physiology.” pp. 327-55. Cambridge Univ. Press, London and New York. Evans, L. T . , and Wardlaw, I . F. 1976. Adv. Agron. 28, 301-349. Filmer, D. L., and Cooper, T. A. 1970. J. Theor. Biol. 29, 131-145. Fincham, J. R. S . 1966. “Genetic Complementation,” p. 143. Benjamin, New York. Fincham, J. R. S. 1969. Proc. Int. Bot. Congr., 11th 60. Fincham, J . R . S. 1972. Heredify 28, 387-391. Flavell, R. B., and Barratt, D. H. P. 1977. Ann. Bot. 41, 1323-1331. Flavell, R . B., Odell, M., Rimpau, J., and Smith, D. 1978. Heredity 40, 439-455. Fleming, A . A. 1975. Crop Sci. 15, 570-573. Fleming, A. A , , and Palmer, J. H. 1975. Crop Sci. 15, 617-620. Fukusawa, H. 1967. Am. N a t . 101, 41-46. Fukusawa, H., Mito, K . , and Fujiwara, M. 1957. Bot. M a g . 70, 251-257. Gelvin, S . , Heizamann. P . , and Howell, S. H. 1977. Proc. Nurl. Acad. Sci. U.S.A. 74,3193-3197. Chose, R., Abrol, Y. P., and Sinha, S. K . 1974. Plant Sci. Lett. 2, 173-176. Giarnordoli, M. 1974. In “Fertilization in Higher Plants’’ (H. F. Lisken, ed.), pp. 221-232. North-Holland Publ., Amsterdam. Gibbons, G . C., Strobaek, S . , Haslett, B . B., and Boulter, D. 1975. Experientia 31, 1040-1041. Gibbs, S . P., and Prole, R. J. 1973. J. Cell Biol. 59, 318-328. Gibbs, S . P., Mak, R . , and Slankis, T. 1974. J. Cell Sci. 16, 579-591. Gibson, P. T., and Schertz, K . F. 1977. Crop Sci. 17, 387-391.

188

H. K. SRIVASTAVA

Giles, K. L. 1974. Plant Cell Physiol. 15, 281-285. Giles, N. H. 1978.Am. Nut. 112, 641-657. Gillespie, J. H. 1976. Am. Nut. 110, 809-821. Gillham, N. W. 1974.Annu. Rev. Genet. 8, 347-391. Gillham, N. W. 1978. “Organelle Heredity.” Raven, New York. Gilyazetdinov, S. Y., Yakhin, I. A , , and Vakhitov, V. A. 1977. Tsitol. Genet. 11, 114-118. Gour, V . K., and Singh, C. B. 1977. Can. J . Genet. Cytol. 19, 187-188. Gowen, J . W. 1952. “Heterosis.” Iowa State College Press, Ames. Gran, D. G., and Possingham, J. V. 1972. Nature (London) New Eiol. 235, 142-149. Griffiths, D. F., Lancashire, W. E., and Zanders, E. D. 1975. FEES Lett. 53, 126-130. Grimes, G.W., Mahler, H. R., and Perlman, P. S. 1974. J . Cell Eiol. 61, 565-574. Grimwood, B. G. 1972.Ph.D. Thesis, University of Arizona. Grobe, B., and Arnold, C. G. 1975. Protoplusma 86, 291-294. Grun, P. 1976. “Cytoplasmic Genetics and Evolution.” Columbia Univ. Press, New York. Hagemann, R. 1964. “Plasmatische Vererbung. ” Fischer, Jena. Hagemann, R. 1976.In “Genetics and Biogenesis of Chloroplasts and Mitochondria,” pp. 33 1-338. North-Holland Publ., Amsterdam. Hagemann, R . H., Leng. E. R., and Dudley, J. W. 1967. Adv. Agron. 19, 45-86. Hanna, W. W., Burton, G. W., and Powell, J. B. 1978.J . Hered. 69, 273-274. Hanson, W. D., Moreland, D. E . , and Shriner. C. R. 1975. Crop Sci. 15, 62-66. Hart, G. E. 1978, Cereal Res. Commun. 6, 123-124. Hart, G. E., and Langston, P. J. 1977. Heredity 39, 263-277. Hartwel, L. H . 1970.Annu. Rev. Genet. 4, 373-382. Hatch, M. D. 1976.Physiol. Eiochem. Cull. Plants USSR 8, 873-883. Hatch, M. D.. and Slack, C. R. 1966.Eiorhem. J . 101, 103-Ill. Harvey, P. H., and Levings, C . S. 1972. Adv. Agron. 24, 1-27. Heber. U. B., and Gottschalk, W. 1964.Z. Narurforsch. 136, 26-27. Heichel, G. H., and Musgrave, R. B. 1969. Crop Sri. 9, 483-486. Hemnann, F. 1971. FEES Lett. 19, 267-269. Hermann, R. G . , Kowallik, K. V., and Bohnert, H. J. 1974. Arm Eiol. Ser. A 14, 91-110. Higgins, M. L., and Shockman, G. D. 1971. C . R . C . Crit. Rev. Microbiol. 1, 29-72. Highfield, P. E., and Ellis, R. J. 1978.Nature (London) 271, 420-424. Highkin, H. R.,Boardman, N . K . , and Goodchild, D. J . 1969. Plant Physiol. 44, 1310-1320. Hobson, G. F. 1971. Biochern. J . 124, 10-11. Holder, A. A. 1976. Carlsberg Res. Commun. 41, 321-324. Holder, A. A. 1978. Carlsherg Res. Commun. 43, 421-427. Holson, 0. 1967. Annu. Rev. Genet. 1, 245-268. Honda, S. I., Kwanyven, P., and Wildman, S. G. 1971. Plunta 97, 1-15. Horak, A., and Zalik, S. 1975. Can:J. Eor. 53, 2399-2404. Horak, I., Coon, H. G., and Darwin, I. B. 1974. Proc. Narl. Acad. Sci. U.S.A. 71, 1828-1832. Hraska, S. 1978. Polnohospodarstvo 24, 22-27. Hsam, S. L. K., and Larter, E. N. 1974. Can. J . Gin. Cytol. 16, 333-340. Hubby, J., and Narise, S. 1967. Genetics 57, 291-300. Hull, F. H. 1945.J . Am. SOC. Agron. 37, 134-145. Hutchinson, C. A., Newbold, J . E., Potter, S. S., and Edgell, M. H . 1974. Nature (London)251, 536-538. Imamaliev, A. A., Rakhmankulov, S. A., and Azizkhodzhaev, A. 1975. Fizilol. Rosr USSR 22, 923-928. Israelstam, G. F., and Fukumato, K . 1977. Eiol. Plant 19, 53-58. Izhar, S., and Frankel, R. 1973. Theor. Appl. Genet. 43, 13-17.

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

189

Jacob, F., and Monod, J. 1961. J. Mol. Biol. 3, 318-356. Jensen, R. G., and Bahr, J . T. 1977 Annu. Rev. Plant Physiol. 28, 379-400. Jinks, J. L. 1964. “Extrachromosomal Inheritance. ” Prentice-Hall, New York. Johnson, G. B. 1974. Science 184, 28-37. Johnson, V. A,, and Mattern, P. J. 1975. “Improvement of the Nutritional Quality of Wheat Proteins through Increased Protein Content and Improved Amino Acid Balance. ” Report, Research from Jan. 1, 1973-March 31, 1975. AID/csd-l208 Washington, D.C. Jones, D. F. 1917. Genetics 2, 466-479. Kameya, T., and Takaha, N. 1971. Jpn. J . Genet. 46, 153-157. Kapuler, A. M., and Berstein, H. 1963. J. Mol. Biol. 6, 443-451. Karmanenko, N . M. 1976. In “Metody Priemy Povyscheniya Zemostui Kosti Osymykh Zern. Kul’tur,” pp. 365-371. Kolos, Moscow. Kazim, A. A . 1974. Diss. Abstr. 35 (16B). Kennedy, R. A , , and Laetch, W. M. 1973. Planfa 115, 113-124. Kern, J. J., and Atkins, R. E. 1972. Crop Sci. 12, 835-839. Kestler, D. P., Mayne, B. C . , Ray, T. B., Goldstein, L. D., and Black, C. C. 1975. Biochem. Biophys. Res. Commun. 66, 1439- 1446. Khanna, R. 1974. Ph.D. Thesis. Indian Agricultural Research Institute. New Delhi. Khanna, R., and Sinha, S. K. 1974. Plant Biochern. J. 1, 32-36. Khanna, R., and Sinha, S. K. 1975. Exp. Bot. 20, 45-57. Khokhlova, L. P., Denisova, G. A., Suleimanov, I. G.. Pankratova, S. I., Stupishina, E. A,, and Khazhina, R. G. 1975. Ref. Z . 12, 77-79. Kihara, H., and Tsunewaki, K. 1964. Seiken Ziho 16, 1-14. Kirk, J. T. D., and Tilney-Bassett, R. A. E. 1978. “The Plastids: Their Chemistry, Structure, Growth and Inheritance,” 2nd ed. Elsevier, Amsterdam. Kittock, D. L., and Low, A. G. 1968. Agron. J. 60, 286-288. Klepper, L. A. 1975. Proc. Int. Winter Wheat Crops Conf., 2nd pp. 334-345. Kloppstech, K., and Schweiger, H. G. 1974. Plant Sci. Lett. 2, 101-105. Knoth, R., and Hagemann. R. 1977. B i d . Zentrabl. 96, 141-150. Knyaseva, V . P., and Romanova, E. B. 1977. Nauchtr. N I1 Kartaf. Khra 27, 41-44. Koehn, R. K., Perez, J . E., and Menitt, R. B. 1971. Am. N a f . 105, 51-69. Kolodner, R., and Tewari, K . K. 1975. Biochim Biophys. Acta 402, 312-390. Konarev, V. G. 1973. T r . Prik. Bot. Genet. Sel. Ser. 52, pp. 5-28. Konarev, V. G. 1976. I n “Heterosis in Plant Breeding’’ (A. Janossy and F. G. H.Lupton, eds.), pp. 81-94. Elsevier, New York. Konarev, V. G., and Tuterev, S. L. 1971. I n “Biochemical and Cytochemical Methods of Nucleinacid Determination in Plants.” Kolos, Moscow. Konarev, V. G . , Arhmetov, R. R., and Gilyasetdinov, S. Y. 1971. Selsk. Biol. 6, 653-662. Konzak, C. F. 1977. Adv. Genet. 19, 407-585. Kouame, B. 1978. Oleagrneux 33, 267-275. Ku, S. B., Gutierrez, M., Kanai, R., and Edwards, G. E. 1974. Z . Pflanz enK. 72, 320-337. Kung, S. D. 1976. Science 191, 429-434. Kung, S. D. 1977. Annu. Rev. Plant Physiol. 28, 401-437. Kung, S. D., and Williams, J . P. 1969. Biochim. Biophys. Acfa 195, 434-445. Kung, S. D., Tornber, J. P., and Wildman, S. G . 1972. FEBS Left. 24, 185-188. Kung, S. D., Gray, J . C., Wildman, S. G . , and Carlson, P. S. 1975. Science 187, 353-355. Kuroiwa, T., Kawano, S., and Hizume, M. 1976. Exp. Cell Res. 97, 435-440. Kuroiwa, T . , Kawano, S., and Hizume, M. 1977. J. Cell Biol. 72, 687-694. Kutzelnigg, H., and Stubbe, W. 1974. Subcell. Biochem. 3, 73-89. Laetch. W. M. 1974. Annu. Rev. Plant Physiol. 25, 27-52.

190

H . K. SRIVASTAVA

Lahib, A., and Kader, J . C . 1977. Phv/ochenri,s/r~16, 1481-1484. Laisk, A. H. 1977. “Kinetika Photosinteza i Photodyanija C,-Rastemii. Nauka, Moscow. Larson, D. A. 1965. Am. J. Bor. 52, 139-154. Laughnan, J . R . , and Gabay, S. J . 1975. In “Genetics and Biogenesis of Mitochondria and Chloroplast” ( W . W. Birky, P. S. Perlman, and T . J. Byers, eds), pp. 330-349. Ohio State Univ. Press, Columbus. Lax, C., and Fogel. S. 1978. Generics 90, 501-516. Lazarus, C. M . , and Turner, G. 1977. Mol. Gen. Gene/. 156, 303-31 I . Lee, S. L. J. 1976. Ph.D. Dissertation, University of Florida, Gainsville. Leech, R. M. 1976. ln “Perspectives in Experimental Biology” (N. Suderland, ed.), pp. 145-162. Pergamon, Oxford. Lehninger, A. 1976. “Biochemistry.” Worth, New York. Lerner, I . M. 1958. “The Genetic Basis of Selection.” Wiley, New York. Levings, C. S., and Pring, D. R. 1976. Science 193, 158-160. Levings, C. S . , and Pring, D. R. 1977. J. Hered. 68, 350-354. Lewis, D. 1954. Heredirv 8, 333-356. Logodinova, E. 1962. Sytnp. Gene!. Wheu! Breed., Bulg. Acad. Sci. Sofiu pp. 459-467. Lombardo, G., and Gerola, F. M. 1968. Plonra 82, 105-1 10. Lopez, A., and Grabe, D. F. 1973. Proc. A S S C J OH. ~ . Seed Anal. 63, 106-1 16. Lorimer, G. H., and Andrews, T. J. 1973. Nature (London) 243, 359-360. Lorimer, H. G . Woo. K. C.. Berry, J . A . , and Osmond, C . B. 1977. Irr “Photosynthesis” (J. Combs, and T . W. Goodwin, eds.), pp. 31 1-322. Biochem. SOC.of London. Luck, D. J . L.. and Reich, E. 1964. Proc. Nor/. A m ( / . Sci. U . S . A . 52, 931-938. Lupton, F. G. H. 1968. “Genetica Agraria, Eucarpia” (G. C . Chisci and G . Haussmann. eds.), pp. 457-464. Milano. Lupton, F. G. H. 1976. In “Heterosis in Plant Breeding” (A. Jamossy and F. G . H. Lupton. eds.), pp. 71-80. Elsevier. New York. Lurquin. P. F. 1977. Prog. Nucleic Acid Res. Mu/. B i d . 20, 161-207. Maan, S. S . 1975. I n “Genetics and Breeding of Durum Wheat’’ (G. T . Scarascia-Mungnoza. ed.), pp. 117-137. Bari. Rome. Maan, S. S. 1976. CropSci. 16, 757-761. Maan, S. S. 1977. Euphyricu 26, 601-613. McArthur. J . A . , Hesketh, J . D., and Baker, D. N. 1975. In “Crop Physiology,” pp. 297-326. Cambridge Univ. Press, London and New York. McDaniel. R. G. 1969. Crop Sci. 9, 823-827. McDaniel, R. G. 1971. lrr “Barley Genetics 2” (R. A. Nilan, ed.), pp. 323-337. Washington State Univ. Press, Pullman. McDaniel, R. G . 1972. Norure (London) New Biol. 236, 190-191. McDaniel, R. G. 1973. Proc. / t i / . Whetrr Genet. S w n p . . 4th Agric. Exp. Sta. Columbia pp. 541-546. McDaniel, R. G . 1975. Barley Geirer. News Lerr. 5 , 29-30. McDaniel, R. G., and Grimwood, B. G. 1971. Comp. Biochein. Phvsiol. 38, 309-314. McDaniel, R. G.. and Sarkissian, I. V. 1966. Science 152, 1640-1642. McDaniel, R. G.. and Sarkissian, I. V. 1968. Generics 59, 465-475. McDaniel. R. G.. and Sarkissian, I. V. 1970. Physiol. Plant 23, 335-342. McGavin. S. 1968. J. Mol. B i d . 37, 239-242. Machold, 0.. and Aurich, 0. 1972. Biorhiin. Biophys. Acfu 281, 103-112. McKee. H. S . 1962. “Nitrogen Metabolism in Plants.” Oxford Univ. Press, London and New York. Mackey, J. 1970. Herediras 66, 165-176. Mackey, J. 1976. / t i “Heterosis in Plant Breeding” (A. Janossy and F. G. H. Lupton, eds.), pp, 17-33. Elsevier, New York. ”

INTERGENOMIC INTERACTION, HETEROSIS. A N D CROP YIELD

191

Magalhaes, A. C . , Neyra, C. A.. and Hagemann, R. H. 1974. P l m t Phvsiol. 53, 411-415. Margulius, L. 1970. “Origin of Eukaryotic Cells.” Yale Univ. Press, New Haven, Connecticut. Marton, I. 1962. J. Exp. Eor. 13, 325-333. Marx, J . L. 1978. Noture (London) 199, 517-518. Mascarenhas. J . P. 1975. Eot. Rev. 41, 259-314. Mason, J . R . , and Turner, G . 1975. Mol. Gen. Genet. 143, 93-99. May, C . E., Vickery. R. S., and Driscoll, C. J . 1973. Proc. / J U . Wheat Genet. Symp., 4th pp. 843-849. Maynard. D. N., Barker, A. V . , Minotti, P. L., and Peck, H . N . 1976. Adv. Agron. 28, 71-118. Melchers. G.. Sacristan, M. D., and Holder, A . A . 1978. Carlsherg Res. Commun. 43, 203-218. Meyer, B., and Stubbe, W. 1974. Ber. Dtsch. Eot. Ges. 87, 29-38. Michaelis, G . , Petrochilo, E., and Slonimski, P. P. 1973. Mol. Gerr. Genet. 123, 51-65. Miflin. B. J . , and Hagemann, R . H. 1966. Crop Sci. 6, 185-187. Miller, R . J . , and Koeppe, D. E. 1971. Science 173, 67-69. Miller, S . , Pearcy, R., and Beryer. E. 1975. Eiochern. Genet. 13, 175-188. Mitcheson, J . M. 1971. “The Biology of the Cell Cycle.” Cambridge Univ. Press, London and New York. Morgensen, H. L. 1972. Am. J . E o t . 59, 931-941. Mounolou. J . C . , Jakob, H.. and Slonimski, P. P . 1966. Eiochem. Biophys. Res. Cotnmun. 24, 218-244. Mukai, Y . , and Tsunewakik. 1975. Seiken Ziho 15-26, 67-78. Mukhamadiev, B., and Zalensky, 0. V. 1972. Eot. Z h . 57, 260-265. Muresan, T . , Serban, M.. and Popov, D . 1976. Rev. Zourn. Eiochern. 13, 101-106. Nagy. A. H., Bokany, A , , Bacs, R . , Doman, N . G . , and Faludi-Daniel, A. 1972. Photosynrhetica 6, 7-12. Nass, M. M. K. 1976. In “Handbook of Genetics” (R. C. King, ed.), Vol. 5, pp. 477-533. Plenum, New York. Nasyrov, Y. S. 1975. “Photosynthesis and Genetics of Chloroplasts. Nauka, Moscow. Nasyrov, Y. S . 1977. tnt. Congr. Photosvnth. Res., 4th, Reading pp. 268-269. (Abstr.) Nasyrov, Y. S . 1978. Annu. Rev. Plant Phvsiol. 29, 215-237. Neyra, C. A., and Hagemann, R. H. 1974. Plant Physiol. 54, 480-483. Nilan, R . A. 1964. “The Cytology and Genetics of Barley, 1951-1962.” Washington State Univ. Press, Pullman. Nishimura, M., and Akazawa, T . 1973. Eiochem. Eiophvs. Res. Cornmun. 54, 842-845. Nosberger. V . J . 1970. Thesis “Separatabdruck Ans. Schweizerische Landwirtschaftliche Forschung,” Vol IX, No. 3/4, S, pp. 235-256. Oakley. K. M., and Clark-Walker, G. D. 1978. Genetics 90, 517-530. Ohba, K . , Iwakwa, M., Okada, Y., and Murai, M . 1971. Silvae Genet. 20, 101-107. Ohno, S. 1970. “Evolution by Gene Duplication. ” Springer-Verlag, Berlin and New York. Osumi. M., Shimoda, C . , and, Yanagishima, N. 1974. Arch. Microbiol. 97, 27-38. Ovchinnikova, M. F . , and Yakovlev, A . 1978. Sel. Sern. 2, 77-79. Palilova, A. N., Turbin, N. V., and Serpokrylova, L. S . 1966. Dokl. Akad. Nauk USSR 169, 677-678. Paolillo, D. J . 1974. /ti “Dynamic Aspects of Plant Ultrastructure” (A. W. Richards, ed.), pp. 504-31. McGraw-Hill, New York. Penning De Vries. F. W. T . , Brunsting, A . H . M . , and van Law, H. H. 1974. J. Theor. Eiol. 45, 339-317. Pinon, H . , Barat, M., Tourte, M., Dufresne, C., and Mounolou, I. C. 1978. Chromosoma 65, 383-89. Planchon, C . 1968. C . R . Acad. Sci. ( P a r i s ) 201 1-2014. ”

192

H. K. SRlVASTAVA

Planchon, C. 1976. In “Heterosis in Plant Breeding” (A. Janossy and F. G . H. Lupton, eds.), pp. 95-104. Elsevier, New York. Possingham, J. V . 1976. J. Microsc. Biol. Cell. 25, 283-288. Possingham, J. V., and Rose, R . J. 1976 Proc. R . Soc. London Ser. B 193, 295-305. Possingham, J. V., and Smith, J . W. 1972. J. Exp. Bot. 23, 1050-1059. Poulsen, C. 1977. Thesis, Copenhagen University. Poulsen, C. 1978. Brookhaven Symp. Biol. Photosynth. Carbon Ass. (30), 214-229. Poulsen, C . , Strobaek, S., and Haslett, B. G . 1976. In “Genetics and Biogenesis of Chloroplasts and Mitochondria” (E. A. Bucher, ed.), pp. 17-24. Elsevier, Amsterdam. Poyton, R. D., and Kavangh, J. 1976. Proc. Natl Acad. Sci. U.S.A. 73, 3947-3951. Pring, D. R., and Levings, C. S. 1978. Genetics 89, 121-136. Pring, D. R.. Levings, C. S., Hu, W. W. L., and Timothy, D. H. 1977. Proc. Nail. Acad. Sci. U.S.A. 74, 2904-2908. Raff, R. A., and Mahler, H. R . 1972. Science 177, 575-582. Rakhmankulov, S. A. 1975. Ref. Zh. 3, 55.118. Rakhmankulov, S. A., Zaprudes, E. G., and Azizkhodzhaev, A. 1976. Ref. Zh. 6, 55.300. Rao, A. P., and Fleming, A. A . 1978. Crop Sci. 18, 935-937. Rao, M. K., and Koduru, P. R. K. 1978. J. Hered. 69, 327-330. Rao, P. N. 1975. Theor. Appl. Genet. 45, 363-367. Raven, P. H . 1970. Science 169, 641-646. Rawson, J . R. Y., and Haselkom, R. 1973. J. Mol. Biol. 77, 125-132. Renner, 0. 1919. Z. Bot. 11, 305-380. Ressler, P. M . , and Emery, D. A. 1978. J. Hered. 69, 101-106. Rhoades, M. M. 1950. Proc. Natl. Acad. Sci. U.S.A. 36, 624-629. Ridley, S. M., and Leech, R. M. 1970. Nature (London)227, 463-465. Ries, S. K., and Emerson, E. H. 1973. Agron. J. 65, 884-886. Ries, S. L., Ayers, G., Went, V . and Everson, E. H . 1976. Can. J . Plant Sci. 56, 823-906. Robertson, W. D. 1971. Proc. Int. Symp., Barley Genet. 2nd pp. 220-242. Robins, E., and Gonata, N. K. 1964. J. Cell Biol. 21, 429-463. Roodyn, D. B., and Wilkie, D. 1968. “The Biogenesis of Mitochondria.” Methuen, London. Roos, E. E., and Sarkissian, I. V. 1968. Crop Sci. 8, 683-686. Rose, R. J. Gran, D. J . , and Possingham, J. V. 1974. Nature (London) 251, 641-642. Rose, R . J., Cran, D. G . , and Possingham, J . V. 1975. J. Cell Sci. 17, 27-41. Roupakias, D. G., and Kaltsikes, P. J . 1977. Can. J . Genet. Cyrol. 19, 39-49. Roy, H., Patterson, R., and Jagendorf, A. T . 1976. Arch. Biochem. Biophys. 172, 64-73. Ruebenbauer, T . 1967. Genet. Pol. 8, 1-41, Ryan, R., Grant, D . , Chiang, K. S., and Swift, H. 1978. Proc. Natl. Acad. Sci. U.S.A. 75, 3268-3272. Saccone, C., and Kroon, A. M . 1976. “The Genetic Function of Mitochondria1 DNA.” NorthHolland Publ., Amsterdam. Sage, G. C . M. 1973. J. Agric. Sci. 81, 125-129. Sage, G. C. M., and Hobson, G. E. 1973. Euphatica 22, 61-69. Sager, R. 1954. Proc. Natl. Acad. Sci. U.S.A. 40, 356-363. Sager, R. 1977. Adv. Gene!. 19, 87-341. Sager, R., and Ishida, M. R. 1963. Proc. Nail. Acad. Sci. U . S . A . 50, 725-730. Sager, R., and Lane, D. 1972. Proc. Natl. Acad. Sci. U . S . A . 69, 2410-2413. Sagi, F., Barabas, Z . , and Kertesz, Z. 1976. In “Heterosis in Plant Breeding” (A. Janossy and F. G. H . Lupton, eds.), pp. 115-122. Elsevier, New York. Sakano, K., Kung, S. D.. and Wildman, S. G. 1974. Mol. Gen. Genet. 130, 91-97. Sanadi, D. R., Lam, K. W . , and Kurup, C. K. 1968. Proc. Narl. Acad. Sci. U.S.A. 61, 277-285.

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

193

Sarkissian, I. V . 1972. Z . Pfanzenz. 67, 53-64 Sarkissian, I . V . , and Huffaker. R. C. 1962. Proc. Narl. Acacl. Sci. U . S . A . 48, 735-743. Sarkissian, I . V . , and McDaniel, R. G. 1967. Proc. Nut/. Acad. Sci. U.S.A. 52, 1262-1266. Sarkissian, I . V . , and Srivastava, H. K . 1967. Generics 57, 843-850. Sarkissian, I . V . , and Srivastava, H. K . 1969. Proc. Nail. Acad. Sci. U.S.A. 63, 302-309. Sarkissian, I . V . , and Srivastava, H. K . 1971. Biochem. Genet. 5, 69-80. Sarkissian, I . V . , and Srivastava, H. K . 1973. In “Genes, Enzymes and Populations” (A. M. Srb. ed.), pp. 53-60. Plenum, New York. Sassen, M. M. A. 1964. Acta Bot. Neerl. 13, 175-181. Savchenko, N. I . , Belons, V . E., and Seraya, L. V. 1968. Cjltol. Genet. Kiev 2, 226-231. Scarascia, M. G. T. 1977. Gener. Agrar. 31, 31-142. Schatz, G., and Mason, T. L. 1974. Annu. Rev. Biochem. 43, 51-87. Schiff, J . A. 1975. In “Muter. Resp. Konf. Molod. Uchenyh. Dushanbe Donish pp. 63-91. Schlanger, G., and Sager, R. 1974. J. Cell B i d . 63, 301-302. Schlessinger, M. J . , and Leventhal, C. 1963. J . Mol. B i d . 7, 1-12. Schneiter, A. A., McDaniel, R. G., Dobrenz, A. K . , and Schonhorst. M. H. 1974. Crop Sci. 14, 82 I -824. Schneiter, A. A., McDaniel, R. G., Dobrenz, A. K . , and Schonhorst, M. H. 1976. Agron J . 68, 51 1-51 3. Schnepf, E., and Brown, R. M. 1971. In “Origin and Continuity of Cell Organelles” (J. Reinhert and H. Ursprung, eds.), pp. 299-322. Springer Publ., New York. Scholl, R. L. 1974. Crop. Sci. 14, 296-350. Schwartz, D. 1964. Proc. Narl. Acad. Sci. U . S . A . 51, 602-605. Schwartz, D. 1975. Genetics 207-212. Schwartz, D., and Dayhoff, M. 0. 1978. Science 199, 395-402. Schwartz, D., and Laughner, W. J . 1969. Science 166, 626-627. Scrowcroft, W . R. 1977. Adv. Agron. 29, 39-81. Semeniuk, P. 1976. J. Hered. 67, 61-62. Shull. G. H. 1908. Rep. Am. Breed. Assoc. 5 , 51-59. Shull, G . H. 1911. Am. Nut. 45, 234-252. Shull, G . H. 1952. In “Heterosis” (J. Gowen, ed.). Iowa State College Press, Ames. Shulyndin, A. F. 1972. Sel. Semenovod. 37, 19-25. Shulyndin, A. F . , and Maksimov, N. G . 1972. Ref. Zh. 10, 55-142. Siegel, A. 1975. In “Modification of the Information Content of Plant Cells” (R. Markham, D. R. Davies, D. A. Hopwood, and R. W. Home, eds.), pp. 15-26. North-Holland Publ., Amsterdam. Singh, R., Hubby, J . , and Lewontin, R. C. 1974. Proc. Natl. Acad. Sci. U.S.A. 71, 1808-1810. Singh, R., Lewontin, R. C., and Felton, A. 1976. Generics 84, 609-629. Sinha, S . K . , and Khanna, R. 1975. Adv. Agron. 27, 123-170. Sirks, M. J . 1938. Bot. Rev. 4, 113-131. Sisodia, N . S . , and McGinnis, R. C. 1970. Crop Sci. 10, 161-162. Slack, C. R., and Hatch, M. D. 1967. Biochem. J . 103, 660-665. Slater, M., and Schaechter, M. 1974. Bacreriol. Rev. 38, 199-221. Smillie, R. M . , Scott, N. S., and Bishop, D. 1973. In “The Biochemistry of Gene Expression in Higher Organisms” (J. K. Pollak and J. W. Lee, eds.), pp. 479-503. Reidel, Amsterdam. Smillie, R. M., Nielsen, N . C . , Hennington, K . W . , and von Wettstein, D. 1977. Austr. J. Plant Physiol. 4, 439-449. Smirnova, L. F., Palilava, A. H., and Gardzievich, R. P. 1975. Ref. Zh. 7, 55-139. Smith, H. H. 1972. f r o g . Exp. Tumor. Res. 15, 138-164. Smith, H. H., Kao, K. N., and Cambatti, N. C. 1976. J. Hered. 67, 123-125.

1 94

H. K. SRIVASTAVA

Soler, L. M. G., and Jouve, B. M. 1971. J . Genet. Losoespanholar 8, 68. Spencer, D., and Whitfield, P. R. 1969. Arch. Biochem. Biophys. 132, 477-478. Srivastava, H. K . 1972. Indian. J . Genet. Plant Breed. 32, 215-228. Srivastava, H. K. 1974. Indian J . Exp. Eiol. 12, 79-81. Srivastava, H. K. 1975. Int. Congr. Sahrao, 2nd. New Delhi Sect. 12 (b), 30-39. Srivastava, H. K. 1980. Theor. Appl. Genet. 56, 25-29. Srivastava, H. K. 1981. “Monographs on Theoretical and Applied Genetics, Vol. 7: Heterosis” (R.Frankel, ed.) (in press). Srivastava, H. K . , and Balyan, H. S . 1977. J . Hered. 68, 338-340. Srivastava, H. K., and Sarkissian, I. V. 1969. Genetics 61, s 57. (Abstr.) Srivastava. H. K . , and Sarkissian, I. V . 1970. Genetics 66, 497-503. Srivastava, H. K.. and Sarkissian, I. V. 1972. Eiochem. Genet. 7 , 347-354. Srivastava, H. K.. Sarkissian, 1. V., and Shands, H. L. 1969. Genetics 63, 61 1-618. Stainer, R. Y . 1974. I n “Evolution in Microbial World,” pp. 219-240. Cambridge Univ. Press, London and New York. Starnes, W. J.. and Hadley, H. H. 1965. Crop Sci. 5 , 9-1 1 , Stem, C. 1948. Genetics 33, 215-219. Steward, F. C., and Krikorian, A. D. 1971. “Plants, Chemicals, and Growth” Academic Press, New York. Stringer, C. D., and Hartman, F. C. 1978. Eioph.vs. Res. Corninun. 80, 1043-1048. Strobaek, S., Gibbons, G. C., Haslett, B., Boulter, D., and Wildman, S . G. 1976. Corlsberg Res. Coinmun. 41, 335-343. Stubbe, W. 1957. Z. Vererbungsl. 90, 188-298. Stubbe, W. 1964. Genetics 35, 28-33. Suernoto, H. 1973. Proc. I n / . Wheat Genet. Synip.. 4th pp. 109- 113. Sugiyam, T., and Akazawa. T. 1970. Biochemistry 9, 4499-4454. Tanof. K . D. 1975. Annu. Rev. Genet. 9, 355-385. Taylor, F. R. J . 1974. Tuson 23, 229-258. Tewari, K. K . , and Wildman, S. G . 1968. Proc. Nail. Acad. Sci. U . S . A . 59, 569-576. Tewari, K. K., and Wildman, S. G. 1970. I I I “Control of Organelle Development” (P. L. Miller, ed.), pp. 147-179. Academic Press, New York. Thomas, R. L., and Grafius, 1. E. 1976. I n “Heterosis in Plant Breeding” (A. Janossy and F. G. H. Lupton, eds.). Elsevier, Amsterdam. Thomberger, J . P. 1975. Annu. Rev. Plant Physiol. 26, 127-158. Tilney-Bassett, R. A. F. 1973. Heredity 30, 1-13. Tilney-Bassett. R. A. E. 1975. I n “Genetics and Biogenesis of Mitochondria and Chloroplasts” (C. W. Birky, P. S . Perlman, and T. J . Byers, eds.), pp. 268-308. Ohio State Univ. Press, Columbus. Tolbert, N . E. 1971. Annu. Rev. Plant Physiol. 22, 45-74. Torres, M. A., and Hart, G. E. 1976. Eiochein. Genet. 14, 595-609. Treat, L. 1978. Genetics 88, 100-101. (Abstr.) Treharne, K. J . 1972. I n “Crop Processes in Controlled Environments” (A. R. Ress, K . E. Cockshull. D. W .Hand, and G. Hurd, eds.), pp. 285-303. Academic Press, New York. Tsunewaki, K., and Endo, T. 1973. Proc. Int. Wheat Genet. Symp., 4th pp. 391-397. Turbin. N . V., Atranshenok, N. V., Palilova, A. N., and Lyulkina, E. I . 1968. Dokl. Akad. Nauk. USSR 182, 699-701. Turbin, N . V., Palilova, A. N., Atranshenok, N . V., and Lyulkina, E. I. 1970. Dokl. Akad. Nauk USSR 191, 1167-1168. Tzagoloff, A., Foury, F.. and Akai, A. 1976. Mol. Geri. Genet. 149, 33-42.

INTERGENOMIC INTERACTION, HETEROSIS, AND CROP YIELD

195

Usmanov, P. D., Abdullaev, K . A., Pinkhasov, Y. I . , and Bikasiyam, G. R. 1975. Genetika 11, 22-29. Uzzel, T., and Spolsky, C. 1974. Am. Sci. 62, 334-343. van Went, J. L. 1974. I n “Fertilization i n Higher Plants” (H. F. Linskens, ed.), pp. 81-88. North-Holland Publ., Amsterdam. van Winkle-Swift, K. P. 1976. Ph.D. Thesis, Duke University, Durham, North Carolina. van Winkle-Swift, K . P. 1978. Genetics 88, s 103. (Abstr.) Vary, M. J., Stewart, P. R . . and Linnane, A . L. 1970. Arch. Eiochem. Eiophys. 141, 430-434. Vecher, A. S . . Krylova, H. Y . , and Bulko, 0. P. 1975. Dokl. Akacl. Nauk USSR 19, 64-85. Vecher, A. S., Krylova. H . Y . , and Bulko, 0. P. 1977. Dokl. Akad. Nauk USSR 20, 558-560. Vershinin, A. V., Sokolov, V. A,, and Shumnyi, V. K. 1976. Genetika 12, 52-58. von Wettstein, D., and Kristiansen. K. 1973. Earley Gener. News Lett. 3, 113-117. von Wettstein, D., Henningsen, K . W., Boyton. I . E.. Kannangara, G. C., and Nielsen, 0. F. 1971. In “Autonomy and Biogenesis of Mitochondria and Chloroplasts“ (N. K. Boardman, A. W. Linnane, and R. M. Smillie, eds.), pp. 205-223. North-Holland Publ., Amsterdam. von Wettstein, D., Kahn, A,. Nielsen, 0. F., and Cough, S. 1974. Science 184, 800-802. von Wettstein, D., Poulsen, C., and Holder, A. A. 1978. Theor. Appl. Genef. 52, 193-197. Wagner, R . P. 1969. Science 163, 1026-1031. Wagner, R. P. 1972. J. Anirn. Sci. 35, 1280-1287. Wallace, D. C., Pollack, Y . , Bunn, C. L., and Eisenstandt, J . M. 1976. I n Vitro 12, 758-776. Walks, S. B. 1971. I n “Structure and Function of Chloroplasts” (M. Gibbs, ed.), pp. 51-68. Springer-Verlag. Berlin and New York. Warmke, H. E., and Lee, S. L. J. 1977. J. Hered. 68, 213-222. Watrud, L. S., Baldwin, J . K.,Miller, R.J., and Koeppe, D. E. 1974. Plant Physiol. Ann. Suppl. p. 219. Watson, C. V . , Nath, J . , and Nanda, D. 1977. Eiochem. Genet. 15, 1 113-1 124. Whaley, W. G . 1952. I n “Heterosis” (J. Gowen, ed.), pp. 98-1 13 Iowa State College Press, Ames. Wildman, S. G.. Chen, K . , Gray, J. C., Kung, S. D., Kwanyven, P., and Sakano, K . 1975. I n “Genetics and Biogenesis of Mitochondria and Chloroplasts’’ (C. W. Birky, P. S. Perlrnan, and T. J . Byers. eds.). Ohio State Univ. Press, Columbus. Wildner, G. F., and Henkel, 1. 1976. Eiochem. Eiophys. Res. Commun. 69, 268-275. Wilkie. D. 1964. “The Cytoplasm in Heredity.” Wiley, New York. Williams, W. 1959. Nature (London) 184, 257-530. Williamson, D. H., Johnston, L. H., Richmond, K . M. V . , and Game, J. C. 1978. I n “Mitochondria.” De Gruyter, Berlin. Wilson, D. 1972. J. Exp. But. 23, 517-525. Wilson, D., and Cooper, J. P. 1969. Heredity 24, 633-249. Wolf, G . . and Rempan, J. 1977 Nature (London) 265, 470-471. Zelitch, I . 1973. Plant Physiol. 51, 299-305. Zelitch. I. 1975. Science 188, 626-633. Zelitch, I . , and Day, P. R. 1973. Plant Physiol. 52, 33-37. Zelitch, I., Oliver, D. J . , and Berlyn, M. B. 1977. I n “Biological Solar Energy Conservation” (A. Mitsui, S. Miyachi, A. S . Pietro, and S. Tamura, eds.). Academic Press, New York. Zoble, R. G . , Fishbeck. F. K., and Laizko, E. 1972. Plant Physiol. 50, 790-791.

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY. VOL. 34

THE DILUTION EFFECT IN PLANT NUTRITION STUDIES W. M. Jarrell and R. 6. Beverly Department of Soil and Environmental Sciences, University of California-Riverside, Riverside, California

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........ . . . . 197 11. System for Expressing Results ...................... . . . . . . . . . . . . . . . . . 199 111. Mechanisms .................................................. 200

E. F. G. H.

Water . . . . . . . . . . . . . . . ............... Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Oxygen . . . . . . . . . . . .................................... Time . . . . . . . . . . . . . ....................................

VI. Concentration Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....................... A. Nutrient Toxicity and Synergism . . . . . . . B . Heavy Metal Toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Soil Solution Salinity.. D. Pathogens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Practical Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII. Summary and Future Research Needs . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

215

217

219 219 22 1 222

I. INTRODUCTION The “dilution effect” and its inverse, the “concentration effect,” have been referred to in numerous studies of plant nutrition and soil fertility to explain 197

Copyright @ 1981 by Academic Press. Inc. All rights of reproduction in any form reserved.

ISBN 0-12403734-7

198

W.M. JARRELL AND R. B. BEVERLY

results that arise when the concentration’ of an element in plant tissue is decreased or increased due to a change in environmental conditions. “Environmental conditions,” in this context, include changes in the soil environment due to the addition of inorganic and organic materials and water to soil, as well as temperature and light; the application of living organisms such as rhizobia and mycorrhizal fungi; and the inclusion of toxic materials such as heavy metals. An additional instance in which “dilution” is often alluded to is the change in nutrient concentration as a function of time. It has often been found, especially for N , P, and K , that young plants contain higher concentrations than do older plants (Loehwing, 1953). A survey of the literature suggested that no extensive articles had been written on the dilution effect. Hughes et al. (1978) and Barea et a/. (1980) both found that P additions decreased the concentration of N in plants, although because of increased yield greater total N accumulation was noted. Numerous articles have alluded to it, either to explain results or to describe basic plant behavior. For example, many articles in the Revised Edition of Soil Testing and Plant Analysis (Walsh and Beaton, 1973) refer to this phenomenon. Relevant statements from several of these articles follow: It is not unusual to find that the addition of certain nutrients reduces the amounts of other nutrients in the plant (Aldrich, 1973, p. 215). When nutrients such as N, P, or K are added, it is difficult to predict whether or not the concentration of a given element in the plant part will increase. remain unchanged, or be decreased. Much depends upon the influence of the other nutrients available in the soil and the direct and indirect effects the applied nutrient has on increased growth and yield (Munson and Nelson, 1973, p. 236). Analysis of plant tissue usually reveals only one deficiency at a time. A second nutrient, or even a third nutrient, may be i n short supply but, due to reduced growth caused by the primary nutrient deficiency, all other nutrients will accumulate in the tissue (Ulrich and Hills, 1973, p. 286). The problem area plants appeared to be N-deficient. Normal plants had few nodules on the roots and the plant tissue N levels were very low. Concentrations of certain elements in the normal plants were lower than corresponding nutrient levels in the problem plants. This was, in part, the result of dilution resulting from more rapid plant growth (Small and Ohlrogge, 1973, p. 324). When the total plant uptake (concentration times dry matter) was plotted versus time, the curves

‘Terminology relating to the quantities of nutrients in plants often relies on ambiguous terms (Leaf, 1973). In particular, “uptake” and “content” are not very precise terms. They may refer to either the total amount in the plant or to the concentration in the tissue. Throughout the course of our discussion, the term “conc&tration” will denote a mass or molar ratio such as milligrams or moles per kilogram. Further, “total accumulation” will be used when refemng to the total quantity of nutrient in the plant, either the whole plant, the above-ground portion of the plant, or some tissue such as leaves or stems. Careful use of terms should minimize at least some points of confusion. “Uptake” will refer to the process of elemental movement into and within plants and not to quantity-based terms such as concentration or total accumulation.

DILUTION EFFECT IN PLANT NUTRITION STUDIES

199

were fairly smooth. Therefore, most of the variations in element concentration were due to concentration or dilution associated primarily with change in dry matter production (Jones and Eck, 1973, p. 356). The basic principle of the use of plant analysis is that the chemical composition of the plant reflects the nutrient supply in relation to growth. We must, however, recognize that the chemical composition of any plant is a “result” of the interaction of nutrient supply and plant growth. Any factor that limits growth, be it light, moisture, temperature, or some nutrient, may cause other nutrients to accumulate in the plant (Martin and Matocha, 1973, p. 394). Samples taken at early stages of growth have high concentrations of N, P, K , and S. The concentration of such nutrients declines as the plant matures and approaches the bloom stage because of the dilution with carbohydrates and other structural solids (Martin and Matocha, 1973, p. 397). Generally, the “dilution effect’’ caused by the rapid growth of this grass will reduce the instances when “luxury consumption” of K will occur (Martin and Matocha, 1973, p. 412). Generally, if a particular element that is limiting growth is added in a fertilizer treatment and a subsequent growth response occurs, an increase in both that element concentration and content in tissues occurs. However, if a dramatic growth response follows treatment, it is possible that the addition of the limiting element may result in a lower concentration of that element in tissues. . , . On the other hand, the levels of other elements (nonlimiting growth) in the tissues may decrease in concentration due to dilution effects but increase in content due to increased biomass (Leaf, 1973, p. 445).

These citations represent to one degree or another the common conception of the “dilution” and “concentration” effects held by those who work with plant tissue sampling. In most instances, reference is made to changes induced by supplying a deficient nutrient to the plant; in one case, by removal of an environmental restriction such as nonoptimal light, temperature, or water. Responses due to microorganisms such as rhizobia or mycorrhizae have not generally been considered in the past. As noted, the concentration of nutrient in a plant tissue is a single-point resultant of plant history, in particular the integration of two dynamic processes, nutrient uptake and transport and dry matter accumulation (Lundegardh, 1966; Martin and Matocha, 1973). Since the concentration is a resultant of the quantitative manner in which growth and accumulation vary, it would seem reasonable that a complete consideration of the crop’s nutrient status could be separated into considerations of these fundamental processes. This article will analyze “dilution” and “concentration” effects in tissue on the basis of the relative rates of elemental uptake and dry matter accumulation.

II. SYSTEM FOR EXPRESSING RESULTS Because the concentration in tissue is a ratio of two quantities, we believe it would clarify much of the following discussion if absolute responses were re-

200

W. M. JARRELL AND R. B. BEVERLY

ported along with relative (concentration) responses. We propose, then, that a sequence of three symbols represent response. The first position would indicate change in total quantity (mass or moles) of the element per plant, the second the change in total yield, dry matter, fresh weight, or other relevant measure of plant growth, and the third the change in concentration. This method produces a series of 11 potential response patterns (Table I). Cases 3 and 7 are classical dilution effects, although the interpretation of causes of these results are quite different. Cases 2, 6, and 10 are unchanged concentrations, with case 6 being a trivial (no change in any parameter) response. Cases 8 and 11 are antagonisms. Case 4 may be called luxury consumption and is most frequently encountered when the element of concern is part of the treatment. Case 9 is definitely a concentrating effect, while case 5 is another form of the concentration effect, but not often delineated separately. There are data in the literature that describe results illustrating each of these response patterns. Although this table systematizes the discussion, it would be best if this would assist in the assignment of mechanisms to explain each of these results. The following section summarizes the types of mechanisms that may act singly or together to produce the observed response patterns.

111. MECHANISMS In this section a summary is presented of the physical, chemical, and biological reasons that might account for the observed changes in the rate of nutrient uptake and the rate of dry matter accumulation as functions of time. A . NUTRIENT ACCUMULATION

1. Factors leading to greater uptake of the nutrients in question: (1) Higher concentration in the external solution (e.g., Epstein and Hagen, 1952). (2) Increased root growth, e.g., better capability to extract available nutrients (Drew, 1975). (3) Faster rate of movement to roots through mass flow with transpirational water or through diffusion (Nye and Tinker, 1977). (4) Increased root activity, e.g., more photosynthate for energy (Pearson and Steer, 1977). (5) Greater transport to the tops (Pitman, 1975). (6) Greater demand for the element.

20 1

DILUTION EFFECT IN PLANT NUTRITION STUDIES

Table I General Representation of Changes in Total Elemental Accumulation, Dry Matter Yield, and Concentration as Affected by Imposed Treatments Change in total element accumulation

Case No.

Change in yield

Change in concentration

1

t

2 3

0

1

t

4 5

t t

0

6

0

0

0

I

.1

t

1

8

.1

0

9

1

1

.1 t

10

1 1

1

t

Comments Synergism “Dilution effect” Synergism “Concentration effect No response “Dilution effect” Antagonism “Concentration effect” ”

II

(7)

2.

0

.1

Antagonism

Positive effect on uptake mechanisms, e.g., stabilized membranes (e.g., Van Stevenick, 1965).

Factors leading to no change in nutrient uptake:

No direct interaction between the modified factor and the nutrient element in question. ( 2 ) Canceling interactions.

(1)

3 . Factors leading to negative interactions between the element and the factor modified in the external environment: (1) (2) (3) (4)

(5)

(6) (7)

Lower concentration in the external solution (coprecipitation, pH, redox potential changes). Direct competitive and noncompetitive inhibition of uptake (e.g., Randall and Vose, 1963). Decreased root growth. Decreased rate of movement to the root either by a reduction in the rate of diffusion or a reduction in the contribution of mass flow to uptake. Smaller demand for nutrients by the plant. Membrane breakdown. Reduction in transport of nutrients to the plant tops.

W . M . JARRELL AND R. B. BEVERLY

202

(8)

Accelerated rates of metabolic processes due to increased concentrations of inorganic constituents. B. DRYMATTER ACCUMULATION

1.

Factors leading to increased dry matter accumulation:

(1) (2) (3)

(4) (5) (6) (7)

Enhanced photosynthesis (e.g., Gaastra, 1962; Terry and Ulrich, 1973). Lowered respiration (e.g., Terry and Ulrich, 1973). Improved translocation of photosynthate to point of incorporation (Wardlaw, 1980). Greater turgor potential (e.g., Hsiao, 1973). Less disease-, pest-, or temperature-related decreases in yield potential (Ellingboe, 1980). Lower rate of senescence. Better hormonal balance (Kriedemann et al., 1976).

2.

Factors leading to no change in dry matter accumulation:

3.

(1) No direct effect on plant metabolism. (2) Compensating effect (e.g., more water coupled with great salinity). Factors leading to decrease in dry matter accumulation: Blockage of metabolic pathway by “toxic” concentrations of an element. Decrease in photosynthesis. Increase in respiration (Zelitch, 197 1). Poor photosynthate translocation. Water limiting-lowered turgor pressure. Hormonal imbalance. Accelerated senescence, hastened maturity, decreased growth period. Greater disease-, pest-, or temperature-related loss of yield potential.

It may be very important to consider how combinations of the above mechanisms may cause changes in concentrations observed in tissue.

IV. TREATMENTS In the past the dilution effect has primarily been associated with the application of fertilizerhutrient materials. For the purposes of this review, several general

DILUTION EFFECT IN PLANT NUTRITION STUDIES

203

categories of treatments are considered, including the following: ( a ) Single nutrient/fertilizer/salt additions: urea, potassium nitrate, calcium phosphate, potassium chloride, metal chelates. ( b ) Multiple material additions: saline water, sewage sludge, fly ash. (c) pH modifications: lime, elemental sulfur, sulfuric acid additions to soil. ( d ) Organisms: Rhizobium, mycorrhizae, plant pathogens. ( e ) Water-related variables: irrigation frequency, drought, relative humidity. cf)Temperature: mulching, season, greenhouse control. ( g ) Light: season, shading, greenhouse control.

Because most work has been done with a, b, and to some extent c, these categories will be developed in the greatest detail. Where data are available, d, e, and f will be discussed as well. Relatively little data of the form required are available for g. It has become trite among workers in the field to say that interactions between nutrients themselves and between nutrients and their environment are complex. There is no readily discernible means by which all factors may be simultaneously taken into account in the manner in which they affect plant growth. It would be highly involved and ultimately not very fruitful to consider every article in which a dilution or concentration effect has been measured. The approach taken within this article will be rather to summarize by example the types of situations in which these effects have been observed. Decisions about the relative importance of specific mechanisms must be left to the judgment of the experimenter. The effect of a chemical or environmental treatment on the concentration of a nutrient in the plant will be considered in two categories, noninteractive and interactive. Noninteractive effects arise because the plant grows larger or smaller than the control plant, i.e., these differences are due primarily to changes in the biomass of the individual plant. Interactive effects may occur for a variety of reasons but deal with direct effect on the nutrient uptake mechanisms, without considerations of plant growth patterns or rates. Of these two, the interactive effects have been most carefully studied in soil-plant nutrition work, and apparently affect the following: ( a ) Solubility of the nutrient in the soil solution. (b) Movement by diffusion or mass flow to the root, e.g., soil water content (Nye and Tinker, 1977). (c) Active or passive movement of the ion through cell membranes, e.g., competition for carrier sites. ( d ) Changes in driving force, e.g. catiodanion balance. (e) Translocation from the root to the shoot and mobility between plant parts, e.g., substitutions in various portions of the transport pathway.

Noninteractive effects encompass the following:

204

W . M . JARRELL A N D R. B. BEVERLY

( a ) Direct dilution because of greater biomass. (b) Extension or shortening of roots, e.g., root length density. ( c ) Availability of energy to the root for uptake processes, e.g., rate of net photosynthesis, carbon assimilation, and translocation from top to roots. Intermediate sorts of effects in which there was no clear separation between interactive and noninteractive types would include changes in the hormonal balance of the plant.

V. DILUTION EFFECTS The situations considered in this section are those in which a deficiency of some type (nutrient, water, oxygen) is overcome by a change in the plant’s chemical, physical, or biological environment. The concentrations of other elements in the leaf tissue are then measured as the plant dry matter increases due to treatment. Such changes help indicate how the plant is managing the supply of nutrients which are available to it, when a greater demand is placed upon this supply by a larger plant. The discussions in this article will emphasize primarily research contributions in which total dry matter data are reported so that the dynamic effects of plant growth and nutrient uptake can be separated out in part at least and placed within the framework of the “types” presented in Table I. However, where it is felt to be important, the information from experiments dealing with changes in concentration only will also be discussed. A. DILUTION OF NUTRIENT APPLIED

In this situation, a nutrient element is applied to a crop, the yield of the crop increases, but upon chemical analysis the average concentration of the element in some or all plant tissue is lower than in a deficient control plant (Piper, 1942; Steenbjerg, 1951; Steenberg and Jacobsen, 1963). This result is probably most surprising because of the implicit assumption, mentioned earlier in this article, that the adequacy of the plant’s supply of a given nutrient is directly related to the tissue concentration of that nutrient (Lundegardh, 1966; Martin and Matocha, 1973). However, as Bates (1971) has summarized (see Table 11), there are several explanations, consistent with physiological considerations, that may account for this. For example, plants may lose the potential for growth or response under acute nutrient stress and be unable to respond even though they have accumulated “adequate concentrations of the limiting element in their tissue (Hiatt and Massey, 1958). ”

205

DILUTION EFFECT IN PLANT NUTRITION STUDIES Table I1 Examples of C-Shaped Yield-Nutrient Concentration Curves" Nutrient

Culture

cu

Solution

cu Mn P

Soil Soil Soil

cu cu

Soil Soil

Tissue Oatsb (mature) (Avena sativa) Oat (whole plant) Oat (whole plant) Barley (straw) (Hordeum vulgare) Barley (straw) Barley (grain)

-

-

Mn

Sand

Zn

Field

Zn Mg

Field Soil

Tomato (lower stems) (Lycopersicum esculentum) Corn (whole plant) (Zea mays) Corn (whole plant) Oat (straw)

Zn

Solution

Sugar beet (young blades)'

Zn B

Solution Solution

so,-s

Soil

S' Zn

Solution Solution

P

Reference Piper (1 942) Steenbjerg (1945) Poulson (1950) Steenbjerg (1951) Prevot and Ollagnier ( 1956)

(Beta vulgaris) Sugar beet (mature petioles) Birch (roots)" (Betula spp.) Grass (leaves) (Lolium multiflorum) Ryegrass (stems) Alfalfa (stems)" (Medicago sativa)

Hewitt (1956) Hiatt and Massey (1958)

Jakobsen and Steenberg ( 1 964) R o d and Ulrich ( 1964)

Ingestad (1954) Saalbach and Jude1 ( 1966)

Ulrich (1968) Lo and Reisenauer ( 1968)

From Bates ( 197 I ) bRye in the same experiment did not give a C-shaped curve 'Mature blades did not give a C-shaped curve. Leaves did not give a C-shaped curve. ' A C-shaped curve was obtained with organic or total S but not with SO,-S. (I

Whatever the actual physiological cause, when the growth-limiting element is supplied, the relative rate of dry matter accumulation increases more rapidly than the rate of nutrient accumulation, resulting in lower final concentrations in treated plants. In all cases, even though the concentration of the element in the tissue has decreased, the total accumulation, as calculated by the product of concentration and dry matter yield, has increased significantly. Thus we represent the behavior as TtL, to indicate that plant growth has proceeded more rapidly than nutrient accumulation. Piper (1942) and Steenbjerg (1951) both noted this effect on cereals with Cu. Steenbjerg found 16.6 mg Cu/kg tissue in untreated control plants, while on a

206

W . M . JARRELL AND R. B . BEVERLY

+

whole plant basis (straw grain) average concentrations of Cu-treated plants were 6.4-14.3 mg Cu/kg. Untreated grain had 3.2 mg Cu/kg, while treated grains ranged from 0.7 to 5.4, mean k SD = 4.1 k 1.7. Copper concentrations of straw, on the other hand, ranged from 8.3 to 14.4, with mean k SD = 1 1 .O 2 2.2. In every case total uptake was greater than controls. So it is evident that higher concentrations could be achieved in grain with treatment, while in straw, or with all. The harvest index (mass ratio of grain yield to straw yield) for control plants was 0.U8.3 = 0.012, while for the best yield (1.02 g CuSO,.SH,O/pot) it was 54.4/47.3 = 1.15, nearly 100 times as great (Fig. I ) . Thus the grain provided a much greater sink for translocatable Cu in treated than in untreated plants. By stunting the plant early during its growth period, there may be a synergistic negative effect due to the accumulation of elements to toxic concentrations in the tissue. This may even include the element in question that was initially deficient (see Section VI). Gupta et al. (1976) found that a seed treatment with Mo may have significantly increased yield, of both onions and cauliflower, with a concomitant drop

5LT

+ m \

za

I

1.25.-

I

0

0

0 0

I .oo.-

0 0

0

0

LT

c3

0 . 7 5 .-

X

0

0 0

W

D Z

0

0

0

GRAIN STRAW

- 0.50.-

I-

m W

> a r

0.25 0

0 0

0

Cu IN S T R A W ( m g kg-') FIG. 1. Relationship between harvest index and [Cu] in straw (0) or grain (0). From Steenbjerg (1951).

DILUTION EFFECT IN PLANT NUTRITION STUDIES

207

in Mo concentrations of leaf tissues. In both cases there were small but probably not significant decreases in total uptake as measured by [Mo] x yield, but cauliflower heads and onion bulbs were not considered in this case, and total uptake of treated plants was probably greater. Some of the data of Thomas and Mather (1979) on Fe application to sorghum suggest a dilution effect both in response to N , P, and K and to Fe applications. In the first crop, application of Fe with NPK increased yields substantially relative to -Fe treatments, but the tissue concentration of Fe dropped from 40 to 30 mg Fe/kg in leaves. With Fe it is probably especially important to interpret total analysis data skeptically. Active Fe may be some variable fraction of the total Fe (Katyal and Sharma, 1980). Apparently the mineral or biochemical environment inside the cell may be more critical in determining the adequacy of Fe supply than is the total analytical concentration. The dilution effect as shown in these examples is fundamentally interesting and may help in understanding biological problems. In a practical sense it is probably less significant in established agricultural settings. An exception to this case may be where virgin land is being converted to agriculture, or where new crops are introduced. But generally, few plants are as extremely deficient as the controls in these examples. In addition, since a yield response is obtained when these results are produced, there is little question about the limiting factor. However, where one is attempting to diagnose a given nutritional deficiency, or where a response is seen to the application of a complex (multielement) material, these problems become more real. B. DILUTION OF OTHER ELEMENTS

This is by far the most common instance in which the “dilution effect” has been invoked to explain results. In most cases one analyzes a wide spectrum of elements in the plant after a change in the application rate of one or more other elements. In this situation, the crop is responding to the limiting element. Dry matter production increases. If uptake of some other element proceeds more slowly than dry matter accumulation, concentration will decrease. The concentration of this other nutrient may decrease below levels of adequacy (however defined) and ultimately produce a deficient plant. It would be highly advantageous to be able to predict the second, third, etc. most limiting element simply by analyzing a single sample. However, much work needs to be done in order to predict response in a reasonable fashion. Here, as in other sections, it is very important to distinguish between interactive and noninteractive effects. If the dilution in concentration is classic, that is, only because more dry matter is produced, the relationship may be termed

208

W . M. JARRELL AND R. B. BEVERLY

noninteractive. However, if the elements interact directly at some uptake site or in the soil, such that uptake and/or translocation is partially inhibited by the added nutrient, then a direct interaction would be involved. It is often very difficult, if not impossible, to separate these two mechanisms when a plant yield response is obtained due to treatment. When yield is not changing, the effects may be separated a little more easily, although it is still not completely clear what the mechanism would be. A number of examples will be presented that demonstrate the range of observations of this type that have been made in the past. Goh et al. (1979) measured levels of a wide variety of elements in ryegrass after treatment with N and S fertilizers, with the primary goal of looking at cationhion balance. At one level of N, where yield increases were obtained, added S tended to decrease N concentrations in the plant. Sulfur additions decreased both concentration and total accumulation of Se in alfalfa at two field sites, indicating that a direct interaction may have occurred (Westerman and Robbins, 1974). Where yields did not increase with S treatment, there was a decrease in Se total accumulation and Se concentration (&O&). Where yield responses were recorded, total uptake tended to increase while concentraa more common observation of the dilution effect. The fact tion decreased (tf&), that total accumulation decreased where yields were unchanged or slightly increased suggests a direct interaction, but this would not be clearly indicated by results where the tt&pattern occurred. One of the most common dilution effects or interactions observations has been that involved with P X Zn interaction, frequently expressed as “P-induced Zn deficiency” (Thorne, 1957). Upon addition of P fertilizers to soils the concentration of Zn in tissue has often been observed to decrease. In terms of the symtypes of dilution effects have been bolism used in this article, both fT& and observed. Burleson ef al. (1961), for instance, found decreases in total Zn accumulation by beans when P fertilizer was applied, even though total yield changed little (.lo&).Where growth responses were observed with Zn and P additions, added P decreased both total Zn accumulation and Zn concentrations. Safaya and Singh (1977) found that as P was increased slightly at low Zn plant yield increased and Zn concentration decreased slightly (&ti). As P was increased further, yield and total Zn uptake decreased again, but concentration tended to increase (i&f). At high Zn levels the first increment of P produced a 7T.l response, typical dilution effect response, but additional P caused a change to &lo.Such results suggest that at low Zn the added P was significantly decreasing the availability of P, but at higher concentrations this was not at all clear. Schultz et al. (1979) found that application of K produced a Mg response of OT& in alfalfa and of fT0 in white clover. This would suggest that perhaps the

DILUTION EFFECT IN PLANT NUTRITION STUDIES

209

clover and alfalfa behave differently in their K-Mg relations, with the clover able to maintain its Mg uptake much better where K is applied. Hughes et al. (1979) found that application of P to nonmycorrhizal red raspberries produced a dilution of N , K , Ca, Mg, Cu, B , and Zn in tissue, but total accumulation increased (?ti).Manganese concentration remained constant (?TO). In mycorrhizal plants, dilution with added P only occurred for N, K , Ca, Mg, and B, with other nutrients showing no significant change. In summary, there are numerous situations where an increase in dry matter accumulation in response to the application of a nutrient element is accompanied by a decrease in the concentration of other elements within the plant. In some instances it is possible to separate out interactive and noninteractive types of effects. However, the background levels of nutrients present are very significant in determining how plants respond. C . CHANGE IN pH

The favorable effects of optimizing pH on plant growth have been welldocumented. In addition, there is some information available on the effects of pH on nutrient uptake by crops. The pH of a soil can apparently affect both the concentration (solubility) of nutrients in the soil solution and the uptake of nutrients from solutions of constant ionic concentration. Since the effects of pH on plant growth are numerous, it is usually difficult to separate out those due to the increased availability of a single nutrient. However, in some instances improved growth may be due primarily to the increased availability of a single nutrient. Experiments conducted in soil generally do not allow one to discriminate between increased solubility and an increased ability of the plant to absorb ions. Carefully conducted solution culture experiments allow one to examine plant behavior where pH but not nutrient ion activity is varied. With tomatoes grown in the greenhouse (Jones and Fox, 1978), raising soil pH from 5.1 to 6.3 significantly increased yield and total Mn accumulation but decreased Mn concentration (?TJ). Above pH 6.3, both total Mn accumulation and Mn concentration decreased as pH increased (404). With A1 a precipitous drop in total accumulation was observed as pH increased from 5.1 to 5.4 (JOJ); total accumulation was roughly constant until pH 6.3 was exceeded, after which the pattern occurred. In flowing solution culture it was found that over the range of pH that increased plant growth rates, in nearly all cases increased pH increased both total accumulation and concentration of the macronutrients N , P, K , Ca, Mg, and S (Islam et al., 1980). Only for S were concentrations diluted or unchanged. For

210

W . M. JARRELL AND R . B. BEVERLY

micronutrients total accumulation increased in all cases, but Fe and B concentrations especially were decreased in several instances. Apparently nutrient uptake is more severely impaired by high proton activity than is dry matter accumulation in nearly all cases. Phosphorus and molybdenum are generally taken up in larger quantities from neutral than from acid soils. In some cases greater nutrient availability should contribute to the increase in growth effected by pH modifications. Decreased toxicity may have the same effect. Wherever possible, the “secondary” effects of pH should be separated from primary effects. D. SYMBIOTIC MICROORGANISMS

There are two important classes of microorganisms whose presence is well known to improve plant growth: nitrogen-fixing root nodule bacteria and actinomycetes, and mycorrhizal fungi. In terms of concentrations of a wide range of elements in tissue, the effects of mycorrhizal fungi have been much more thoroughly examined than the effects of nitrogen-fixing organisms. The primary mode of growth stimulation attributable to mycorrhizal fungi has been improved accumulation of P in above-ground tissue. Concentrations of nearly every element have been measured in an attempt to correlate the presence of mycorrhizae to increased concentrations of many nutrients in plants. The current emphasis has been placed on P and to an extent the micronutrients Zn and Cu (Gerdemann, 1968; Mosse, 1973; Sanders et al., 1975). Other elements have occasionally been found to increase, decrease, or remain at constant concentration after treatment with mycorrhizal fungi. These observations bring to mind questions about the types of stimulatory or inhibitory effects mycorrhizae have on elemental uptake and dry matter accumulation. Because the system is complex, it is difficult to sort out the mechanisms responsible for observed responses. One cannot limit them to a single element as can be done with the application of chemical salts. The mechanism by which mycorrhizal fungi improve rates of nutrient uptake appears to be the reduction in length of the diffusion pathway for ions of low solubility in soil (Nye and Tinker, 1977; Russell, 1977). This appears to be especially true for P, Cu, and Zn. However, there are other nutrients for which diffusion may be a significant factor in determining availability. Potassium uptake may be predominantly affected by mass flow or by diffusion, depending upon the solution concentrations (Russell, 1977). The diffusion coefficient of K in soil is somewhat greater than that of P, but still frequently limits uptake (Drew et al., 1969; Drew and Nye, 1970). In an extremely good article that is relevant for any discussion of the dilution

DlLUTlON EFFECT IN PLANT NUTRITION STUDIES

21 1

effect, Powell (1 975) investigated the dilemma that other researchers had encountered with the effect of mycorrhizal fungi on K uptake. He related that numerous investigators had found lower K concentrations in mycorrhizal than in nonmycorrhizal plants (Gerdemann, 1964; Holevas, 1966; Ross, 1971; Deal et al., 1972; Kleinschmidt and Gerdemann, 1972). Others had found increases (Mosse, 1957; Baylis, 1959), while in some cases no change was encountered. In every instance total accumulation of K was greater in mycorrhizal plants. The response of K in plants to mycorrhizal fungi infection may thus be expressed as ???, ?TO, or ?ti, the latter a definite dilution effect. Powell grew Griselinia littoralis plants under a high- and a low-K regime with and without mycorrhizae. At low K concentrations (K availability very limiting) plant growth increased 42% and total accumulation increased 23%, but concentration decreased (?Ti). At high K, no significant differences were noted in either growth, total K uptake, or K concentration. As Powell concludes, the observed change in K concentration response to mycorrhizae at two soil availability levels suggests that mycorrhizae will enhance growth in low-K soil only where K is extremely limiting. Menge el al. (1981) found that mycorrhizal infection significantly increased K concentration (see Table 111) in leaves of citrange grown on 54% of the 26 and decreased concentration in 19%of California soils used in their study (t??), the cases (?ti,Table 111). Only Cu and Na were affected in a completely consisTable I11 Effect of Gfomus fascicuhlus upon the Mineral Concentration in Leaves of Troyer Citrange Growing in 26 California Citrus Soilso Soils (%) in which mycorrhizae resulted in significant change in leaf concentrationb Mineral nutrient P Ca Mi? K Na Zn Mn cu Fe

Increase

Decrease

81 12 4 54 0 46 35 100

15 35 62 19 100 31 35 0 21

8

=From Menge et al. (1981). *Leaf concentrations of mineral nutrients were significantly @ = 0.01) increased or decreased. In the remaining percentage of soils, the leaf concentrations of mineral nutrients were not significantly affected by G. fasciculatus.

212

W . M . JARRELL AND R . B. BEVERLY

tent manner, with all mycorrhizal plants having higher Cu and lower Na concentrations than nonmycorrhizal plants. Generally lower concentrations of Ca, Mg, and Fe were found in mycorrhizal plants, while Zn and Mn were roughly evenly divided between increased and decreased concentrations. Of particular interest was a dilution of P in tissue in 15% of the cases. In these situations especially the plant may be responding to nutrients other than P, increasing their uptake and hence total yield more rapidly than P uptake rates are increased. Recent articles on mycorrhizal response that provide the types of data needed to evaluate dilution effects include those by Snowball et al. (1980), Chambers et al. (1980), Islam et al. (1980), Hughes et al. (1978, 1979), and Barea et al. (1980). In an approximate manner, the article by Timmer and Leyden (1980) also allows one to estimate dilution effect patterns. In numerous instances increased growth with mycorrhizae has been found although P concentrations did not increase (e.g., Islam et al., 1980; Barea et al., 1980). Barea et al. (1980) found N concentration to decrease in alfalfa with soluble P treatments, while mycorrhizae + soluble P brought N concentrations back up to control levels. Total uptake increased with both treatments because yield was greater. Although their published data do not allow one to calculate directly total elemental accumulation in the plant, Timmer and Leyden (1980) found that added P produced strong Cu deficiency in sour orange, apparently because of a direct dilution effect. On mycorrhizal plants Cu deficiency symptoms were only evident in the largest plants. As more P was applied, plants took up more Cu, but tissue concentrations decreased with mycorrhizal plants. One very consistent result noted from numerous studies has been lower Na concentrations in tissue of mycorrhizal plants (e.g., Menge et al., 1980). This response has been found to follow both and 4T.l patterns. Currently no success has been achieved in experimentally determining whether the effect is direct or indirect. The stabilizing effect of mycorrhizally contributed P on membrane integrity (Ratnayake et al., 1978) may be important. Further work should be done, especially in relation to soil salinity. Although there has been much work recently on biological N2 fixation by symbiotic microorganisms, apparently very little of it has been concerned with the effects of nodule N2 fixation on the accumulation of other nutrient elements by plants. Apparently the relatively simple experiments of supplying plants with some N 0 3 - N and allowing them to fix variable proportions of their N needs, and then examining the uptake of other nutrients, have not been carefully undertaken, with a few exceptions. Zaroug and Munns (1980), for example, found that inoculation tended to decrease P concentrations in plants at low S, compared with S N plants. At high rates of P, for comparable yields, inoculated plants tended to have lower concentrations of P in their tissue.

DILUTION EFFECT IN PLANT NUTRITION STUDIES

213

Application of NaN0, tended to increase concentrations of K in both roots and shoots of plants compared with inoculated controls (Chambers et al., 1980). Since K and NO, may be associated in transport processes within the plant (e.g., Ben-Zioni et al., 1971), use of NO, fertilizers may facilitate the uptake of K , an effect not found where plants are fixing their own N . No references were found that gave complete, detailed summaries of nutrient composition of nodulated versus nonnodulated plants for a wide spectrum of elements. It certainly appears that more work should be done to determine how the substitution of fertilizer N for symbiotically fixed N may affect both the internal nutrient requirements of the plant and the ability of the plant to absorb nutrients from solution. This situation is particularly intriguing because we are providing the nutrient, N, through two very different mechanisms. In addition, we have the difference between the energy required to actively accumulate NO3 and reduce it, compared with the energy required for N, reduction. Since N fixation is felt to require relatively more energy, the plant may be less able to accumulate other nutrients whose uptake is energy-dependent. E. WATER

A few studies have examined the effects of drought on the elemental composition of plant tissue. Most of these have not systematically examined changes in total uptake compared with concentration when drought stress is eliminated. In dry soils, movement of ions to the plant root is restricted by low rates of mass flow and diffusion (Nye and Tinker, 1977). As adequate water is applied, the relative movement to the plant root should be facilitated. Water use efficiency (WUE, g dry matter/g water transpired) should be considered in evaluating results of water and of temperature variables (see next section). With greater WUE (low water supply, for example) less soil solution is absorbed per unit of dry matter produced. Therefore movement of nutrient ions to roots by mass flow would be concurrently decreased. Begg and Turner (1976) noted that the uptake rates of N and P are frequently increased more rapidly than plant growth rates when adequate water is supplied to crops. Boron is another element whose total accumulation and concentration are both enhanced when a water deficit is removed, and this may even result in an increase in apparent B toxicity where B is applied and soil water conditions are good (e.g., Gupta e t a / . , 1976). In these cases concentration effects may become important (see Section VI). Bassiri and Nahapetian (1977) measured grain concentrations of several nutrients by ten varieties of wheat when grown dryland and irrigated. The effects of irrigation on nutrient accumulation and concentration, averaged across all varieties, gave the following patterns: P, Mg, ?ti;Ca, k7.1; Fe, ?TO; and Zn, ?TO. Apparently increased soil water content and uptake actually had a very

t??;

W .M. JARRELL AND R. B. BEVERLY

2 14

depressing effect on Ca accumulation in grains. Because leaf tissue concentrations are not reported, it is not known if Ca accumulation by the whole plant was actually decreased. Maintenance of high soil water potentials (SWP) tended to decrease or not change concentrations of Ca, Mg, and K in tomato leaves but increased fruit concentrations of these nutrients (Pill and Lambeth, 1980). Total accumulation was not reported. F. TEMPERATURE

Because it affects the rate of reactions, temperature must be considered an important factor in controlling the uptake of nutrients, as well as plant growth. This variable produces such a multitude of effects on the soil-plant system that it is difficult to separate out cause and effect in many instances. In soil, of course, the solubility of ions may be affected by temperature. Root growth and development of mycorrhizae are sensitive to temperature. Diffusion rates and mass flow (water requirements) increase with increasing temperature in most instances (Nye and Tinker, 1977). And rate of biological reactions dependent upon enzyme kinetics are, of course, highly sensitive to temperature. The appearance of a dilution effect as temperature is raised or lowered to its optimum depends upon how temperature affects the relative rates of reaction. Passive processes should, of course, show less temperature dependence than active processes. Labanauskas et al. (1964) found differences in the distribution of K , C1, and B between tops and roots of citrus with a 5.3”C temperature change, but no significant differences in either dry weight or total accumulation of nutrient elements in the tops. Applications of Zn and P to flax were found (Moraghan, 1980) to affect one another in analogous ways at three temperatures (Table IV). Although definite dilution effects were found, total accumulation of Zn was either increased or not Table IV Zinc-Phosphorus “Interaction” at Three Temperatures” Interaction High P (120 rng kg-I), P response increasing Zn High Zn (8 rng kg-I), Zn response increasing P “From Moraghan (1980)

7°C

15°C

24°C

ttl

ttJ

0t.l

tTJ

t7.l

ot J

DILUTION EFFECT IN PLANT NUTRITION STUDIES

215

affected by P applications, and vice versa. Temperature can obviously affect the type of total accumulation-concentration pattern obtained. G . OXYGEN

Lack of oxygen has a very profound general effect on the respiration activity of the plant root (Marschner, 1972). By decreasing the rates of reaction, the accumulation of many ions may be decreased. Hopkins et al. (1950) found that total accumulation of P and K was increased by increasing O2 from deficient to ambient levels. Phosphorus accumulation was more dependent upon p o z than was K . Concentrations of P and K in barley showed marked increases as p o n was raised from 0 to 5%, while Na concentrations generally decreased (Letey et al., 1962). Patterns for P, K, and Na as O2 increased from 0 to 5% were 7t.T.tft, and 07 J, respectively. Thus even though concentration decreased, evidence for increased resistance to Na uptake was not clear because approximately equal quantities were absorbed at all levels of 0,. Increased quantities and concentrations of N, P, K , Ca, Mg, Zn, Cu, Mn, B, and Fe were found in citrus leaves as soil O2 was increased from 0.8 to 21% (Labanauskas et al., 1972). Concentrations in roots were decreased as 0,was increased, suggesting that translocation processes were quite sensitive to aeration. Concentrations of N, Na, and C1 in stems decreased as 0, increased, although the total quantities of N and Na did not change significantly (0t.l). Similar results were reported in an earlier experiment (Labanauskas et al., 1965). It appears that nutrient uptake rates are more severely affected, in a general way, than is plant growth as O2is decreased in the soil. H. TIME

Young plants are frequently found to contain higher concentrations of many nutrients, especially N, P, and K , than do older plants (e.g., Loehwing, 1953; Mitchell and Reith, 1966; Bassett et al., 1970; Munson and Nelson, 1973; Lea et al., 1979a,b; Page et al., 1978). The decrease in concentration with time has often been referred to as a “dilution” of elements, and attributed to a faster rate of dry matter accumulation than of nutrient uptake. Generally the approach in this type of study has been to sample diagnostic tissues or whole plants through the growing season. One interesting study found that concentrations of lead (Pb) in ryegrass, white clover, and cocksfoot were greatest in late November and considerably lower in spring (Mitchell and Reith,

216

W . M . JARRELL AND R . B. BEVERLY

1966). Whether this response was due to differences in growth rates during these periods, in transport to plant tops, or in uptake rates was not made clear. But obviously warm temperatures would increase rates of dry matter production, and if Pb uptake rates were not correspondingly increased the Pb concentration in tissue would drop. In many studies concentrations of some elements are found to decrease, others to remain the same, and still others to increase. A major problem with many of these has been a failure to characterize the changes in soil solution composition andor nutrient availability over time. Thus decreases in concentration over time may be attributed to either physiological or ecological factors. In most studies where tissue concentrations decreased with time, the concentration of elements in the root zone was not held constant, and more nutrient was probably available early in the growing season than late. Whether this phenomenon is related to most plant requirements or to the conditions under which the experiment was conducted remains to be seen. I. LIGHT

Relatively few plant nutritional studies have been conducted in which light intensity was a controlled variable. As Gupta (1979) stated, “The faster the plant grows-for example, under high light conditions-the faster it will develop deficiency symptoms in a particular growth period. (p. 291). In most cases the increase in plant growth caused a dilution of the concentration of B in the plant (e.g., MacInnes and Albert, 1969). Along with increasing the growth of the plant, raising light intensity may increase the root : shoot ratio because more energy can be translocated down to the roots (Russell, 1977). And more energy should be available for active ion uptake as well. ”

VI. CONCENTRATION EFFECTS When the plant is exposed to excessively high concentrations of elements, excessive temperatures, excessive soil water, or pathogenic organisms, total dry matter accumulation may decrease. If the rate of uptake of a given nutrient does not decrease more rapidly than the growth rate drops, its concentration in the tissue will increase. When discussing concentration effects, the situation analogous to the Steenbjerg effect would arise where the concentration of the toxic element or elements actually decreased in the tissue after application of the toxic element. To the authors’ knowledge, such results have not been reported. This would

DILUTION EFFECT IN PLANT NUTRITION STUDIES

217

require the rate of uptake to decrease due to an adverse effect on plant metabolism, while yield was decreased at a slower rate. This phenomenon could occur where root activity was very adversely affected, such that movement of ions to tops was decreased. Here yields might show only a small decrease but overall translocation to tops may be decreased. However, in most instances when the growth of the plant is stunted, nutrients may continue to accumulate in tissue and reach higher concentrations than in control plants. The reduction in plant size may accelerate the toxic effect because the toxic ion will accumulate to higher levels because of a “concentration effect” and have an increasingly detrimental effect on growth. This synergistic effect may be important in many cases of slowed growth due to the application of toxic chemicals. It may also be an important secondary effect when plants are stunted due to poor lighting, low COz, or other non-nutrient-related problems, and potentially toxic elements are present at marginally toxic levels for the unstunted plant. A . NUTRIENT TOXICITY A N D SYNERGISM

A concentration effect may occur in two contrasting situations: where yield is decreased due to a nutrient imbalance caused by over-application of a nutrient element; and where a synergism develops between two elements, and elemental uptake rates increase more rapidly than plant growth. Excessive applications of K are known to decrease uptake of Mg, and vice versa, while NH4 affects the uptake of both (e.g., Martin and Matocha, 1973). Generally these responses can be represented as 404, or T f f . Likewise, the P-induced Zn deficiency may cause yield to decrease. In most instances, both the total accumulation and the concentration of the critical element decreases (404 or 4T.l). Numerous instances have been recorded in which the concentration of P in plants has increased as plants responded to applied N (tff for P) (Grunes, 1959). This response has been variously attributed to increased growth, improved energy supply to the roots, and generally better health of the plant. Increasing N can decrease the root : shoot ratio under high P, but may increase it under low P (Davidson, 1969). However, yield was also decreased with increasing N and the effect was (for P) OJt, a more classical concentrating effect.

JtL,

B. HEAVYMETALToxicirv

Numerous studies have been conducted in which the elemental composition of plant tissue has been monitored as a function of quantities of heavy metals added to the system. In this case it is hoped to find whether or not the metals adversely affect the accumulation of other ions by plants. Adding metals in the proportions they are found in sewage sludge to three soils

W . M . JARRELL AND R . B. BEVERLY

218

produced different kinds of total uptake responses for different metals (see Fig. 2, taken from Mitchell et ul., 1978). It is apparent from this curve that Zn uptake increased linearly with rate of application, while Cu and Cd increased at a rate below that of application. At extreme toxicity total accumulations of Zn, Cd, and especially Cu were greatly decreased, while Ni showed a small, possibly insignificant drop. This suggests that Ni uptake on a per plant basis continued at a rate nearly equal to that of a plant 30 times as large, while rates decreased substantially for other added metals. Apparently different mechanisms may control Ni uptake at extreme toxicity than control Zn, Cu, and Cd. Analysis of data from Khalid and Tinsley (1980) suggests that additions of Ni to ryegrass consistently increases concentration, although total accumulation first shows an increase, then a decrease (from 74.T to &&?).While both concentrations and total uptake of Mn and Zn decreased with Ni additions, Fe concentrations consistently increased as the plant was stunted. Nickel application obviously retarded Zn and Mn uptake more rapidly than it slowed growth, while it may not have interacted with Fe at all. However, since no root metal concentrations were measured, differential translocation problems may have developed as well. C . SOILSOLUTION SALINITV

As soil solution salt content is increased beyond a specific point, the yield of plants decreases. Accompanying this decrease one generally observes higher

z 0 ta

1.00

g

0.80

2 3

8a 21 0.60

is +LL

0 0.40 W

1

5

0.20

-I W [L

0

I

100

I 200

I

I

1

I

1

I

300

400

500

600

700

800

SLUDGE E Q U I V A L E N T ( M T ha-') FIG.2. Pattern of total elemental accumulation per plant as the sludge equivalents of metal added increase ( I sludge equivalent equals 20 mg Cdkg, 786 mg Cu/kg, 68 mg Ni/kg, and 2068 mg Zn/kg sludge). From Mitchell er al. (1978).

1

900

DILUTION EFFECT IN PLANT NUTRITION STUDIES

219

concentrations of the salinizing ions, e.g., Ca, Na, C1, and/or SO4. Concentrations of other ions are sometimes decreased (e.g., N or P) and sometimes increased. Bernstein et al. (1974) measured elemental concentration changes in leaves of beet, broccoli, cabbage, lettuce, carrot, and onion as solution osmotic potentials decreased from 0 to -4 bars (equal equivalents of NaCl and CaCl,). In general, accumulation of Na, Ca, and C1 increased more rapidly than dry matter yield, resulting in patterns of fl? for these elements. Nitrogen generally followed a 4 l O response, while for cabbage, carrot, lettuce and onion, P, K, and Mg generally showed patterns. Uptake rates of these nutrients were apparently more severely affected by high salt than was dry matter production. However, the published data do not allow definite conclusions to be drawn. Patel et al. (1975), growing sorghum in sand culture, found that raising NaCl-CaCl, (equi-equivalent) salinity from 0 to 40 mmho cm-I conductivity produced the following patterns of uptake: for Na and C1, tit; for Ca, OJ? for P, Mg, and K , 11.1,although Mg was most drastically affected. Thus the ions added produced a concentration effect only for themselves, and tended to substantially restrict the uptake rates of other nutrients. In one of the few studies to examine the effects of salinity on micronutrient metal uptake by plants, Maas et al. (1972) imposed NaCl levels up to 100 meq liter' on tomato, soybean, and squash yield and nutrient composition. In tomato, there appeared to be a slight concentrating effect for all three microelements measured, Fe, Mn, and Zn (Table V). The other crops behaved differently, in at least one case producing no change in total accumulation and a definite concentrating effect (Fe for soybean). In all cases Na and C1 were increased in terms of both total accumulation and concentration.

4.14

D. PATHOGENS Diseases that stunt the growth of plants but do not interfere greatly with nutrient uptake may cause accumulation of elements in plant tissue. For example, leaves infected with beet yellows virus accumulated higher NO,-N concentrations in their tissues than did noninfected plants (Ulrich and Hills, 1973).

VII. PRACTICAL IMPLICATIONS There has been considerable emphasis placed upon plant tissue testing in recent years as a means of determining the extent to which plants are adequately supplied with nutrients. Although there has been some success in diagnosing

W .M. JARRELL AND R. B. BEVERLY

220

Table V Effects of NaCl Solution Concentration on Microelement, Na, and CI Concentrations, and Total Accumulation in Tomato, Squash, and Soybean Crops" Treatment NaCl concentration increased from 0 to 100 meq liter-'

Crop Tomato

Soybean

Squash

Element

Fe Mn Zn Na CI Fe Mn Zn Na CI Fe Mn Zn Na

c1

Response

i l t i l t

l i t t i t t l t

O l t l i 0 i l 0

t i t t i t

i l 0

4.1.1 $ 4 0

t i ?

t i t

'Based on data of Maas ef al. (1972).

single-element deficiencies, other potentially limiting elements often cannot be predicted. A full understanding of how tissue elemental concentrations may change as plant yield increases in response to treatment would greatly assist in making recommendations. Even where plant analysis proves useful in isolating the most limiting nutrient, there are frequently problems in making recommendations for quantities of fertilizers to apply. If the size or growth rate of the sampled plant is known relative to the size expected at that stage for plants not limited by nutrient deficiencies, one may estimate the quantity of nutrient required to produce optimum-sized plants. In addition, such information would supply one with a basis for estimating whether or not the nutrients would become deficient as plant growth rates increased. Some nutrient analysis techniques purport to rank the required elements according to their relative degree of adequacy within the plant (i.e., DFUS, or Diagnosis and Recommendation Integrated System). Most report results in terms of adequacy in relation to norms established in single-variable experiments. It would appear to be important to begin systematic analysis to determine if the composition of treated plants could be predicted from the composition of untreated (e.g., deficient) plants. Such a system might require soil analysis information as well as input on cultural practices. However, the results would be

DILUTION EFFECT IN PLANT NUTRITION STUDIES

22 1

much more useful in terms of producing the optimum nutrient balance for the crop. One potentially important issue may be that the size as well as the age of the plant should be taken into account in assessing whether or not a plant is adequately supplied with a given nutrient. This would give an indication of the relative yield at the time of sampling compared with “reference, ” nondeficient plants. In addition, one could calculate total quantities of nutrients in the plant at time of sampling, which could assist in predicting whether other elements are present in sufficient quantities to tolerate some dilution. Taking the entire plant is, of course, impractical with tree or vine perennial crops. In this case, for young plants, nondestructive measurements such as trunk diameter, height, and average leaf size may be helpful in assessing its relative size. On mature trees, leaf area, leaf volume, leaf weight, length of new growth, or increase in two- or three-year-old branch caliper may all be of some use in looking at relative growth patterns. Comparing such measurements with “expected” norms may be extremely helpful in making recommendations, both for predicting the specific elements required and quantities that should be applied.

VIII. SUMMARY AND FUTURE RESEARCH NEEDS It would appear that, although reference is frequently made to “dilution effect” or at least to a dilution in concentration due to increased dry matter production, there is much to be gained by considering how this dilution occurs. A case could also be made for considering situations in which concentrations increase due to a loss of plant dry matter. We propose that data on total uptake and total dry matter yield be considered wherever possible, and that consideration of these factors be coupled with consideration of concentrations. This approach should remove much of the ambiguity in the current approach to discussions of nutrients in plants. We would suggest that the means for including information on the total nutrient accumulation and total dry matter be expressed as a sequence of the symbols T, J, or 0, to summarize whether values increase, decrease, or show no change. For experiments involving more than one level of imposed treatment, in which yield may show nonconsistent response over the range of treatment values, we propose that only that part of the response curve over which yield is consistently increasing, decreasing, or unchanging be considered as a unit. In instances where total nutrient uptake is difficult to calculate, we suggest that this be estimated by the product of concentration and yield. Although admittedly artificial, this may still give a better indication of plant nutrient status than by relying on concentration alone.

222

W. M. JARRELL AND R. B. BEVERLY REFERENCES

Aldrich, S. R. 1973. I n “Soil Testing and Plant Analysis” (L. M. Walsh and J . D. Beaton, eds.), pp. 213-221. Soil Sci. SOC.Am., Madison, Wisconsin. Barea, J . M., Escudero, J . L., and Azcon-G. de Aguilar, C. 1980. Plant Soil 54, 107-1 17. Bassett, D. M., Anderson, W. D., and Werkhoven, C. H. E. 1970. Agron. J. 62, 299-303. Bassiri. A,, and Nahapetian, A. 1977. J. Agric. Food Chew. 25, I 1 18-1 122. Bates, T. E. 1971. SoilSri. 112, 116-130. Baylis, G . T. S. 1959. New Phytol. 58, 274-280. Begg, J . E., and Turner, N. C. 1976. Adv. Agron. 28, 161-217. Ben-Zioni, A., Vaadia, Y., and Lips, S . H. 1971. Phj~siol.Plunr. 24, 228-290. Bemstein, L., Francois, L. E., and Clark, R. A. 1974. Agron. J. 66, 412-421. Burleson, C. A , , Dacus, A. D., and Gerard, C. J. 1961. Proc. Soil Sci. Soc. A m . 24, 365-368. Chambers, C. A., Smith, S. E., and Smith, F. A . 1980. New Phytol. 85, 47-62. Davidson, R. L. 1969. Ann. Bot. 33, 571-577. Deal, D. R., Botthroyd, C. W., and Mai. W. F. 1972. Phytopathology 62, 172-175. Drew, M. C. 1975. New Phyrol. 75, 479-490. Drew, M. C., and Nye, P. H. 1970. Plant Soil 38, 545-563. Drew, M. C., Nye, P. H., and Vaidyanathan, L. V. 1969. Plant Soil 30, 252-270. Ellingboe, A. H. 1980. I n “The Biology of Crop Productivity” (Peter S. Carlson, ed.), pp. 203229. Academic Press, New York. Epstein, E., and Hagen, C. E. 1952. Plant Physiol. 27, 457-474. Gaastra, P. 1962. Nerh. J. Agric. Sci. 10, 31 1-324. Gerdemann, J . W . 1964. Mycologia 61, 342-349. Gerdemann, J . W. 1968. Annu. Rev. Phvtopathol. 6, 397-418. Goh, K . M., Haynes, R. J . , and Kee, K . K . 1979. N . 2. J. Agric. Res. 22, 319-328. Greenway, H., Hughes, P. G . , and Klepper, B. 1969. Physiol. Plant. 22, 199-207. Grunes, D. L. 1959. Adv. Agron. 11, 369-396. Gupta, U. C., MacLeod, J . A., and Sterling, J . D. E. 1976. Soil Sci. SOC. A m . J. 40, 723-726. Hewitt, E. J . 1956. I n “Plant Analysis and Fertilizer Problems’’ (T. Wallace, ed.), pp, 16-27. Am. Inst. Biol. Sciences. Hiatt, A. J., and Massey, H. F. 1958. Agron. J. 50, 22-24. Holevas, C. D. 1966. J. Horr. Sci. 41, 57-64. Hopkins, H. T., Specht, A. W . , and Hendricks. S . B. 1950. Plunt Physiol. 25, 193-208. Hsiao, T. C. 1973. Annu. Rev. Plunr Phvsiol. 24, 519-570. Hughes, M., Martin, L. W.. and Breen, P. J . 1978. J. Am. Soc. H o r r . Sci. 103, 179-181. Hughes, M., Chaplin, M. H.,and Martin, L. W. 1979. HortScience 14, 521-523. Ingestad, T. 1964. I n “Plant Analysis and Fertilizer Problems’’ (C. Bould, P. Prevot, and J . R. Magness, eds.), pp. 169-173. Am. SOC.Hort. Sci. Islam. A. K . M. S., Edwards, D. G . , and Asher, C. J . 1980. Plan/ Soil 54, 339-357. Islam. R.. Ayanaba. A., and Sanders, F. E. 1980. Plant Soil 54, 107-1 17. Jakobsen, S. T., and Steenbjerg, F. 1964. I n “Plant Analysis and Fertilizer Problems” (C. Bould, P. Prevot, and J. R. Magness, eds.), pp. 174-188. Am. Soc. Hort. Science. Jones, J . B.. Jr., and Eck, H. V. 1973. I n “Soil Testing and Plant Analysis” (L. M. Walsh and J. D. Beaton, eds.), pp. 349-364. Soil Sci. Soc. A m . , Madison, Wisconsin. Jones, J . P., and Fox, R. L. 1978. Soil Sci. 126, 230-236. Katyal, J. C., and Sharma, B. D. 1980. Plunt Soil 54, 105-119. Khalid. B. Y . , and Tinsley, J . 1980. Plunt Soil 55, 139-144. Kleinschmidt, G . D., and Gerdemann, J . W. 1972. Phytoputhology 62, 1447-1453. Kriedemann, P. E., Loveys, B. R . , Possingham, J . V., and Satoh, M. 1976. I n “Transport and

DILUTION EFFECT IN PLANT NUTRITION STUDIES

223

Transfer Processes in Plants” (1. F. Wardlaw and J. B. Passioura, eds.), pp. 401-414. Academic Press, New York. Labanauskas, C . K . , Stolzy, L. H . , Klotz, L. J., and DeWolfe, T. A. 1965. Proc. SoilSci. Soc. Am. 29, 60-64. Labanauskas, C . K., Stolzy, L. H . , and Handy, M . F. 1972. Proc. Soil Sci. Soc. A m . 36,454-457. Lea, R., Tierson, W . C., Bickelhaupt, D. H., and Leaf, A. L. 1979a. Plant Soil 51, 515-533. Lea, R . , Tierson, W . C., Bickelhaupt, D. H., and Leaf, A . L. 1979b. Planr Soil 51, 535-550. Leaf, A . L. 1973. I n “Soil Testing and Plant Analysis” (L. M. Walsh and J. D. Beaton, eds.), pp. 427-454. Soil Sci. SOC. Am., Madison, Wisconsin. Letey, J., Stolzy, L. H., Valoras, N., and Szuszkiewicz, T. E. 1962. Agron. J. 54, 538-540. Lo, S. Y., and Reisenauer, H . h4. 1968. Agron. J . 60, 464-466. Loehwing, W . F. 1953. I n “Mineral Nutrition of Plants” (E. Truog, ed.), pp. 343-358. Lundegardh, H. 1966. “Plant Physiology.” American Elsevier, New York. Maas, E. V., Ogata, G . , and Garber, M. J. 1972. Agron. J. 64, 793-795. Maclnnes, C. B., and Albert, L. S. 1969. Phvsirs 44, 965-967. Marschner, H. 1972. Trans. Comm. V . V I , Int. Sor. Soil Sci. pp. 541-555. Martin, W . E., and Matocha, J. E. 1973. In “Soil Testing and Plant Analysis” (L. M. Walsh and J. D. Beaton, eds.), pp. 393-426. Soil Sci. SOC. Am., Madison, Wisconsin. Menge, I. A , , Jarrell, W.M . , Labanauskas, C . K . , Huszar, C., Johnson, E. L. V., and Siebert, D. 1981. Soil Sci. Soc. A m . J . 45, (in press). Mitchell, G . A , , Bingham, F. T . , and Page, A. L. 1978. J . Environ. Qual. 7, 165-171. Mitchell, R. L., and Reith, J. W. S. 1966. J. Sri. Food. Agric. 17, 437-440. Moraghan, J. T . 1980. Soil Sci. 129, 290-296. Mosse, B. 1957. Nature (London) 179, 922-924. Mosse, B . 1973. Annu. Rev. Phytopathol. 11, 171-196. Munson, R. D., and Nelson, W . L. 1973. I n “Soil Testing and Plant Analysis” (L. M. Walsh and J. D. Beaton, eds.), pp. 223-248. Soil Sci. SOC. Amer., Madison, Wisconsin. Nye, P. H., and Tinker, P. B. 1977. “Solute Movement in the Soil-Root System.” Univ. of California Press, Berkeley. Page, M. B., Smalley, J. L., and Talibudeen, 0. 1978. Plant Soil 49, 149-160. Patel, P. M., Wallace, A., and Wallihan, E. F. 1975. Agron. J. 67, 622-625. Pearson, C . J . , and Steer, B. T . 1977. Planta 137, 107-1 12. Pill, W . G., and Lambeth, V. N. 1980. J. Am. Soc. Hort. Sci. 105, 730-734. Piper, C. S. 1942. J . Agric. Sci. 32, 143-178. Pilman, M. G . 1975. I n ‘ I o n Transpon in Plant Cells and Tissues” (D. A. Baker and J. L . Hall, eds.), pp. 267-308. North-Holland Publ., Amsterdam. Poulsen, J. F. 1950. Tidskr. Plunteavl 53, 413-442. Powell, C. L. 1975. I n “Endomycorrhizas” (F. E . Sanders, B. Mosse, and P. B. Tinker, eds.), pp. 461-468. Academic Press, New York. Prevot, P., and Ollagnier, M. 1956. In “Plant Analysis and Fertilizer Problems” (T. Wallace. ed.), pp. 172-192. IRHO, Paris. Randall, P. J . , and Vose, P. B. 1963. Plant Physiol. 38, 403-409. Ratnayake, R., Leonard, R. T . , and Menge, J. A. 1978. New Phytolo. 81, 543-552. Reyes, D. M., Stolzy, L. H . , and Labanauskas, C. K . 1977. Agron. J . 69, 647-650. Rosell, R. A., and Ulrich, A. 1964. S o i l S r i . 97, 152-167. Ross, J . P. 1971. Phyroparhology 61, 1400-1403. Ross, J . P., and Harper, J. A . 1970. Phytopathology 60, 1552-1556. Russell, R. S . 1977. “Plant Root Systems.” McGraw-Hill, New York. Saalbach, E., and Judel, G. K . 1966. Agrochimica 10, 114-125. Safaya, N. M., and Singh, B. 1977. Plant Soil 48, 279-290.

224

W. M. JARRELL AND R. B. BEVERLY

Sanders, F. E., Mosse, B., and Tinker, P. B. (eds.) 1975. “Endomycorrhizas.” Academic Press, New York. Schultz, I., Turner, M. A , , and Cooke, J. G. 1979. N. Z. J . Agric. Res. 22, 303-308. Small, H . G., Jr., and Ohlrogge, A. J. 1973. In “Soil Testing and Plant Analysis” (L. M. Walsh and J. D. Beaton, eds.). pp. 315-327. Soil Sci. SOC.Am., Madison, Wisconsin. Snowball, K., Robson, A. D., and Loneragan, J. F. 1980. New Phytol. 85, 63-72. Steenbjerg, F. 1945. Tidsskrij for Planteavl 49, 158-174. Steenbjerg, F. 1951. Plant Soil 3, 97-109. Steenbjerg, F., and Jakobsen, S. T. 1963. Soil Sci. 95, 69-88. Terry, N., and Ulrich, A. 1973. Plant Physiol. 51, 43-47. Thomas, J . D., and Mathers, A. C. 1979. Agron. J . 71, 792-794. Thomas, J. R., and Langdale, G. W. 1980. Agron. J. 72, 449-452. Thorne, D. W. 1957. In “Advances in Agronomy” (A. G. Norman, ed.), p. 31. Academic Press, New York. Timmer, L. W., and Leyden, R. F. 1980. New Phytol. 85, 15-23. Ulrich, A. 1968. Plant Soil 29, 274-284. Ulrich, A. and Hills, F. I . 1973. In “Soil Testing and Plant Analysis” (L. M. Walsh and J. D. Beaton, eds.), pp. 271-288. Soil Sci. SOC.Am., Madison Wisconsin. van Steveninck, R. F. M . 1965. Physiol. Plant. 18, 54-69. Walsh, L. M., and Beaton, J. D. (eds.). 1973. “Soil Testing and Plant Analysis.” Soil Sci. SOC. Amer., Madison, Wisconsin. Wardlaw, I. F. 1980. I n “The Biology of Crop Productivity” (Peter S. Carlson, ed.), pp. 297-339. Academic Press, New York. Westermann, D. T., and Robbins, C . W. 1974. Agron. J. 66, 207-208. Zaroug, M. G . , and Munns, D. N. 1980. Plant Soil 55, 251-259. Zelitch, 1. 1971. “Photosynthesis, Photorespiration, and Plant Productivity. ” Academic Press, New York.

ADVANCES IN AGRONOMY, VOL. 34

DESIGNING “LEAFLESS” PLANTS FOR IMPROVING YIELDS OF THE DRIED PEA CROP C. L. Hedley and M. J. Ambrose Department of Applied Genetics, John lnnes Institute, Norwich, England

1. General Introduction . . , . , , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 11. Comparative Responses of Peas to the Crop Environment . . . . . . . . . . . . . . . . . . . . . . . 229

111.

IV.

V.

VI.

VII.

A. Effects of Planting Density on the Leafed and “Leafless” Phenotype . . . . . . . . . . B. Consequences of Growing “Leafless” Peas at High Planting Densities . . . . . . . . . Attaining Maximum Biological Yield per Unit Area . . . . . . . . A. Identifying Plants that Are Tolerant of Interplant Competiti B. The Relationship between Initial Seed Size and Planting Density . . . . . . . . . . . . . . C. Effect of Plant Growth Rate on the Accumulation of Biological Yield by the Crop . D. The Influence of Plant Morphology on Crop Development Attaining the Maximum Economic Yield per Unit A r e a . . . . . . . . . . . . . . . . . . . . . . . . . A. Relationship between Inter- and Intraplant Competition . . . . . . . . . . . . . . . . . . . . . . B. The Influence of Flowering Time on Yield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Effect of Interplant Competition on the Composition of Economic Yield . . . . . . . . D. Interactions between Yield Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Improving the Efficiency of the Pea Fruit , , . , , , , . , . , . , . , . , . . . . . . . . . . . . . . . . . . . A. The Pod . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. The Seed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Plant Ideotype for Improving Yields of Dried Peas . . . . . . . . . . . . . . . . . . . . . . . . . . A. The Plant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Breeding Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

229 235

241 246 25 I 252 252 253 257 262 265 265 267 272 272 273 274 275

I. GENERAL INTRODUCTION The principal uses of dried peas and the problems associated with breeding for such uses have been reviewed by Snoad (1980). At present the main market for dried peas (Pisurn sutivurn) in the United Kingdom is as canned processed peas and as ground pea flour for incorporation into convenience foods. In addition, waste seeds, not suitable for human consumption, have an outlet for use as 225

Copyright @ 1981 by Academic Press, Inc. All nghts of reproduction In any form reserved.

ISBN 0-12-wO734-7

226

C . L. HEDLEY AND M . J . AMBROSE

animal feed. There is also a growing interest in using peas as a substitute for soya as a source of protein, either for compounding or for direct use. The dried pea, in common with the field bean (Viciafuba), the other major grain legume grown in Britain, suffers from relatively low yields and poor stability of yield compared with cereals. The poor performance of the dried pea crop is mainly a consequence of poor standing ability, resulting in increased susceptibility to disease and difficulties with harvesting. The proportion of the crop that may be lost after lodging, through staining or chitting, has been shown to be 13-47% (Proctor, 1963) and may be as high as 80% (see Davies, 1977). Average yields per unit land area of dried peas in the United Kingdom have not increased significantly for the past 20 years. This has been partly due to the lack of breeding activity, which has been more geared to the selection of genotypes suited for harvest by mobile viners when the seed is immature, and marketed as frozen or canned “garden peas.” The majority of agronomic and crop physiological studies over this period have also been more concerned with the specific problems of the vined crop rather than with those of the dried pea crop. Existing dried pea varieties have been derived from genotypes previously selected for growth on a horticultural or garden scale and not for growth as an agricultural crop. Usually such varieties would have been selected for their yielding ability as supported spaced plants and there is no reason to suppose therefore that these varieties will be suited to growth as crop plants. There is good cause for suggesting that such plants will be unsuitable, since the selection criteria used would have taken no account of the interplant interactions that occur within the crop. The effect of planting density on the yield of leafed peas is well established, although many of these studies have concentrated on the yield of immature seeds from the vined crop (e.g., Bleasdale and Thompson, 1960, 1961, 1962, 1963, 1964; Gritton and Easton, 1968; Meadley and Milboum, 1970; Nichols and Nonnecke, 1974; Salter and Williams, 1967). There are fewer reports for the dried pea crop (e.g., Cruzat et a!., 1976; Kruger, 1977; Proctor, 1963; Reynolds, 1950) and in general these suggest that yields of dried peas plateau over a range of population densities. A major step in producing a plant model more suited to the crop environment was made by Snoad (1972, 1974), who introduced the st gene for reduced stipule size and the uf gene, which substitutes tendrils for leaflets, into genetic backgrounds derived from leafed commercial varieties (Fig. 1). Plants with this genetic constitution (afufstst) have acquired the descriptive name of “leafless. The development and reasoning behind this new plant model have been discussed more fully elsewhere (Snoad and Davies, 1972; Davies, 1976, 1977). The main advantage of the “leafless” pea is its improved standing ability due to its greater number of tendrils (Fig. 2); the risk of lodging is therefore reduced and the crop can be more easily harvested. The improved canopy structure may also allow the crop to dry more rapidly with a reduced risk of disease.

”

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

227

FIG. 1. Comparison of a leafless pea, with genes for converting leaflets to tendrils (anand for reducing stipule size ( s r ) , with a conventional leafed pea (.4fAfSrS1).

Current evidence suggests that the yield of leafless peas is not restricted by photoassimilate source limitation (Hedley and Ambrose, 1979). Comparisons between near-isogenic leafed and “leafless” lines grown as spaced pot plants, however, have shown that the yield per plant of the mutant is reduced relative to that of the leafed plant (Harvey and Goodwin, 1978). Monti and Frusciante (1978), however, found no significant difference in yield per plant between comparable near-isogenic lines when grown in randomized plots. More detailed

228

C.

L. HEDLEY AND M . J. AMBROSE

FIG. 2.

Part of a microplot of leafless peas.

physiological studies using near-isogenic lines demonstrated no disadvantage in the efficiency of the mutant form for net COP fixation, or for translocation of photoassimilates to the developing pods (Harvey, 1972, 1974). These latter observations made on individual “leafless” plants are in accord with comparative yield trials, which have shown leafless peas to be capable of yielding as well as leafed varieties (Snoad and Gent, 1976). The unique problem of how to predict the effects of the “leafless” phenotype on characters that can only be derived from leafed peas, has been partially overcome by incorporating the leafless character in as wide a range of genetic backgrounds as possible. The question still remains, however, of how to select from large numbers of single plant segregants, plants that will produce good yields when grown in a crop environment. One method of overcoming this problem is to identify plant characteristics that suit an individual for growth within a crop. If this could be achieved, then a crop plant ideotype could be formulated for the breeder. It is essential that such a plant model be not too specific in its environment requirements, a criticism that has been made of the

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

229

original concept of the ideotype as formulated by Donald (1968). Davies (1976) suggests that since the level of adaptability is genetically determined, stability of performance in a range of environments could become a feature of a model. This assumes, however, that no disadvantageous interaction exists between stability to environment and other features of the crop plant model. Our approach to designing an efficient dried pea crop plant is to base our studies on observations made on plants grown in the crop environment. Information gained from various plant models grown in such environments is then used as a guide to useful plant characteristics that can be incorporated into the ideotype. These observations are also used to highlight areas where more detailed physiological and biochemical studies are required. We feel that there is a greater chance of feeding useful information into the breeding program using this approach than if we based our studies on noninteracting spaced or pot-grown individuals. In this report we outline some of our observations on the behavior of different plant models within simulated crops. Following on from these observations we describe more specific studies geared towards improving the efficiency of the plant when grown in a crop. All of the crop studies have utilized the leafless phenotype. Where necessary, leafed peas have been used for comparative purposes and also where genetic variation for a plant character was not available in a leafless background.

II. COMPARATIVE RESPONSES OF PEAS TO THE CROP ENVIRONMENT T H E LEAFEDA N D “LEAFLESS”PHENOTYPE

A . EFFECTSO F PLANTING DENSITY ON

The effect of the leafless character on a plant population grown as a crop has been determined from comparative studies of leafed and leafless genotypes. Three pairs of genotypes were selected, each composed of a leafed and a comparable leafless genotype. The paired genotypes either were near-isogenic except for the Af and Sr loci, had a common parental line, or were selected because they were similar for other morphological and physiological characters. The genotypes were compared for planting densities in the range 16-400 plants/m” To reduce error variation, seed was pregerminated, and selected uniform plants were planted out on a square planting pattern at the various planting densities. This particular planting pattern was chosen so that each plant within a given density had a similar amount of available growing space and because there is evidence that yields of peas increase to a maximum as the planting pattern approaches a square (Bleasdale and Thompson, 1960; Cruzat ef a l . , 1976; Gritton and Eastin, 1968).

230

C. L. HEDLEY AND M . J . AMBROSE

A detailed analysis of the data revealed an overriding difference between the two phenotypes compared with differences between genotypes within a phenotype. These differences were so marked for biological and economic yield that it was possible to combine data from genotypes within a phenotype and then make comparisons between the two phenotypes. For a given planting density the two phenotypes produced canopies that differed in their ability to intercept light. This was demonstrated by comparing the two near-isogenic lines for the amount of light intercepted at soil level and at the level of the first flowering node (Fig. 3a and b). The measurements were made at a stage when seeds were developing in the early pods. Canopies composed of the leafed line intercepted almost 90% of the light before soil level (Fig. 3a), even at the lowest planting density (16 plant/m'). Canopies composed of the leafless line, however, only approached this level of light interception at the highest planting density (400 plants/m2), light interception declining progressively as planting density decreased. Differences were even more marked at the level of the first flowering node (Fig. 3b), with the leafed genotype intercepting about

100

100

80

.-5

60

c

n

: /J

c e

40

40

-

20

-

C

c

s .?20

k

4

8 0

I,

1

.

1

b

1: x"

t

-

-

.

-

I

0 - 1

u *

'

'

Square of Distance B e t w e e n Plants (cm'x lo-' )

FIG.3. Comparison of percentage light interception for two near-isogenic pea lines over a range of planting densities. Conventional leafed (JI 1194, AfAfStSr. 0-0) and leafless (JI 1198, afafssrsr, 0

4 ) at soil level (a) and at the level of the first flowering node (b). Numbers in parentheses are the planting densities (plants per square meter).

I

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

23 1

70% of the light at densities down to 25 plants/m2 and more than 50% at 16 plants/m” In contrast, none of the leafless canopies exceeded 30% light interception at the first flowering node level. A similar increase in light penetration has been shown for the super okra leaf mutant of cotton (Kerby et al., 1980), which is perhaps the only leaf variant that is comparable to the leafless character. It is very evident therefore that light will penetrate the leafless canopy and reach photosynthetic structures below the level of the first flowering node. Photosynthetic structures produced early in the leafless plant’s development therefore may remain above the C02 compensation point for longer and compensate to some extent for the poor light interception of the crop above the first fruiting node. Photosynthetic activity of the lower leaves of soybeans (Johnson et al., 1969), beans (Phaseolus vulgaris; Crookston ef al., 1975), and alfalfa (Brown ef al., 1966) have all been shown to be increased as a result of increased irradiation. Such compensation in the leafless pea canopy, however, may only be significant at population densities in excess of 44.4 plants/m2, since 50% or more of the light falling on canopies composed of lower plant populations fails to be intercepted. The extent to which photoassimilate produced by the photosynthetic structures below the first flowering node is incorporated into developing fruits is unknown. It has been shown that the major photosynthetic contribution to the fruits of spaced plants is derived, in leafless and leafed peas, from the tendrils or leaflets subtending each developing pod (Harvey, 1974). It has been suggested that the photosynthetic potential per unit area of the tendrils of the leafless mutant may be higher than for the leaflets of the leafed phenotype (Harvey and Goodwin, 1978). If this proves correct, then tendrils subtending pods would be more efficient and compensate to some extent for the poor light interception at the top of the crop. One consequence of more light penetrating the leafless canopy to soil level is an increase in soil temperature compared with corresponding leafed canopies (Table I). At 100 plantslm’ a soil temperature of 29.8”C in the leafless canopy was reduced to 24.2”C in a corresponding leafed population. This temperature difference increased to 8°C when the populations were reduced to 25 plants/m2. The physiological consequences of this increased soil temperature are not known. It is likely, however, that there will be a higher rate of evaporation from the soil surface of the “leafless” crop, although this may be more than compensated for by a decrease in the rate of water utilization by leafless plants (Harvey, 1980). Although soil temperatures differed between the two phenotypes, air temperature within the canopies were much more similar, and at 100 plants/m2 were not significantly different. Greater air movement in the leafless canopy may compensate for the increased level of radiation penetrating the crop, although at present there has not been any research on this aspect of leafless canopy structure. The differences in light interception between leafed and leafless canopies

C. L. HEDLEY AND M. J . AMBROSE

232

Table I Comparison of Temperatures (“C) through the Crop Canopies of JI 1194 (Leafed) and JI 1198 (“Leafless”) at Two Planting Densities Planting density (plantsh’) Height within canopy (cm)

25

I00

JI 1194

JI 1198

JI 1194

JI 1198

25.6 25.2 25.4

24.9 24.9 26.4

24.6 25.6 26.2

27.2 27.5 28.7

24.2

29.8

27.5

35.5

~

30 20 10 1 cm below

soil level

correlate with the differences between the two phenotypes in total above-ground biological yield per unit area (Fig. 4a). The biological yield of the leafed canopies decreased marginally as the planting density was increased. The leafless phenotype, however, had a biological yield per unit area at the lowest planting density (16 plants/m2) that was approximately half that attained by the corresponding leafed canopy. At higher planting densities the biological yield per unit area of the leafless phenotype increased progressively and at densities in excess of 100 plants/m2 exceeded that attained by the leafed canopies at any of the population densities. The responses observed on a unit area basis are obviously a reflection of the responses of individual plants within the population. As the space available to individuals is increased (plant density decreased) so plants will take advantage of increased resources and grow larger. There will therefore be a tendency for low plant populations to compensate by each plant contributing more biological yield to the total biological yield per unit area. This explanation, although true, is an oversimplification of the effect of planting density on the individuals within the population. The “average” plant response, as determined by dividing the response per unit area by the number of individuals, does not convey the differences between plants within each population, induced by interplant competition. The use of average plant values, while ignoring these complex plant-toplant interactions, does, however, indicate how genotypes (or phenotypes) are, in general, responding to particular environments. Over the range of planting densities, the average leafed plant (Fig. 4b) had the capacity to compensate for a 25-fold increase in the available space [25 cm2/plant (400 plants/m2)to 625 cm’/plant (16 plantdm’)] by a 30-fold increase in biological yield. The average leafless plant (Fig. 4b), however, could respond to a similar increase in space by only a 12-fold increase in biological yield. The biological yield per unit area of the leafless phenotype was therefore greatly

233

DESIGNING “LEAFLESS” PLANTS FOR DWED PEA CROP

reduced at low population densities. An alternative view is that there was a steep reduction in the biological yield of average leafed plants as the space available was reduced, whereas the effect on average leafless plants was less severe. The result of this differential effect on average plants of the two phenotypes was for the biological yield per unit area of the leafless phenotype to exceed that of the leafed when the space available was less than 100 cm2/plant (density greater than 100 plants/m2). Up to a density of approximately 100 plants/m2, the effect of planting density on the economic yield per unit area (seed dry weight per square meter) of the two phenotypes (Fig. 5a) was similar to the effect on biological yield (Fig. 4a). Seed weights per unit area of the leafed phenotype decreased marginally between 16 and 100 plants/m2, whereas those of the leafless phenotype increased progressively up to this density. As with the biological yield, the economic yield per unit area of the leafless phenotype exceeded that of the leafed at densities of approximately 100 plants/m2 and greater. The economic yield per unit area of both phenotypes, however, was significantly reduced between 100 and 400 plants/m2. 7-

5-

6

4 -

b

h

c

I

2

7 5 0,

3-

Y c

Y

c

N

0

E

k4

2 2 -

\ 0

3

1 -

2

l

”

7

”

6

“

5

1

4

A

A

( 16

25

3

2 A 44

1 A

0 A

100 400)

0-

7

6

5

4

3

2

A

A

A

( 16

25

44

1 A

loo4oo)

Square of Distance Between Plants ( c m 2 x

FIG.4.

0 A

Biological yield in grams per square meter (a) and grams per plant (b), of leafed

(0-0) and leafless ( 0 4 )phenotypes, over a range of planting densities. Each phenotype is the mean response of three genotypes. Numbers in parentheses are the planting densities (plants per square meter).

234

C. L. HEDLEY AND M. J . AMBROSE

4

3

N I

b

2

N*

E

2

1

a

l

"

7

6

"

5

"

2222%

7

l

4

3

2

1

0

A

A

A

A

( 16

25

44

100 400)

A

6

A

(16

5

4

3

2

1

0

A

A

A

25

44

looroo)

Square of Distance Between Plants ( an2 x

lo-'

)

Economic yield in grams per square meter (a) and grams per plant (b), of leafed phenotypes, over a range of planting densities. Each phenotype is the mean response of three genotypes. Numbers in parentheses are the planting densities (plants per square meter).

FIG.5.

(0-0) and leafless (0-0)

The effect of planting density on the seed yield of average plants (Fig. 5b) was more extreme than the effect on biological yield (Fig.4b). This was most marked for average leafed plants that had a 45-fold reduction in seed yield per plant when space per plant was reduced 25-fold [625 cm2/plant (16 plants/m2) to 25 cm2/ plant (400 plants/m2)] compared with a 30-fold reduction for biological yield. Average leafless plants had a 16-fold reduction in seed yield per plant compared with a 12-fold reduction for biological yield. Although the degree of response of the two phenotypes to planting density was very different, there was a remarkable similarity between them in the efficiency with which biological yield was partitioned into economic yield (Fig. 6), often termed the harvest index (Donald, 1962). Both phenotypes partitioned approximately 50% of their biological yield per unit area into seed at planting densities in the range 16- 100 plants/m2,this figure declining to nearer 35% at 400 plants/m' . The reason for the decline in the harvest index at the highest planting density, however, was different for the two phenotypes. In the leafed phenotype the decline was due almost solely to a reduction in economic yield per unit area, whereas in the leafless phenotype it was due partly to a reduction in economic yield and partly to a continued increase in biological yield per unit area, at this high planting density. There was also a close similarity between the two phenotypes in the proportion

A

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

235

of total biological and economic yield produced by basal and axillary branches (Table 11). Although there was a great deal of variation for this character between the three genotypes that constituted each phenotype group, the overall trends were similar in each case. It was apparent that on average, widely spaced plants of both phenotypes branched profusely and contributed 40-50% of the biological and economic yield per unit area. This proportion was reduced progressively as the planting density was increased and the average contribution made by branches at 100 plants/m2 was reduced in both phenotypes to 5-10%. At the very dense population of 400 plantdm’ all of the measurable biological and economic yield was derived from the main stem of the plants. B. CONSEQUENCES OF GROWING “LEAFLESS”PEASA T HIGH PLANTING DENSITIES

In the previous section it was demonstrated that “leafless” crops have only a limited ability to produce compensatory increases in biological and economic yields at planting densities below 100 plantdm’. From the observations on light 60

-

50

.

4O

-

K a!

-

‘0 C

c H

z

a!

0

X 3 0 .

s

20.

I

1

I

7

6

5

4

A

A

(16

25

3

2 A 44

1 A

0 A

100 4 0 0 )

Square of Distance Between Plants ( c m 2 x lom2) FIG. 6 . Percentage harvest index for leafed (0-0) and leafless (0-0) phenotypes over a range of planting densities. Each phenotype is the mean response of three genotypes. Numbers in parentheses are the planting densities (plants per square meter).

236

C. L. HEDLEY AND M. J. AMBROSE

Table II Effect of Planting Density on the Percentage Contribution of Branches to Total Biological (BY) and Economic (EY) Yield of Leafed and Leafless Peas Planting density (plantslm') 16

25

44.4 100

400

Leafed"

Leaflessb

% BY

% EY

46 23 10.7 5.4 0

BY

% EY

46 20 11.5 7.0

44.1 15 14.3

41.4 31.5 16.8 8.0

0

0

%

5.1

0

aMean of three leafed genotypes. *Mean of three leafless genotypes.

penetration it is likely that this is due to a low leaf area index, which has been shown by Watson (1947a,b) to be a major factor governing crop productivity. As a consequence the biological and economic yield per unit area of leafless crops will decrease to unacceptable levels at low planting densities, and it is a necessity therefore to grow the plants in dense populations. As with other crops, when leafless peas are grown at high planting densities there will be an increase in the competition between adjacent plants for light, water, and minerals. As well as affecting the growth and development of individual plants within the crop, interplant competition will also affect the yield of the crop as a whole per unit area. It has been demonstrated for a number of crops (Holliday, 1960a-c) that the effect of planting density on biological yield per unit area is asymptotic (a plateau being formed over a large range of planting densities), while the effect on economic yield is parabolic (yield decreasing at densities on either side of an optimum). Harper (1961) has suggested that this differential response occurs because the weight of individual plants can almost exactly compensate for changes in the number of plants per unit area, but the allocation of assimilatory products to organs of the plant changes, often to the detriment of the seed output: The concept of an asymptotic biological yield curve and a parabolic economic yield curve in response to increased planting density appears to hold for leafless peas. Our evidence suggests, however, that unlike other crops, including leafed peas, the planting density giving maximum biological yield may be higher than the optimum density for economic yield (Fig. 7). Donald (1963) has suggested that the maximum biological and economic yields occur at the same density when the main limiting factor to production is light, and that the relationship may not hold if water is limiting during seed production. It is possible therefore that leafless crops at high planting densities may be limited by water before competi-

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

237

tion for light becomes critical and that economic yield, produced late in the development of the crop, is disproportionately affected. In addition, our data suggest that the maximum biological yield per unit area of the leafless phenotype will exceed that of the leafed pea. The basis for this difference is likely to be a reduced competition between the photosynthetic surfaces of the leafless canopy for light. Donald (1961) suggests that competition for light within a crop, unlike competition for water and minerals, does not occur between plants but between leaves. If a leaf is shaded by another, then the depression of the photosynthetic rate will be the same whether the superior leaf is on the same plant or another. Once leaves at the base of a canopy become so shaded that they are below the compensation point, they will die. When the rate of death of these lower leaves equals the rate of appearance of new leaves, then the leaf area index of the crop will become static. The improved light penetration through the leafless canopy, even at high planting densities, would allow tendrils at the base of the crop to remain above the compensation point for longer. The attainment of the static leaf area index would therefore be delayed and a higher biological yield per unit area would be achieved. The most obvious effect of planting density on individual “leafless” plants within a crop is the great reduction in plant size as planting density is increased. As with other crops, once biological yield per unit area has reached its 76 -

7

0

5 -

c

K

N

4 -

E 3m

\

2 -

7

6 (1:

5

4

3

2

A

A

25

44

1 A

0 A

100 4 0 0 )

Square o f Distance B e t w e e n Plants (cm2 x Fic. 7. Effect of a range of planting densities on the biological (A-A), economic (0-0). and vegetative (0-0) yield of the leafless phenotype, derived from the mean of three genotypes. Numbers in parentheses are the planting densities (plants per square meter).

238

C. L. HEDLEY AND M . J . AMBROSE

maximum, the mean weight per plant is inversely proportional to the density. Kira et al. (1953) have shown that in soybean, once competition between plants has occurred, the logarithm of the mean individual plant weight plotted against the logarithm of the reciprocal density is linear and sloping. The slope becomes steeper with increased plant age until it is 45",and the relationship extends over the whole range of densities. At that time a constant final yield, irrespective of plant number, is attained. Similar results were obtained when Kira ef al. (1953) applied the same transformations to data derived from other sources (e.g., Donald, 1951). The basis of our studies with leafless peas is to identify plant characteristics that suit the plant to the crop environment. It was mentioned earlier that the use of mean plant values in studies at high planting densities masks the complex responses of individual plants. It is essential therefore to understand the relative responses of individuals within the crop and the effect of specific plant characters on the relative performance of individuals. There have been very few reports where the responses of individual plants within a crop have been measured. Koyama and Kira (1956) have shown for a number of species that a population, initially of nearly uniform plant weight, will move progressively toward a skew distribution. As growth continues there will be an increasing proportion of small plants and a decreasing proportion of large plants within the population. Donald (1963) suggested that crowding accelerates this process and that it is due to increased variability of relative growth rate in crowded communities. An increase in the variation between plants occurs as competition becomes intense and is reflected in an increase in the coefficients of variation (CVs). For example, Stem (see Donald, 1963), using subterranean clover, found that the CV for plant weight was similar at all densities for 90 days and then increased sharply at the higher densities. Certain plant characters respond more to increased competition than others. It has been shown for Zea mays (Edmeades and Daynard, 1979) that the CV for plant dry weight and for ear components increases with plant density, whereas the CV for plant height and leaf area per plant were little affected. Hozumi et al. (1955), in one of the few experiments specifically designed to study the effect of competition on individual plants, found that yellow dent corn plants within a row oscillate for weight and shoot length between negative and positive relative to the overall plant mean. It was apparent that if a plant grew vigorously, its neighbors were suppressed, and if its growth were retarded, the neighbors were favored in their growth. At very high planting densities distributions for plant size become so skewed that self-thinning occurs. Donald (1 963) demonstrated that self-thinning in wheat occurs to a density greater than that giving the highest grain yield per unit area. This suggests that within a dense population survival of individuals has precedence over total seed production per unit area. How variation between individual plants in the population relates to the yield

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

239

per unit area of the crop and to stability of crop yield is not clear. It is assumed that populations where the yields of individuals are distributed normally are to be preferred to those where the distribution is skewed and the CVs are high. In other words, a crop where all of the individuals yield something is to be preferred to a crop where a few individuals yield much while others yield little or nothing. It is the degree of competition between plants at high planting density that determines how much individuals differ from each other. The extent to which individuals interact within a crop is dependent on the competitiveness of the individuals, the planting density, and the environment. Since leafless plants must be grown at high planting densities, we can only reduce the interaction between individuals by selecting genotypes that are less competitive or more tolerant of high population densities. This is a similar conclusion to that derived for wheat by Donald (1968), who states that, “The individual plant within the community will express its potential for yield most fully if it suffers minimum interference from its neighbours. ” Neighboring plants should therefore be weak competitors, and the ideotype itself must be of low competitive ability. It is not, however, sufficient to define a crop plant solely by its tolerance of other individuals at high planting density. An ideal crop plant should also have a high efficiency for partitioning its assimilate into economic yield. Donald (1968) has stated for wheat that, “The successful crop plant should be of low competitive ability relative to its mass and of high efficiency relative to its environmental resources.” Therefore, in the following sections of this article we discuss those characteristics of leafless peas that may be incorporated into the ideotype to make it more tolerant to high planting density and maximize biological yield per unit area. We then define, to the best of our knowledge and experience, characters that will maximize the efficiency with which biological yield may be partitioned into economic yield.

111. AlTAINING MAXIMUM BIOLOGICAL YIELD PER UNIT AREA A. IDENTIFYING PLANTSTHAT ARETOLERANT OF INTERPLANT

COMPETITION

The main conclusion from the previous section was that the most suitable leafless crop plant will be tolerant of its neighbors at high planting densities. “Average” plant responses can be used to indicate those characteristics that will be advantageous to a genotype in a competitive environment. Average plants of strongly competitive genotypes will greatly increase their biological yield when grown at low planting density. When such plants are grown at high planting

240

C. L. HEDLEY A N D M. J . AMBROSE

density, however, they will compete so vigorously that the yield per average plant will be drastically reduced. The density response for an average strongly competitive plant will therefore be extreme and the slope of this response will be steep (Fig. 8a). Weak competitive genotypes with the same duration of growth as strong competitors, will not take full advantage of the resources available at low planting densities and average plants will therefore have lower biological yields than strong competitors. When such genotypes are grown as dense populations, their reduced aggressiveness in competition for resources will reduce the interaction between individuals, and the yield per average plant will be less affected and may be higher than that of the strong competitor. The density response for an average weakly competitive plant will therefore be less extreme and the slope of this response will be shallow (Fig. 8a). The effect of density on an average individual is determined from measurements made per unit area of the population. The population-density interactions for strong and weak competitors (Fig. Bb), however, will be the inverse of those for the average individual within a population (Fig. 8a). Yields per unit area from a monoculture composed of a genotype that is tolerant of high planting density will therefore show a steep positive response to planting density (Fig. 8b), whereas a population composed of a genotype that is intolerant of competition will show a shallow or even negative response to increased density (Fig. 8b). Therefore, in theory a genotype's ability to tolerate competition can be determined from a comparison of yields per unit area of populations grown at a low (noncompetitive environment) and at a high (competitive environment) planting density. This comparison can be made to test the relative effect of specific genotype characteristics on tolerance to competition. Comparisons using strongly

a

b ,I

"/

BIl Unit A r m

Plant

low

High Plant Density

, I

low

,, ,,

,,

High

FIG. 8. Biological yield (BY) for weak (WC)and strong (SC) competitors at high and low planting densities (a) per plant and (b) per unit area.

24 1

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

a

b

Strong C o m p e t i t o r s

W e a k Competitors

,, I

I

,

0 0 0

,

0

,, 0

0

,, ,, I

,

, ,, ,, ,,, ,, , , 0

L

low

I

High

I

tow

I

High

Plant Density

FIG. 9. Biological yield per unit area of (a) strong and (b) weak competitors that are relatively high- (H) or low- (L)yielding over all planting densities.

and weakly competitive environments will also distinguish between genotypes that are overall high or low yielding over a range of planting densities, as well as being strong (Fig. 9a) or weak (Fig. 9b) competitors. B. THERELATIONSHIP BETWEEN INITIALSEEDSIZE A N D PLANTING DENSITY

The effect of seed size on the response of a genotype to planting density is well known, especially through the work of Black (1957). Black compared crops of subterranean clover for the effect of initial seed size on individual plant weight after 194 days of growth. When grown as spaced plants (weakly competitive environment) the ratios of the “average” plant weights between the crops were similar to the ratios between the initial seed weights. At a high population density (strongly competitive environment), however, the average plant weight from each crop was similar and not related to the initial seed size. The similarity between average plants grown in the strongly competitive environment was due to the fact that plants derived from the small seeds were affected by competition later than those derived from large seeds. The increased duration of noncompeti-

242

C.

L.HEDLEY AND M . J . AMBROSE

tive growth for plants derived from small seeds resulted in these plants growing to the same size as those derived from the larger seeds. This relation between planting density and seed size has been used for determining the optimum planting density for leafed pea varieties for many years. The recommended planting density for large-seeded varieties has always been lower than that for small-seeded types (Gane et al., 1971). We have demonstrated similar relationships for a number of leafless genotypes that differed for seed size (Fig. 10a and b). We selected two genotypes that had small seeds (BS 20 and BS 41), two that had large seeds (BS 42 and BS 151), and two that had mediumsized seeds (BS 22 and JI 1198). The six genotypes were grown in microplots on a square planting pattern at two planting densities: 16 and 100 plants/m' (weakly and strongly competitive environments, respectively). The total above-ground biological yield per unit area was determined by oven-drying plants taken from the center of each microplot. The biological yield per unit area of the two large-seeded genotypes was similar at the two planting densities (Fig. 10a). This similarity was reflected in an approximate fivefold decrease in the "average" a

26r BS 42..

I

4

24

-

22

.

20

-

b

18 *

;16-

'B

3

0

'c

14-

312 -

M

U

E m a

\

10

-

8 -

1

a

Seed Size

6 -

large -medium

4.

---small

2-

*-**a

t

I

16

High 100

low

0-

I

low 16

Plant Density ( plants/m')

FIG. 10. Biological yield (a) per unit area and (b) per plant of six leafless pea genotypes that differ for seed size, grown at high andslow planting densities.

1

High 100

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

243

plant weight between the low and high plant densities (Fig. lob). These two genotypes therefore behaved as strong competitors. The biological yield per unit area of the two small-seeded genotypes responded steeply to increased planting density (Fig. 10a). At the low planting density the biological yield was approximately half that of the large-seeded types, whereas at high planting density the biological yield of the small- and large-seeded types was similar. The steep population-density interaction for the small-seeded genotypes resulted in only a threefold difference in average plant weight between the two densities (Fig. lob). The small-seeded genotypes therefore behaved as weak competitors. The genotypes with medium-sized seeds showed responses that were intermediate between the large- and small-seeded types. The association of seed size with tolerance to high planting density is symptomatic of a closer relationship between plant growth rate and plant competition. The relative growth rate (RGR) is a measure of the efficiency with which cellular systems, plants, or parts of plants will grow. In a bacterial suspension or in a plant meristem it is a measure of the rate of cell division, while in the development of a seedling it is a measure of the combined rate of cell expansion and cell division. Assuming that a developing seedling has ample light, minerals, and water, then initially its RGR will be linear on a logarithmic scale and the slope will be dependent on the interaction between the inherent rate of cell division and expansion and the environment in which the seedling is developing. In order to main:ain this environmentally determined RGR, the growth rate (GR), which is a function of the mass of cells that are dividing or expanding at a particular time, must increase exponentially. The environmental resources required to maintain this exponential GR will be utilized at a rate equivalent to the GR; therefore if the GR is low, demand for resources will be less than when the GR is high. It is the demand that a seedling makes on environmental resources that determines the competitiveness of the seedling. The GR, after a given duration of time, can therefore be used as a measure to compare seedlings for their competitiveness; seedlings with a low GR will be weaker competitors than those with a high GR. The apparent relationship between seed size and competitiveness exists because there is usually a good correlation between the size of the seed and the GR of the seedling. We have studied the relationship between seedling growth rate and initial seed size in a wide range of both leafed and leafless peas. Seedlings were grown in spaced pots in a greenhouse. At regular intervals seedlings were selected at random and the dry weight was determined. Measurements were only made over the first 30 days of growth from sowing and were stopped before the formation of flowers. The plant GR of all the genotypes was exponential over this time period and the RGR was therefore linear when the growth curve was expressed as a logarithm. A representative sample of these data is presented in Fig. 11. The relative growth rates were not significantly different for any of the genotypes,

244

C . L. HEDLEY AND M. J . AMBROSE

2.0 J1463(leafed)

1.0 1.6

1*4 102 1.0

s

t

0.8

D '-

5

0.6

?

0-4

n t

c

0.2

0

0

0 P

D 0 4

BS 52 ( l e a f l e s s )

-0.2

- 0.4 - 0.6

J I 463

- 0.8

JI 2 J I 197

- 1.0

BS52

-

-

054

-

28

30

375 127 64

283 227

-192

-1.4 1

12

I

14

16

18 20 2 2 24 26 Days from Sowing

32

FIG.11. Comparison of growth curves (log,) for a range of leafed and leafless pea genotypes that differ for weight per seed.

including those with the leafless phenotype. There were, however, large GR differences between the genotypes, as determined from the difference in the height of the regression lines above the x axis (Fig. 1 1 ) . The difference in GR correlated, as expected, with seed size, but this correlation only held when

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

245

genotypes within a phenotype were compared and not when leafed and leafless genotypes were compared. Genotypes with the leafless phenotype had, in general, lower growth rates than those of genotypes with the leafed phenotype irrespective of seed size. The poor correlation between seed size and growth rate in comparisons between leafed and leafless genotypes was not due to a changed relationship between the weight of the embryonic axis and the weight of the seed, since a good correlation exists between the weight of the embryonic axis and the weight of the seed irrespective of the phenotype (Fig. 12). Although comparisons between the developing leafed and “leafless’ ’ phenotype demonstrated no difference in RGR, it is apparent that the two phenotypes must have differed for RGR earlier in development, to give the observed differences in GR. The differences may be due to the composition of the embryonic axes. A difference in the proportion of

-

50

0 c X

W

r 40 0

c .

c

s

2 30

n

‘0 Q, Q, IA

A

+

20

0

x *

0

A

10

I

I

I

I

I

I

I

1

2

3

4

5

6

7

E m b r y o n i c A x i s D r y Wt.tmgJ FIG. 12. Relationship between the dry weight of the seed and the dry weight of the embryonic axis from a range of leafed and leafless genotypes. Leafed: JI 956 (A), JI I194 ( x), J1 321 ( 0 )JI, 813 (W). Leafless: BS 21 (0).I1 1198 (+), BS 5 (*), BS 4 (A). Each point is the mean of ten seeds.

246

C. L. HEDLEY A N D M . J . AMBROSE

dividing to expanding cells in the embryonic axis will give an initial difference in the RGR of the germinating seedling-the higher the proportion, the greater will be the initial RGR. Such differences would become less significant as seedlings develop because the proportion of growth attributed to dividing compared with expanding cells becomes progressively less significant. The leafed embryonic axis has leaf meristems that are absent from the leafless phenotype. It is possible that these additional meristems contribute the dividing cells by which the RGR could be initially increased.

c. EFFECTO F P L A N T GROWTHR A T E O N THE ACCUMULATION OF BIOLOGICAL YIELDB Y

THE

CROP

The high planting densities required to attain an acceptable yield per unit area of leafless peas can be interpreted as a requirement to increase crop growth rate (CGR), especially early in the development of the crop. By definition a crop growing at a high CGR will have a high leaf area index, and it is the integral of the leaf area index over the growth period that is related to the biological yield of the crop (Donald, 1961). In this respect a low-growth rate crop will resemble a late-sown crop, where the integral of the leaf area index will be reduced by the reduction in the growing season. This is one reason why pea yields decline with sowings made progressively later in the season (Kruger, 1973; Milboum and Hardwick, 1968; Proctor, 1963). Crop growth rate is determined by the individual plant growth rate and planting density and can be raised therefore by increasing either or both of these components. Although genotypes with the leafless phenotype have inherently lower growth rates than those of leafed peas, there is still variation attributable to differences in seed size. As discussed previously, large seeds will give rise to plants that have high growth rates and small seeds will give rise to plants that have lower growth rates. We have studied the interaction between plant growth rate, as determined by initial seed weight, and planting density, which is the other component of CGR. The crop growth rates of two genotypes, one large-seeded (BS 151) and the other small-seeded (BS 5 ) , were compared when grown at relatively high (100 plants/m') and low 16 plants/m') planting densities. As with all of our crop experiments, young seedlings selected for uniformity were planted on the square at both densities, to eliminate any intra- and interrow effects and to maximize yield per unit area of each treatment. Large plots of each genotype at each density were used and the whole experiment was replicated three times. Plants were removed at frequent intervals from random positions within the plots, dryweighted, and used to determine the change in biological yield per unit area of each crop with time.

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

247

Initially environmental resources were not limiting and the individual plants comprising each crop developed independently of planting density. The relative growth rate of each crop was therefore directly related to the relative growth rate of an individual spaced plant. Since it was demonstrated previously (Section III,B) that pea genotypes grown free of competition have similar relative growth rates, it follows that during this early phase of crop development the crop relative growth rates will be similar. A comparison of the linear regressions obtained by plotting the logarithm of the biological yield per unit area against time (Fig. 13a and b) confirmed that there was no significant difference between the relative growth rate of the large- and small-seeded genotypes at either planting density. Differences between the growth of the crops were therefore due solely to differences in crop growth rate attributable either to initial seed size or to planting density. Because the regressions were parallel, the differences in growth rate can be determined from the height of the regression lines from the x axis. The difference between the regressions of the large- and small-seeded crops was a constant and independent of planting density. By extrapolating the regression lines to zero time the initial difference between the genotypes could be determined (Fig. 13a and b). The regression lines cut the y axis, not at an initial weight equivalent to that of the seed, but at a weight that equates with the weight of the embryonic axis. This confirmed the suggestion from the previous section that the embryonic axis must be the unit determining differences in growth rate. The difference in weights of embryonic axes derived by extrapolation (approximately 3 mg for BS 5 and 5 mg for BS 151) are very similar to those determined by dissecting dried seed of the two genotypes (2.4 and 4.5 mg, respectively). This difference in the initial embryonic axis weight determined that the genotype regressions were separated by 5 days. The difference between the regressions of the high- and low-density crops within each genotype was proportional to the initial weight or number of embryonic axes per unit area. The difference between 16 and 100 axes per unit area resulted in a difference of 18 days between the regression lines for crops growing at the two planting densities. The period of exponential growth during which plants grew free of interplant competition ended when resources for plant growth became limiting. Interplant competition was then initiated, resulting in a progressive decline in the relative growth rate of all the crops (Fig. 13a and b). Competition was initiated when a specific biological yield per unit area had been attained that was independent of genotype and planting density and varied only with the time from sowing. The crop growth rates were therefore similar at this specific biological yield per unit area. Within a given density the plant growth rate, when competition was initiated, was therefore similar for both genotypes (Fig. 13c and d). The plant growth rates differed between planting densities and were higher in the lower density by an amount that was proportional to the decrease in plant number per

248

C. L. HEDLEY AND M. J . AMBROSE

-

13

C.

12 11 10 9

8 7 6 5 4 41 3

20

40

60

80

1 100 120

3 20

40

60

80

loo

120

Days From Sowing

d

11

C

10

-C.

-c.

9 8

-C.

7 6

5 4 3 2

FT

I

i 20

40

60

80

.

100 120

1 0

I

' i 20

Days From Sowing

40

60

80

100 120

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

249

unit area. More simply, when competition was initiated in the high-density crops, it was between plants growing at relatively low growth rates, while in low-density crops competition was between plants growing at relatively high growth rates. Competition was first observed in the large-seeded crop (BS 151) grown at the high planting density, occurring after 58 days from germination. The same biological yield per unit area was attained by the small-seeded genotype (BS 5 ) after 63 days. At the low planting density competition was initiated between plants of the large- and small-seeded genotypes after 76 and 81 days, respectively. It is apparent from this experiment that crop growth rate can be varied equally well by initial seed weight, or more correctly the weight of the embryonic axis, and by planting density. In theory it is therefore possible to produce an identical crop growth rate by sowing large seeds at low planting density or by sowing small seeds at a proportionally higher planting density (Fig. 14). Although it has been suggested that such crops would be similar (Donald, 1961), it is apparent that the effect on individual plants will be very different (Fig. 14). Competition would occur between the plants of both crops after the same time from germination. The large-seeded plants, however, would have high growth rates when competition was initiated, whereas the growth rate of the small-seeded plants would be low. The consequences of these different crop structures are more significant to the partitioning of biological yield into economic yield and will be discussed in Section IV. In the experiment (Fig. 13a-d) the duration of crop growth was similar for the two genotypes at both planting densities and so the amount of growth that occurred after competition had been initiated within the crops was limited not only by diminishing resources, but also by a finite time in which the plants could utilize these resources. The average plant weight of the two genotypes at 100 plants/m’ was approximately 6 g/plant, which gave a crop yield of about 600 g/m2 for both crops. The similarity between the final yield of the two genotypes at this planting density suggests that both crops had become limited by competition at this maximum biological yield and that this yield was therefore an environmentally determined optimum. The two genotypes differed for final biological yield at the low planting

FIG. 13. Crop (a, b) and plant (c, d) growth curves of two leafless pea genotypes that differed for weight per seed (BS 5 , 130 mg; BS 151, 300 rng) grown at two planting densities. (a) BS 5 (0)and BS 151 ( + ) at 100 plants/rn2; (b) BS 5 (0) and BS 151 (63) at 16 plants/m’; (c) BS 151 plants at 100 (+) and 16 plantsh’ (@ (d) BS I); 5 plants at 100 (0)and 16 plantdrn’ (0).Horizontal arrows indicate the biological mass at the onset of interplant competition (C). Symbols 0, 0, +, @ with vertical arrows indicate the time from sowing to the onset of interplant competition. fl with vertical arrows indicates flowering time.

C. L. HEDLEY AND M. J. AMBROSE

250

Biomass Ceiling Onset of Plant Competition

t T

Time

b

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

C

/

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

//

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

0

0 0

0

'...................../(............"'....*. /

, 8

,'

Small

High

/

0

t

t

T

T

Time FIG. 14. (a) Theoretical growth curves of two contrasting crops growing at the same crop growth rate. One crop composed of a large-seeded genotype sown at low planting density and the other of a small-seeded genotype sown at high planting density. (b,c) Growth curves of single plants from each of the two crops. T is the time of onset of interplant competition.

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

25 1

density and in both cases the biological yield was less than the potential optimum for the environment. An average plant of BS 151 at low planting density attained a weight of about 27 gm, giving a crop weight of about 430 g/m‘. The smallseeded genotype (BS 5) attained an average plant weight of 16 g, giving a crop weight of about 260 g/m’. The initial low crop growth rates of the low-density crops ensured that, within the finite growing season, neither genotype would be able to utilize all of the resources available. The final biological yield was therefore lower than for the higher-density crops. Similarly the lower initial growth rate of BS 5 compared with BS 151 ensured that the final biological yield of this small-seeded genotype would be less than that attained by BS 151. The differences in flowering time between the two genotypes may also have contributed to the difference in biological yield; this will be discussed more fully in Section IV,B.

D. THEI N F L U E N C E OF P L A N T MORPHOLOGY ON

CROP

DEVELOPMENT

Although crop and plant growth rates are paramount in determining the biological yield of the leafless crop, differences between the morphology of genotypes that may have similar growth rates must be considered in any crop plant ideotype. It has been suggested earlier that branches are usually suppressed in leafless peas at high planting densities (Table 11). Branches are produced by some genotypes, however, during the early development of the crop when plants are growing free of interplant competition. In these genotypes, once resources become limiting, competition will occur not only between plants but also between the branches within a plant. This often results in only one of the branches continuing to grow and produce flowers and eventually seed. Nonreproductive branches add to the biological yield of the crop but add little to economic yield and therefore reduce the efficiency of assimilate partitioning into seed. Selection against such early branching would therefore be advantageous. Variation can also be found for the erectness of the stem and for tendril characters such as size and the degree to which tangling occurs. There appears to be an advantage in selecting plants that are more erect and have less tangled tendrils. Such plants have improved standing ability and the degree of interference between neighboring plants is reduced. These attributes improve the harvestability of the crop and may well be important in reducing the plant-toplant variability that occurs in dense populations, during the later stages of crop development. These characters are difficult to quantify, although they can be easily recognized when genotypes are grown in populations.

252

C. L. HEDLEY AND M. J . AMBROSE

IV. ATTAINING THE MAXIMUM ECONOMIC YIELD PER UNIT AREA A . RELATIONSHIP BETWEEN INTER-A N D INTRAPLANT COMPETITION

Attaining a high biological yield per unit area is only one component in the equation for increasing crop seed yields. It is equally important to partition a high proportion of this biological yield into economic yield. In most crops, as planting density is increased beyond the density giving maximum grain yield, the efficiency of partitioning into economic yield within the crop (harvest index) decreases (Donald and Hamblin, 1976). As discussed earlier, it is necessary to grow leafless peas at a high planting density in order to attain a high crop growth rate and a sufficiently high biological yield per unit area. In most crops there appears to be an inverse relationship between the competitiveness of a genotype and crop harvest index (Donald and Hamblin, 1976), and so the relatively noncompetitive genotype more suited to high population densities will also maintain its harvest index. Most of the observations leading to this conclusion, however, have been made on reproductively determinate crops such as cereals, and a more complex relationship may exist in peas, where vegetative and reproductive growth continues in parallel. The competitiveness of an individual in a population is an expression of the sink demands of those parts of the individual that are growing rapidly. The ideal crop plant model will therefore be a compromise between the ability of an individual to utilize environmental resources at a rate that does not cause undue interference to neighbors, and the development of sinks within the plant the demands of which do not exceed this resource rate limit. As interplant competition increases, either between leaves for light or between plants for minerals or water, so the resources available to each individual for partitioning into developing sinks is decreased. Once the resources available to each individual are reduced, intraplant competition will increase between developing sinks, which in the case of peas may be vegetative or reproductive. In competitive genotypes, partitioning favors vegetative growth to the detriment of seed yield, and yields are low because of a decreased harvest index. In noncompetitive plants the balance may be much more in favor of reproductive growth and the harvest index may be high. If the reproductive sinks of a noncompetitive plant are too successful, however, in attracting plant resources and exceed the resource limit of the plant as a whole, then the growth of the plant may be prematurely curtailed before all of the available environmental resources have been utilized. Seed yields per unit area may then be low because of an insufficient biological yield per unit area. Thus our ideal leafless crop plant must incorporate an ability to produce a high

253

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

biological yield per unit area with a high harvest index at high planting density. It will be necessary therefore to ensure that the components constituting the reproductive sinks are not so demanding of the plants’ resources that growth ceases prior to the attainment of maximum biological yield per unit area. B . THEINFLUENCE OF FLOWERING TIMEO N YIELD

The experiment in which a large- and a small-seeded genotype were compared at a high and a low planting density (Fig. 13a-d) demonstrated that genotypic differences in flowering time had little effect on crop or plant relative growth rates, while resources were nonlimiting. The final economic yield per unit area and the crop harvest index differed both between genotypes and between treatments in this experiment (Table 111). Economic yield per unit area increased in both genotypes with increased planting density, but the small-seeded earlierflowering genotype (BS 5 ) showed the greatest response, increasing from 170 to 320 g/m2 compared with from 220 to 300 g/m2for BS 15 1. Although BS 151 had the higher yield at the low planting density, the final crop harvest index of this genotype (0.49) was significantly lower than that of BS 5 (0.58). The lower economic yield of BS 5 at this planting density can be attributed therefore to an insufficient biological yield. It is not possible to determine how much of this reduced biological yield could be attributed to a lower crop growth rate, dependent on seed size, and how much was due to a curtailment in plant growth induced by competition between reproductive and vegetative sinks in the later stages of crop development. The two genotypes had similar biological yields per unit area at the high planting density and so the relatively higher economic yield of BS 5 (320 g/m2) compared with BS 151 (300 g/m’) can be attributed almost entirely to the higher crop harvest index of BS 5 (0.54) compared with BS 151 (0.47). It is likely that the earlier flowering time of BS 5 contributed to its relatively high harvest index. It is not possible, however, to separate this effect from the effect of interplant Table 111 Effect of High and Low Planting Densities on the Economic Yield and Harvest Index of Large- (BS 151) and Small-Seeded (BS 5 ) Leafless Genotypes

Genotype BS BS BS BS

151 5 151 5

Planting density (plantdrn’)

Economic yield Wm’)

16 16 100 100

220 170

Harvest index 0.49 0.58

300

0.47

320

0.54

254

C. L. HEDLEY AND M. J. AMBROSE

competition. BS 151, by virtue of its large initial embryo mass, is by definition a strong competitor and as discussed earlier, there appears to be an inverse relationship between the competitiveness of a genotype and harvest index. In addition, it is not possible to determine the effect that the yield component differences that exist between the two genotypes will have on harvest index. In particular the large difference in seed size could result in a marked difference in intraplant competition during the development of economic yield. The relationship between seed phenotype and its role as the main component of yield will be discussed later. The obvious interaction between flowering time and seed size, either through the role of the seed as an embryo plant or as the harvestable yield of the plant, was very marked when we compared two leafless genotypes that differed for flowering time (as determined by the node to first flower). The two genotypes could both be classified as small-seeded, although they differed significantly for seed size (Table IV). In this comparison, however, the larger-seeded genotype (BS 21) flowered earlier than the smaller-seeded genotype (BS 23). These two genotypes were compared by growing them at high (100 plants/m') and low (16 plantdm') planting densities. The average plant response of the larger-seeded early-flowering (BS 2 1) genotype differed only by about twofold in biological yield between the two environments, whereas the biological yield of the smaller-seeded late-flowering genotype (BS 23) differed by a factor of approximately 5 (Fig. 15). The biological yield per unit area for BS 21 therefore increased sharply between the planting densities, while BS 23 showed a much less marked response. According to previous definitions of strong and weak competitors, the reponse of the larger-seeded BS 2 1 would classify this genotype as a weak competitor and the smaller-seeded BS 23 would be classified as a strong competitor. This is contrary, therefore, to the discussions in Section 111, where increasing seed size was related to increased competitiveness. The classification of BS 21 and BS 23 is supported by the difference in harvest index of the two genotypes at low planting density (Table V). The crop harvest index of BS 21 was 0.62 with an economic yield of 114 g/m2 compared with 0.53 and 234 g/m', respectively, for BS 23. This type of comparison tends to highlight the problem of studying the effect Table IV Weight per Seed and Node to First Flower of Two Leafless Genotypes Genotype

Weight per seed (mg)

Node to first flower

BS 21 BS 23

I95 I64

8-9 12

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

255

a

b

2826 24 22 20

/ /

I

BS 2 3 - Small I

-

I

/ /

/

18

/

*

/ /

2

-

16-

BS21

/

0

/ /

t l 4 -

m

/

I

12

-

10

-

.....

I I

/ / I

I /

8 . 6 -

4 . 2 0-

low 16

High 100

Plant Density

low 16

(

High

plants/m2)

FIG.15. Biological yield (a) per plant and (b) per unit area of two leafless genotypes that differ for seed size and flowering time (BS 21, large, early; BS 2 3 , small, late), grown at high and low planting densities.

of a character such as flowering time on a complex character such as economic yield, which is determined by the interaction of many plant characters such as seed size, simultaneity of fruit development, and the degree of reproductive determinacy. Before an accurate assessment of the effect of flowering time on yield can be made, the effect of these interacting components must be reduced. This can only be achieved by synthesizing genotypes that are similar in all

100

C. L. HEDLEY A N D M. J . AMBROSE

256

Table V Economic Yield and Harvest Index of Two Leafless Genotypes Grown at High and Low Planting Densities

Genotype BS BS BS BS

21 23 21 23

Planting density (plants/m')

Economic yield (dm')

16 16 100 100

114 234 302 296

Harvest index 0.62 0.53 0.55

0.57

respects with the exception of flowering time. Although at present such genotypes do not exist, it is possible to speculate on their likely responses, based on a knowledge of the leafless pea plant and crop. If two leafless pea crops were compared that varied only for flowering time, many of the features of crop development would be similar (Fig. 16). The crops would be sown at a high planting density, which would be the same for both crops because seed size would be similar, as would the crop relative growth rate and growth rate while resources are nonlimiting. Likewise the potential ceiling biological yield and the duration of crop growth would be similar. If it is now assumed that one genotype flowered early, prior to any resource limitation, while the other flowered late, when interplant competition for some limiting resource was occumng, then certain consequences can be predicted. The early reproductive nodes of the early-flowering genotype would be produced when environmental resources were not limiting plant growth. Intraplant competition between these developing sinks and the vegetative part of the plant would therefore be at a minimum and pods set at this time would be more likely to produce mature seed. The reproductive sink demand must not be so excessive, however, that the plant as a whole cannot grow and continue to utilize environmental resources until the potential ceiling biological yield has been attained. Economic yield must therefore increase slowly and a high harvest index should be reached when the crop has attained this potential maximum biological yield per unit area. Economic yield development in the late-flowering genotype would occur when plant growth rates and interplant competition were high. The maximum duration of reproductive growth would be reduced and so the rate of reproductive development would be increased. The overall effect of increased interplant competition and a high rate of reproductive sink development would be to increase the intraplant competition both between vegetative and reproductive growth and between the yield components comprising reproductive growth. This high level of intraplant competition would increase the abortion of yield components and could well result in a decreased crop harvest index. It may be concluded that the early-flowering type would be the better crop

257

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

/fl-*

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

Biomass Ceiling

.~. ..................

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

I Flowerinq

LOG, D r y wt./ Unit arec

I I I I 1

Onset of Plant Competition

!

I

I

I I I 4

I

b

d

Time

C FIG. 16. Crop development showing the effect of flowering time on the duration of reproductive growth: (a) duration for late flowering; (b) duration for early flowering; (c) duration of total crop growth.

model as long as the rate at which economic yield develops allows the maximum biological yield per unit area to be achieved. c .

EFFECTOF

INTERPLANT COMPETITION ON THE COMPOSITION OF

ECONOMIC YIELD

The reproductive potential of an average pea crop plant is derived from the number of flowering nodes, the number of flowers produced at each flowering

258

C . L. HEDLEY AND M . J . AMBROSE

node, the number of ovules per pod, and the maximum weight per seed. The yield that is realized by the plant is only a proportion of this potential because each of the components of the potential yield becomes modified by the plant during development. The modified components contributing to the actual yield of an average pea crop plant are the number of flowering nodes producing pods, the number of pods maturing at each flowering node, the number of ovules per pod producing mature seed, and the mean weight per seed. Potential and actual yield components within a genotype can be ranked according to their instability to environment. Of the potential yield components only the number of flowering nodes shows any significant variation with environment. This character is a major component of plant size and so any factor within the environment that affects the plant size, after flowering has been initiated, will affect the number of reproductive nodes. The number of flowers produced at each flowering node and the number of ovules per pod have both been shown to have a high degree of heritability (Snoad and Arthur, 1973a,b) and are relatively unaffected by changes in environment. Likewise it may be assumed that the maximum weight per seed will be relatively unaffected by environment and will be under close genetic control. Although genes have been identified that affect seed size (Wellensiek, 1925a; Gottschalk, 1976), the genetics of seed weight, or size, are poorly understood and confused by maternal interactions (Snoad and Arthur, 1974a). Unlike the potential yield components, those contributing to actual yield all vary to a greater or lesser extent with environment. The number of flowering nodes producing pods will vary because of the close relationship with the number of flowering nodes produced, which, as stated, is largely dependent on plant size. The number of pods that mature at each flowering node is also extremely variable with environment. In addition, considerable genotype-environment interaction has been found for this character among leafed varieties (Snoad and Arthur, 1974b), with the most extreme response being shown by multipodded varieties. A similar response has been found for a multipodded leafless genotype when compared with single- and double-podded types (Hedley and Ambrose, 1979). The number of ovules per pod producing mature seeds has also exhibited significant variation between sites and varieties, and significant genotypeenvironment interaction (Snoad and Arthur, 1974b). Mean weight per seed is a more stable component to certain environments and we have little evidence of any significant effect of planting density on this character, although this does not exclude possible differences between the CVs. There is, however, evidence of site-to-site variation for mean seed weight (Snoad and Arthur, 1976) and for induced variation in mean seed weight due either to surgical reduction of the seed number developing per pod or to reduction of the number of developing pods (D. M. Harvey, personal communication). When environmental resources are nonlimiting, both the potential yield and

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

259

the proportion of this potential yield that is realized as actual yield are determined by the plant genotype. When plants are grown in crops, however, both the potential and the actual yields of an average plant may be greatly modified by the effect of interplant competition for diminishing resources. At low planting densities, when interplant competition is not too severe, there appears to be a close relationship between the potential and the actual yield. The proportion of the potential yield realized is therefore relatively stable. Data are presented in Table VI for the effect of different planting densities on the three representative genotypes of the leafless phenotype described in Section II,B. At low planting densities (16-44.4 plants/m’) the proportion of the potential yield realized was maintained at approximately 35%, even though the economic yield per average plant decreased over this density range (Fig. 5a). This relationship between potential and actual yield indicates that it is changes in potential yield that determine changes in actual yield at these low planting densities. It follows, since all of the potential yield components are relatively stable to environment, with the exception of the number of flowering nodes, that this potential yield component is responsible for declining yields as planting densities increase over this range. This conclusion is supported by the stability of three of the actual yield components, the proportion of flowers producing pods, the proportion of ovules producing seed, and the mean weight per seed, over these low planting densities (Table VI). At the high planting densities (100 and 400 plants/m2) the proportion of the potential yield realized decreased to 29 and 13%, respectively (Table VI). This Table VI Effect of Planting Density on the Proportion of Potential Economic Yield and Potential Economic Yield Components Realized per Unit Area for Leafless Peas“ Planting density (plants/m’)

16

25

44.4

100

400

% Potential

yield realized

35

36

35

29

13

61

62

64

48

35

51

57

55

50

36

I00

100

93

88

85

% Flowers

producing pods % Ovules

producing seed .P Weight per seed as a proportion of iweight at 16 plants/m2 Harvest index

0.52

Mean of three leafless genotypes.

0.51

0.52

0.47

0.35

260

C. L. HEDLEY A N D M . J . AMBROSE

decline was accompanied by a decrease in all of the components contributing to actual yield, i.e., fewer flowers produced pods, fewer ovules produced seed, and there was also a small decline in the mean weight per seed. It is apparent from Table VI that the difference between potential and actual yield, over all planting densities, can be attributed to a stepwise reduction in the proportion of each potential yield component that is realized as actual yield. There appears to be some system within the plant for determining, at each stage in yield development, what the potential will be and what proportion of that potential will be achieved. Increased competition at high planting densities tends to decrease this proportion for each yield component. This effect was also observed when leafless genotypes were subjected to different shade treatments during development (Hedley and Ambrose, 1979). In this experiment shading was used to reduce the light falling on pot-grown leafless genotypes to either 30 or 80% of an unshaded control. In the high-shade treatment (80%) there was an obvious carbon source limitation that resulted in a significant reduction in the rate of dry weight increase per plant, in the node number per plant, and in the economic yield. The porportion of the potential yield realized by these plants, however, was similar to that in the unshaded controls, and was accompanied by a similar stepwise reduction in the proportion of each potential yield component producing actual yield. A similar stepwise reduction has been observed by Hardwick and Milboum (1967) for the conventional vining pea Dark Skinned Perfection. The data presented in Table VI are for the effect of increasing planting density on the proportion of potential yield components realized by an “average” crop plant. This disguises, however, the effect of planting density on the plant-toplant variation within populations for these components. The extent of variation can be expressed as CVs or the proportional relationship between the mean for each population and its standard deviation. We have studied plant-to-plant variation by comparing two “leafless” genotypes-BS 151, which has large seeds, and BS 4, which has medium-sized seeds-grown at a range of planting densities. The effect of population density on the CVs for flower number per plant (which incorporates the potential yield components for the number of flowering nodes and the number of flowers per node) and seed number per plant (which incorporates the actual yield components for the number of nodes with pods, the number of pods per node, and the number of ovules per pod producing mature seeds) are presented in Table VII. The CVS for flower number per plant for both genotypes remained stable over the whole planting density range. The CVs for seed number per plant increased at the higher planting densities in both genotypes. This increase in CV, however, occurred at 100 plants/m2 in the large-seeded BS 151 and at 400 plants/m2 in the smaller-seeded BS 4. This differential response could be predicted from the relationship between seed size and competitiveness discussed earlier. The different effect of planting density on the CVs of these two compound components may be a reflection of the extent of

DESIGNING “LEAFLESS” PLANTS FOR DFUED PEA CROP

26 1

Table VII Effect of Planting Density on the Coeflicientsof Variation for Flower Number and Seed Number per Plant for Two Leafless Pea Genotypes Coefficients of variation Planting density (plantdrn’) 16 25

44.4 100

400

Flowerdplant

Seeddplant

BS 151

BS 4

BS 151

BS 4

30 34 31 38 38

33 30 39 30 33

41 42 38 61 I22

28 45 34 31 62

interplant competition when the flowers and seeds are developing; interplant competition would be less during flower development than during seed development. Similar population effects are hidden in the crop harvest index changes induced by increasing planting density (Table VI). Although the crop harvest index is stable, at about 0.52, over the low planting densities and decreases at higher planting densities, there is no indication in this value of the vanation between individual plants in the population. To gain some insight into the plant-to-plant variation for partitioning within a population, we removed 1 m2 of plants from an agriculturally grown crop of the leafless pea BS 4.One square meter from this crop contained 203 plants, all of which were individually dry-weighted before the biological and economic yields were determined. When the seed weight for each individual is plotted against its biological yield, it is apparent that there is a relatively constant relationship irrespective of plant size (Fig. 17), and plant size ranged from less than 1 g/plant to more than 12 g/plant. It was also evident that more than 30% of the plants that contributed to the total biological yield, contributed nothing to the economic yield of the crop. These plants with few exceptions had the lowest biological yield. When the harvest index of each individual that contributed to economic yield was calculated and plotted against the seed weight per plant (Fig. 18), it became apparent that some plants with a high economic yield had high harvest indices (in excess of 0.6) compared with the overall crop harvest index of 0.5 1 . These high harvest indices were similar to those attained by plants when grown free of interplant competition. There was, however, considerable variation in the economic yield and harvest index of plants over the whole biological yield range. It may be assumed that these plants have fallen short, for some reason, of the potential maximum harvest index. There was also a relationship between how small an individual was and the extent to which it fell short of this potential maximum.

262

C. L. HEDLEY AND M. J . AMBROSE

10.0

8.0

-

m

CI

$ 6.0

0

P

\ c

.- . . ..

s m u

.-

4.0

. .. ..

2.0

*. 2 .o

. 4.0

6.0

Biological Y i e l d / P l a n t

8.0

10.0

12.0

(9)

FIG. 17. Relationship between the biological yield and seed weight of single plants from a I-m2 area o f a leafless pea crop.

From these few observations it would appear that interplant competition affects partitioning within the crop in two ways. First, it accentuates the differences between plant sizes within the population, such that very small plants are present that may not actually flower or, if they flower, cannot support the development of any seed. Second, very few of the plants that produce seed attain the efficiency of partitioning of a similar plant grown without interplant competition. D. INTERACTIONS BETWEEN YIELDCOMPONENTS

It is well known for many species that yield components interact and that these interactions are often negative. For example, there is a negative correlation in

263

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

tomato plants between the number of fruits per plant and the weight per fruit (Williams, 1959). Similarly in crested wheatgrass there is a negative correlation between seed size and seeds per spikelet and between seeds per spike and plant size (Dewey and Lu, 1959). Studies with soybeans have demonstrated that in100

80

60 w

X

*

Q)

‘0

**

**

c

5

H

e

em*

Q)

1

*

a*

-c

**

** * * * . )

40

m

*.**a

e

**

*I*#a

a

*

**

*

* *

*

*

**

a

* *

* *

a

a *

*

O

I

be

20

* *

2.0

1

I

1

4 .O

6.0

8 .O

Seed W e i g h t /Plant

(9)

FIG.18. Relationship between the seed weight and the percentage harvest index of single plants from a I-m’area of a leafless pea crop.

264

C. L. HEDLEY AND M . J . AMBROSE

creases in seed size can compensate for a reduction in pod number of 22%, such that maximum yield is maintained (McAlister and Krober, 1958). Adams (1967), using the navy bean (Phaseolus vulgaris), found low or zero correlations between yield components in well-spaced noncompetitive plants. Negative correlations were only found when plants were grown in competitive environments. Comparisons between high-, medium-, and low-yielding navy bean lines revealed that negative correlations between yield components were independent of actual yield levels (Adams, 1967). It may be assumed that similar interactions occur between the yield components of peas. In addition, genetic variation exists within Pisum sativum for all of these yield components and considerable variation exists between genotypes in the final composition of economic yield. Pea genotypes may differ for flowering time from very early-flowering types, flowering at about the seventh node, to late-flowering types that remain vegetative until 20 or more nodes have been formed. Once flowering has been initiated, genotypes may differ in the number of reproductive nodes that the plant may produce, from about four to an almost infinite number. There appears to be a strong interaction between the number of reproductive nodes that a genotypes produces and the number of pods produced at each node. Genotypes that have a reproductive indeterminate habit tend to have one or at the most two pods per node. In such types, the flower development at successive nodes is usually separated in time, such that flowers are fully developed at only one or perhaps two nodes at a time. Genotypes that produce very few reproductive nodes, however, tend to be multipodded at each node or the pods at successive nodes may develop simultaneously, or genotypes may have very few multipodded nodes, which develop in near synchrony. The asynchrony of the indeterminate reproductive habit ensures that competition between fruits developing at each node is kept to a minimum, and that plant growth and yield development are prolonged. Competition between fruits of genotypes with a determinate habit is more intense, and yield development and plant growth are more condensed in time. As well as variation for the number of fruits at a node and for the number of nodes with fruits, genetic variation also exists for the size and number of seeds within the pod. Variation between genotypes for weight per seed ranges from approximately 100 to 500 mg. Potential seed number (ovule number) ranges from about 6 to 12 ovules per pod. In genotypes where the final seed size is large the number of mature seeds per pod tends to be low, and conversely, where the final seed size is small the number of mature seeds tends to be high. Final seed number will also depend, however, on the ovule number, and it is apparent that small-seeded genotypes will only produce high seed numbers per pod if the ovule number is high. In very large-seeded types very few seeds develop to maturity within each pod and seed abortion is common even among seeds that are rela-

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

265

tively well developed (personal observation). The only real evidence of an interaction between seed size and seed number comes from experiments where young developing seeds have been surgically aborted. In such experiments the final size of the remaining seeds increases significantly (D. M. Harvey, personal communication).

V. IMPROVING THE EFFICIENCY OF THE PEA FRUIT A. T H E P O D

When plants are grown in conditions where interplant competition is strong, small improvements in the efficiency with which assimilates are utilized within the plant may be important in determining the final yield of the crop. Since the fruit forms a significant proportion of the total plant weight, consideration of its structure and functions must play an important part in the design of the ideal crop plant. The pod walls of legume fruits appear to serve two main functions, in addition to providing a suitable environment in which seeds can develop. The pod may act as a temporary storage organ for the developing seeds and also as an efficient organ for trapping and recycling C02 respired by the developing seeds. These functions appear to vary in importance depending on the legume species and possibly on the variety within a species. Soybean pods have been shown to be incapable of net C02 uptake from the atmosphere at any time during development (Quebedeaux and Chollet, 1975). The seeds and pod walls, however, may refix 50-70% of the COz respired by these tissues (Sambo et al., 1977). There is some evidence in this species of variation for the role of the pod as a temporary storage organ for the seeds. In a comparison between an early- and a late-maturing variety of soybean, 12.7% of the final seed weight of the early variety was derived by redistribution of assimilate from the pod, whereas in the late variety this proportion was only 1.8% (Thorne, 1979). The pods of Phaseolus vulgaris have also been shown to be inefficient at fixing atmospheric C02 compared with the leaf, but taking into account its recycling capacity the carbon fixation amounted to 26% of a similar area of leaf (Crookston et al., 1974). The pod of this species appears to act as a storage organ, but it is not known whether this is a temporary storage for later redistribution to the seed or whether it acts as a competitive sink and its assimilate not used during seed maturation. It has been suggested that variation may occur for this storage function between those varieties of Phaseolus vulgaris used for their edible pod and those grown for dried seed (Crookston el al., 1974).

266

C. L. HEDLEY AND

M. J . AMBROSE

Unlike other legumes, the illuminated pod of the lupin (Lupinus albus) can make CO, gains from the atmosphere for all but the last 2 weeks of its life (Pate et al., 1977). It is also suggested that the pod acts as a temporary reservoir and agent for remobilization of the respiratory products of the seed. The pea (Pisurn sarivurn) pod is committed to exporting assimilates, derived from CO, fixation, to the developing seeds (Lovell and Lovell, 1970). This carbon is mainly derived from fixation of COBrespired by the seeds into the pod cavity (Flinn and Pate, 1970). Most of the carbon required for seed development originates from the subtending leaf and stipule (Flinn and Pate, 1970), the photosynthetic activity of which is modulated by the growth rates of the pod and the seed (Flinn, 1974). Although the pea pod is capable of a net uptake of CO, from the atmosphere only during the very early stages of pod development (Harvey et al., 1976), its role in refixing and recycling carbon to the seed is substantial and accounts for up to 20% of the fruits assimilate requirement (Flinn et al., 1977). Both the photosynthetic carbon fixation enzyme, ribulose- 1,5-bisphosphate (RuBP) carboxylase and phosphoenolpyruvate (PEP) carboxylase, often associated with dark C 0 2 fixation mechanisms, have been demonstrated within the pod tissue of peas (Hedley er al., 1975). The distribution of these enzymes within the pod wall emphasizes the refixation role of the pod, the inner epidermis being rich in both carboxylase enzymes (Atkins el al., 1977). At high light intensities these enzymes are capable of fixing 60% of the CO, released by the seeds, although this proportion is reduced at the light levels present in the canopy (Atkins et al., 1977). It is likely that the type of pod that will be most efficient in the “leafless” canopy, with its increased light penetration and improved standing, will be very different in structure and physiology from that more suited to the leafed pea crop. A wide range of genetic variation is available for pod type in Pisum sativum. Large differences exist between genotypes for pod size and pod growth rate (Hedley and Ambrose, 1980), and also for pod wall thickness (Wellensiek, 1925b) and for the presence or absence of schlerenchyma layers (White, 1917). Comparisons, using genotypes that differed for wall structure or chlorophyll content, have been made to assess variation for carboxylase activity (Price and Hedley, 1980). The activity of both carboxylase enzymes was shown to vary with pod type and with pod age, the activity of the photosynthetic enzyme (RuBP carboxylase) correlating with the chlorophyll concentration. Yellow-podded types had lower activities of this enzyme but also had higher absolute levels of the dark fixation enzyme (PEP carboxylase). The PEP-carboxylase system was shown for all pods to comprise a far higher proportion of the total carboxylase activity than that normally found in leaf tissue exhibiting the C, photosynthetic system (Price and Hedley, 1980). As yet there is little information about the significance of different pea pod types for the seed yield of the plant. Even more important, there is no informa-

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

267

tion about the effect of pod phenotypes when incorporated into “leafless” plants. There is a suggestion from studies using near-isogenic lines for the uf and st genes, that the net CO, uptake of the pods from leafless plants is higher than that for corresponding leafed plants (Harvey, 1978). No clear explanation, however, has been put forward to explain this observation. There is also little information for peas on the role of the pod as a temporary storage organ. Although some studies suggest that pods are committed to the export of assimilate to seeds (Flinn and Pate, 1970), there is no reason to suppose that variation does not exist for this characteristic, especially among pods differing for wall thickness. B . THESEED

Ultimately it is the number and weight of individual seeds that survive through to maturity that determines the yield of a crop. Although the ability of a genotype to tolerate interplant competition is important in partitioning assimilate into economic yield, it is of equal importance for the seeds of such a genotype to tolerate the resulting intraplant competition. As with the competition between plants, if the competition between developing seeds for diminishing resources within the plant is high, then the number of seeds that succeed in developing through to maturity will be low. As discussed in Section IV,B, competition between seeds developing at different nodes will be reduced if a genotype has a reproductive indeterminate habit, and competition within reproductive nodes will be reduced if only a single pod develops at each node. The number and weight of seeds that develop within each pod, however, will be determined by the tolerance of each seed for its neighbors. Consequently, if seed variants that are less demanding of the plant’s resources can be found, then more seeds would be expected to develop successfully. In general, seeds within a pod develop asynchronously with the largest individuals in the center, tapering to very immature seeds at both ends (Fig. 19). This asynchronous growth is initiated very early in the development of the fruit. The cause of the asynchrony is not known, but it does not appear to be due to a lack of fertilization (Linck, 1961). As development continues there is a tendency for the seeds at both ends of the pod to abort, and for only the central seeds to continue development (Linck, 1961). When plants are grown in environments that induce intensive intraplant competition, further abortion occurs and it is the smallest seeds that appear to be most susceptible (personal observation). The interactions between individual seeds within a pod are in many ways similar to those between individual plants within a population. It is possible therefore to apply to the behavior of seeds toward their neighbors reasoning that is similar to that applied to explain the responses of individual plants within the crop (Section 111,C). It can be assumed that there is a finite rate for the translocation of assimilate

268

C. L. HEDLEY A N D M. J . AMBROSE

FIG. 19.

Distribution of seed size within a pea pod

from the plant into the pod. This rate of assimilate input will therefore only supply a finite sink demand from the seeds, the sink demand being equivalent to the sum of the growth rates of all the developing seeds within the pod. Initially the sum of the individual seed growth rates will not exceed the assimilate supply and all of the seeds will begin to develop. As the seeds within the pod develop, the sum of the growth rates may eventually exceed the rate of assimilate input into the pod and competition between seeds will occur. As with plants within a competitive sward, the seeds with the highest growth rate, situated in the center of the pod, will continue to grow at the expense of the seeds with lower growth rates, situated at the two ends of the pod. If this interseed competition occurs at a critical stage in development, then the seeds with the lowest growth rates will abort. In large-seeded types abortion occurs even among quite large developing seeds and accounts for the low numbers of large seeds reaching maturity (personal observations). It is apparent that the key to improved tolerance between seeds within the pod is to ensure that the total seed sink demand throughout development does not exceed the assimilate input into the pod. By definition this will entail maintaining relatively low seed growth rates. This can be achieved by selecting seed types that have a decreased relative growth rate (RGR) and that would therefore, for a given seed size, have an increased duration of growth (Fig. 20a). The overall effect of a reduced RGR will be to reduce the absolute growth rate and hence the

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

269

sink demand of the seed. Relative growth rate is determined by the interaction of the rate of cell division and expansion with environment. Seed phenotypes with a lower RGR would either have decreased rates of cell division and/or cell expansion, or more likely the proportion of seed growth attributed to cell division would be reduced relative to that attributed to cell expansion. This will lower the seed RGR because dividing cells have a higher RGR than expanding cells. Another possibility for improving interseed tolerance, especially later in seed

a

b

Time

I CC 0

m

Q) Q)

In

Time FIG. 20. Theoretical variation for seed growth. (a) Seeds of similar size differing for relative growth rate. (b) Seeds differing for size but with similar relative growth rate. (c) Seeds differing for size and for relative growth rate. L, large seed; S , small seed.

270

C. L. HEDLEY AND M. J . AMBROSE

development, is to select small seeds, which by definition will mature while their growth rates are relatively low (Fig. 20b). For a given ovule number per pod the best seed phenotype will therefore be small-seeded with a low RGR (Fig. 20c). In preliminary studies using six-leafed genotypes, we have found significant differences in the RGR of seeds during the early part of development (Hedley and Ambrose, 1980). We have not, as yet, studied the effect of this variation on competition between seeds, or the effect of interplant competition on the different seed phenotypes. Selection for a character such as RGR will be difficult because the seed is not a genetically homogenous structure. The seed is composed of the embryo, the testa, which is maternal, and the endosperm, which is triploid and composed of two maternal and one paternal genome (Cooper, 1938). The seed phenotype is determined by the development of these tissues and by the developmental interactions between them. The effect of the maternal influence on the development of the seed can best be observed from the differences between reciprocal crosses of genotypes that differ for seed size. In such crosses the resulting F, seeds usually resemble the maternal parent in size (Davies, 1975). Such a maternal control may act via the testa determining or controlling the transfer of nutrients to the embryo (Murray, 1979, 1980). In an attempt to understand the complex development of the seed, we have studied the development of the component parts, in a range of genotypes (Hedley and Ambrose, 1980). The physical relationship between the embryo and embryo sac (the endosperm-filled vacuole formed within the developing testa) is primarily concerned with the rate at which the embryo and embryo sac volumes expand relative to each other (Fig. 21). The initial increases in volume of the embryo and embryo sac are exponential and therefore linear on a logarithmic scale. Variation was found between the genotypes for the slopes of both lines and for the separation in time for the initiation of exponential embryo growth relative to that of the embryo sac. The slope for increase in embryo volume was always considerably greater than that for the volume of the embryo sac. The difference in time between the initiation of the two exponentials, however, determined that initially the absolute volume increase of the embryo sac was greater than that of the embryo. The absolute difference between the two volumes therefore increases initially and liquid endosperm accumulates. The difference in slope between the embryo and embryo sac, however, determines that a point is eventually reached where the absolute rates of volume increase are the same. This point corresponds to the maximum volume of endosperm within the seed. After this point endosperm is absorbed, presumably by the developing embryo. Variation between genotypes in the growth of the seed and in the final seed size are determined by differences in the slopes of the exponentials and the difference in time between their initiation. An understanding of the mechanisms controlling these three variables will be necessary before seeds with specific

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

27 1

MIN. END.

Time FIG. 21. Relationship between the expanding embryo sac (es) and embryo (e), showing point of maximum (max. end.) and minimum (min. end.) endosperm volume. r is the difference in time between the onset of embryo sac and embryo exponential growth.

growth characteristics can be selected. This forms the basis of our present research into seed development. As part of these investigations we have studied the cellular composition of embryos from a range of genotypes. Differences have been found between genotypes in the number of cells within a given embryo volume, suggesting differences in the proportion of cells within the embryos that are expanding. Such variation, as stated earlier, may affect the RGR of the embryo. It is not known, however, if these observations are due to intrinsic genetic differences between embryos or if they are the result of a maternal influence. There is evidence from reciprocal crosses that the cell number of embryos is greatly modified by the maternal parent (Davies, 1975). As well as selecting seed phenotypes that are less demanding per unit time of the assimilate input from the plant, it may also be possible to improve the

272

C . L. HEDLEY AND

M. J . AMBROSE

efficiency with which seeds utilize assimilate. As with the pod wall, both the testa and embryo have significant levels of PEP-carboxylase activity as well as much lower levels of RuBP-carboxylase activity. These enzymes may act to reduce respiratory losses and recycle carbon within the seed (Hedley et ul., 1975). More conclusive evidence of such a system has been found for the developing seeds of lupin (Lupinus ulbus; Atkins and Flinn, 1978). The significance of this system for the developing embryo is not known, but it can be suggested that the provision of C4 acids from such a recycling system will be important in the synthesis of amino acids at a time when protein synthesis is high.

VI. A PLANT IDEOTYPE FOR IMPROVING YIELDS OF DRIED PEAS A. THEPLANT

The preceding sections have described some of the problems associated with the “leafless” phenotype, the main problem being the necessity to grow leafless plants at high planting densities in order to attain a sufficiently high biological yield per unit area. We have also suggested how the efficiency of the plants within the dried pea crop can be improved. Many of these suggestions will apply irrespective of whether or not the crop plant has a leafless phenotype. If we persist with the leafless phenotype (ufafstst), it is evident that the ideotype must be tolerant of high planting density and therefore by definition must be a relatively uncompetitive plant. The relationship between a plant’s competitiveness and its growth rate determines that a plant with a reduced growth rate is required. Unless variants can be found that grow at a reduced relative growth rate, which seems unlikely, plant growth rate can only be reduced by selecting plants that have developed from a small embryonic axis. This will determine that the ideotype will have small seeds, unless the relationship between the size of the embryonic axis and seed size can be broken. A modification to the leafless model, and one that we are now considering, is to incorporate the gene for normal stipule size (St) while maintaining the gene for converting leaflets into tendrils (uf).There is some evidence that this phenotype (ufafStSt) has a higher growth rate than leafless plants of comparable seed size (Snoad, 1981), but it will hopefully maintain some of the improved canopy characteristics of the leafless phenotype. It is possible that such a modification will overcome the absolute requirement for high planting densities and many of the problems, both economic and physiological, that accompany such densities. Even if such a modified model proves successful, we have no reason to suggest

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

213

that other features of our crop ideotype, incorporated to improve plant tolerance to the crop environment, will change. Ideally the plant should have a genetically determined nonbranching habit. Branches do not add significantly to economic yield at high planting densities, although they do utilize environmental resources and increase biological yield per unit area. If such a plant type cannot be identified, then plants that branch late in development should be selected, since such plants will be inhibited physiologically from branching by competition within the sward. The ideotype will be relatively early-flowering so that partitioning of assimilate into reproductive structures is initiated when competition between plants is low. Early flowering should be coupled with a reproductive indeterminate habit. This will allow the ceiling biological yield per unit area to be attained and will increase the duration of assimilate partitioning into economic yield. An increased duration of partitioning will reduce the competition between yield components during development. Each reproductive node will contain a single pod selected for maximum efficiency of carbon refixation and recycling. The seeds will have a low demand per unit time of available resources, and therefore will have a reduced relative growth rate and will mature when small. In addition, seeds that have an improved efficiency for recycling carbon will be incorporated into the ideotype, if such a system can be shown to have a beneficial effect on yield. Any incompatibility between the seed’s dual role as the main reproductive sink and as an embryo plant must be taken into account, and may result in an appropriate compromise. B . BREEDING STRATEGY

The overriding effect of interplant competition on plant characters, such as the number of reproductive nodes and overall plant size, determines that the assessment of an individual’s suitability as a crop plant must be made in an environment akin to that encountered by the plant within the crop. If, however, a segregating population of individuals that differ for competitiveness is grown as a microplot at commercial planting densities, then strongly competitive individuals will thrive relative to weak competitors. Therefore in such an environment the weak competitors would perform much worse and the strong competitors much better than if each were grown in a crop of genetically similar individuals. This presents the breeder with a problem, since for reasons discussed earlier, those individuals within a segregating population that are most likely to make successful crop plants will be weak competitors. In order that such plants not be discarded early in the breeding program, it is essential that selection be delayed until

274

C.

L. HEDLEY AND M.J . AMBROSE

the performance of each segregant can be assessed in a sward composed of like individuals. To overcome this problem, the leafless dried pea breeding program utilizes a technique termed single-seed descent (Snoad, 1980). Using this technique, selection can be delayed until relatively pure breeding lines exist for each original segregant. In practice a preliminary screening of segregants for characters not affected by the crop environment, such as node to first flower and ovule number per pod, occurs and only these selected individuals are used for the single-seed descent program. Although growing such pure-lined segregants in microplots overcomes to some extent the problem of assessing genotypes in heterogeneous populations, estimates of yield per unit area from such a microplot will not reveal the competitiveness of a genotype. Neither does it overcome the problem of possible genotype-microplot interactions. If all genotypes are compared in microplots at the same planting density, then there will be a tendency to select genotypes that suit this environment. Potentially higher-yielding genotypes that optimize at different planting densities may be overlooked. As discussed in Section III,A, a better estimate of an individual’s potential as a crop plant can be gained from a comparison made in a competitive and noncompetitive environment. A simplified system could entail comparing the performance of “edge” plants in the microplot with those in the center.

VII. CONCLUSIONS

In this article we have outlined our studies on designing an efficient dried pea crop plant, based on the “leafless” phenotype. Our approach has been to study existing “leafless” models grown in simulated crop environments, and to use this information to indicate where more detailed plant physiological studies are required. Of necessity, we have concentrated initially on studies of the plant grown in the crop environment, and much of this first review concerns this aspect of our work. These crop studies have highlighted some of the problems associated with the leafless phenotype, as well as indicating some of the characters that can be incorporated into the ideal crop plant. The high planting densities required to attain good yields with the leafless pea make it a necessity to select plants with reproductive sinks that are tolerant of competitive environments. These observations and conclusions have stimulated further studies on the effect of high planting densities on the interactions between yield components. They have also directed us to study ways of improving the efficiency with which assimilates are utilized within the plant. The seed plays a key role in these more fundamental studies because of its dual function as the unit

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP

275

of economic yield and as the unit from which the crop is derived. Therefore we are studying how seeds develop and whether seed phenotypes can be selected that are more suited to the crop plant. In addition, we are investigating more fully the relationship between seed phenotype and the rate of development of the crop. Information from the more fundamental studies of the plant will ultimately be used to produce new plant models that can be used for crop studies. It is often said that fundamental plant physiology has contributed little to the breeding of crop plants. By integrating such studies, however, with observations made within the crop, we hope to be more successful in incorporating plant physiological information into a breeding program.

REFERENCES Adarns, M . W . 1967. Crop Sci. 7 , 505-510. Atkins, C. A,, and Flinn, A. M. 1978. Plant Physiol. 62, 486-490. Atkins, C. A., Kuo, J . , Pate, J. S., Flinn, A. M., and Steele, T. W . 1977. Plunf Physiol. 60, 779-786. Black, N . J. 1957. Ausf. J . Agric. Res. 8, 335-351. Bleasdale, J . K . A , , and Thompson. R. 1960. Rep. Nut. Veg. Res. Sfa. pp. 28-29. Bleasdale, J. K . A., and Thompson, R. 1961. Rep. Nut. Veg. Res. Sta. pp. 36-37. Bleasdale, J. K . A., and Thompson, R. 1962. Rep. Nut. Veg. Res. Sta. pp. 35-36. Bleasdale, J . K . A., and Thompson, R. 1963. Rep. Nut. Veg. Res. Sta. pp. 39-40. Bleasdale, J. K . A., and Thompson, R. 1964. Rep. N N I . Veg. Res. Sta. p. 36. Brown, R., Cooper, B., and Blaser, R. E. 1966. Crop Sci. 6 , 206-209. Cooper, D. C. 1938. Bot. Gal. 100, 123-132. Crookston, R. K . , O’Toole, J., and Ozbun, J . L. 1974. Crop Sci. 14, 708-712. Crookston, R. K.. Treharne, K . J.. Ludford, P., and Ozbun, J. L. 1975. Crop Sci. 15, 412-416. Cruzat, M. B., Cafati. C., and Bascur, G. 1976. Agric. Tec. (Chile) 36, 116-121. Davies, D. R. 1975. Planru 124, 297-302. Davies, D. R. 1976. Appl. Biol. 2, 87-127. Davies, D. R. 1977, Sci. Prog. (Oxford) 64, 201-214. Dewey, D. R., and Lu, K . H. 1959. Agron. J . 51, 515-518. Donald, C. M. 1951. Ausf. J. Agric. Res. 2, 355-376. Donald, C. M. 1961. Symp. Soc. Exp. Biol. 15, 282-313. Donald, C. M. 1962. J. Ausf. Insr. Agric. Sci. 28, 171-178. Donald, C. M. 1963. A&. Agron. 15, 1-118. Donald, C. M. 1968. Euphyricu 17, 385-403. Donald, C. M . , and Hamblin, J. 1976. A h . Agron. 28, 361-405. Edmeades, G. 0 . . and Daynard, T. B. 1979. Can. J. Plant Sci. 59, 561-576. Flinn, A. M. 1974. Physiol. Plant. 31, 275-278. Flinn, A. M., and Pate, J . S. 1970. J . Exp. Bor. 21, 71-82. Flinn, A. M., Atkins, C. A , , and Pate, J . S. 1977. Planf Physiol. 60, 412-418. Gane, A. J . , King, i . M . , and Gent, G . P. 1971. “Pea and Bean Growing Handbook.” P.G.R.O., Peterborough, England. Gottschalk, W . 1976. Pisum Newsleft. 8, 18. Gritton, E. T., and Eastin, J . A. 1968. Agron. J. 60, 482-485. Hardwick, R. C., and Milbourn, G. M. 1967. Agric. Prog. 42, 24-31.

276

C. L. HEDLEY AND M. J. AMBROSE

Harper, J. L. 1961. Symp. SOC.Exp. Eiol. 15, 1-39. Harvey, D. M. 1972. Ann. Bot. 36, 981-991. Harvey, D. M. 1974. Ann. Bot. 38, 327-335. Harvey, D. M. 1978. Ann. Eot. 42, 331-336. Harvey, D. M. 1980. Ann. Eot. 45, 673-680. Harvey, D. M., and Goodwin, J . 1978. Ann. Eot. 42, 1091-1098. Harvey, D. M., Hedley, C. L..and Keely, R. 1976. Ann. Eot. 40, 993-1001. Hedley, C. L., and Ambrose, M. J. 1979. Ann. Eot. 44,469-478. Hedley, C. L., and Ambrose, M. J. 1980. Ann. Eot. 46, 89-105. Hedley, C. L., Harvey, D. M., and Keely, R. J . 1975. Nature (London) 258, 352-354. Holliday, R. 1960a. Nature (London) 186, 22-24. Holliday, R. 1960b. Field Crop Abstr. 13, 159-167. Holliday, R. 1 9 6 0 ~Field . Crop Ahstr. 13, 247-254. Hozumi, K., Koyama, H., and Kira, T. J. 1955. J. Inst. Polyfech. Osaka City Univ. Ser. D 6, 121- 1 30. Johnston, T. J., Pendleton, J. W., Peters, D. B., and Hicks, D. R. 1969. Crop Sci. 9, 577-581. Kerby, T. A., Buxton, D. R., and Matsuda, K. 1980. Crop Sci. 20, 208-213. Kira, T., Ogawa, H., and Sakazaki, N. 1953. J. Inst. Poly. Osaka City Univ. 4, 1-16. Koyama, H., and Kira, T. 1956. J. lnst. Polytech. Osaka City Univ. Ser. D 7, 73-94. Kruger, N . S . 1973. Q . J . Agric. Anim. Sci. 30, 25-38. Kruger, N. S. 1977. Q . J. Agric. Anim. Sci. 34, 35-52. Linck, A. J. 1961. Phytomorphology 11, 79-84. Lovell, P. H., and Lovell, P. J. 1970. Physiol. Plant. 23, 316-322. McAlister, D. F., and Krober, D. A. 1958. Agron. J. 50, 675-677. Meadley, J. T., and Milbourn, G. M. 1970. J. Agric. Sci. 74, 273-278. Milbourn, G. M., and Hardwick, R. C. 1968. J . Agric. Sci. 70, 393-402. Monti, L. M.. and Frusciante, L. 1978. Genet. Agric. 32, 365-373. Murray, D. R. 1979. Plant Physiol. 64, 763-769. Murray, D. R . 1980. Ann. Bot. 45, 273-281. Nichols, M. A , , and Nonnecke, I. L. 1974. Sci. H o r t . 2, 113-122. Pate, 1. S., Sharkey, P. J., and Atkins, C. A. 1977. Plant Physiol. 59, 506-510. Price, D. N., and Hedley, C. L. 1980. Ann. Eot. 45, 283-294. Proctor, J. M. 1963. J . Agric. Sci. 61, 281-289. Quebedeaux, B., and Chollet, R . 1975. Plant Physiol. 55, 745-748. Reynolds, 1. D. 1950. Agriculture (London) 56, 527-537. Salter, P. J., and Williams, J. B. 1967. J . H o r t . Sci 42, 59-66. Sambo, E. Y., Moorby, J., and Milthorpe, F. L. 1977. Aust. J . Plarit Physiol. 4, 713-721. Snoad, B. 1972. Annu. Rep. John I m e s Inst., Norfolk No.63, pp. 29-33. Snoad, B. 1974. Euphytica 23, 257-265. Snoad, B. 1980. A D A S Q . Rev. 37, 69-86. Snoad, B. 1981. Sci. H o r t . 14, 9-18. Snoad, B., and Arthur, A. E. 1973a. Euphytica 22, 327-337. Snoad, B., and Arthur, A. E. 1973b. Euphytica 22, 510-519. Snoad, B., and Arthur, A. E. 1974a. Euphytica 23, 105-1 13. Snoad, B., and Arthur, A. E. 1974b. Theor. Appl. Genet. 44, 222-231. Snoad, B., and Arthur, A. E. 1976. Theor. Appl. Genet. 47, 9-19. Snoad, B., and Davies, D. R . 1972. Span 15, 87-89. Snoad, B., and Gent, G. P. 1976. Annu. Rep. John lnnes Inst., Norfolk No 67, pp. 35-36. Thorne, J . H. 1979. Agron. J. 71, 812-816. Watson, D. J . 1947a. Ann. Eot. 11, 41-76.

DESIGNING “LEAFLESS” PLANTS FOR DRIED PEA CROP Watson, D. J . 1947b. A n n . Bot. 11, 375-407. Wellensiek, S . J . 1925a. Bihliogruphiu 2, 343-476. Wellensiek, S. J . 1925b. Geneticu 7, 1-64. White, 0. 1917. Proc. Am. Philos. Soc. 56, 487-588 Williams, W. 1959. Nurure (London) 184, 527-530.

277

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY, VOL. 34

LOW-INPUT TECHNOLOGY FOR MANAGING OXISOLS AND ULTISOLS IN TROPICAL AMERICA Pedro A. Sanchez* and Jose G. Salinast *Soil Science Department, North Carolina State University, Raleigh, North Carolina and ?Tropical Pastures Program, Centro lnternacional de Agricultura Tropical, Cali, Colombia

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Acid Soils of the Tropics . . . . . . . . . . . . . , , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Conceptual Basis of Low-Input Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Main Components of Low-Input Technology . , . . . , . . . . . . . . , . , , . , . , , . , . , . , . ............................... 11. Site Selection , , , . . . . . . , , , , . , , . , . . . . , . 111. Selection of Acid-Tolerant Germplasm . . . . . . . . . . . . . . . , . , . , . . . . . , . . . . . A. Annual Food Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..,.............. ......................................... ............. C. Grass and Legume Pastures . . . . . . . . . . . . . . . . . . . . . . . D. Conclusions . . ......................................... 1V. Development and Maintenance of Ground Cover . . . . . . . . . . . . . A. Land Clearing Methods in Rain Forests., , , , , , , . , . . . . . . . . . . . . . . . . . . . . . . . . . B. Soil Dynamics after Clearing Tropical Rain Forests . . . . . . . . . . . . . . . . . . . . . . . . . C. Land Preparation and Plant Establishment in Rain Forests . . . . . . . . . . . . . . . . . . . . D. Land Clearing Methods in the Savannas , . , , , , . . . , . , . , . , . . , , , . . . . . . . . . . . . . E. Crop and Pasture Establishment in Savannas ........................... F. Maintenance of Established Pastures . . . . . . . . . . . . . . . . . . . G . Mulching, Green Manures, and Managed Fallows . . . . . . . . . . . . . . . . . . . . . . . . . . H. Intercropping and Multiple Cropping Systems , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Management of Soil Acidity . . . . ...................... A. Lime to Decrease Aluminum .................................. B. Lime as Calcium and Magnesium Fertilizer , , , , . , . . . . . . . . . . . . . . . . . . . . . . . . . C. Selection of Aluminum-Tolerant Varieties . . . . . . . . . . . . . . D. Selection of Manganese-Tolerant Varieties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Phosphorus Management.. . .................................. A. Rates and Placement Methods of Phosphorus Applicati B. The Need to Improve Soil Fertility Evaluation Procedu C. Use of Less Soluble Phosphorus Sources., , . , , , , . . . D. Decrease of Phosphorus Fixation with Liming . . . . . . . . . E. Selection of Varieties Tolerant to Low Levels of Availabl F. Potential Utilization of More Effective Mycorrhizal Associations . . . . . . . . . . . . . . . , ....... G. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279

280 281 285 293 293 295 2% 300 301 307 308 308 313 318 319 32 1 325 326 330 333 334 334 338 346 35 1 353 354 355 360 365 37 1 372 376 379

Copyright @ 1981 by Academic Press, tnc. All rights of reproduction in any form reserved.

ISBN 0-12-003734-7

PEDRO A. SANCHEZ AND JOS6 G . SALINAS

280

VII. Management of Low Native Soil Fertility .......................... 380 A. Maximum Use of Biological Nitrogen ............................. 380 B . Increase of the Efficiency of Nitrogen and Potassium Fertilization . . . . . . . . . . . . . 382 C. Identification and Correction of Deficiencies of Sulfur and Micronutrients 384 D. Promotion of Nutrient Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386 E. Conclusions . . . . . . . . . . . . . . . . . ...................... 389 VIII. Discussion ..................................................... 390 . . . . . . . . . . . . . . . . . . . 390 ...................... s

1X. Summary . .

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

References . . . . . .

..........

393 395 391 398

I. INTRODUCTION The outcome of the race between world food production and population will largely be determined in the tropics, where most of the world’s undernourished people live. During the decade from 1965 to 1975 food production increased at a slightly faster rate than population in food-deficient countries (IFPRI, 1978). This achievement is due to a number of factors, among which the predominant agronomic one is the development and adoption of high-yielding varieties of several crops with improved agronomic practices. Most of these varieties were selected for their ability to produce high grain yields under conditions of little or no soil or water stress. Not surprisingly, their adoption has been most successful when grown on fertile, high-base status soils with sufficient fertilization and a reliable water supply. The elimination of soil constraints by applications of the necessary amounts of fertilizers and amendments can be considered as high-input soil management technology. Its basic concept is to change the soil to fit the plant’s nutritional demands. This high-input approach is largely responsible for our present world food production levels and undoubtedly must continue where economic conditions permit. The applicability of high-input soil management technologies, however, diminishes in marginal lands where soil and water constraints are not easily overcome at low cost. The rising price spiral of petroleum-related products since 1973 has further limited the economic feasibility of soil management technologies based on the intensive use of purchased inputs, particularly for farmers with limited resources in the tropics. Many research efforts in the tropics are now directed towards developing low-input soil management technology, which does not attempt to eliminate the use of fertilizers or amendments but rather attempts to maximize the efficiency of purchased input use through a series of practices. The basic concept of low-input soil management technology is to make the most efficient use of scarce purchased inputs by planting species or

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

28 1

varieties that are more tolerant to existing soil constraints, and thus decrease the rates of fertilizer applications while attaining reasonable, but not necessarily maximum yields. Although basic knowledge about plant adaptation to acid soil stresses has been available for decades (Levitt, 1978), systematic research for developing technology based on this concept began only a few years ago (Foy and Brown, 1964; Spain et al., 1975; NCSU, 1975; Foy, 1976a: Salinas and Sanchez, 1976; Wright, 1976; Foy and Fleming, 1978; Loneragan, 1978). These efforts have caused considerable controversy and some misinterpretations in the popular literature, such as the belief that “fertilizer-proof” cultivars can be developed and concerns about “mining” the soil of its available nutrients. The purpose of this review is to bring together examples of low-input soil management technology adapted to well-drained, acid, inherently infertile soils of the American tropics classified mainly as Oxisols and Ultisols. These examples are components of overall production systems, but seldom have all the necessary components been developed for one specific farming system. Most of the examples are drawn from tropical America, reflecting the authors’ experience, without resting importance to related work performed in other parts of the world. Soil taxonomy terminology (Soil Conservation Service, 1975), including soil moisture regimes, will be used. A. ACID SOILSOF THE TROPICS

At the broadest possible level of generalization there are three main avenues for increasing food production in the tropics: increasing yields per unit area in presently cultivated regions, opening new lands to cultivation, and expanding irrigated land. The first two require the alleviation or elimination of soil constraints, while the third eliminates water stress as the main constraint. Bentley et al. (1980) have examined these three alternatives and concluded that all three are needed, although the irrigation alternative is limited to relatively small areas and is the most costly of the three. There is little question that increasing productivity in land already under cultivation is the principal avenue for increasing world food production. Recent F A 0 estimates quoted by Dudal(1980), however, show that in order for per capita food production to remain at the present but largely inadequate levels, food production must increase by 60% within the next 20 years. Dudal further estimated that increasing yields on lands already in use is not sufficient; an additional 200 million ha of land must be incorporated into agnculture during the next two decades in order to accomplish this goal. This amount is roughly equivalent to the present cropland area of the United States. Is this possible? The answer is largely dependent on the use made of the acid soils of the tropics.

282

PEDRO A. S h C H E Z AND Josh G.SALINAS

I . Extension and Importance The world is currently utilizing about 40% of its potentially arable land resources (Buringh er al., 1975). The greatest potential for expanding the world’s agricultural frontier lies in the tropical rain forest and savanna regions dominated by acid, infertile soils classified mainly as Oxisols and Ultisols (Kellogg and Orvedal, 1969; National Academy of Sciences, 1977a). These vast regions have a large proportion of favorable topography for agriculture, adequate temperatures for plant growth throughout the year, sufficient moisture year-round in 70% of the region, and for 6-9 months in the remaining 30% (Sanchez, 1977). The paramount limiting factors preventing widespread agricultural development in these areas are low native soil fertility and the limited transportation and market infrastructure. Table I shows the approximate extension of areas dominated by Oxisols and Ultisols in the tropics. As a whole, they account for about 1582 million ha or 43% of the tropical world. The almost equal proportion of Oxisols and Ultisols differ from previous estimates (Sanchez, 1976), as new information shows that there are less Oxisols than previously thought in Africa and Latin America. The sum of areas dominated by Oxisols and Ultisols, however, remains similar to previous estimates. The largest concentration of Oxisols occurs in the South American savannas, the eastern Amazon, and parts of Central Africa. These soils are generally located in old, stable land surfaces, which makes them attractive for mechanized agriculture. Ultisols are scattered over large areas of tropical America, Africa, and Southeast Asia. Many of these regions are being rapidly developed. There are other acid soils with similar properties and potentials included in other rows of Table I: acid, well-drained Inceptisols (Dystropepts); acid volcanic ash soils (Dystrandepts); and acid, well-drained red sands (Oxic Quartzipsamments). Excluded from consideration in this article are acid soils that are poorly drained and have an aquic soil moisture regime. Tropical America, at the broadest level of generalization, can be subdivided into two major regions in terms of farming systems and soil constraints (Sanchez and Cochrane, 1980). About 30% of tropical America (405 million ha) is dorninated by relatively fertile, high-base status soils that support dense populations. The remaining 70% of the tropical portions of the Western Hemisphere is dominated by acid, infertile soils of the orders Oxisols and Ultisols with relatively low population densities and mostly under savanna and forest vegetation. In spite of a widespread belief that Oxisols and Ultisols cannot support intensive and sustained agriculture in the tropics (McNeil, 1964: Goodland and Irwin, 1975), there is ample evidence that they can be continuously cultivated and intensively managed for growing annual crops (Sanchez, 1977; Marchetti and Machado, 1980), pastures (Vincente-Chandler et a / . , 1974), and permanent

283

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS Table I Generalized Areal Distribution of Soils in the Tropics"

Soil associations dominated by: Oxisols Ultisols Entisols Alfisols Inceptisols Vertisols Aridisols Mollisols Andisols Histosols Spodosols Total

Tropical America" (lot' ha) 502 320 124 183 204 20 30 65 31 4

Tropical Africa" (lot' ha)

Tropical Asia'' (IW ha)

10

I 5 3

15 286 75 I23 I69 66 23 9 I1 27 6

1493

I143

810

316 135 282 I98 I56 46 I

Tropical Australia" (lot' ha) -

Total (lot' ha)

Percentage of tropics 23 20 16 15 14 5 2 2

I

833 149 574 559 532 163 87 74 43 36 20

224

3670

100

8 93 55 3 31 33 0 0

1

I 1

Based on tabular data from FAO-UNESCO ( 1 97 I - 1979) with indicated modifications. *From 23"-23"S, updated by senior author. "Areas with more than 150 days of growing season. From Dudal (1980). "Includes temperate portions of India, Bangladesh, and Indochina plus Papua New Guinea. "North of the Tropic of Capricorn. From Sanchez and Isbell (1979).

crops (Alvim, 1976). This is also the case with Oxisols and Ultisols of Hawaii, and Ultisols of southeastern United States and southeastern China where they support large populations. 2 . Major Constraints

The major soil-related constraints of tropical America and its acid, infertile soil region are shown in Table 11, based on preliminary estimates. The most widespread ones in the Oxisol-Ultisol regions are chemical rather than physical, including deficiency of phosphorus, nitrogen, potassium, sulfur, calcium, magnesium, and zinc plus aluminum toxicity and high phosphorus fixation. The main soil physical constraints are low available water holding capacity of many Oxisols and the susceptibility to erosion and compaction of many Ultisols with sandy topsoil texture. Laterite hazards cover a minor areal extent and most of the soft plinthite occurs in subsoil layers in flat topography not prone to erosion. In contrast, the main soil constraints of the high-base status soil region of tropical America are drought stress, nitrogen deficiency, and erosion hazards (Sanchez and Cochrane, 1980). When the chemical soil constraints are eliminated by liming and application of

284

PEDRO A. SANCHEZ AND JOSk G. SALINAS

Table I1 Geographical Extent of Major Soil Constraints in Tropical America (23%-23"S) and in Regions Dominated by Acid, Infertile Soils" Acid, infertile soil region (1043 1 8 ha)

Tropical America (1493 I W ha) Soil constraint N deficiency P deficiency K deficiency High P fixation Al toxicity S deficiency Zn deficiency Ca deficiency Mg deficiency H,O stress > 3 months Low H,O holding capacity Low ECEC" High erosion hazard Cu deficiency Waterlogging Compaction hazard Laterite hazard Fe deficiency Acid sulfate soils Mn toxicity B deficiency Mo deficiency

lW

Percentage of total area

I@ ha

Percentage of total area

1332 1217 799 788 756 756 74 I 732 73 I 634

89 82 54 53 51 51 50 49 49 42

969 1002 799 612 756 745 645 732 739 299

93 96 17 64 72 71 62 70 70 29

626 620 543 310 306 169 I26 96 2

42 41 36 21 20 I1 8 6 0

583 577 304 310 I23 169 81

56 55 29 30 12 16 8

?

?

? ? ?

? ? ?

?

"Source: adapted from Sanchez and Cochrane (1980). "ECEC = Exch. A1 + Exch. Ca + Exch. Mg + Exch. K (Exch.

2 a? ?

=

0 ,? *? ?

exchangeable)

the necessary amounts of fertilizers, the productivity of these Oxisols and Ultisols are among the highest in the world. For example, Fig. 1 shows the annual dry matter production of elephant grass (Pennisetum purpureum) under intensive nitrogen fertilization in Ultisols.of Puerto Rico, where all other fertility constraints have been eliminated. This yield approximates the calculated maximum potential of tropical latitudes of 60 tondhdyr of dry matter according to DeWitt (1967). Another example is shown in Fig. 2, where excellent corn grain yields on the order of 6.3 tonslhdcrop were obtained on a sustained basis in clayey Oxisol from Brasilia, Brazil, when its high phosphorus requirement was satisfied by one broadcast application of 563 kg P/ha and the other chemical soil constraints were corrected by liming and fertilization.

285

2

FIG. 1. Dry matter production of Pennisetum purpureum c v . Napier cut for forage in Ultisols of the udic mountains of Puerto Rico under intensive management. (Source: Vicente-Chandler et at., 1974.)

These management strategies can be very profitable, even at present prices, when the market provides a favorable ratio of crop prices to fertilizer cost. Whenever economics and infrastructure considerations make these high-input strategies profitable, they should be vigorously pursued.

B. CONCEPTUAL BASISOF LOW-INPUT TECHNOLOGY

In the majority of acid soil regions in the tropics, favorable market conditions do not exist, either because fertilizers and lime are expensive or not available at

PEDRO A. S h C H E Z AND JOSE G. SALINAS

286

50r

J

3

f

V

0- 70 140

282

563

880

BASAL BROADCAST P APPLICATION ( kg/ ha 1 FIG. 2. Corn grain yield response to phosphorus applications on an Oxisol (Typic Haplustox) of the Cerrado of Brazil. Cumulative yield of six consecutive crops. (Adapted from NCSU, 1978.)

all, because transportation costs are excessive, or simply because the risks are too high. The first two situations are self-explanatory. The third one is illustrated in Fig. 3, showing the response to phosphorus by Phaseolus vulgaris in a Typic Dystrandept from Popayan, Colombia, with a high capacity to fix phosphorus. The optimum phosphorus application rate according to marginal analysis was 507 kg P/ha, taking into consideration the residual effects for two subsequent crops. When the costs were further analyzed, economists found that farmers needed to invest a total of U.S. $1500/ha/crop to approach these maximum yields and obtain a net profit of U.S. $375/ha (CIAT, 1979). Although this represents a 25% return on the investment, most farmers with limited resources are unwilling to make such an investment, considering the risk due to high variability in yields caused by drought, disease, insect attacks, and unpredictable price fluctuations. Low-input soil management technology is based on three main principles: ( 1 ) adaptation of plants to the soil constraints, rather than elimination of all soil constraints to meet the plant's requirements; (2) maximization of the output per unit of added chemical input; and (3) advantageous use of favorable attributes of acid, infertile soils. It should be emphasized that the elimination of fertilization is not contemplated.

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

287

I. Use of Plants Adapted to Soil Constraints The first basic concept of low-input soil management technology for acid soils is to alleviate or overcome certain soil constraints simply by using species or varieties that are tolerant to them. Among the soil constraints listed in Table I1 more knowledge is available on tolerance to aluminum toxicity, followed by tolerance to low levels of available soil phosphorus. Less information is available on tolerance to manganese toxicity and low levels of other nutrients. Figures 4 and 5 illustrate the concept with two different yield response patterns to liming in two savanna Oxisols. Figure 4 shows the differential response of two upland rice cultivars grown on an Oxisol of Carimagua, Colombia with a pH value of 4.5 and 80% aluminum saturation prior to lime application. The tall variety Colombia 1 produced twice the yield without lime as the short-statured IR 5. Colombia 1 responded positively only to the first increment of lime (0.5 ton/ha) and negatively to higher increments. Spain el al. (1975) attributed this behavior to a nutritional response to the calcium and magnesium content of lime and to lodging at higher lime rates. In contrast, IR 5, bred under high fertility

Economic

Sum of three crops

A

0

z \

u)

C

0

+

Y

u)

0

7L,nrn

I.

J

wP

I I I

I

z a

I

!

I

W

rn

0

I76

352

P APPLIED

528

704

880

( kg P i h a )

FIG. 3. Cumulative response of Phaseolus vulgaris grain yields to basal phosphorus applications and its residual effect to two consecutive crops on a Typic Dystrandept in Popayitn, Colombia. (Source: CIAT, 1979.)

288

PEDRO A. S h C H E Z AND JOS6 G. SALINAS % At SATURATION

Colombia I

0 0.5

2

4

6

LIME APPLIED (ton CaC03- equiv / ha)

FIG.4. Differential response to aluminum saturation and liming by two rice varieties grown on a Tropeptic Haplustox at Carirnagua, Colombia. (Adapted from Spain el a / . , 1975.)

conditions in the Philippines, produced a typical quadratic response to lime, attaining its maximum yield at the highest lime rate, which corresponded to pH 5.5 and 15% aluminum saturation. The maximum yield attained by the aluminum-sensitive IR 5 cultivar was lower than the maximum yield attained by the aluminum-tolerant Colombia 1 cultivar, which required less than one-tenth as much lime. The differential response shown in Fig. 4 shows an overwhelming advantage for the aluminum-tolerant cultivar. Figure 5 illustrates a less dramatic but perhaps more common type of differential response to lime. Two grain sorghum hybrids were grown at different lime rates in a Typic Haplustox near Brasilia, Brazil that had a topsoil pH value of 4.4 and 79% aluminum saturation at the time of planting (NCSU, 1976; Salinas, 1978). The Taylor Evans Y-101 hybrid produced about four times more grain without liming than RS-610. This difference decreased with increasing lime rate and disappeared at the highest lime rate, where both hybrids produced the same maximum yield of 6.8 tonslha. The dotted lines of this figures indicate considerable savings in lime required to obtain 80 and 90% of maximum yields. For 80% maximum yields, the aluminum-tolerant hybrid required 1.3 tons lime/ha and the aluminum-sensitive hybrid required 2.9 tons/ha. For 90% maximum yield, the lime requirements were 2.0 tons/ha for the aluminum-tolerant hybrid and 5.2 tons/ha for the aluminum-sensitive one. The use of aluminum-tolerant cultivars therefore can significantly decrease lime input without a sacrifice in yields at 80 and 90%of the maximum. These two examples illustrate the need for researchers to include more treat-

LOW-INPUTTECHNOLOGY FOR OXISOLS AND ULTISOLS

289

ments at lower input rates than used in the past in order to observe whether differential tolerance exists. If these experiments had not included rates of 0.5 or 1 ton lime/ha, the effects may not have been observed because cultivar differences tend to disappear at high input rates.

2. Maximization of Output p e r Unit o j Fertilizer Input Traditional methods used for determining optimum fertilizer rates are based on marginal analysis, where the optimum level is reached when the revenue of the last increment of fertilizer equals its added cost. This approach is designed to maximize yields and profit per unit area. A major disadvantage of this approach is that the optimum economic fertilizer rates frequently fall in the flatter portion of the response curve, where large increments in the fertilizer input cause small

7t

Taylor Evans Y - 101

1

.-

-

_/--

90% max. yield

I

I I I I

I

I

I I

I

LSD,,o, = 0.35 ton/ha

I

I I I I

I I

I I

0

I

I

I

I

2

L 4

8

LIME APPLIED IN 1972 ( t o n s l h a ) I

I

I

79

42

24 Oh

1

12

I

2

Al S A T U R A T I O N

FIG. 5 . Differential response of two grain sorghum hybrids to liming in Typic Haplustox of the Cerrado of Brazil. (Source: Salinas, 1978.)

290

PEDRO A.

SANCHEZ AND JOSE G. SALINAS

increases in yields. Given the uncertainties associated in predicting yields under tropical conditions, these small yield increases are seldom realistic. A common feature of yield response curves in Oxisol-Ultisol regions is that the amount of fertilizer required to produce 80% of the maximum or optimum yield is considerably lower than the amount required to reach the maximum or optimum point. In Fig. 3 the optimum level of phosphorus application according to marginal analysis is 507 kg P/ha. If only 80% of that optimum yield is desired, this amount decreases to less than half, 242 kg P/ha. Other examples from Oxisol-Ultisol regions presented in Table 111 show that fertilizer or lime rates decrease by 33-76% when the target yield is lowered to 80% of the maximum. This table includes two examples of the effect of phosphorus and lime applications for a sufficiently long period of time to adequately evaluate their residual effects. The reduction in input is on the order of 50-75% in these cases. Consequently, by lowering yield expectations, the cost of input use can be reduced by a considerable amount. Boyd (1970, 1974) in England and Bartholomew (1972) in the United States summarized large numbers of fertilizer response functions from all over the world and concluded that in most instances fertilizer response curves can be characterized by a sharp linear increase followed by a flat horizontal line. In essence, this approach follows Liebig’s Law of the Minimum. Several techniques have been developed to put this principle to practice in interpreting fertilizer response curves (Cate and Nelson, 1971; Waugh et ul., 1975; Waggoner and Norvell, 1979). These methods are now widely used in tropical America. A comparison of the linear approach versus the conventional marginal analysis with quadratic equations is shown in Fig. 6, using a wheat data set from Bolivia. This figure shows a lower recommended nitrogen application rate with the linear plateau model. This rate occurs at a point along the linear portion of the response curve where the efficiency of fertilization is highest, measured in terms of units of crop yield per unit of fertilizer input. One of the authors of this review used previously published data from a series of nigrogen response studies of rice in Peru to compare the two ways of developing fertilizer recommendations (Sanchez er n l . , 1973). The average nitrogen recommendation was 224 kg N/ha according to the quadratic model and 170 kg N/ha according to the linear response and plateau model. The differences in gross returns to fertilization were not significant, but the net return per dollar invested in fertilizer nitrogen was $8.80 in the linear plateau model versus $6.10 with the quadratic model (Sanchez, 1976). Although the applicability of the linear plateau model should be validated locally before using it as the basis for fertilizer recommendations, the concept of recommending rates that will produce the maximum output per unit of fertilizer input at an acceptable yield level is part of low-input technology.

Table I11 Decreases in Recommended Fertilizer and Lime Application Rates When Only 80%of the Maximum Yield Is Desired" Rate to reach

Location

Crop

Brasilia, Brazil Brasilia, Brazil Brasilia, Brazil Brasilia, Brazil Brasilia, Brazil Orocovis, Puerto Rico

Corn (6)* Corn (5) Corn (1) soybeans ( I ) Wheat (1) Elephant grass

Carimagua, Colombia Carimagua, Colombia Carimagua, Colombia Carimagua, Colombia Carimagua, Colombia

Cassava (42) Corn (20) Rice (96) Sorghum (240) Beans (49)

Input

Lime Lime Lime Lime Lime

Reduction of fertilizer rate 80% MY (%)

MY (todhdcrop)

MY (kglha)

80% MY (kdha)

7.0 5.6 4.9 3.2 2.4 53.0

563 8000 249 1200 800 1792

282 2000 60 300 200 746

50 75 76

8.0 3.2 2.8 3.1

6000 6OOo

1700 2200 3500 1800 4000

72 63 42 70 33

1 .o

"Examples from Oxisol-Ultisol regions. R. residual effects; MY, maximum yield. *Numbers in parentheses indicate number of crop harvests.

6000 6000 6000

75 75 58

Source NCSU (1978) NCSU (1978) NCSU (1978) CPAC (1976) CPAC (1976) VicenteChandler ef al. (1964) CIAT (1978) CIAT (1978) CIAT (1978) CIAT (1978) CIAT (1978)

PEDRO A. SANCHEZ AND JOSk G.SALINAS

292

LINEAR RESPONSE AND PLATEAU

2 2

QUADRATIC

I

1 .

2I

I

I

I

0

NITROGEN

60

120 4180 2 4 0 160

APPLIED (kg N/ha)

FIG.6. Determination of nitrogen recommendations for potatoes in a set of field experiments from Bolivia according to the linear plateau response and the conventional curvilinear models. Each dot is the mean of several field experiments in a given crop-crop-soil category. (Source: Adapted from Waugh e t a / . , 1975.)

It should be emphasized that this approach differs from the simple fertilizer trials of the Food and Agriculture Organization (FAO), which also advocate the use of lower fertilizer rates than that suggested by marginal analysis (Hauser, 1974). The difference is that with the linear response and plateau model the yields at recommended fertilizer rates are at the maximum yield, while the F A 0 trials normally consist of low fertilizer rates that seldom approach maximum yields. Both methods emphasize working on the linear portion of the fertilizer response curve that produces maximum output per unit input, but they differ on the expected yield levels. In addition to methods of determining fertilizer recommendations, there are a number of agronomic practices that also increase the efficiency of fertilizer use, such as better fertilizer sources, timing of application, and placement methods. These and other practices will be discussed in other sections of this review. 3 . Advantageous Use of Favorable Attributes of Acid, Infertile Soils

Many Oxisols and Ultisols in thier acid state have several positive agronomic factors that can be used advantageously. By keeping the soil acid, the solubility

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

293

of slowly available rock phosphate is higher than if the soil is limed, and weed growth is decreased considerably as compared with a limed and fertilized soil. Also, the low effective cation-exchange capacity (ECEC) of these soils favors the downward movement of applied calcium and magnesium to the subsoil. Examples of these observations will be discussed in later sections of this review.

c.

MAINCOMPONENTS OF LOW-INPUT TECHNOLOGY

Several concepts or techniques are being developed as building blocks of low-input soil management technology for Oxisols and Ultisols of the tropics. The following is a partial list, some of which can be combined for certain fzrming systems: 1 . Selection of most appropriate lands where, because of soil properties, landscape positions, and market accessibility, low-input technology has the comparative advantage over high-input technology. 2. Use of plant species and varieties that are more tolerant to the major acid soil constraints as well as being adapted to climatic, insect, and disease stresses. 3. Use of low-cost and efficient land clearing, plant establishment, cropping systems, and other practices to develop and maintain a plant canopy over the soil. 4. Manage soil acidity with minimum inputs, with emphasis on promoting deep root development into the subsoil. 5. Manage phosphorus fertilizers at the lowest possible cost with emphasis on increasing the efficiency of cheaper sources of phosphorus and prolonging the residual effects of application. 6. Maximize the use of biological nitrogen fixation with emphasis on acidtolerant Rhizohiurrz strains. 7. Identify and correct deficiencies of other essential plant nutrients.

11. SITE SELECTION The first step is to select the soils and landscape positions most appropriate for low-input technology. This involves avoiding the best lands in terms of high native fertility, irrigation potential, or close proximity to the markets. Most of these favored lands could be managed more effectively with high-input technologies. In tropical America unfortunately, this is not always the case. It is common to find many valleys where the best bottomland soils are under extensive low-input management systems while attempts are made to intensively farm the adjacent acid steeplands. In many cases, this is due to land tenure patterns.

294

PEDRO A. SANCHEZ AND JOSk G . SALINAS

Efforts should be made to intensify production in the soils with less acute chemical constraints. Large-scale evaluation schemes have improved our understanding of the areas suitable for low-input technologies in tropical America. Approximately 6% of the Amazon (30 million ha) is dominated by well-drained, high-base status soils classified as Alfisols, eutric Inceptisols, Vertisols, and Mollisols (Cochrane and Sanchez, 1981). Their higher native fertility gives the comparative advantage to intensive annual food crop production or to acid-sensitive export crops such as cocoa (Theohrorna cacao). In addition, the same study indicates that the Amazon has about 116 million ha of poorly drained soils either in floodplains or swamps, accounting for 24% of the basin. Some of the alluvial floodplain areas are already under intense use, such as many “virzeas” in Brazil and many “restingas” in Peru and Ecuador. Flood hazards, however, limit the production potential of the lower topographic positions. Also to be avoided, but for different reasons, are acid, infertile soils with severe physical limitations, such as shallow depth or steep slopes, and coarse sandy soils classified as Psamments or Spodosols and often called “tropical podzols. ” This latter group has extremely low native fertility and severe leaching and erosion hazards. These three groups cover about 41 million ha or 8.5% of the Amazon (Cochrane and Sanchez, 1981). The Psamments or Spodosols represent only 2.2% of the Amazon and combine the worst physical and chemical soil constraints. The total area to which low-input technology may apply in the Amazon region is therefore on the order of 275 million ha or 57% of the basin-mainly Oxisols and Ultisols with less than 8% slope. In the savanna regions of tropical America it is less difficult to identify the soils to be avoided, but the criteria remain the same. Many of the islands of high-fertility soils are already under intensive production, such as in the Eastern Llanos of Venezuela. Steep and shallow soils are readily recognized in the savanna landscapes. Large areas of seasonally flooded plains, such as parts of the Western Llanos of Venezuela and its extension into Colombia, and parts of the Beni of Bolivia and the Pantanal of Brazil, will require a different management strategy. In the savanna regions of tropical America, the Land Resource Study of Tropical America of Centro Internacional de Agricultura Tropical (CIAT, 1978, 1979) indicates that there are 71 million ha of Oxisols and Ultisols with less than 8% slopes (T. T. Cochrane, personal communication). These lands correspond to approximately 24% of the savanna regions and are primarily where the low-input technology described in this article can be applied. These estimates are conservative since it is possible to produce beef from legume-based pastures on steeper slopes. There are an additional 19 million ha of savanna Oxisols and Ultisols with 8-30% slope that could be used for such a purpose.

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

295

Although the above generalizations provide an overall picture, actual land selection is site-specific. Soil parameters per se are not sufficient for appropriate site selection. Land classification therefore is a more useful tool because it also considers climate, landscape, native vegetation, and infrastructure. The land systems approach used in CIAT’s Land Resource Study appears to be an appropriate method for evaluating the potential of these vast areas. Using a scale of 1 : I million, about 500 land systems have been identified so far, each representing a recumng pattern of climate, soil, landscape, and vegetation (Cochrane, 1979). Soils and climate are classified according to technical systems such as the moisture availability index (Hargreaves, 1977; Hancock et al., 1979) and the Fertility Capability Soil Classification System (Buol et a!., 1975). The data are assembled on computer tapes (Cochrane et al., 1979). Users of these tapes can examine computer-made maps of specific regions pinpointing one or several parameters, such as shallow soils or soils with more than 60% aluminum saturation at a specific depth. A modification of the Land Use Capability Classification System of the U.S. Department of Agriculture (USDA) has been developed in Brazil to take into account the realities of the tropical environment. Ramalho er al. (1978) redefined land capability classes in terns of the high-, moderate-, or low-input use. High input levels mean intensive use of fertilizers, lime, mechanization, and other new technology. “Moderate” input use implies limited fertilizer use and limited use of mechanization. This corresponds to the low-input technology cokept of this article. The “low” input use of Ramalho et al. implies primarily manual labor and few, if any, purchased inputs. This interpretive system has been applied to RADAM soil survey of the Brazilian Amazon (Ministtrio das Minas e Energia, 1973-1979). Consequently, for low-input technology soil management systems it is appropriate to select Oxisols and Ultisols with less than 8% slopes, to avoid the high-base status soils that can be put to more intensive use, and to avoid acid soils with severe physical limitations such as steep slopes, shallow depth, the Spodosols, and poorly drained or seasonally flooded soils.

111. SELECTION OF ACID-TOLERANT GERMPLASM A substantial number of plant species of economic importance are generally regarded as tolerant to acid soil conditions in the tropics. Many of them have their center of origin in acid soil regions, suggesting that adaptation to soil constraints is part of the evolutionary process. Also, some varieties of certain species possess acid soil tolerance although the species as a whole does not. These varieties have probably been selected involuntarily by farmers or plant

296

PEDRO A. SANCHEZ AND JOSE G. SALINAS

breeders because of their superior behavior under acid soil conditions. Examples of such involuntary selection are well documented in the literature (Foy et al., 1974; Silva, 1976; Martini et al., 1977; Lafever et al., 1977). The term “acid soil tolerance” covers a variety of individual tolerances to adverse soil factors and the interactions that occur among them. When mentioned in this article, this term only conveys a qualitative assessment of plant adaptation to acid soil conditions under low fertilizer or lime levels. Quantitative assessments of plant tolerances to acid soil stresses include tolerances to high levels of aluminum or manganese, and to deficiencies of calcium, magnesium, phosphorus, and certain micronutrients, principally zinc and copper. One example of an interaction among this group is that the calcium level of the soil solution can partially attenuate aluminum toxicity in many plant species (Foy and Fleming, 1978; Rhue, 1979). Tolerance to aluminum and low phosphorus stresses occur together in cultivars of wheat, sorghum, rice, and common beans but not in corn (Foy and Brown, 1964; Salinas, 1978). The physiological mechanisms involved, however, are beyond the scope of this article. The reader is referred to review articles in books edited by Wright (1976), Jung (1978), Andrew and Kamprath (1978), and Mussell and Staples (1979) for detailed discussions. Duke (1978) compiled a list of 1031 plant species of economic importance with known tolerances to adverse environmental conditions. Tolerance to “acid soils, “lateritic soils, and “aluminum toxicity” were included. The first two categories were qualitative assessments, and the last one identified only those species for which aluminum tolerance studies have been carried out. Duke’s list, although preliminary and incomplete, illustrates the broad base of acid-tolerant germplasm. A total of 397 species were listed as tolerant either to acid soils, lateritic soils, or to aluminum toxicity. Of these, 143 species met two of these criteria and 29 met all three. This last number reflects the limited number of species for which aluminum tolerance studies have been conducted. Tables IV and VI-VIII list selected species from Duke’s compilation that meet at least two of these criteria. These tables include modifications, additions, or deletions by the authors of this review, based on their own observations. ”

A . ANNUAL FOODCROPS

Table IV lists several of the world’s most important basic food crop species that have a considerable degree of acid soil tolerance. Seven of them-cassava, cowpea, peanut, pigeon pea, plantain, potato, and rice-can be considered acidtolerant species, although there are some acid-sensitive cultivars. The degree of knowledge as to the nature and degree of acid soil tolerance varies with the species. Cassava (Munihot esculentu) is more tolerant to high levels of aluminum and

LOW-INPUT TECHNOLOGY

FOR OXISOLS AND ULTISOLS

297

Table IV Some Important Food Crops Considered to Be Generally Tolerant of Acid Soil Conditions in the Tropics ~~

Generally tolerant species: Cassava (Munihor esculeiirtr) Cowpea (Vignu unguiculatu) Peanut (Aruchis hypoguea) Pigeon pea (Cajutius Cajun) Plantain (Mum purudisiuca) Potato (Solanurn tuherosum) Rice (Oyzu suriwn)

Generally susceptible species with acid-tolerant cultivars: Common bean ( Phaseolus vulgaris) Corn (Zea rnavs) Sorghum (Sorghum bicolor) Soybean (Glycine m a r ) Sweet potato (Ipornoeu baratas) Wheat (Triricum aesrivum)

manganese and to low levels of calcium, nitrogen, and potassium than many other species are (Gomes and Howeler, 1980; Cock, 1981). Although it has high phosphorus requirements for maximum growth, cassava apparently can utilize phosphorus sources that are relatively unavailable through mycorrhizal associations (Cock and Howeler, 1978; Edwards and Kang, 1978). Many cassava cultivars respond negatively to liming because of induced zinc deficiency at high soil pH levels (Spain et al., 1975). The ability of cassava to tolerate acid soil stresses may be due to an interesting mechanism. Cock (1981) observed that cassava leaves maintain an adequate nutritional status in the presence of low nutrient availability (Table V). Rather than dilute its nutrient concentration as in other plants, cassava responds to nutritional stress by decreasing its leaf area index. This is one reason why it is difficult to assess visual symptoms of nutrient deficiency in cassava growing on acid soils. Cowpea (Vigna unguiculata) is the major grain legume species considered to be most tolerant to acid soil stresses and specifically to aluminum toxicity (Spain et al., 1975; Munns, 1978). Under field conditions in Oxisols, cowpea commonly outyields other grain legumes such as soybean and Phaseolus vulgaris beans at high levels of aluminum saturation (Spain et al., 1975). As in other legumes, the acid soil tolerance of the associated rhizobia is an important as the acid soil tolerance of the cowpea plant (Keyser et a/., 1977; Munns, 1978). Peanut (Arachis hypogaea) is also regarded as tolerant to soil acidity (Munns, 1978), although it has a relatively high calcium requirement. Fortunately, small quantities of lime can provide sufficient calcium without altering the soil pH for maximum yields in Oxisols and Ultisols of the Venezuelan Llanos (C. Sanchez, 1977). Plantain (Musa paradisiaca) is one of the most important carbohydrate food sources in many areas of the humid tropics of America and Africa. Its tolerance to aluminum and general adaptability to acid soil stresses has been demonstrated in Ultisols of Puerto Rico (Vicente-Chandler and Figarella, 1962; Plucknett, 1978)

298

PEDRO A.

SANCHEZ AND JOSe G . SALINAS

Table V The Effect of Soil Fertility Level on Leaf Area Index and Leaf Nutrient Concentration of the Cassava Variety M Mex 59 6 Months after Planting''

Fertility level High Medium LOW

Nutrient concentration (%)

Nutrient content per unit of leaf area (rnddrn')

Leaf area index

N

P

K

N

P

K

5.39 3.54 I .65

3.69 3.68 3.52

0.25 0.19 0.18

2.00 I .40 0.73

18.9 20.2 21.7

I .28 1.04

10.3 7.7 4.5

1.11

"Source: Cock (1981).

and Oxisols of the Llanos Orientales of Colombia (CIAT, 1975). This crop, however, has relatively high requirements for nitrogen and potassium. Strong positive responses to nitrogen, phosphorus, potassium, magnesium, and micronutrient applications have been recorded (Caro Costas et ul., 1964; Silva and Vicente-Chandler, 1974; Samuels et al., 1975). The potato (Solanum tuberosum) has long been considered an acid-tolerant crop. Potato growers keep pH values below 5.5 in order to control the common scab organism, Streptomyces scorbies. Definite varietal differences in tolerance to aluminum have been established (Villagarcia, 1973). Disease problems in isohyperthermic temperature regimes are a greater limitation than acid soil constraints. Acid soil tolerance of rice (Otyzu sariva) under flooded conditions is normally not of significance. Except in some acid sulfate soils, the pH of most acid soils rises to 6 to 7 with flooding as a consequence of the chemical reduction of iron and manganese oxides and hydroxides (Ponnamperuma, 1972). Exchangeable aluminum is precipitated at these pH levels, thereby eliminating aluminum toxicity. In nonflooded systems, many rice varieties are quite tolerant to aluminum (as shown in Fig. 4) and/or low available levels of phosphorus (Spain et al., 1975; Howeler and Cadavid, 1976; Salinas and Sanchez, 1976; Ponnamperuma, 1977; Salinas, 1978). Also, varietal differences in tolerance to manganese toxicity and iron deficiency in acid soils have been identified (Ponnamperuma, 1976). In the Oxisol/Ultisol regions of Latin America, upland rice is generally considered to be more tolerant to acid soil stresses than corn is (Salinas and Sanchez, 1976; Sanchez, 1977). Other less common grain legume species are also considered to be tolerant to acid soil stresses in Oxisols and Ultisols of the tropics, although there is little quantitative information about their degree of tolerance. They are pigeon peas (Cajanus Cajun), lima beans (Phaseolus lunatus), winged beans (Psophocurpus tetragonolobus), and mung beans (Vigna rudiatu), according to Munns (1978).

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

299

Table IV also lists five species for which certain cultivars have been identified as acid soil-tolerant, but the species as a whole is not. Great variability exists in Phaseolus vulgaris beans, some cultivars being tolerant to aluminum toxicity and/or low phosphorus levels and some highly sensitive to both stresses (Spain et a / . , 1975; Whiteaker et a / . , 1976; Salinas, 1978; CIAT, 1977, 1978, 1979, 1980). In this species, disease and insect stresses, particularly in isohyperthermic temperature regimes, are more yield-limiting than soil constraints. Although corn (Zea mays) is considered by some investigators to be generally acid-tolerant (Rhue, 1979), lime response trials in the tropics tend to demonstrate the opposite. Nevertheless, several hybrids and composites possess a marked degree of aluminum tolerance and/or tolerance to phosphorus stress (Fox, 1978; Salinas, 1978). As a species, grain sorghum (Sorghum bicolor) is poorly adapted to acid soil conditions. Most of the varietal improvement work on this crop has been conducted in neutral or calcareous soils. Fortunately, cultivar differences in terms of aluminum tolerance do exist (Brown and Jones, 1977a). An example is shown in Fig. 5 adapted from Salinas (1978). Brown and Jones (1977a) have also reported marked cultivar differences to copper stress but none to manganese toxicity. Cultivar differences in tolerance to phosphorus stress also exist (Brown ef al., 1977). As a species, soybean (Glycine max) is probably less tolerant to overall acid soil conditions than most of the previously mentioned ones. Considerable varietal differences in tolerance to aluminum exist (Sartain and Kamprath, 1978; Muzilli et a / . , 1978; Miranda and Lobato, 1978) as well as intolerance to manganese toxicity (Brown and Jones, 1977b). Unlike the other grain legumes, rhizobia strains associated with soybeans tend to be more aluminum-tolerant than the plant (Munns, 1980). Aluminum tolerance in some sweet potato (Ipomoea baratus) cultivars has also been identified (Munn and McCollum, 1976; Toma, 1978). Some varieties grown in Pu&to Rico are quite tolerant to aluminum and manganese toxicity (Perez-Escolar, 1977). Wheat (Triticurn uestivum) is probably the species most thoroughly studied in terms of acid soil tolerance. It is an important crop in Oxisol-Ultisol regions of Latin America with isothermic or thermic soil temperature regimes. Varietal differences appear to be related to the soil acidity status where they were developed (Silva, 1976; Foy et al., 1974). For example, the well-known shortstatured CIMMYT wheat varieties, which were selected on calcareous soils of northern Mexico, perform poorly in Oxisols of the Cerrado of Brazil in comparison with varieties that were developed in Brazil, in spite of the latter’s inferior plant type (Salinas, 1978). Acid soil tolerance in such wheat cultivars is related to a joint tolerance to aluminum toxicity and low available soil phosphorus (Salinas, 1978; Miranda and Lobato, 1978). Other studies also show that

PEDRO A. S h C H E Z AND JOS6 G. SALINAS

300

aluminum-tolerant wheat varieties perform well at higher percent aluminum saturation levels than aluminum-tolerant soybean varieties in Oxisols (Muzilli et al., 1978). B. PERENNIAL A N D TREECROPS

Table VI lists some of the tropical fruit crop species considered to be tolerant to acid soil stresses. Some species like pineapple and cashew are well known for their adaptation to acid soils. Like the annual food crops, some species are severely affected by other constraints. For example, bananas are hampered by diseases and high potassium requirements; the citrus species are less productive in isohyperthermic temperature regimes than in cooler climates; mango requires an ustic soil moisture regime for high productivity. Some important perennial crops and forestry species adapted to acid soils in the tropics are listed in Table VII. Arabica coffee is very tolerant to aluminum but is sensitive to manganese toxicity (Abrufia et a l ., 1965). It prefers an isothermic soil temperature regime and an udic soil moisture regime. Robusta coffee is better adapted to isohyperthermic regimes but produces lower-quality coffee than arabica coffee. Among other perennial crops, rubber and oil palm are very well adapted to Oxisol-Ultisol regions, particularly those with udic isohyperthermic regimes (Alvim, 1981; Santana et a l . , 1977). Sugarcane is also generally tolerant to acid soil conditions (Abrufia and Vicente-Chandler, 1967) but requires large quantities of nitrogen and potassium to support high production levels. Table VI Some Important Fruit Crops Considered to Be Generally Tolerant to Acid Soil Conditions in the Tropics Name

Species

Source

Banana Carambola Cashew Coconut Granadilla Grapefruit Guava Jackfruit Lime Mango Orange Pineapple Pomegranate

Musa sapiensis Averrhoa caratnhola Anacardium occidentale Cocos nucifera Pussiflora edulis Citrus paradisi Psidium guajava Artocarpus heterophyllus Citrus aurantiifolia Manguifera indica Citrus sinensis Ananas cotnosus Punica grunutum

Authors Duke (1 978) Duke (1978) Duke (1978) Duke (1978) Duke (1978) Authors Duke (1978) Duke (1978) Duke (1978) Duke ( I 978) Duke (1978) Duke (1978)

30 1

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS Table VI1 Some Important Perennial and Forest Crops Considered to Be Tolerant to Acid Soil Conditions in the Tropics Name

Species

Source

Brazil nut Coffee Eucalyptus Gmelina Guarana Jacaranda Oil palm Peach palm Pepper, black Pine Rubber Sugarcane

Bertholleriu excelsa Coffea arubicu Eucalyptus grandvllorit Gmeliria urhoren Puulliniu cupanci Dalbergia nigru Elaeis guirieerisis Guilielrna gasipaef" Piper nigrurn P irius carihea Hevea hrusiliensis Sacchorurri officiriarurri

Duke (1978) Duke ( I 978) Alvim (1981) Alvim (1981) Alvim (1981) Alviin (1981) Duke ( 1 978) Alvim (1981) Duke (1978) Alvim (1981) Duke (1978) Duke ( 1978)

"Known as "pejibaye," "chontaduro," "pijuayo." and "pupunha.

"

Although many native wood species of the Amazon are tolerant to acid soil conditions, some of the most promising forestry species are imported from other regions. Gmelina arborea, Pinus caribea, Da lbergia nigra, and certain species of Eucalyptus have proven to be well adapted to Oxisols and Ultisols of the Brazilian Amazon without liming (Alvim, 1981). Other species native to the Amazon, such as Brazil nut (Bertholletia excelsa), guarana (Paullinia cupana), and peach palm (Guilielrna gasipaes), also have significant commercial potential. Several important tropical perennial crops are not included in the above list. Noteworthy among them are cocoa (Theobrorna cacao) and Leucaena leucocephala, a legume species with potential for grazing, browse, and firewood (National Academy of Sciences, 1977b). Neither of these two species are aluminum-tolerant (Alvim, 1981; Hill, 1970). Therefore, they are not adapted to acid soils with low inputs. Breeding for aluminum tolerance, however, is proceeding in both species. In the case of legume, selection for acid-tolerant Rhizobiurn strains is considered to be equal in importance to plant selection (CIAT, 1979; Munns, 1978). C . GRASSA N D LEGUME PASTURES

Extensive work on screening grass and legume pasture species for acid soil tolerance has been conducted in Australia and Latin America (Andrew and Hegarty, 1969; Andrew and Vanden Berg, 1973; Spain et al., 1975; Andrew,

302

PEDRO A .

SANCHEZ

AND JOSB G. SALINAS

1976, 1978, Helyar, 1978; CIAT, 1978, 1979, 1980, 1981; Spain, 1979). A fundamental difference of the work in the two continents is that aluminum toxicity is infrequent in the tropical pasture regions of Australia, while the opposite is the case in tropical pasture regions of Latin America (Sanchez and Isbell, 1979). The predominant acid soil stresses in tropical Australia are low phosphorus, sulfur, molybdenum, and to a lesser extent manganese toxicity. Aluminum toxicity, low phosphorus availability, and high phosphorus fixation are more important in tropical America. I . Alutninutn Tolerance

A wide range of CIAT’s forage germplasm bank is tolerant to high levels of exchangeable aluminum simply because much of it has been collected from acid, infertile soil regions of tropical America (Schultze-Kraft and Giacometti, 1979). An example of differential tolerance to aluminum of four common tropical grasses is shown in Fig. 7 from a solution culture study of Spain (1979). Brachiaria decutnbens even shows a slight positive response to the first increment of aluminum. Parzicurn tnaxirnurn exhibits strong tolerance up to one-half the aluminum concentration tolerated by Brachiaria decurnbens. In contrast, Cenchrus cifiaris, one of the most widespread tropical grasses in ustic but not acid areas of Australia, is severely affected by aluminum. This excellent grass is well adapted to nonacid soils, but to grow well in Oxisol-Ultisol regions it is

\o Hyparrhenio rufo

Cenchrus

0

I

I

1

J

0.5

I

2

4

Al IN SOLUTION

( ppm)

FIG. 7. Differential tolerance to aluminum in culture solution by four tropical grasses. (Source: Spain, 1979.)

LOW-INPUTTECHNOLOGY FOR OXISOLS AND ULTISOLS

8

C

GRASSES

3

\"

---Q--

2 \

-,d 3

v)

c

2'P'

z 0

- - - --- - - - - 4Hyparrhenia

-

a

0

t

303

rufa

Digitaria decumbens ,-*Grain sorghum _/-@ --#

--##T

/--

rT ---

I-

a

I

Zornia latifolia 728 St ylosa nt hes ca pitat0 1019 Desmodium ovalifolium 350

-------

____-c

,-a Centrosema plurnieri 470

D

l

Pueraria phaseoloides 9900

.' 'B

0

0.5

2

6

LIME APPLIED (tonslha) 1

1

90 85

I

I

60

15

% Al SATURATION FIG. 8. Field response to lime applications by several grass and legume forage species in an Oxisol of Carimagua, Colombia. Mean of four to five cuts for the grasses and first cut for legumes. (Adapted from Spain, 1979.)

necessary to completely neutralize the exchangeable aluminum by liming to about pH 5.5. A list of tolerant species is shown in Table VIII. Figure 8, also adapted from Spain (1979), shows responses to lime applications in an Oxisol of Carimagua, Colombia, with pH 4.5 and 90% aluminum saturation before liming. Acid-tolerant grasses such as Andropogon gayanus, Brachiaria decumbens, and Panicum maximum and the legumes Stylosanthes capitata and Zornia latifolia produced maximum growth either at 0 or 0.5 ton

304

PEDRO A. SbrNCHEZ AND JOSE G . SALINAS

lime/ha. The 0.5 ton/ha rate did not alter soil pH or aluminum saturation but provided calcium and magnesium to the plants. Their performance is clearly superior to aluminum-sensitive species such as grain sorghum and Centrosema plumieri, a legume clearly not adapted to acid soils. It is also relevant to point out that some aluminum-tolerant species do not grow vigorously in acid soils. This is the case of pangola grass (Digitaria decumbens), shown in Fig. 8. 2 . Low Levels of Available Soil Phosphorus

Phosphorus is the single most expensive input needed for improved pastures in Oxisol-Ultisol savannas (CIAT, 1979). It is not, however, the only nutrient that is deficient in these soils, but its correction is usually the most expensive one. No improved pastures are likely to be established or maintained without phosphorus fertilization in these savannas. In order to increase the efficiency of phosphorus fertilization, it is possible to select plants that have a lower phosphorus requirement for maximum growth than those presently used. Fortunately, aluminum tolerance and “low phosphorus tolerance” often occur jointly because the latter seems associated with the plant’s ability to absorb and translocate phosphorus from the root to the shoot in the presence of high levels of aluminum in the soil solution and/or root tissue (Salinas, 1978). Several promising grass and legume species require a fraction of the available soil test phosphorus levels required by annual crops and much less than other pasture species. For example, the general soil test critical level used for crops in Colombia is 15 ppm P by the Bray I1 method (Marin, 1977). Promising aluminum-tolerant ecotypes of Stylosanthes capitata, Zornia latifolia, and Andropogon gayanus require 1/3-1/5 of that amount to attain maximum yields. This information is shown in Table XXXIII of Section V1,D. It should be noted that adapted grasses such as Andropogon gayanus and Brachiaria decumbens require higher critical levels of Bray I1 available soil phosphorus (5-7 ppm P) than adapted legumes like Stylosanthes capitata and Zornia latifolia (3-4 ppm P) require for near maximum growth (CIAT, 1979). The commonly held view that fertilization of grass-legume mixtures should be based on the legume’s higher nutritional requirement does not apply to these species. This has been proven in the field by Spain (1979), where, in addition to phosphorus, there was a higher need for potassium in the grasses than in the legumes. Field responses during the establishment year show significant differences in the levels of phosphorus fertilization needed for near maximum growth on an Oxisol with about 1 ppm available P (Mehlich 2 method) prior to treatment applications (Fig. 9). Andropogon gayanus required 50 kg P,O,/ha to reach maximum yields, while Panicum maximum required 100 kg P,O,/ha and Hyparrhenia rufa required 200 or perhaps more. The latter species, very widespread in

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

305

Table VIII Some Important Pasture Species Adapted to Oxisols and Ultisols of the Tropics" Species

Observations

Andropogon gayanus Brachiaria decumbens Brachiaria humiciicola Digitaria decumbens Hyparrhenia rufu Melinis minutiflora Panicurn maximum Pennisetum purpureum Paspalurn notatum Paspalurn plicatulutn

Grasses Well adapted; new release in tropical America Well adapted: spittlebug susceptible Very Al-tolerant, low palatability Adapted, but requires high fertility Adapted, high K requirement, low productivity Adapted but low productivity Adapted. somewhat higher nutritional requirement Adapted for cut forage, high nutrient requirement Low productivity Disease susceptibility in some areas

Desmodium heterophyllum Desmodium gyroides Desmodium ovalifofium Calopogonium mucunoides Centrosema pubescens Galactia striata Pueraria phaseoloides Stylosanthes capitata Stylosanthes guianensis Stylosanthes scubra Stylosutithes viscosa Zornia latifolia

Legumes Prefers udic soil moisture regime Shrub for browse High tannin in ustic climates Persistent but low palatability Insect attack problems Productive in certain systems only Not for long dry season Savannas only Only few cultivars have anthracnose tolerance Promising for isothermic savannas Promising for isothermic savannas Promising for isohyperthermic savannas

"Source: CIAT (1978, 1979, 1980) and authors' observations

Latin America, performs poorly in Oxisol-Ultisol regions because of a generally higher requirement for phosphorus and potassium and a lower tolerance to aluminum than the other two (Spain, 1979). These differences are quite significant at the animal production level. At levels of inputs where other grasses produce good cattle liveweight gains, Hyparrhenia rufa produced serious liveweight losses at Carimagua, Colombia (Paladines and Leal, 1979). It may be argued that the use of pastures requiring less phosphorus may provide insufficient phosphorus for animal nutrition. There is no evidence in the CIAT work that this is so (CIAT, 1978, 1979), but if it were, it is probably cheaper to apply only the phosphorus fertilizer required for maximum plant growth to the soil and supplement the rest directly to the animals via salt licks.

3 . Water Stress The ability to grow and survive the long dry seasons of ustic environments under grazing is a necessary requirement for acid-tolerant forage species because

306

PEDRO A. SANCHEZ AND JOSE G . SALINAS

0

50

P APPLIED

100

200

( k g P2 O,/ha)

FIG. 9. Differential response to phosphorus fertilization of three grass species during the establishment year in an Oxisol of Carimagua, Colombia: (0)Andropogori gayanus 621, (A)Panicurn maximum 622, (W) Hyparrhenia rufa 601. Sum of three wet season cuts. All treatments received 400 kg N/ha. (Source: CIAT, 1979.)

irrigating pastures is prohibitively expensive in most Oxisol-Ultisol regions. Because of their aluminum tolerance, roots of adapted forage species are able to penetrate deeply into acid subsoils and exploit the residual moisture that is available. This is in sharp contrast with aluminum-sensitive crops that suffer severely from water stress, even during short dry periods, because their roots are confined to the limed topsoil (Gonzalez er a l . , 1979). Adapted legume species are generally more tolerant to drought stress than the grass species. Also, legumes are able to maintain a higher nutritive value during the dry season than the grasses. For example, Zornia latifolia 728 contained 24% protein in its leaves at the height of the Carimagua dry season, while accompanying grasses contained about 5% protein (CIAT, 1979). Among the adapted grasses, Andropogon is more tolerant to drought stress than Brachiaria decumbens or Panicum tnaximum (CIAT, 1979). Its pubescent leaves also permit dew drops to remain on the leaves longer than in B . decumbens or P . maximum. It is common to get one's pant legs wet while walking through an Andropogon pasture at about 1O:OO A . M . in the Llanos or in the Amazon, when swards of the other two species are already dry. 4 . Insect and Disease Attacks

Most of the adapted legume species have their center of origin in Latin America and therefore, have many natural enemies. Anthracnose caused by Collectotrichum gloesporoides is a most devastating disease of legumes (CIAT,

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

307

1977, 1978, 1979). Stem-borers of the genus Caloptilia also attack several Stylosanthes species (CIAT, 1979). Spittlebug attacks caused by Deois incompleta and other species have destroyed thousands of hectares of Brachiaria decumbens pastures in udic regions of tropical Brazil. The solution to these problems is varietal resistance since applications of insecticides or fungicides to these pastures are likely to be uneconomical. Screening for tolerance to these and other pathogens has provided ecotypes that combine the adaptation to adverse soil conditions with pathogen resistance. Examples of these to date are several ecotypes of Andropogon gayanus, Stylanthes capirata, and Desmodium ovalifolium. Several promising ecotypes of Stylosanthes guianensis, a legume extremely well adapted to acid soil constraints, unfortunately have succumbed to insect and disease attacks (CIAT, 1978, 1979). As in other plant improvement programs, the search for new ecotypes that combine tolerance to pathogens with other desirable characteristics is a continuing activity. It is interesting to note that plant protection problems increase in importance after the soil constraints are allekiated by plant selection or fertilization in Oxisol-Ultisol regions. This may be a consequence of elimination of a previously limiting factor, or of a pathogen buildup as new plants are grown on many hectares for the first time in a new environment. This observation applies both to pastures and to annual food crops,. Tolerance to disease and insect attacks, however, varies with ecological conditions and therefore the degree of tolerance of each promising cultivar must be validated locally.

5 . Tolerance to Burning Accidental burning is common in savanna regions and intentional burning may be a necessary management practice in cases where grasses approach maturity rapidly and lose their nutritive value. Consequently, the adapted pasture species must be able to regrow after burning. Studies in Quilichao, Colombia show that Andropogon gayanus, Panicum maximum, Brachiaria decumbens, and Brachiaria humidicola regrow rapidly after burning (CIAT, 1979). Later CIAT work shows that regrowth after burning depends very much on soil moisture conditions at the time of burning. The Brachiarias, for example, are very susceptible to burning when the surface soil is moist.

D. CONCLUSIONS There is a broad germplasm base of acid-tolerant annual crops, permanent crops, tree crops, and pasture species adapted to tropical conditions in Latin America. In addition, selection of breeding programs can provide acid-tolerant varieties from generally sensitive species. The degree of quantification of these differences, however, is very limited. A more systematic classification of what

308

PEDRO A. SANCHEZ AND JOSE G . SALINAS

are the critical tolerance levels of each important variety or species is needed. Such a plant classification system could be linked with present technical soil classification systems in order to better match plant characteristics with soil constraints.

IV. DEVELOPMENT AND MAINTENANCE OF GROUND COVER The choice of farming systems is extremely varied and very dependent on market demands or opportunity, farming tradition, and government policies. The prevalent farming systems in Oxisol-Ultisol regions of tropical America can be grouped into four major categories: shifting cultivation (primarily in the forested areas), extensive cattle grazing in both forested and savanna regions, permanent crop production systems, and intensive annual crop production systems. The extent of the last two is very limited. These systems are described in a review by Sanchez and Cochrane ( 1 980). Regardless of the farming system or the plant species used, a basic principle of low-input technology is to develop and maintain a plant canopy over the soil for as long as possible in order to decrease erosion, compaction, and leaching hazards. The main technology components are land clearing methods, crop and pasture establishment techniques, mulching, the use of managed fallows, intercropping, and multiple cropping systems. Some of the advances in developing these technology components are discussed in this section. A . LANDCLEARING METHODSI N RAINFORESTS

The choice of land clearing method is the first and probably the most crucial step affecting the future productivity of farming systems in rain forest areas. Several comparative studies conducted in the humid tropics of Latin America confirm that manual slash-and-burn methods are superior to different types of mechanical clearing because of the fertilizer value of the ash and because of less soil compaction and topsoil displacement compared to mechanized land clearing. 1 . Nutrient Additions by the Ash

The nutrient content of ash has been directly determined upon burning a 17-year-old secondary forest on Typic Paleudult from Yurimaguas, Peru. The data of Seubert et al. (1977) in Table IX show significant beneficial effects of ash on soil chemical properties (Fig. lo), which resulted in higher yields of a

LOW-INPUTTECHNOLOGY FOR OXISOLS AND ULTISOLS

309

Table IX Nutrient Contribution ofAsh and Partially Burned Material Deposited on an Ultisol of Yuiimaguas, Peru, after Burning a 17-Year-Old ForestO

Element

N

P K Ca Mg Fe Mn Zn

cu

Coinposition I .72% 0.14% 0.97% I .92% 0.41% 0.19% 0.19% 132 ppm ’79 ppm

Total additions (kdha) 67 6 38 75 16 7.6 7.3 0.3

0.3

“Source: Seubert ef a / . (1977)

wide variety of crops during the first 2 years after clearing (Table X). There is considerable variability among sites in the quantity of ash and its nutrient composition because of differences in soil properties, clearing techniques, and the proportion of the forest biomass actually burned. Silva (1 978) estimated that only 20% of the felled forest biomass was actually converted to ash after burning a virgin forest on an Oxic Paleudult of southern Bahia, Brazil. Silva also analyzed the ash composition of the burned parts of individual tree species and observed wide ranges (0.8-3.4% N; 0-14 ppm P; 0.06-4.4 meq Cd100 g; 0.11-21.03

MONTHS AFTER CLEARING

FIG. 10. Effects of two land clearing mi:thods on changes in topsoil (0-10 cm) properties in a Typic Paleudult of Yurimaguas, Peru: ( 0 )slash-and-bum method; (0)bulldozer clearing. (Source: Seubert et a/., 1977.)

310

PEDRO A. SANCHEZ AND JOSE G. SALINAS

Table X Effects of Land-Clearing Methods on Crop Yields at Yurimaguas".b

Crop Upland rice (3)

Corn (1)

Soybeans (2)

Cassava (2)

Panicurn tnnxitnum (6 cuts/yr)

Mean relative yields

Fertility level" 0 N PK NPKL 0 NPK NPKL 0 NPK NPKL 0 NPK NPKL 0 NPK NPKL

Slash and bum (tonslha)" I .3 3.0 2.9 0.1 0.4 3. I 0.7 I .0 2.7 15.4 18.9

25.6 12.3 25.2 32.2

0 NPK NPKL

Bulldozed (tondha)" 0.7 1.5 2.3 0.0 0.04 2.4 0.2 0.3 1.8 6.4 14.9 24.9 8.3 17.2 24.2

Bulldozed Burned (%)

53 49 80 0 10

76 24 34 67 42 78 97 68 68

75 37 47 48

"Source: Seubert e/ a / . (1977). bYield is the average of the number of harvests indicated in parentheses. "50 kg N/ha, 172 kg P/ha, 40 kg K/ha, 4 tons lirne/ha (L). "Grain yields of upland rice, corn, and soybeans; fresh root yields of cassava; annual dry matter production of Panicum maximum

meq Mg/100g, and 34-345 meq K/100 g). This information suggests the presence of certain species that can be considered accumulators of specific nutrients. The fertilizer value of the ash is likely to be of less importance in high-base status soils. Cordero (1964) observed that increases in phosphorus and potassium availability caused by burning the biomass on an Entisol of pH 7 in Santa Cruz, Bolivia, did not increase crop yields. The soil was already high in these elements. Information on ash composition from different soils and clearing methods therefore will contribute significantly to our understanding of soil dynamics and its subsequent management. 2 . Soil Compaction Conventional bulldozing has the clearly detrimental effect of compacting the soil, particularly coarse-textured Ultisols. Significant decreases in infiltration

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

31 1

rates, increases in bulk density, anld decreases in porosity have been recorded on such soils in Surinam (Van der Weert, 1974), Peru (Seubert et ul., 1977), and Brazil (Silva, 1978) after mechanized land clearing. Table XI shows the decreases in infiltration at three sites. The slash-and-bum method had a moderate effect on infiltration rates, but bulldozing decreased them by one order of magnitude. Comparisons between sites are difficult because of differences in the time span used in measuring. The Manaus example illustrates the compaction observed in degraded pastures in parts of the Brazilian Amazon. 3. Topsoil Displucement

The third major consideration is the degree of topsoil carryover, not by the bulldozer blade, which is normally kept above the soil, but by dragging uprooted trees and logs. Although no quantitative data are available, topsoil removal from high spots and accumulation in low spots are commonly observed. The better forest regrowth near windrows of felled vegetation suggests that topsoil displacement can result in major yield reductions (Sanchez, 1976). For example, La1 et al. (1 975) in Nigeria observed that corn yields decreased by 50% when the top 2.5 cm of an Alfisol was removed. No comparable data, however, is available from acid soils of tropical America. Nevertheless, the yield decreases shown in Table X are undoubtedly associated with topsoil displacement. 4 . Alternative Lund Clearing Methods

The detrimental effects of bulldozer land clearing are generally well known to farmers and development organizations in parts of the Amazon. Government credits for large-scale mechanized land clearing operations have been sharply reduced in the Brazilian Amazon since 1978. Also, the practice of completely Table XI Effects of Clearing Methods on Water Infiltration Rates in Ultisols from Yurimaguas, Peru; Manaus and Barrolandia (Bahia), Brazil"

Clearing method Undisturbed forest Slash and burn ( I year) Bulldozed ( 1 year) Slash and bum and 5 years in pasture

Yurimaguas Peru (cm/hr) 26

Manaus, AM Brazil (cdhr)

Barrollndia, BA Brazil (cdhr)

15

24 20 3

10 0.5 0.4

"Sources: NCSU (1972). Seubert ef a / . (1977). Schubart (1977), and Silva (1978)

312

PEDRO A. SANCHEZ AND JOSG G . SALINAS

destroying the forest versus its partial harvest before burning is being considered. Silva (1978) provided the first quantitative estimate of the possible benefits of such a practice. He compared the two extremes, the slash-and-bum method and bulldozing, with treatments that include the removal of marketable trees first, followed by cutting and burning the remaining ones. All the advantages of burning on soil fertility were observed in this latter treatment, with no significant differences from the conventional slash-and-bum method (Silva, 1978), but with a valuable increase in income. The lack of difference is probably due to the small proportion of the total biomass that is actually burned. Indeed many farmers in the Amazon harvest wood first, some of them developing profitable lumber mills in the process of clearing land for pasture establishment. The pressures for opening new lands in some areas of the Amazon are so intense that it is now necessary to develop technology that minimizes the detrimental effects of mechanized land clearing on soil properties. Research comparing presently available mechanized land clearing technologies has not been conducted in this region on a systematic fashion. Bulldozers equipped with a “KG” blade that cuts tree trunks at ground level by shearing action could cause less topsoil displacement since the root systems remain in place. “Tree pusher” attachments on tractors reduce energy requirements for felling and may decrease compaction by machinery. A heavy chain dragged by two bulldozers should also minimize compaction. With these three techniques the felled vegetation could be burned and the remaining material could be removed by bulldozers equipped with a root rake at a later time. A large-scale unreplicated study on Typic Acrorthox near Manaus showed little difference in chemical or physical soil properties when some of the above combinations were compared with conventional bulldozing (UEPAE de Manaus, 1979). The slash-and-bum treatment provided superior chemical properties and better pasture growth than the mechanized land clearing treatments. Work on Alfisols of Nigeria with totally different physical and chemical properties shows that the clearing of land with bulldozers equipped with a shear blade, followed by burning and removal of residues with a root rake, was the least damaging mechanized system (IITA, 1980). One type of low-input technology that has produced few satisfactory results is the partial clearing of tropical rain forests. Strips are cleared by the slash-andbum method in order to plant shade-tolerant corps such as cocoa or certain pastures, or to enrich the forest with valuable timber species. Experiments have been conducted in Manaus, Brazil, by various organizations, but the results have been disappointing. No data are available as such experiments have not been published. Apparently, it is difficult to provide sufficient sunlight for vigorous plant establishment without eliminating the forest canopy. Leaving a few trees untouched, however, is often done, particularly when they are of value or to provide shade for pasture. Hecht (1979) has identified several legume tree and

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

313

shrub species that should be allowed to regrow after clearing for pastures because of their capacity to provide browse forage for cattle. Many of the failures of large-scale farming operations observed by the authors in the humid tropics can be directly attributed to improper land clearing methods. Research on alternative mechanized land clearing methods that involve burning is needed. B. SOILDYNAMICS AFTER CLEARING TROPICAL RAINFORESTS

When a tropical forest is cleared and burned several changes in soil properties generally occur within the first year: Large losses of biomass nitrogen and sulfur occur upon burning by volatilization, soil organic matter decreases with time until a new equilibrium is reached; the pH of acid soils increases, aluminum saturation levels decrease, exchangeable bases and available phosphorus increase; and topsoil temperatures increase (Sanchez, 1973). The following discussion is based on a recent review of the subject by the senior author (Sanchez, 1979). Most of the available data is based on sampling nearby sites of known age after clearing at the same time. This technique confounds time and space dimensions and increases the already considerable variability between sites. Fortunately, there are five studies in which changes in soil properties were followed with time in humid tropical America; Yurimaguas, Peru; Manaus, Belem, and BarrolCndia, Brazil; and Carare-Opon, Colombia. Most of them, however, are limited to what happens during the first year, but one covers an 8-year period. Nevertheless, they illustrate the differences that take place within sites as a function of time.

I . Soil Organic Matter Salas and Folster (1976) estimated that 25 tons C/ha and 673 kg N/ha were lost to the atmosphere when a virgin forest growing on an Aeric Ochraquox in the middle Magdalena Valley of Colombia was cut and burned. These figures were derived by measuring the biomass changes before and after burning, but before the first rains. These losses accounted for only 11-16% of the total carbon and about 20% of the total nitrogen in the ecosystem (Salas, 1978). Consequently, assertions that most of the carbon and nitrogen in the vegetation is volatilized upon burning deserve scrutiny. Another unknown factor is whether or not a proportion of the volatilized elements is returned back to nearby areas via rainwash. The influence of burning on the thin organic-rich layer consisting of littertopsoil interphase was also determined by Salas (1978). The C/N ratio of this

3 14

PEDRO A. S h C H E Z AND JOSE G . SALINAS

material increased from 8 to 46 within 5 months, suggesting that the volatile losses were rich in nitrogen. The literature has conflicting information about the losses of soil organic matter when the cropping phase begins. Larger losses will occur in soils with higher initial organic matter contents (Sanchez, 1976). This effect, however, is attenuated by the topsoil clay content. Turenne (1969, 1977) found an inverse relationship between organic carbon losses and clay contents in Oxisols of French Guiana. Another supposedly detrimental effect of burning is a decrease in soil microbiological activity. Silva’s (1978) southern Bahia study reports no significant differences caused by various degrees of burning on fungal flora, but decreases in the bacterial and actinomycetal population during the first 30 days after the conventional burn. Figure 11 shows the time trend in cellulose decomposition activity. Burning actually had a stimulating effect on the decomposing microflora, probably because of the increase in phosphorus and other nutrients, plus the higher soil temperatures incurred upon exposing the soil surface to direct sunlight. No such effect was observed in the bulldozer clearing, probably because of topsoil displacement and soil compaction. The partial sterilization effect in the conventional burn may account for the lower microbiological acitvity observed during the first 25 days after burning. The dynamics of organic carbon during the first 4 years of continuous upland rice-corn-soybean cropping on an Ultisol from Yunmaguas, Peru, without fertilization or liming, are shown in Fig. 12. There was an actual increase in organic carbon contents 1 month after burning, probably a result of ash contarni-

DAYS AFTER CLEARING AND BURNING

FIG.11. Effects of degrees of burning intensity on microbial activity as measured by cellulose decomposition rates as a function of time after burning a rain forest on an Ultisol of southern Bahia, Brazil. A--.-A, conventional slash-and-bum method; A-A, harvest-valuable trees plus slash-and-bum method; 0-0, bulldozer clearing (no burning). (Adapted from Silva, 1978.)

315

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

Exch. Al (meq/IOOml)

Al Saturation (%)

2 .o I.6

Exch. Mg (meq/100 m l )

0.8 0.4 1

I

Exch. K (meq / 100 ml)

0.2

0. I -

0

I

6

12

24

36

0 48 I

6

12

24

36

48

MONTHS AFTER CLEARING

FIG. 12. Changes in chemical properties of an Ultisol (0- 10 cm) continuously cropped to upland rice, corn, and soybeans (8 crops), without fertilization at Yurimaguas (1972-1976). (Compiled from data by Seubert el ul., 1977; Villachica, 1978; and Sanchez, 1979.)

nation. This increase was followed by a plateau for the first 6 months, then a sharp decrease was observed after the first rice crop was harvested, and finally an equilibrium was reached at the end of the first year. The annual decomposition rate during the first year was on the order of 30%, but a new equilibrium was attained the second year of cropping (Villachica, 1978). This high decom-

3 16

PEDRO A. S h C H E Z AND JOSh G . SALWAS

position rate resulted in a very large increase in inorganic nitrogen in the topsoil during the first 6 months at Yurimaguas (80 kg N/ha in the top 50 cm), which quickly disappeared because of leaching and/or crop uptake (Seubert et al., 1977). This “nitrogen flush” probably contributes to the initial lush growth of the first crop after burning.

2 . Initial Increases in Nutrient Availability The changes in topsoil properties before clearing and after burning in several properly sampled time studies are summarized in Table XII. This table shows the general trends and deviations thereof. Soil pH values increase after burning but not to neutrality. Exchangeable Ca + Mg levels doubled, tripled, or quadrupled, but there was considerable variability among nearby clearings on the same soil as shown by the two Yurimaguas sites. This particular difference was attributed to an initially higher base status in site I1 and a better-quality bum than in site I. Exchangeable potassium also increased, but the effect was short-lived because of rapid leaching. This probably explains why there were no increases in the Yurimaguas Chacra I1 and Belem sites, which were sampled at 3 and 12 months after burning. Exchangeable aluminum decreased in proportionate amounts to increases in Ca + Mg, suggesting a straight liming effect. An exception to this statement occurred in the southern Bahia site, which had relatively low exchangeable aluminum contents. Aluminum saturation decreased in all but one case to levels below that considered as critical for crops such as corn (60%). Available phosphorus also increased with burning, surpassing the generally accepted critical level for annual crops (10-15 ppm P with the modified Olsen, Bray 2, or Mehlich extractants). Regardless of site differences, there is no question that the fertility of acid soils improved considerably after burning.

3 . Fertility Decline Pattern The positive effects noted above begin to reverse with time. Figure 10 illustrates the changes occurring within the first 10 months after clearing in Yurimaguas without fertilization. Silva (1978) has reported almost identical results at the other end of the continent, in southern Bahia. Inorganic nitrogen (not shown) and potassium are the first elements to be depleted, while the others show a slower decline. Figure 12 shows the changes occumng in topsoil properties during the first 4 years in Yurimaguas. Equilibrium values were attained with pH and organic carbon after the first year. Exchangeable aluminum began to increase after the original decline, attaining preclearing levels within a year. This is attributed to the rapid organic matter decomposition rate during the first year, which released H+ ions and aluminum compounds bound to organic matter into the soil solution. This, in turn, released A13+ ions from the clay minerals (Villachica, 1978). Consequently, the residual “liming” effect of the ash was short-lived. Increases in exchangeable calcium remained relatively stable with

3 17

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

Table XI1 Summary of Changes in Topsoil Chemical Properties Before and Shortly After Burning Tropical Forests in Ultisols and Oxisols of the Amazon Yurimaguas" (2 sites) Soil property

Manad

I

Timing

11

(x 7 sites)

Manaus" ( I site)

Belh" Barrolbndia" (9. 60 sites) Bahia ( 1 site)

Months after burning:

1

3

0.5

4

pH (in H,O)

Before: After:

4.0 4.5

4.0 4.8

3.8 4.5

4. I 5.5

4.8 4.9

4.6 5.2

Before: After:

0.41 0.88

1.46 4.08

0.35 1.25

0.92 5.44

I .03 1.97

I .40 4.40

A

0.47

2.62

0.90

4.52

0.94

3.00

Before: After:

0.10 0.32

0.33 0.24

0.07 0.22

0.08 0.23

0.12 0.12

0.07 0.16

A

0.22

(0.07)

0.15

0.15

0.00

0.09

Before: After:

2.27 1.70

2.15 0.65

1.73 0.70

1.81 0.10

I .62 0.90

0.75 0.28

A

(0.59)

(1.50)

(1.03)

(1.71)

(0.72)

(0.45)

Before: After:

81 59

52 12

80 32

64

58

2

30

34 5

Available P (pprn) Before: (Olsen in Peru, Mehlich in After: Brazil)

5

15

-

2

6.3

1.5

16

23

-

5

1.5

8.5

II

8

-

3

1.2

7.0

Exch. K (meq/100 g)

Exch. A1 (meq/100 g)

Al saturation (%)

A

12

1

"Calculated from data by Seubert el al. (1977) and Villachica and Sanchez (unpublished data). *Calculated from data by Brinkmann and Nascimento (1973). "Calculated from data by UEPAE de Manaus (1979). "Calculated from data by Hecht (unpublished data). "Calculated from data by Silva (1978). 'Exch. = exchangeable.

time. Exchangeable magnesium and potassium, however, decreased after 6 months of cultivation. Available phosphorus levels remained close to the critical level of 15 ppm P (modified Olsen) in this particular trial. Crop performance data (Villachica, 1978; Sanchez, 1979) show that nitrogen

PEDRO A. SANCHEZ AND JOSE G.SALINAS

318

and potassium became deficient with 6 months after clearing. Aluminum reached toxic levels for corn at 10 months after clearing. At that time phosphorus, magnesium, copper, and boron became deficient and crop yields without fertilization approached zero. When potassium fertilizers were added, a K/Mg imbalance resulted; this necessitated further magnesium additions. Zinc approached deficiency levels at the end of the second year and sulfur and molybdenum deficiencies were observed sporadically (Villachica, 1978; Sanchez, 1979). The Yurimaguas results indicate that most of the rapid changes occur during the first 2 years after clearing, after which equilibria are established.

c.

LANDPREPARATION A N D P L A N T ESTABLISHMENT IN RAINFORESTS

In traditional slash-and-burn clearings, land preparation is usually limited to removal of some logs for firewood or charcoal. The first plantings consist of poking holes in the ground with a pointed stick called “espeque” or “tacaqo,” followed by dropping seeds or simply inserting cassava stakes or plantain rhizomes. This zero-tillage system protects the soil against erosion by a tangled mass of logs and branches, numerous tree stumps, and a mulch of ash and unburned plant material. Since fertilizers are seldom needed for the first planting, there is little need for tillage. Trials in Yurimaguas, Peru, showed no significant differences in upland rice yields between the “tacarpo” no-till plantings and rototilling followed by row seeding after clearing a rain forest by the slash-andbum method (Sanchez and Nureria, 1972). Plant spacing, however, had a marked effect on yields. Table XI11 shows that decreasing spacing between the “tacarpo” holes from the conventional pattern of 50 X 50 cm to 25 X 25 cm increased rice yields. The incidence of weeds decreased dramatically. Closer spacing plus a change from the traditional tall-statured Carolino variety to the short-statured blast-tolerant IR4-2 variety has resulted in a 76% yield increase (0.95 to 1.67 tons/ha) on farmer field trials in the Yurimaguas region (Donovan, 1973). This simple low-input technology has improved the traditional shifting cultivation system. To change from shifting to continuous cultivation in this region, however, fertilization is definitely needed (Sanchez, 1977). Oversowing pasture species on land cleared by the slash-and-bum method is a common practice in the Amazon. The high initial fertility favors rapid pasture establishment and ground cover development. Toledo and Morales (1979) reported successful pasture establishment in Ultisols of Pucallpa, Peru, with Brachiaria decumbens and Panicum maximurn. They also reported that grasslegume associations may be difficult to establish because the most aggressive species may tend to dominate. To avoid this difficulty it is recommended to plant each species in single or double rows.

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

319

Table XI11 Effect of Planting Method, Spacing, and Seed Density on 1R8 Upland Rice Yields on Aeric Tropaqualf in Yurimaguas, Peru" Planting method and spacing Rototilled, row seeding (25-crn rows) No till, "tacarpo" holes 25 x 25 cm No till, "tacarpo" holes 50 x 50cm LSD.05 "

Seed density (kdha)

Grain yields (tons/ha)

50

5.93

35

5.68

18

4.25 0.31

Source: Sanchez and Nurena (1972).

For many of the pasture species adapted to acid soil conditions, better establishment is obtained when the seeds encounter a corrugated soil instead of a highly pulverized one (Spain, 1979). This is attributed to the need of small pasture seeds to be sheltered and to avoid desiccation during germination. Planting deeper than 1 or 2 cm is likely to retard establishment or prevent it altogether. Because of the initial high fertility level of the topsoil after burning, the development of a plant canopy after slash-and-bum land clearing is seldom a problem in the humid tropics. The critical issue is the nature of such a cover. With good management it consists of vigorous crops or fast-growing pastures; with poor management or adverse weather conditions, weeds and jungle regrowth may constitute the principal components of the canopy. In either case, the soil is likely to be protected from erosion hazards. With mechanized land clearing, however, the situation is totally different. The absence of burning keeps the soil in its original acid, infertile state (Fig. 10) and some degree of compaction can be expected. Tillage is usually necessary to correct compaction and to incorporate moderate quantities of fertilizer and lime that the first crop or pasture may need. Although weed competition is likely to be less than with slash-and-bum clearing, jungle regrowth does take place in bulldozed areas.

D. LANDCLEARING METHODSI N

THE

SAVANNAS

The absence of a closed tree canopy in savanna regions poses a wide variety of alternatives for transforming the native savanna into agricultural production systems. Unlike in rain forests, a significant production system-extensive cattle grazing with essentially zero soil management-exists in native savanna. Native

320

PEDRO A. S h C H E Z AND JOSG G.SALINAS

savanna vegetation, however, is far from uniform. Five physiognomic types are recognized in the Cerrado of Brazil: 1. “Campo limpo” (clean field): a continuous grass canopy without arboreal vegetation; a treeless savanna. 2. “Campo sujo” (dirty field): a continuous grass canopy with widely scattered small bushes. 3. “Campo cerrado”: a continuous grass canopy overlain by a discontinuous arboreal canopy sufficiently scattered so that it is possible to drive a jeep through it. 4. “Cerrado” (in the strict sense): a two canopy savanna where it is impossible to drive a jeep through. 5. “Cerradiio”’: a dominant and almost closed canopy of taller trees of the same species as before underlain by a discontinuous grass canopy.

These physiognomic types are related to topsoil fertility parameters in welldrained areas (Lopes and Cox, 1977b). Treeless savannas also occur on shallow soils and on poorly drained areas, although with a different species composition in the latter case. Large areas of the Llanos Orientales of Colombia are of the campo limpo type. Land clearing and crop establishment techniques are related to the above physiognomic types. Duque et al. (1980) described the different clearing techniques practices in the Cerrado of Brazil, for areas intended for either crop production or improved pastures. For campo limpo and campo sujo areas the common technique is to bum the native savanna, remove by hand whatever shrubs exist, and plow. For campo cerrado, cerrado, and cerradiio the usual procedure is to fell the arboreal vegetation with two bulldozers dragging a 25-m heavy chain. A third machine piles the woody residues in windrows along the contour, providing some protection against erosion. Part of this material is gradually removed for charcoal production. The areas between the windrows are burned to eliminate the grass canopy. The effects of land clearing practices in savannas are not well documented but appear to be less marked than those reported in rain forests. The amount and composition of ash produced by annual burning of native Oxisol savannas has not been measured, but because of less biomass the amount is estimated to be a fraction of that produced after burning rain forests. Consequently, the changes in chemical soil properties with clearing are probably minor. Topsoil displacement due to bulldozer clearing is also less pronounced because of the low density and generally smaller size of the arboreal vegetation. Unlike the rain forests, where mineral cycling has concentrated nutrients in an organic-rich topsoil layer, organic matter and nutrient distribution is more uniform with depth in the savannas (Sanchez, 1976). Consequently, topsoil displacement will cause less damage in

LOW-INPUTTECHNOLOGY FOR OXISOLS AND ULTISOLS

32 1

deep, largely uniform savanna Oxisols than in Ultisols and Oxisols under rain forest vegetation. E. CROPA N D PASTURE ESTABLISHMENT I N SAVANNAS

In Cerrado Oxisols, the preferred tillage implement for crop establishment is the disk harrow. Duque ef al. (1980) recommend avoiding the moldboard plow and deep disking because they cause compaction. Root fragments should be picked up after each disking operation during the first or second year after clearing. A second operation, either disking or rotovating, is normally conducted to incorporate lime and broadcast fertilizer applications. Planting upland rice, soybeans, corn, or other crops is normally accomplished with grain drills equipped with attachments for band placement of fertilizer. Excessively deep and frequent tillage operations are common in certain Oxisols and Ultisols of savanna regions where peanuts and sorghum are grown extensively. These practices result in a very pulverized seedbed that easily washes away during heavy rainstorms. Some of these regions are irrigated with center-pivot systems that are often poorly managed. Such technologies often have a clearly detrimental effect on soil properties. Conventional pasture establishment methods in savanna regions commonly consist of one or two passes with a disk harrow, followed by seeding with a grain drill equipped with a fertilizer attachment (Spain, 1979). These operations are accomplished during the rainy season, but the cost is generally high (CIAT, 1979). Regardless of the quality of tillage, the soil is left exposed for a considerable amount of time until the crop or pasture canopy is established. This critical period corresponds to the beginning of the rain season when high-intensity rainstorms occur. Although Oxisols are among the least erodible soils in the world (El Swaify, 1977), sheet erosion is an important constraint in the savannas. Given the fairly uniform distribution of organic matter and nutrients in many of the savanna Oxisols, it has been argued that sheet erosion is unimportant. This argument loses its validity when phosphorus and lime are incorporated into the topsoil. Also, some Oxisols have umbric epipedons with higher organic carbon contents than those of the underlying oxic horizon. This is usually the case with many savanna Ultisols as well. In such cases, erosion will significantly decrease the effective cation exchange capacity of the topsoil due to organic matter, thus increasing potential leaching losses. A series of low-input land preparation techniques are being developed in order to reduce costs and erosion hazards. Four techniques are described in this section: the introduction of improved pastures in native savanna, its gradual replacement,

322

PEDRO A. ShJCHEZ AND JOSE G . SALINAS

low-density pasture establishment methods, and crop-pasture relay intercropping.

I . Improving the Native Savanna Unlike in the rain forests, where partial clearing is not promising, gradual improvement of the native savanna appears promising. Oversowing pasture species on undisturbed native savanna, however, is usually unsuccessful (Spain, 1979). Some degree of soil disturbance is necessary for the small pasture seeds to have contact with sufficient moisture for germination. Light disking or sod seeding in rows 50 cm apart has successfully established acid-tolerant legumes in c a r p o lirnpo savannas of the Brazilian Cerrado and improved the nutritional quality of the sward considerably (CIAT, 1980). After 1 year of disking and sod seeding, improved legume species with 14% protein content were well established in the native savanna, which contained only 4% protein (CIAT, 1980).

2 . Gradual Displacement of Native Savanna with Improved Pastures A second low-cost alternative is to plant improved pasture species in strips without disturbing the native savanna between the strips. A trial conducted by J . M. Spain and colleagues in Carimagua, Colombia, is showing promising results (CIAT, 1980). Grass-legume pastures were established in 60-cm-wide strips prepared with spring tines or with a field cultivator to a 12-cm depth, followed by phosphorus and potassium applications. The area between strips, about 2.5 m wide, received four levels of native savanna vegetation control. Several grass and legume species were able to invade and gradually displace the native savanna strips. The most successful species were the legumes Desmodium ovalifolium and Pueraria phuseoloides and the creeping grasses Brachiaria humidicola and B . decumbens. Table XIV summarizes the results. Spain’s work shows that the native savanna can be gradually replaced by such strip plantings, at a much lower cost, while limiting erosion hazards to a fraction of the land.

3. Low-Density Seedings

In Oxisol savannas, weed growth after land preparation is normally slow due to the extremely low native soil fertility, as long as the soil is not limed or fertilized. Taking advantage of this situation, Spain (1979) developed a lowdensity plenting system with considerable savings in seed costs and initial fertilizer applications. After the land is prepared with one or two passes of an offset disk harrow, grass and/or legume seeds are planted in holes spaced about 3 m, giving a population of 1000 plants/ha during the rainy season. The plants

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

323

Table XIV Ability of Different Forage Species to Invade and Displace Fertilized Native Savanna with Different Degrees of Control and Tillage in Oxisols of Carimagua, Colombia" Species capable of Treatment of native savanna

Invading

Displacing

Bum only

D. ovalifolium P . phaseoloides B . radicans

D . ovalifolium P . phaseoloides

Chemical control

D . ovalifoliurn P . phaseoloides B . hu,niclicoltr B . radicans

D. ovalifolium P . phaseoloides B . humidicola

Tine tillage to 12 cm

D. ovalifoliurn P. phaseoloides B . hurnirlicolo B . decurnbens

D . ovalifolium P. phaseoloides B . humidicola B . decumbens A . gayanus

A . gciwnus

B . rudicans

Complete seedbed preparation

D . ovalifolium P . phaseoloides B . humiclicola

B . tleeurnberis A . gci>wnu.s

B . radicans

D . ovalifolium P . phaseoloides B . humidicola B . decumbens A . gayanus B . radicans

"Source: CIAT (1980)

received a localized high rate of phosphorus and potassium, but on a per hectare basis the highest rates used were 9 kg P,O,/ha and 1.5 kg K,O/ha. One man equipped with a shovel can plant and fertilize 1 ha in 1 day (Spain, 1979). These plants grow vigorously during the rainy season due to their high soil fertility status and the absence of competition from weeds or plants of their own species. Stoloniferous species cover the ground within 8 months, at the beginning of the next rainy season (CIAT, 1979). Tussock-type grasses such as Andropogon guyanus and Panicurn muximum produce seed at the end of the rainy season. At Carimagua, the seeds aligned themselves in the furrows left by the disk harrow and germinated with the first rainy season showers, starting ahead of the weeds. The new seedlings had to be fertilized shortly after emergence, otherwise they would have died because of acute phosphorus and potassium deficiencies. With such a system, pastures in Carimagua were ready for grazing within 9 months after planting, which is about 3 months later than with conventional land preparation. The details are explained more thoroughly in reports by Spain (1979) and CIAT (1978, 1979, 1980). Although this system does not

324

PEDRO A. SANCHEZ AND JOSk G. SALINAS

reduce the fertilizer requirements relative to conventional plantings, the seed costs are greatly reduced (from U.S. $34 to $3/ha; CIAT, 1979). Since seed of improved pasture species is generally scarce, the use of vegetative propagation is an additional advantage. 4 . Use of Crops as Precursors of Pasture Establishmeni

The previously described low-density planting system is not likely to be successful in savanna areas that have been previously fertilized or in recently cleared rain forest areas where vigorous weed and jungle regrowth takes place. In many of these areas a feasible alternative is to grow crops as precursors of pasture establishment, using the land preparation and fertilization practices required by the crops, but interplanting pasture species so that when crops are harvested, the pasture is established (Kornelius et al., 1979; Toledo and Morales, 1979). In effect, pasture establishment costs are largely paid for by the cash crop. Results with an Orthoxic Palehumult in Quilichao, Colombia, shown in Table XV, describe some of the relationships involved. When cassava and Stylosanthes guianensis were planted simultaneously, cassava yields were slightly decreased and stylo production was halved, but a stylo pasture was ready for grazing after the cassava harvest. When cassava was interplanted with a mixture of Brachiaria decumbens and S . guianensis, crop yields were adversely affected by the vigorous grass growth. Although the relative yield totals were identical to the previous Table XV Crop and Pasture Production in Monoculture and Row-Intercropped Systems Planted Simultaneously on an Ultisol from Quilichao, Colombia, Fertilized with 0.5 todha of Dolomitic Lime and 100 kg P,OJha of Triple Superphosphate" Species

Crop Cassava (roots) Cassava (roots) Beans (grain) Beans (grain)

Crop yields (tondha)

Pastureb (No. of cuts)

Pasture (dry matter. tons/ha) RY"

Monoculture

Intercropped

RY

Sum ofRY

(%)

Monoculture

Intercropped

(%)

(%)

S.g. (3)

45.6

38.2

84

2. I

1 .o

48

132

+

42.4

17.0

40

7.0

6.4

92

130

B.d.

S.g. (3) S.8. (1)

I .08

1.08

100

0.80

0.37

40

146

+

I .22

1.24

102

1.70

0.93

55

157

B.d.

S.g. ( I )

"Adapted from CIAT (1979). bS.g., Stylosanthes guianensis 136; B.d.. Brachiaria decumbens. "RY, relative yields; RY = (Intercropped/MonocuIture) x 100.

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

325

case, this combination seriously decreased cassava yields and is therefore not promising. When a crop with short growth duration is used, the results are different. Table XV also shows the same two pasture species planted at the same time with Phaseolus vulgaris. Bean yields were not affected by the presence of either the legume alone or the grass-legume mixture, although pasture growth was retarded by the presence of the bean crop. Nevertheless, a pasture was already established after the bean harvest. Intercropping between pastures and crops is extremely site-specific and weather-dependent. The actual systems to be used must be validated locally, particularly in terms of relative planting rates, row spacing, crop varieties, and fertility levels. On the same location in Colombia, the first upland rice-pasture experiment failed because rice growth was so vigorous that pastures could not compete with it. A second trial with different relative planting dates and spacings produced excellent association of short-statured upland rice with Brachiaria decumbens and Desmodium ovalifoliuin (CIAT, 1979). It is likely that pastures established in such a manner will enjoy a higher initial and residual soil fertility level than pastures established in the conventional manner. If managed in conjunction with other conventionally established pastures, they could serve as a source of protein or energy for cattle herds. F. MAINTENANCE O F ESTABLISHED PASTURES

After the pasture is established, management is aimed at maintaining its initial productivity and botanical composition by manipulating stocking rates, grazing pressure, fertilization, and weed control. Unfortunately, most of the existing information in Oxisol-Ultisol regions is limited to stocking rates and grazing pressure, with little experience of maintenance fertilization rates and weed control. It is generally believed that maintenance fertilization rates should be less than half of the establishment rates of all nutrients applied. Soil tests and field trails can identify the most economical rates and what their frequency of application should be, either every year or every 2 years. Also, these techniques would identify nutritional deficiencies or imbalances that arise with time. Unfortunately, soil testing services for maintenance pasture fertilization are very scarce in tropical America. Pasture degradation in the Amazon has received considerable attention. Accdrding to Hecht (1979) most of the Panicum maximum pastures in the Brazilian Amazon are in some stage of degradation. In the Paragominas area of the State of Para, Hecht (1981) reports that about 70% of the cattle ranches went out of business because of degraded pastures. The main causes of degradation are the use of a grass species with relatively high nutritional requirements, no

326

PEDRO A .

SANCHEZ AND JOSC G.SALINAS

fertilization, no legumes, and often excessively high stocking rates. The costs of controlling jungle regrowth becomes too high when the Panicurn maximum population decreases; then the fields are gradually transformed into a secondary forest. Serriio and co-workers (1979) have found that phosphorus deficiency is the limiting factor that sets this process in motion. Phosphorus availability was high immediately after burning the forest, remained above the critical level for up to 4 years, and then declined. The correction of this problem is relatively simple. Se r r b et al. (1979) recommended cutting the jungle regrowth with machetes and burning in the field, then broadcasting 50 kg P,O,/ha, half as simple superphosphate and half as phosphate rock. Under these conditions, the Panicum maximum population increased from abour 25% to 90%. Broadcasting legume seeds is being incorporated into the system. It is likely that potassium, sulfur, and other nutrients may also become limiting with time. Frequent monitoring of soil properties is essential to identify these constraints and correct them quickly. The use of better adapted species that are more tolerant to aluminum toxicity and low levels of available phosphorus could also improve this particular system. The grasses Brachiaria humidicola and Andropogon gayanus and the legume Desmodium ovalifolium appear more promising for these areas than Panicurn maximum. Hutton (1979) asserted that the main reason for pasture degradation in Ultisol-Oxisol regions in Latin America is lack of soil fertility maintenance. This is a correct statement, and it underscores the need to establish critical levels of soil test or tissue analysis for the main species grown in this region, particularly for phosphorus, potassium, calcium, magnesium, sulfur, zinc, boron, copper, and molybdenum. The present lack of such information is a major limiting factor preventing the maintenance of productive pastures in the region. G . MULCHING,GREENMANURES,A N D MANAGEDFALLOWS

In crop production systems, better soil cover protection can be obtained by the use of mulches and green manures. The possibility of using managed fallows as opposed to the typical secondary forest fallow may also improve soil protection. 1 . Mulching

A major component of low-input technology in the subhumid (ustic) forest region of West Africa is the use of crop residues as mulches to maintain soil physical properties (Lal, 1975). Impressive results have been obtained by the International Institute for Tropical Agriculture in Nigeria showing the advantages of mulching for sustained crop production. Most of this work, however, has been

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

327

conducting on Plinthic or Oxic Haplustalfs characterized by a sandy, gravelly topsoil underlain by clayey, gravelly subsoils, often with soft plinthite. Unlike most Oxisols and Ultisols of tropical America, the dominant soils of West Africa's forest region have more acute physical constraints than chemical ones. Limited research on mulching conducted in Oxisol-Ultisol regions of tropical America has provided less positive results than those reported in West Africa. The effect of a 10-cm-thick Melinis minutiflora grass mulch on corn growth on Oxisols of the Cerrado of Brazil provided only slight yield increases (Bandy, 1976; NCSU, 1976). Table XVI shows the results during the rainy season, which included a considerable period of drought stress at about tasseling. Mulching decreased topsoil temperatures by 2-3"C, decreased evaporation losses by 4-7 mm daily during the water stress period, and reduced water stress in the plant as evidenced by a lower leaf water potential (Bandy, 1976). The resulting average yields, however, were only 6% higher with mulch than without. The experiment also continued during the dry season with an imgation pattern that simulated the water stress periods encountered during the previous rainy season. A black plastic mulch treatment was also included. Corn yields were the same without mulch as with the Mefinis minutiflora mulch, but a significantly higher yield was obtained with the black plastic mulch (Table XVI). This was attributed to vigorous and deeper root development associated with higher soil temperatures caused by the black plastic mulch during the cool dry season in Brasilia (Bandy, 1976; NCSU, 1976). Consequently, the benefits of a grass mulch were not sufficiently attractive to recommend it as a practice. The black plastic mulch is probably too expensive to justify its use. Mulching with Panicum maximum has been extensively evaluated on Typic Paleudults at Yurimaguas, Peru. The overall effect on crop yields, summarized in Table XVII, is not clear. Valverde and Bandy (1981) indicate that mulching is

Table XVI Effects of Mulching on Corn Yields on a Typic Haplustox near Brasilia, Brazil"," Grain yields (tondha)

Treatment No mulching Melinis minutiflora mulch Black plastic mulch

Rainy season

Dry season (irrigated)

6.16 6.54

5.93 5.99 6.75

-

"Sources: Bandy (1976) and NCSU (1976). "Means of varieties and other management treatments per season.

328

PEDRO A. SANCHEZ AND JOS6 G . SALINAS

almost always detrimental to upland rice, since the plants remain greener into maturity and are subject to more fungal attacks. Mulching is especially advantageous to corn when severe drought stress occurs. Since corn is planted during the drier part of the year, it is subjected to more drought stress than rice. Therefore the differences encountered are also related to the amount of rainfall during the cropping season. There were no overall trends on the effect of mulching on the three grain legumes included in this study. Most of the comparisons summarized in Table XVII as well as the Brasilia results shown in Table XVI were conducted at a generally high level of fertilizer inputs. A study conducted at lower input levels by Wade (1978) in Yurimaguas showed a definitely positive effect of mulching on crop yields. Table XVIII shows the relative yields of five consecutive crops that were either left bare or were covered with a Panicum maximum grass mulch or a Pueraria phaseoloides legume mulch. These treatments did not receive fertilizers or lime. The results are compared with a bare plot that received sufficient fertilizer and lime applications to overcome most fertility constraints (120 kg N/hdcrop and 70 kg K,O/ hdcrop, 4 tons lime/ha/yr, and 45 kg P,O,/ha/yr). The yields obtained with this treatment were considered maximum. Crops mulched with Panicurn maximum produced an average of 54% of the maximum yields without chemical inputs. The beneficial effect of the Pueraria phaseoloides mulch was even greater, producing 80% of the maximum yield, again without inorganic inputs. The Panicurn maximum mulch decreased maximum topsoil temperatures by an average 2°C on dry, hot afternoons. It also increased available soil moisture, prevented surface crusting, and reduced weed growth. Both mulch materials had no effect on soil chemical properties, but because of higher yields than in the bare unfertilized plots, they promoted greater nutrient uptake by the crops. Table XVII Overall Effect of Mulching with Panicurn marimum on Crop Grain Yields in Typic Paleudults of Yurimaguas, Peru"

Number Crop

Upland rice Corn Soybeans Peanut Cowpea Mean yields

of harvests 7 4 6 4 1

With mulching (tons/ha)

Without mulching (tondha)

2.10 3.94 2.34 2.96 0.64

2.71 3.56 2.29 2.88 0.74

-

-

-

20

2.56

2.49

"Source: Valverde and Bandy (1981).

LOW-INPUTTECHNOLOGY FOR OXISOLS AND ULTISOLS

329

Table XVIII Overall Effect of Mulching and Green Manure Incorporations in Unfertilized Treatments Relative to the Yields Attained in the Bare, Fertilized Treatments in Five Consecutive Crops".*

Treatments (all unfertilized) Bare soil Grass mulch Grass incorporated Kudzu mulch Kudzu incorporated

First crop, soybeans (1.10)

Second crop, cowpeas (0.74)

Third crop, corn (4.17)

9 14 33

59 103 90 91 77

33 57 70 72 88

-

109

Fourth crop, Fifth crop, peanuts rice (2.88) (2.74)

55 52 69 63 79

64 94 94 90 99

Mean effect

44 64 71 80 90

"Sources: NCSU (1976) and Wade (1978). *Numbers in parentheses are the actual yields in tons per hectare, which were made equal to 100. Values in table are percent of yields in bare, high-NPKL treatments. Yurimaguas, 1974-1975.

2 . Green Manuring Table XVIII also includes treatments in which Panicum maximum and Pueraria phaseoloides were incorporated as green manures after every crop harvest. The yields obtained average 71 and 90% of the maximum, respectively. This suggests an almost equivalent substitution of legume green manure for inorganic fertilization and liming. The incorporation of these green manures also increased soil moisture retention, and reduced bulk density and soil compaction. The kudzu green manure supplied considerable quantities of nitrogen, potassium, calcium, and magnesium to the soil. The addition of bases decreased aluminum saturation and provided a more favorable environment for plant growth. As a result, nitrogen, phosphorus, potassium, calcium, and magnesium uptake of the four crops increased (Wade, 1978). It appears that kudzu green manure can be substituted for fertilizers in Yurimaguas to obtain moderate yields of continuous crops. This is essentially a trade-off between nutrients supplied as fertilizers and the use of green manures. Taking account of the labor involved in incorporating kudzu, the cost of adding 1 kg N/ha as urea is approximately equal to the cost of adding the same amount of nitrogen as kudzu. The trade-off of labor for purchased input appears attractive, but has the disadvantage of the hard work involved in incorporating the green manure: labor shortage at peak periods. Farmers in Yurimaguas seem more interested in obtaining credit to purchase fertilizers and machinery than in carrying and incorporating kudzu with a hand hoe. It should be pointed out that the aforementioned green manure treatments were not grown in situ but were collected from adjacent areas. If grown in situ, green manures would compete with growing an additional crop at the same time. Experience from West Africa

330

PEDRO A. SANCHEZ AND JOSE G . SAUNAS

indicates that farmers would rather grow an additional crop and use fertilizers if available than grow a green manure crop (Sanchez, 1976). Intercropping green manures with cereal crops could be a better alternative because no time is wasted in growing the green manure crop. Agboola and Fayemi (1972) have shown the beneficial effects of such practice in Western Nigeria. 3 . Managed Fallows

A further extension of the green manuring concept would be to substitute the conventional secondary forest fallow with one that could improve soil physical and chemical properties in a shorter period of time. Promising results have already been produced in Alfisols of Nigeria (Jaijebo and Moore, 1964; Juo and Lal, 1977), and the potential of planted kudzu fallow is presently being studied at Yurimaguas with promising results.

H. INTERCROPPING

A N D MULTIPLE CROPPING SYSTEMS

Various forms of intercropping are widely used by farmers in the OxisolUltisol regions of tropical America. They range from intercropping annual food crops to combining annual crops with pastures, permanent crops, or both. These patterns are generally more complex in the udic than in the ustic soil moisture regime. Intercropped systems other than the use of crops as precursors of pasture establishment are not widespread in the savannas. In udic rain forest areas, intercropping is practiced both by shifting cultivators and by large-scale plantations. Unlike other sections of this review, most of the technology described is based on farming rather than research experience. I . Intercropping Food Crops

Traditional shifting cultivators almost invariably intercrop. In the Amazon, a marketable cash crop is planted right after clearing, usually upland rice or corn. Shortly afterward, cassava and plantains are interplanted either in rows or at random with an average spacing of about 2 x 2 m for cassava and 3 x 5 m for the plantains. When the grain crop is ready for harvest the cassava canopy takes over; with time it is gradually replaced by a plantain canopy that can last as long as 2 years depending on the rate of soil fertility depletion and the presence of nematode attacks. Finally the degrading plantain canopy is gradually replaced by a secondary forest fallow, from which occasional plantain bunches may be harvested. There are many variations of the above theme, some of which have been described by Pinchinat et al. (1976) in a review of multiple cropping systems in

LOW-INPUTTECHNOLOGY FOR OXISOLS AND ULTISOLS

33 1

tropical America. Variations include other annual food crops such as cowpeas, pigeon peas (Cujunus cajan), yams (Dioscorea sp.), malanga (Xanthosoma sp.), yautia (Colocasiu esculenta), and a wide variety of vegetable crops. The traditional intercropping pattern has the advantage of keeping a continuous crop canopy over the soil, imitating the regrowth of a forest fallow eventually becoming one. Soil exposure to erosion and compaction hazards is limited, and the use of acid-tolerant species such as rice, cassava, and plantain permits a better utilization of the available soil nutrient supplies. The more nutrientdemanding crops, such as corn, or the most valuable, such as rice, are normally grown first to capitalize on the fertilizer value of the ash. Research has shown that intensifying intercropped systems can produce higher annual yields than when individual crops are grown in monocultures. In an Ultisol of Yurimaguas, Peru, Wade (1978) also developed a row-intercropped system that produced nine consecutive crops in 21 months. A virgin rain forest was cleared by the slash-and-burn method and the first upland rice crop was grown without fertilization. Following the rice harvest, corn was planted in rows 2 m apart and soybeans in three rows 50 cm apart between the corn rows. Forty-five days later, cassava cuttings were inserted in the corn rows as 1-m spacing. Soybeans were harvested at 91 days and corn at 105 days. Cassava grew vigorously in the former corn rows and cowpeas were planted where the soybeans had been. The four crops were harvested in 266 days. A second cycle started 1 month afterward. Corn was planted the same way, but upland rice replaced soybeans as the companion crop. Cassava was planted again in the corn rows, this time 67 days after the corn seeding. Corn was harvested at 105 days and rice after 140 days. Five days after the rice harvest, peanuts were planted where the rice had been and matured 96 days later. There was enough time to grow a crop of cowpeas where the peanuts had been before the cassava canopy closed in. The crop yields shown in Table XIX, include a comparison of monocultures grown in separate stands at the same time. Although the yields of individual crops were always lower under intercropping than as monocultures, the total market value of 1 ha of intercropping was 20-28% higher than if the same hectare were split among the four or five crops grown as monocultures. Intercropping produced more protein and energy per hectare than the monocultures. Also, intercropping increased nutrient uptake and the efficiency of nitrogen fertilizer used (Wade, 1978). The annual fertilizer application was moderate for the very acid soil conditions: 1 ton limelha, 45 kg Nlha, 100 kg P,O,/ha, 45 kg Klha, 10 kg Slha, 0.5 kg Blha, and 0.5 kg Molha. Although this intensive intercropping system does not require high levels of purchased inputs, it requires intensive hand labor. Therefore its value may be limited to small areas near the farmhouse, while less labor-demanding systems could be used on a larger scale.

PEDRO A. SANCHEZ AND lOSk G. SALINAS

332

Table IX Intensive Intercropping Systems Producing 4-5 CropdYr as Compared with Growing the Same Crops under Monoculture in a Typic Paleudult at Yurimaguas, Peru“,b

Year I :

Corn

Soybeans

Cassava

Cowpeas

Total market value (U.S. $/ha)

Intercropped Monoculture

1.54 3.35

0.83

11.7 16.8

0.54 1.05

1055 879

Grain or tuber yields (tons/ha)

1.15

Percentage over monoculture

-

20

Grain or tuber yields (tondha)

Year 2:

Rice

Soybeans

Cassava

Peanuts

Cowpeas

Intercropped Monoculture

2.01 2.38

0.52 1.19

8.0 22.9

2.62 3.05

0.24 0.47

1996 1558

28

-

“Source: adapted from NCSU (1975, 1976) and Wade (1978) *Tall crops spaced in 2-m rows.

Other intercropping systems can be even more efficient. Leihner (1979; CIAT, 1980) reported that when cassava was interplanted with cowpeas or peanuts in an Orthoxic Palehumult of Quilichao, Colombia, at their normal planting densities, neither crop suffered significant yield declines. This was apparently due to less interspecific competition between the early-maturing grain legumes and the later-maturing cassava. Planting cassava in double rows spaced at 2-3 m with 50 cm between rows has increased yields significantly and enhanced the advantages of intercropping throughout Brazil (Oliveira, 1979). These and other refinements may further increase the value of intercropping acid-tolerant annual crops in Oxisol-Ultisol regions. 2 . Intercropping Annual with Perennial Crops

Planting of acid-tolerant perennial crops such as rubber, oil palm, guarana, and timber species requires an alternative soil cover until the trees produce a closed canopy. Several variations of the “taungya” agroforestry system are presently being practiced in the Amazon. Corn, cowpeas, and sweet potatoes are grown between rows of rubber, oil palm, and guarana for 2-5 years until the tree canopy fully develops (UEPAE de Manaus, 1978; Andrade, 1979). Although no data on the relative yields of annual and perennial crops are available, there seems to be little interspecific competition for the first 2-3 years. In addition to producing food while a plantation is being established, the soil be-

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

333

tween the tree rows is protected from erosion during most of the year, except for intervals between harvesting of annual crops and the planting of the subsequent one.

3 . Intercropping Pastures with Tree Crops When a legume or legume-grass pasture is grown under young tree crops, the soil is better protected than with annual crops. Many combinations exist in tropical America (Thomas, 1978). Pueruria phaseoluides is used as understory for rubber, Gmelina arborea or Dulbergia nigra plantations in Brazil presumably supplying nitrogen to the tree crops. In some cases, cattle graze the Pueruria with little apparent detriment to rubber production under careful management. When the trees are planted at less than optimal density, certain grasslegume pastures persist and produce beef and milk. This is the case for Brachiaria humidicola-Desmodium ovalifolium pastures under a planted stand of laurel (Cordia afeodora), a fast-growing fuelwood species in nonacid alluvial soils of the Ecuadorian Amazon (Bishop, 1981). The value of agroforestry as a low-input soil management component is now widely acknowledged (Mongi and Huxley, 1979). Research data on agroforestry, however, are difficult to find. The lack of data to accompany these most interesting combinations underscores the need of systematic research aimed at understanding soil dynamics and improving soil management in agroforestry systems. The potential of some annual crop-pasture-permanent crop successions in acid soils of the humid tropics of tropical America is indeed tremendous. There is little doubt that the most stable production system in this environment is the one that produces essentially another tree canopy. It is also the one that require the least chemical inputs because a nutrient cycle between the soil and the trees is reestablished. Acid-tolerant food crops like rice, cassava, soybeans, peanuts, cowpeas, plantains, and others must be grown in order to provide food, but they can gradually be replaced by pastures or better by perennial crops. Oil palm, for example, can produce 5 tons/ha/yr of oil without lime and with modest fertilizer applications in Oxisols and Ultisols (Alvim, 1981). This is three to five times the oil production potential per hectare of other oil crops, including soybeans. Palm oil can be directly used as fuel in diesel engines with minor modifications. Mass production of totally renewable bioenergy could accompany increased food crop and livestock production in Oxisol-Ultisol regions. 1.

CONCLUSIONS

The desirable goal of keeping the soil covered by a plant canopy during most of the year can be accomplished by various low-input technology components in

334

PEDRO A . S b C H E Z AND JOSE G . SALINAS

Oxisol-Ultisol regions. Some, like low-density pasture seedings, take advantage of acid soil infertility in suppressing weed growth. An understanding of changes in chemical and physical soil properties with time is helpful for designing or improving continuous farming systems in acid, infertile soil regions. It would be ideal from the ecological point of view if this review could stop at this point. Unfortunately, few of the above systems would remain productive unless fertilizers and lime were added to partially overcome critical acid soil constraints. The remaining sections of this review address this issue.

V. MANAGEMENT OF SOIL ACIDITY Soil acidity constraints are largely eliminated in the northern temperate regions of the world by liming to increase the soil pH to near neutrality. This strategy does not work in most Oxisol-Ultisol regions because of the different chemistry of low-activity clay minerals, which often results in yield reductions if such soils are limed to near neutrality (Kamprath, 1971). In addition, lime transportation costs are often very high in many savanna and rain forest areas. Nevertheless, the main soil acidity constraints (aluminum and manganese toxicities, calcium and magnesium deficiencies) need to be alleviated in order to have successful agriculture in these regions. The importance of these constraints has been indicated in Table 11. Aluminum toxicity and calcium and magnesium deficiencies occur in about 70% of the acid, infertile soil region of tropical America, and on approximately half the territorial extension of tropical America as a whole. Three main strategies are used to attenuate acid soil stresses without massive lime applications: (1) lime to reduce aluminum saturation below toxic levels for specific farming systems; ( 2 ) lime to supply calcium and magnesium and to promote their movement into the subsoil; and (3) use of plant species and varieties tolerant to aluminum and manganese toxicities. A . LIMETO DECREASE A L U M I N U M SATURATION

There are three major considerations when adding lime to decrease aluminum saturation: determination of how much, if any, lime should be added, consideration of the quality of lime used, and promotion of the longest residual effect.

I . Lime Rate Determination The diagnosis of aluminum toxicity in acid soils of tropical America has been based on exchangeable (Exch.) aluminum extracted by 1 N KCI since the

335

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

1960s (Mohr, 1960; Cate, 1965; Kamprath, 1970; Salinas, 1978). Liming recommendations are commonly derived from the following formulas, where the lime requirement is expressed in either milliequivalents of calcium or tons of CaC03 equivalent per hectare: meq Cd100 g soil tons CaC0,-eq/ha

x meq Exch. Al/l00 g

(1)

1.65 x meq Exch. AVIOO g

(2)

= 1.5 =

Lime applications based on these formulas usually neutralize most of the exchangeable aluminum and raise the soil pH to the range 5.2-5.5. Figure 13 shows the relationship between pH and exchangeable aluminum levels in acid soils of Panama (Mendez, 1973). The very low levels of exchangeable bases common to these soils must be taken into consideration, along with the amounts of exchangeable aluminum present (Olmos and Camargo, 1976; Freitas and Silveira, 1977). Percent aluminum saturation [Exch. AV(Exch. Ca + Mg + K + Al) x 1001 expresses these relationships well. Lopes and Cox (1977a) suggested that in most cases the

A W

1

2 W

A

3-

c3

z a A

$ 2 W

A A

I -

A

A

A

A A

I

,

SOIL p H FIG. 13. Exchangeable A1 at different pH values in nine Oxisols and Andepts from Panama. (Source: Mendez, 1973.)

336

PEDRO A. SANCHEZ AND JOSE G . SALINAS

percentage aluminum saturation should be considered first, since soils having the same level of exchangeable aluminum but different degrees of aluminum saturation would have different crop responses to liming at the same lime rates. Moreover, Evans and Kamprath (1970), Kamprath (1971), and subsequent workers, including Spain (1976), have indicated that for many crops the liming requirements based only on exchangeable aluminum may overestimate the lime rates because of varying degrees of plant tolerance to aluminum. From the pioneer work of liming an acid soil of tropical America by Menezes and Araujo in Brazil 30 years ago (Coimbra, 1963) until a recent experiment established 8 years ago also in Brazil (Gonzalez et a l . , 1979), the common approach has been to lime the soil for optimum crop response. This criterion can be interpreted as changing the soil to satisfy the plant’s demands. This approach is difficult to apply in many areas of tropical America due to economic constraints. It may also be noted that Kamprath (1971) has reported that excessive liming may have a detrimental effect on plant growth, for example, lime-induced zinc deficiency in cassava (Spain, 1976). Therefore, it is important to determine the most appropriate formula to convert exchangeable A1 into the amount of lime for specific soil-crop systems. Cochrane e f al. (1980) developed a formula for determining the amount of lime needed to decrease the aluminum saturation level of the topsoil to the desired range: Lime required (tons CaC0,-eq/ha) = 1.8[A1 - RAS(A1 + Ca

+ Mg)]/100,

(3) where RAS is the critical percent aluminum saturation required by a particular crop, variety, or farming system to overcome aluminum toxicity, and Al, Ca, and Mg are the exchangeable levels of these cations expressed in meq/100 g. When compared with actual field data, the predictability of this equation is excellent (Cochrane et al., 1980). An additional advantage is that it requires no soil analysis beyond the 1 N KCl extraction of aluminum, calcium, and magnesium as well as the information about crop tolerance to aluminum in terms of percent aluminum saturation. The adoption of such a formula could lead to the more effective use of lime and considerable savings in the quantities applied as well as in cost. 2 . Use of Quality Liming Materials In addition to how the amounts of lime to be applied are determined, the quality of the liming material deserves consideration. Unfortunately, usually little attention is given in Oxisol-Ultisol regions of tropical America to particle size and chemical composition of lime, other than whether the lime is calcitic or dolomitic (Lopes, 1975). Characterization studies of local lime deposits such as that conducted by Guimariies and Santos (1968) for the State of Para in the

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

337

Brazilian Amazon should be encouraged. The ideal liming material should be in the carbonate form with all of it passing a 10-mesh sieve and 50% passing a 100-mesh sieve. Coarse CaCO, sources seldom produce the desired yield responses for the first crop because they are slow to react. In order to compensate, farmers often apply higher than recommended rates, which may cause overliming problems for later crops (Camargo et a l . , 1962; Jones and Freitas, 1970). In parts of the Amazon, most of the lime sources are exploited for construction purposes and hydrated lime, Ca(OH),, is produced. This liming material is extremely reactive and produces short-lived residual effects (NCSU, 1975, 1976). The alternative for better utilization of Ca(OH), as a lime source is smaller and more frequent application rates (Wade, 1978). A better alternative is to request the lime producers to grind the limestone to the appropriate size and thus keep it in the carbonate form. Since magnesium is frequently limiting in Oxisols and Ultisols, dolomitic lime sources are preferred. A Ca:Mg ratio of 10:l in the liming material is generally considered adequate, although there is very little data to support this assertion. 3 . Residual EfSects of Lime The beneficial effects of liming acid soils are usually expected to last for several years. However, the residual effects are often shorter in tropical than in temperate regions because of higher rainfall and higher temperatures (Lathwell, 1979). The estimation of the residual effects of liming acid soils becomes a main soil management concern for udic tropical rain forests and ustic savanna regions. The length of the residual effect will also depend on the ecosystem. In general, acid soils in tropical rain forests will have shorter residual effects of lime than in savanna regions because of faster release of aluminum from organic matter complexes and higher base removal by plants in year-round crop production systems, and perhaps higher leaching losses in the rain forests (Villachica, 1978). Figure 14 shows the changes in topsoil exchangeable aluminum, calcium, and magnesium within 4% years after applying lime on an Oxisol from Carimagua, Colombia, on which seven annual crops were grown consecutively. There was an increase in exchangeable aluminum with time at all but the high lime rate, probably caused by leaching of bases, release of H+ ions from organic matter, and residual acidity of nitrogen fertilization. The losses were on the order of 1-2 tons lime/ha for the 4%-year period. Howeler (1975) considered an annual application of 200-500 kg lime/ha/yr to be sufficient to maintain an adequate level of calcium and magnesium in this soil under continuous cropping and reverse the above increases in exchangeable aluminum. Table XX summarizes the residual effect of a Brazilian long-term liming

338

PEDRO A. S h C H E Z AND JOSk G. SALINAS 4 h

Q,

0

0 3 \

FT

E

-

2

Y

a

S 0

,,fL

--

-€r --A

I

X

w

--\

0 3 I

I

I

1

35 44

II

I 55

O35

55

44

0.4r

Lime applied (tons/ha)

0

0

0

112

A A

2 6

W

"35

44

55

MONTHS AFTER LIME APPLICATION FIG.14. Residual effect of lime (CalMg = 10 : I ) in an Oxisol of Carimagua, Colombia, from January 1972 to August 1976. (Source: Gualdron and Spain, 1980.)

experiment after seven consecutive crops (five of corn, one of sorghum, and one of soybean). After 6% years, soil pH decreased at all lime rates, probably because of the residual acidity of nitrogen fertilizers. Exchangeable aluminum increased with time and exchangeable calcium and magnesium decreased. Aluminum saturation levels increased by about 20% from initial values for the 0, 1, and 2 ton/ha rate. The grain yields indicate an excellent residual effect, with the 1 ton/ha rate still providing over 80% of the maximum yield of soybean in the seventh successive crop. This is probably associated with the relatively high aluminum tolerance of the soybean variety used (UFV-1). B. LIMEAS CALCIUM A N D MAGNESIUM FERTILIZER

The traditional emphasis on NPK fertilization in tropical America (with the welcome addition of sulfur in recent years) has distracted attention from the

339

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

widespread calcium and magnesium deficiencies found in Oxisol-Ultisol regions. In high-input systems, traditional fertilizer sources such as ordinary superphosphate and dolomitic lime often satisfy the plant's nutritional requirements for the three secondary elements. In low-input systems with plants tolerant to high levels of aluminum saturation and low available phosphorus growing on low-effective cation exchange-capacity (ECEC) soils, the correction of calcium and magnesium deficiencies requires direct attention.

I . Availability of Calcium and Magnesium The principal factors affecting the availability of calcium and magnesium in Oxisols and Ultisols are the level of these nutrients in exchangeable form, ECEC, exchangeable aluminum levels, texture, and clay mineralogy (Kamprath and Foy, 1971). Exchangeable calcium and magnesium levels in Oxisols and Ultisols are usually very low. The range encountered in savannas of Brazil, Colombia, and Venezuela is on the order of 0.1-0.7 meq Ca/100 g and 0.06-0.4 meq Mg/100 g for the topsoil (Lopes and Cox, 1977a; Salinas, 1980; C. Sanchez, 1977). Calcium and magnesium levels in the subsoil are usually lower and sometimes are undetectable in Oxisol subsoils (Ritchey et al., 1980). Exchangeable calcium and magnesium levels in rain forest Oxisols and U1tisols are somewhat higher, particularly in the topsoil. The examples previously shown in Table XI1 indicate a range of 0.4-1.46 meq Cd100 g in the topsoil prior to clearing and burning. The same data set indicates a range of 0.07-0.33 Table XX Residual Effects of Lime Applications on an Oxisol of Brasilia in Terms of Changes in Topsoil Properties and Relative Grain Yields at 6 and 66 Months afkr Application" Lime applied in 1972 (tons/ha) 0 I 2 4 8

Exch."Al pH ( I : 1 H,O) (meq/100 g)

Exch. Ca + Mg (meq/100 g)

A1 saturation

(S)

Relative grain yields (%)

6"

66"

6

66

6

66

6

66

6

66

4.7 5.0

3.9 4.2 4.3 4.8 5.2

1.1

1.5

0.6

0.9 0.5 0.2 0.0

1.1

1.1

1.0 0.4 0.1

1.5 3.1 4.4

0.3 0.6 1.0 2.1 4.0

63 45 25 6 2

80 61 46 15

53 85 88 100

50 93 88 89

2

93

100

5.1

5.6 6.3

"Compiled from NCSU (1974). Gonzalez (1976), Gonzalez et al. (1979), CPAC (1979). and Miranda er al. (1980). "Exch., exchangeable. "Months after lime incorporation. Yields refer to the first crop (corn) and the seventh consecutive crop (soybeans). Maximum yields were 4.0 and 2.1 tons/ha, respectively.

340

PEDRO A. SANCHEZ AND JOSk G . SALINAS

meq Mg/100 g. Consequently, topsoil exchangeable calcium levels seem higher in the rain forest than in savanna regions, but exchangeable magnesium levels show no differences. Decreases with depth of these two elements is sharper in the rain forest than in the savannas, but the levels remain within the detectable range. The dynamics of these two nutrients as a result of burning rain forests has been described in Section IV,B. The low ECECs of most Oxisols and Ultisols pose some advantages and disadvantages to the supply of calcium and magnesium. The first disadvantage is the rapid leaching during periods of intense rainfall. During such periods temporary anaerobic conditions may actually occur and inhibit calcium and magnesium uptake by roots. During the dry season, drought stress may accentuate calcium and magnesium deficiencies. The concentration of these elements in tissue samples of Melinis multiflora and native savanna species decreased significantly during the dry season at Carimagua (Lebdosoekojo, 1977). Plants are therefore faced with a difficult situation: probably adequate calcium and magnesium supplies during part of the rainy season, rapid leaching losses during periods of intense rainfall, and low availability of both nutrients during the dry season because of water stress (Gualdron and Spain, 1980). Nevertheless, both native and introduced plants in Oxisol savannas appear to do better in terms of calcium and magnesium than the low soil levels and the adverse moisture-dependent relationships would infer. Rodriguez (1 975) indicated that some species may have more efficient calcium and magnesium uptake mechanisms than those presently understood. Aluminum competes with calcium in the soil solution for exchange sites. Aluminum toxicity therefore can be decreased by calcium additions (Millaway, 1979). In cocoa, the presence of aluminum decreases calcium uptake but not its translocation to the aereal plant parts (Garcia, 1977). Reduction in root development under high aluminum concentrations could be due to calcium deficiency, which hinders the development of tap roots (Zandstra, 1971). In general, soils dominated by 1 : l clays require a lower level of base saturation for adequate availability of calcium and magnesium to plants than soils dominated by 2:l clays (Kirkby, 1979). This is an advantage of Oxisols and Ultisols because of the predominance of 1 : I clays. 2. Fertilizer Requirements

Information about the rates of lime application to satisfy calcium and magnesiuh fertilization requirements is scanty. Table XXI summarizes the experience in Oxisols of the Llanos Orientales of Colombia, with levels in a range on the order of 0.1-0.4 meq/100 g for both elements. In some cases the response of 0.5 ton/ha of dolomitic lime is due to magnesium. Spain (1979) reported this situation for the establishment and mainte-

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

34 1

Table XXI Estimated Lime Requirements for Main Crops and Pasture Species in the Well-Drained Savanna Oxisols of the Colombian Llanos Orientales

Species Crops Rice (tall-statured) Cassava Mango Cashew Citrus Pineapple Cowpea Plantain Corn Black beans Tobacco Peanuts Rice (short-statured) Pastures Andropogon gayanus Panicum marirnurn Brachiaria decurnbens Stylosanthes capitcita Zornia lotifolia Desmodium ovalifoliurn Pueraria phaseoloides Pennisetum purpureum

Lime rate (tons/ha)

0.25-0.5 0.25-0.5 0.25-0.5 0.25-0.5 0.25-0.5 0.25-0.5 0.5 -1.0 0.5 -1.0 1.0 -2.0 1.0 -2.0

1.5 -2.0 1.5 -2.0 2.0 + 0.4 I .5 1.1 0.5 0.5

0.5 I .o 2.6

Source

Calvo et al. (1977) Spain et al. (1975) Spain et al. (1975) Spain et a / . (1975) Spain et a / . (1975) Spain et al. (1975) Spain et al. (1975) Spain et al. (1975) Spain et al. (1975) Spain et al. ( I 975) Spain et al. (1975) Alvarado (undated) Alvarado (undated) Salinas and Delgadillo (1980) Salinas and Delgadillo (1980) Salinas and Delgadillo (1980) Salinas and Delgadillo (1980) Spain (1979) Spain (1979) Salinas and Delgadillo (1980) Salinas and Delgadillo (1980)

nance phase of two pasture legumes, Desmodiutn ovalifolium and Pueraria phaseoloides, in Carimagua, Colombia. A straightforward magnesium response also accounted for most of the lime response by the first crop of corn in a long-term experiment at Brasilia (NCSU, 1974). In rain forest Ultisols of Yurimaguas, Peru, where dolomitic lime is not available, Villachica (1978) recommended magnesium application rates on the order of 30 kg Mg/ha/crop to overcome magnesium deficiencies and prevent K/Mg imbalances. Recent studies show that tropical grasses also differ in their calcium requirements (CIAT, 1981). Figure 15 shows the field response of seven grass species grown in an Oxisol of Carimagua as a function of calcium concentration in plant tissue. The critical internal calcium requirements ranged from 0.32 to 0.60%. Figure 15 also shows the corresponding levels of aluminum saturation, calcium saturation, and lime requirement, according to the formula of Cochrane et ul. (1980). This information suggests that these species should be classified not only

342

PEDRO A . SANCHEZ AND JOSE G . SALINAS P.moximum 604 D. decumbens 659 P. purpurwrn 658

Critical Critical Lime required as a Ca source

4 W

K

60 2.6

74 I .4

I.5

M. mlnutif lora 608

86 0.4

8.humidicola 6013 B. decumbens 606

2 0 p , , ,

I

'

I

11'

0.30 0.8 I . 2 ' o ; d 'ol8'112' ';&'ole' CALCIUM CONCENTRATION IN PLANT TISSUE Critical Ca Sat. (Yo) 13 9 17 Critical A l Sot. (YO)85 89 77 Lime required 0.4 0 1.1 as a Ca source

O

(%I

FIG.15. Critical calcium concentrations in the tissue of seven tropical grasses grown on a Carimagua Oxisol under field conditions. Lime requirement calculated from formula by Cochrane et a / . (1980). (Source: Salinas, unpublished results.) ~~

according to their tolerance to aluminum but also according to their different calcium requirements. 3. Downward Movement of Calcium und Mugnesium Regardless of whether liming is practiced to decrease aluminum saturation andlor to supply calcium and magnesium, its beneficial effects occur mainly at the depth to which it is incorporated, because lime does not move appreciably in soils. The subsoil of most Oxisols and Ultisols is usually quite acid and often presents a chemical barrier to root development, either because of aluminum toxicity, extreme calcium deficiency, or both. It is common to observe roots of annual crops almost exclusively confined to the limed topsoil, with little penetration into the acid subsoil in savanna Oxisols (Gonzalez, 1976; Bandy, 1976) and rain forest Ultisols (Bandy, 1977; Valverde and Bandy, 1981). Such plants suffer from water stress when drought periods occur in spite of having ample soil moisture stored in the subsoil. Large yield losses occur when temporary droughts occur at critical growth stages during the rainy season in Oxisol regions (Wolf, 1977).

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

343

A major objective of low-input technology is to promote root development into these acid subsoils as an alternative to the more expensive supplemental inigation systems. Three strategies have been devised to overcome this problem: (1) deep lime applications i n Oxisols, (2) promotion of the downward movement of calcium and magnesium, and (3) the use of aluminum-tolerant cultivars and species. Although incorporating the same rates of lime application into the top 30 cm rather than the top 15 cm does not appear to be a low-input technology, it has increased corn yields through various seasons in an Oxisol near Brasilia (NCSU, 1974; Salinas, 1978; Gonzalez et al., 1979). This practice is feasible in wellgranulated Oxisols that can be tilled to a depth of 30 cm without major increases in tractor fuel consumption. In Ultisols with a marked textural change within the top 30 cm, this practice is not recommended because it may create severe soil physical problems (Sanchez, 1977). This suggests that not only chemical but also physical soil parameters should be considered when defining the most appropriate liming practice. Olmos (1 97 1) presented experimental results that demonstrate significant differences among various kinds of acid soils because of subsoil aluminum. Figure 16 shows the changes in pH, calcium, magnesium, potassium, and aluminum saturation throughout the profile of a Tropeptic Haplustox. Aluminum toxicity levels that inhibit root penetration are found within the first 80 cm. Below this depth aluminum saturation decreases to values less than 60% (Salinas and Delgadillo, 1980). A major advantage of many acid, infertile soils is that their chemical and PH 6

EXCH. Ca EXCH.Mg EXCH. K Al SATURATION (rneq/100g I (rneq/tWg I (rneq/t00g) (%I 7 0 0.1 0.2 0.3 0.4 05 0 0.050.10 0.150 0.05 OX)0150 50 60 70 80 90

----T---r---llfi"

I

'

' 1 '

loo-

g

120-

4

140

'

r'

'I

-

160180

200

-

-

FIG. 16. Acidity profile of an Oxisol of Carimagua, Colombia. (Source: Salinas and Delgadillo, 1980.)

344

PEDRO A. SANCHEZ AND JOSE G . SALINAS

physical properties permit the downward movement of calcium and magnesium into subsoil layers, thereby decreasing acid soil stresses at depth and increasing root development. Downward movement of calcium and magnesium applied as lime is of little or no practical significance in other soils dominated by highacitivity clays. As mentioned before, lime does not move appreciably in soils, but exchangeable calcium and magnesium do so in low-ECEC Oxisols and Ultisols accompanied by anions such as sulfates or nitrates (Pearson, 1975; Ritchey et al., 1980). The first evidence of this phenomenon in tropical Latin America was reported by Pearson et al. (1962) after applying about 800 kg N/ha/yr as ammonium sulfate to intensively fertilized grass pastures in Puerto Rico. The prob0

0.5

1.0

IS

20

15

30 45

60 e

5

Y

75

90

90

W Q

I00

1

9oL

o/o

AI SATURATION

FIG.17. Residual effects of lime incorporation on changes in soil properties with depth 40 months after lime application to the top 15 cm in a Typic Haplustox from Brasilia, Brazil. Numbers on top of curves are lime rates in tonslha. (Source: NCSU, 1976.)

345

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

able presence of large concentrations of accompanying anions promoted rapid movement of basic cations into the subsoil. Within the last 3 years similar observations have been reported for Oxisols of the Brazilian and Colombian savannas and for Ultisols of the Peruvian Amazon, but at much lower levels of lime and fertilizer inputs (Salinas, 1978; NCSU, 1978; Villachica, 1978; Ritchey et al., 1980; Gualdron and Spain, 1980). Figure 17 shows the changes in soil properties with depth 40 months after applying lime to the top 15 cm of a Brazilian Oxisol and cropping it continuously for 5 years. Subsoil acidity was gradually ameliorated, particularly when high rates of lime were used. With rates of 2 and 4 tons/ha, the critical level of 60% aluminum saturation for corn was reached at a depth of about 30 cm. With 8 tons lime/ha, this level was reached at a depth of about 80 cm. Crop rooting volume did increase as the aluminum toxicity bamer was gradually pushed down (Bandy, 1976). Laboratory column experiments and field observations with the same soil have confirmed the previous results. Ritchey ef al. (1980) showed significant calcium movement down to depths of 180, 75, and 25 cm when CaCl,, CaSO,, and CaCO,, respectively, were mixed with the top 15 cm of an Oxisol column and the equivalent annual rainfall leached through (Fig. 18). Under field conditions, gypsum included in simple superphosphate increased subsoil pH and calcium plus magnesium levels, while aluminum saturation decreased at depths of 75-90 cm 3-4 years after application (Fig. 19). Corn roots growing in the improved subsoil environment were able to take up water and withstand droughts (Ritchey et al., 1980). EXCHANGEABLE Ca oO

-5

60

=

90

g

120

J -

150

v)

180

-

1.0

1.5

4.40

5

0.5

1.0

(meq/ IOOg) I.5 0

7

0.5

1.0

1.5

30

! i 0

0.5

210

Calcium Carbonate

A

B

C

FIG. 18. The effects of various anions on the distribution of calcium after leaching with the equivalent of 1200 mm rainfall in a reconstructed virgin Oxisol profile 0-135 cm. Calcium as carbonate (A), sulfate (B),or chloride (C) was added to the 0-15 cm layer and incubated 3 weeks 0 (A) and initial value (C). (Source: before leaching began. Ca (kg/ha): 0, 800; 0 , 2000; 0, Ritchey et a / . , 1980.)

346

PEDRO A. SANCHEZ AND JOSE G . SALINAS EXCHANGEABLE C o t Mg (rneq/IOOg)

120 L

FIG. 19. Effect of varying rates of simple superphosphate (as kg P-ha) on Ca profile as sampled 4 years after application. (Source: Ritchey et u l . . 1980.)

+ Mg in the soil

It is interesting to observe that considerable increases in subsoil calcium and magnesium can be attained with moderate applications of lime (1 -2 tons/ha) and simple superphosphate (70 kg P/ha).

c.

SELECTION OF

ALUMINUM-TOLERANT VARIETIES

The main component of managing soil acidity is the selection of productive varieties that are tolerant to aluminum toxicity. Screening a large number of ecotypes either in culture solution, in the greenhouse, in the field, or a combination of the three is the preferred procedure. This requires close cooperation between soil scientists and plant breeders. Among the nutrient culture solution screening techniques, the hematoxylin test proposed by Polle er al. (1978) appears to be very useful. Results of culture selection or greenhouse screening, however, must be validated in the field with a representative range of the cultivars screened. Examples of such correlations are given by Spain et al. (1975), Howeler and Cadavid (1976), Salinas (1978), and Salinas and Delgadillo (1980). The latter two studies include the joint tolerance to aluminum and phosphorus stresses, because they tend to occur together (Salinas, 1978). Cultivars can then be classified by the critical aluminum saturation level required for attaining 80% of the maximum yield. For a specific site, this parameter can be reported in terms of lime requirement using the formula of Cochrane er al. (1980), incorporating the required percent aluminum saturation (RAS).

347

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

I. Screening of Annual

Crops

Figure 20 shows an example of 10 wheat varieties screened in this fashion on a Typic Haplustox from Brasilia, Brazil. The results are presented in a modified Cate-Nelson diagram (Cate and Nelson, 1971) plotting percent maximum yield against aluminum saturation, with the critical aluminum saturation level indicated by a vertical arrow. Critical levels ranged from 22 to 60% aluminum saturation, which for that soil represents a lime requirement of 0.5-1.6 tons CaC0,-eq/ha. Figure 21 shows similarly obtained data on five upland rice varieties. Critical aluminum saturation levels ranged from 22 to over 70%, and lime requirements from 0.2 to 1.4 tons CaC0,-eq/ha. These results confirm the existence of wide differential tolerance to aluminum in both rice and wheat. The rice variety Pratio Precoce was not affected by aluminum within the range tested, while the sensitive varieties Flotante and Batatais showed a decreasing linear yield response to increasing aluminum saturation. The general trend shows that wheat varieties bred in Brazil exhibit greater tolerance to both stress factors than do varieties bred in Mexico, such as Sonora 63, INIA 66, and CIANO. Brazilian varieties were selected under acid soil conditions, while the Mexican ones were selected in calcareous soils. Among Brazilian varieties, the two developed closest to the Cerrado, (IAC-5 in Campinas and BH 1146 in Belo Horizonte, were more tolerant to the aluminum and low PARAGUAY -214

SONORA - 6 3

*

1

I+;

CIANO

I N I A -66

AMAZONAS

h

60

I I

I

.

I

Lime Required 1.0 (IOnSlhO)

BH - I 146

TOAOPI

IAS-55

20

I .o

I.6

1.0

IAS-20

42

60

1.0

IAC-5

W 60 [L

40 20 ‘ 0 20

50

Lime Required 08

20

42 10

60

20

50 05

08

ALUMINUM SATURATION

05

(%I

FIG. 20. Critical aluminum saturation levels of 10 wheat varieties grown on a Brazilian Oxisol. “Lime required” refers to the formula of Cochrane et al. (1980). (Source: Adapted from Salinas, 1978.)

PEDRO A. SANCHEZ AND JOSE G. SALINAS

348

401 20

I

!

64 Lime Required ( t o n l h a l

0

20

40

ALUMINUM

>?O 02

05

SATURATION

(%I

FIG.21. Critical aluminum saturation levels of five rice varieties grown on a Brazilian Oxisol “Lime required” from Cochrane e? al. (1980). (Adapted from Salinas, 1978.)

phosphorus than those developed in Rio Grande do Sul (IAS-20 and IAS-55), where the soils, although acid, are generally more fertile than in the Cerrado. Some variability is also observed among the Mexican varieties. These results suggest good possibilities of combining the aluminum tolerance of the Brazilian varieties with the short-statured, lodging-resistant plant type of the Mexican varieties. A third field study conducted in Oxisols in the State of Parana, Brazil, compared the differential aluminum tolerance of 10 soybean cultivars. Muzilli et al. (1978) defined the critical aluminum saturation level as that required to obtain 80% of the maximum yield. This procedure is quite similar to that reported by Salinas (1978) in Figs. 20 and 21 since the modified Cate-Nelson plot diagrams show that yields at critical aluminum saturation levels were on the order of 70-80% of the maximum. Table XXII shows the classification of Muzilli et al. None were classified as tolerant, which Muzilli et al. defined as having a critical level of more than 25% aluminum saturation. These critical levels may vary with location and management, and particularly

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

349

Table XXII Classification of Soybean Cultivars according to Critical Aluminum Saturation Levels (Required for 80% Maximum Yields) in Oxisols of Parana, Brazil" Critical Al saturation

Category

Cultivar

Very susceptible

Andrews Cobb

Moderately susceptible

Florida Bragg Sant 'ana Hutton Santa Rosa UFV- I Vicoja Bossler

level (9%) 9 10

13 15 17 18 18 21 22

22

"Source: Muzilli et ul. (1978)

with the availability of calcium, magnesium, and phosphorus in the soil during the experiment. For example, one soybean variety, Improved Pelican, was tested in Yurimaguas, Peru (NCSU, 1976), using the same procedure as in the Brazilian experiment. Improved Pelican showed a critical aluminum saturation level of 40%, which was not approached by soybean cultivar in Parana. Nevertheless, such studies clearly show which cultivars are more tolerant. The Parana study suggests that the cultivars Bossier, ViGoja, and UFV-1 should be used instead of Andrews, Cobb, or Florida, as far as aluminum tolerance is concerned.

2. Screening of Pasture Species A somewhat different approach has been followed by Salinas and Delgadillo (1980) in their systematic screening of grass and legume ecotypes for adaptation to aluminum and phosphorus stress. Both absolute and relative yields are considered since growth vigor during the establishment phase is an important consideration in the selection of superior pasture ecotypes. Salinas and Delgadillo considered a 50% maximum yield level as an index of survival, 50-79% maximum yield as moderate tolerance, and 80% of the maximum yield or more as high tolerance under conditions of high aluminum and phosphorus stresses. The 50% limit is consistent with biologic toxicology (Matsumura, 1976; Liener, 1969; Lal, 1980), while the 80% limit was set as the point beyond which the response curve is nearly flat. Table XXIII, adapted from Salinas and Delgadillo (1980), summarizes the behavior of six grass and nine legume ecotypes at different levels of aluminum

350

PEDRO A. SANCHEZ AND JOSk G. SALINAS

Table XXIII Differential Tolerance Rating of Pasture Grass and Legume Species under Field Conditions in an Oxisol of the Colombian Llanos OrientalesO Tolerance categories”

No lime Species and CIAT No. Grasses/soil test levels (ppm Bray 11): Brachiaria hurnidicola 692 Andropogon gayanus 621 Melinis minutiflora 608 Brachiaria decuinhens 606 Panicum muximum 604 Penisetuni purpureurn

Legumeskoil test levels: St.vlosanthes capitata 1078 Stylosanthes guianensis I84 Centroserna hybrid 438 Stylosanthes capirata 1405 Stylosanthes cupirata I0 I9 Desinodium ovalifolium 350 Desmodium heterophyllum 349 Macroprilium sp. 506 Leucaena leucocephala 734

0 kg P/ha

0.5 ton lime/ha

5 tons lime/ha

227 kg P/ha

I7 kg P/ha

227 kg P/ha

17 kg P/ha

H M H S S S

M H H S M M

M M M M M M

227 kg P/ha Maximum dry matter yield 92% Al 90% Al 89% Al 86% Al 81% A1 26% Al 22% Al 1.7 P 2.1 P 11.7 P 2.3 P 14.8 P 1.5 P 18.3 P (tons/ha) I7 kg P/ha

M

H

H

M S

M

M

M S S S

H S S S

S S S

S

M M M H H

3.33 7.35 3.09 3.58 5.86 6.98

92% Al 92% Al 92% Al 86% Al 86% Al 27% Al 27% Al 1.6P 2 . 6 P 24.1P 2 . 6 P 2 4 . 1 P 1 . 6 P 2 4 . 1 P

M S S S S S

X X X

M M M M M S X X X

H M H H M M S M S

M H M M M M S S S

M H H H M H S M S

M M S M M M M S H

H H M M M M H M M

4.04 2.66 2.04 2.88 2.67 3.68 2.41 2.96 1.56

“Adapted from Salinas and Delgadillo (1980) and ClAT (1980). “ X , dead; S, surviving ( - G O % maximum yield); M, moderate (50-80% maximum yield); H, highly (>80% maximum yield).

and phosphorus stress at Carimagua, Colombia. The unamended Oxisol topsoil had 93% aluminum saturation and 1.7 ppm available P (extracted by the Bray I1 method). Treatments included lime rates of 0.5 ton/ha to supply calcium and magnesium and 5 tons/ha to neutralize most of the exchangeable aluminum. This latter rate decreased aluminum saturation to about 25%. Two phosphorus rates were included: 17 kg P/ha as minimal and 227 kg P/ha to attenuate and overcome most of the high phosphorus fixation capacity of the soil. The field design was a factorial of four lime rates x three phosphorus levels. Plant tolerance was classified as high (H) when the relative yield exceeded 80%, moderate (M)

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

35 1

between 50 and 79%, surviving (S) between 1 and 49%, and dead (X) for those that did not survive. Table XXIII shows a marked differential response among grass and legume ecotypes. The tolerance rating varied with different levels of aluminum and phosphorus stresses. In the case of the grasses there was an overall positive growth response as the stresses were gradually eliminated, except for a decrease in yields by Brachiaria humidicola and Andropogon gayanus at high lime and phosphorus levels. Brachiaria humidicola and Andropogon gayanus showed the greatest overall tolerance, P enniserum purpureum the least. The absolute yields show that Andropogon gayanus was the most productive overall. Also this species attained over 80% of the maximum yield with 86% aluminum saturation and 2.3 ppm P, a result of adding 0.5 ton lime/ha to supply calcium and magnesium and 17 kg P/ha. Panicum maximum showed less overall tolerance but relatively high absolute yield. Under Carimagua conditions this species required relatively high levels of lime and phosphorus to reach 80% of its maximum yield. As a group, the legumes listed in Table XXIII are generally more tolerant to acidity and low phosphorus than the grasses, except for Desmodium heterophyllum, Macroptilium sp., and Leucaena Ieucocephala. These ecotypes died unless 0.5 ton lime/ha and some phosphorus were added. Stylosanthes showed generally better performance than other genera. Such ratings do not guarantee the success of a tolerant ecotype under grazing conditions. Persistence and productivity of the pasture also depends on many other plant attributes, including regrowth capacity, tolerance to defoliation, trampling, drought, insect, and disease stresses. Nevertheless, the tolerance ratings give a clear estimate on the inputs needed to overcome acid soil constraints. D. SELECTION OF MANGANESE-TOLERANT VARIETIES

Manganese toxicity is another constraint present in certain Oxisols and U1tisols. Although its geographical extent is not known (Table 11), it is believed to be less common than aluminum toxicity. Manganese toxicity occurs in soils high in easily reducible manganese, usually with fairly high organic contents than can cause temporary anaerobic conditions. Manganese is very soluble at pH values lower than 5.5, particularly under anaerobic conditions where Mn4+ is reduced to Mn2+. Temporary anaerobic conditions may occur in well-drained Oxisols and Ultisols due to rapid decompositions of organic matter and/or temporary waterlogging during periods of heavy rainfall. Examples of such soils are Cot0 clay, a Tropeptic Eutrorthox from Puerto Rico (Pearson, 1975), and some Orthoxic Palehumult soils at CIAT’s Quilichao station in Colombia. Unlike aluminum toxicity, manganese toxicity can occur at pH levels as high as 6.0

352

PEDRO A. S h C H E Z AND JOSk G. SALINAS

(Simar et al., 1974). The lime levels commonly needed to raise the pH of manganese-toxic Oxisols and Ultisols to about 6 is usually very high. For example, to raise the pH from 4.6 to 6.0 in the Ultisol at CIAT's Quilichao station it is necessary to apply pure CaCO, at a rate of 20 tons/ha. (CIAT, 1978). Consequently, the main strategy is to select tolerant varieties. Unlike aluminum toxicity, symptoms of manganese toxicity occur in the leaves because this element tends to accumulate in the aerial parts, while excess aluminum accumulates in the roots (Foy, 1976b). Manganese toxicity symptoms include marginal chlorosis, induced iron deficiency, distortion of young leaves, and localized spots where manganese accumulates (Vlamis and Williams, 1973; Foy, 1976b). In general it seems that legumes are more susceptible to manganese toxicity than grasses (Lohnis, 1951; Hewitt, 1963). Australian scientists have found important differences in tolerance to manganese excess among the main pasture legume species. Table XXIV shows Andrew and Hegarty's ranking of manganese tolerance of major Australian tropical legumes. Souto and Dobereiner (1969) also found similar differences in manganese-toxic Oxisols of the State of Rio de Janeiro, Brazil. Their results shown in Table XXV suggest that Centroserna pubescens is relatively tolerant, while Pueraria phaseoloides is sensitive. Ongoing work by Salinas (unpublished) shows the opposite results, according to visual observation in Ultisols of Quilichao, Colombia. Australian scientists are breeding specifically to incorporate manganese tolerance into Macroptiliurn atropurpureurn because the widespread variety Siratro is quite sensitive to manganese toxicity (Hutton et al., 1978). Table XXIV Differential Response of Nine Forage Legumes to Manganese Toxicity in Australia"

Species Centrosema pubescens Stylosanthes humilis Lotononis bainesii Macroptilium lathyroides Leucaena leucocephala Desmodium uncinatum Medicago sativa Glycine wightii Macroptilium atropurpureum

Regression coefficientb -0.0023 -0.0038 -0.0039 -0.0066 -0.0077 -0.0080 -0.0102 -0.0128 -0.0159

Tolerance rating

2 3 4 5

6

I 8

Internal critical level (ppm Mn) 1600 1140 1320 840 550 I160 380 560 810

OSource: Andrew and Hegarty (1969). blndicates magnitude of dry matter production decreases with increasing manganese levels

353

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS Table XXV Differential Response of Five Tropical Forage Legumes to Manganese Toxicity in Rio de Janeiro State, Brazil" Regression coefficient

Species Stylosunthes guiunensis Glycine wightii Centrosema pubescens Macroptilium arropurpureum Pueraria phaseoloides

-0.014 -0.091

Tolerance rating

:I

I Tolerant

-0.162 -0. I97

4

-0.210

5 Sensitive

Source: Souto and Dobereiner ( I 969).

Very little has been accomplished in establishing external (soil) or internal (foliar) critical levels of manganese toxicity. Andrew and Hegarty (1969) have developed internal critical levels shown in Table XXIV but they do not follow their tolerance rankings. Based on preliminary work at CIAT, more than 100 ppm 1 N KC1-extractable Mn within the top 50 cm of the soil could be considered as a tentative indication of manganese toxicity (Sanchez and Cochrane, 1980). This figure needs local validation before it can be considered as an external critical level for manganese toxicity. E. CONCLUSIONS

Although about 70% of the land area of the Oxisol-Ultisol regions of tropical America posses severe acid soil constraints, it is not necessary to lime these soils to neutrality or even to pH 5.5 in order to obtain sustained crop and pasture production. Estimates of long-term world food production need should not include heavy rates of lime applications for the 750 million ha of tropical America with serious aluminum toxicity, calcium deficiency, and magnesium deficiency constraints. At the same time, statements that sustained agricultural production is possible without liming in most Oxisols and Ultisols are misleading. The existence of very aluminum-tolerant varieties of forage species and crops may eliminate the need to decrease the aluminum saturation level of the soil by liming, but in most cases the plants require fertilization with calcium and magnesium. This can be accomplished by small lime applications or by fertilizers containing sufficient amounts of these two essential nutrients. Small lime applications are probably less expensive per unit of nutrient than calcium and magnesium fertilizers. A very positive attribute of many Oxisols and Ultisols of tropical America is the relative ease of movement of calcium and magnesium into the subsoil. It is

354

PEDRO A. S h C H E Z AND JOSE G.SALINAS

possible to take advantage of what is normally considered a soil constraint-low ECEC. Together with a favorable soil structure and plenty of rainfall, low ECEC favors the gradual amelioration of the chemical properties of the subsoil. This, in turn, favors deeper root development and less chance of drought stress.

VI. PHOSPHORUS MANAGEMENT Phosphorus deficiency is one of the most widespread soil constraints in tropical America. Approximately 82% of the land area of the American tropics is deficient in phosphorus in its natural state (Table 11). In the Oxisol-Ultisol savannas and rain forests the estimate increases to 96% of the area (Sanchez and Cochrane, 1980). Phosphorus deficiency problems are compounded by widespread high phosphorus fixation capacity. Soils with high phosphorus fixation capacity can be defined as those that require additions of at least 200 kg P/ha in order to provide an equilibrium concentration of 0.2 ppm P in the soil solution (Sanchez and Uehara, 1980). Acid soils that fix such amounts of phosphorus can be identified as those with loamy or clayey topsoil textures with a sesquioxide/ clay ratio of 0.2 or greater, or by the dominance of allophane in the clay fraction of the topsoil (Buol et af., 1975). About 53% of tropical America’s land surface is dominated by soils with such high capacity to fix phosphorus. In the OxisolUltisol regions this figure increases to 72%, but high-fixing soils are less extensive in the Amazon jungle than in the savannas (Cochrane and Sanchez, 1981). Figure 22 shows some examples of phosphorus sorption isotherms according to the Fox and Kamprath (1970) procedure. Among Oxisols and Ultisols phosphorus fixation generally increases with clay content because of its direct relationship with surface area, where the iron and aluminum oxides and hydroxides largely responsible for phosphorus fixation are located (Pope, 1976; Lopes and Cox, 1979; Sanchez and Uehara, 1980). High phosphorus fixation is considered one major reason why vast areas of arable lands in tropical American savannas are largely underutilized (Leon and Fenster, 1980). The relatively high unit cost of phosphorus fertilizers coupled with the widespread deficiency and fixation constraints require the development of low-input technologies that can make most efficient use of applied phosphorus in these soils. Salinas and Sanchez (1976), Fenster and Leon (1979a,b), Leon and Fenster (1979a,b, 1980), and Sanchez and Uehara (1980) have suggested similar strategies in order to develop sound phosphorus management for crops and pastures on the acid, infertile soils of tropical America. The strategy now consists of six major components, five of which are relatively well established: (1) determination of the most appropriate combination of rates and placement methods to enhance initial and residual effects; (2) improvement of soil fertility evaluation

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

I I I

I400

355

% Clay

I200

-E

1000

CL

Y

n W m

800

600

a 400

200

0

0.00I

0.01

0.05

0.2

I .o

P IN SOLUTION (ppm) FIG. 22. Examples of phosphorus sorption isotherms of Oxisols and Ultisols in some research centers in tropical America. (Sources: Sanchez, 1976; CIAT, 1978; Sanchez and Isbell, 1979.)

procedures for making phosphorus recommendations; (3) use of less costly sources of phosphorus such as phosphate rocks, either alone or combined with superphosphate; (4) use of moderate amounts of lime to increase the availability of soluble phosphorus sources; (5) selection of species and varieties that can grow well at lower levels of available soil phosphorus; and (6) exploration of the practical possibilities of mycorrhizal associations to increase phosphorus uptake by plants. These strategies are discussed in the following sections. A . RATESA N D PLACEMENT METHODS OF PHOSPHORUS APPLICATIONS

Extensive research has been conducted in tropical America to determine the optimum crop responses to phosphorus fertilization in Oxisols and Ultisols (Kamprath, 1973). Most of it, however, is limited to broadcast applications of superphosphates and their incorporation into the topsoil. Although this applica-

356

PEDRO A. SANCHEZ AND JOSfi G.SALINAS

tion method usually produces large yield responses, such as the one shown in Fig. 2 (Section I,B), the high rates required and the placement method are not necessarily the most efficient way to apply phosphorus.

I . Annual Crops A long-term experiment conducted on a high-fixing Typic Haplustox of the Brazilian Cerrado provides a comparison of banded versus broadcast superphosphate applications for a sufficient period of time to adequately evaluate the residual effects. Figure 23 (drawn from data by NCSU, 1974, 1975, 1976, 1978; CPAC, 1978, 1979, 1980; Yost et al., 1979; and Miranda et al., 1980) shows the results of different rates and placement of triple superphosphate on nine consecutive corn harvests during a 7-year period. Contrary to conventional opinion, banding was inferior to broadcast applications for the first crop. This soil was so deficient in phosphorus that root development was restricted to topsoil areas that had received phosphorus fertilization. With subsequent crops this effect disappeared as the banded applications were mixed with the rest of the topsoil by tillage operations. Considering the long-term effects, the highest average grain yield of 6.3 tons/ha was obtained by broadcasting a massive application of 1280 kg P,O,/ha and incorporating it into the topsoil prior to the first planting. The residual effect was sufficient to keep the available soil phosphorus level above the critical level of 10 ppm P for corn (by the Mehlich 2 extraction) for 7 years. Economic calculations by Miranda et al. (1980) also indicate that this high-input strategy is the most profitable among the ones studied in this experiment, assuming an annual interest rate of 25% on credit to buy the fertilizer and an average price:cost ratio where 6.7 kg of corn are needed to pay for 1 kg of P,05 as triple superphosphate. The high capital investment and the implications on world fertilizer supplies suggest that other alternatives be pursued. Splitting the 1280 kg P,O,/ha rate into four 320 kg P,O,/ha banded applications to the first four crops produced 97% of the maximum yield; therefore the efficiency of fertilizer use was not affected. This alternative, however, has the disadvantage of initially low yields, but has the advantage of spreading the purchase of phosphorus over 4 years. A similar gradual buildup by banded applications for 4 years to reach a total of 640 and 320 kg P,O,/ha produced 74 and 51% of the maximum yield, respectively. These treatments performed similarly to initial broadcast applications of 640 and 320 kg P,O,/ha (Fig. 23B). The trade-offs are higher initial crop yields with broadcast applications instead of gradual yield increase and a more effective residual effect with the banded applications. Combinations of broadcast and banded applications, shown in Fig. 23C, show more promise. An initial broadcast application of 320 kg P,O,/ha followed by

357

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

0

Q)

p1

0 320

1280 6.2

97

0 160

640 4.7

74

320 3.2

51

x4

x 4 0 80 x4

::I, '0

I

,

,

2

3

5

6

75

320 80x4 640

79

I

I

4

80 80x9 800

7

8

9

CONSECUTIVE CORN CROPS ( I972 -1979) FIG. 23. Residual effects of different rates and placement methods of superphosphate applications on nine consecutive corn crops on a Typic Haplustox from Brasilia, Brazil. (A) Broadcast in 1972; (B) banded crops 1-4; (C) broadcast plus banded. Solid lines, successive applications; dashed lines, no additional P applications. (Compiled from NCSU, 1975, 1976, 1978; CPAC, 1979, 1980; and Yost er d.,1979.)

358

PEDRO A. SANCHEZ AND JOSB G . SALINAS

four banded applications of 80 kg P,O,/ha produced 79%of the maximum yield as an average of the nine crops. Miranda et al. (1980) reported that the economic return to this strategy was similar to broadcasting 1280 kg P,O,/ha once, but the total amount of phosphorus added was reduced to one-half. Another possibility is to broadcast a minimum amount of 80 kg P20,/ha and apply the same quantity in bands to every crop, including the first one. This strategy produced 75% of the maximum yields, but the total investment in phosphorus during the nine crops increased to 800 kg P,O,/ha. Both broadcast and banded combinations have the additional advantage of higher yield stability than either all broadcast or all banded applications. In retrospect, a more effective treatment may have been an initial broadcast application of 160 kg P20,/ha followed by banding 80 kg P,O,/ha to all crops. This would have reduced the total investment to 640 kg P/ha for the nine crops, produced 75-80% of the maximum yield, and avoided large initial capital investments. Considering the high phosphorus fixation capacity of this soil (780 ppm P or 3545 kg P20,/ha to reach 0.2 ppm in soil solution, shown in Fig. 22 as the Oxisol from Brazil), the broadcast-banded application strategies are examples of how to decrease fertilizer phosphorus inputs by a more judicious combination of rates and placement methods, with sufficient time to evaluate the residual effects. 2 . Pastures Phosphorus fertilizer rates and placement considerations are fundamentally different in the case of pastures in these high-fixing soils. The main reasons are the lower phosphorus rates required by acid-tolerant pastures, the lack of subsequent tillage operations that mix applied phosphorus within the topsoil, and a nutrient recycling mechanism via animal excreta under grazing. Figure 24 shows a completely different response pattern of adapted pasture species to broadcast superphosphate applications in an Ultisol from Quilichao, Colombia with a phosphorus fixation capacity similar to that of the Oxisol from Brasilia mentioned in the previous example. Figure 22 indicates that the amount of phosphorus added to maintain 0.2 ppm P in the soil solution is similar in both soils (650 ppm P for the Quilichao Ultisol and 760 ppm P for the Brasilia Oxisol). Annual crops grown on the Quilichao Ultisol require about 400 kg P,O,/ha to approach maximum yields. Pasture species like Panicum maximurn, Andropogon gayanus, and Centrosema pubescens require about 80 kg P205/haas one broadcast incorporated application to approach maximum dry matter production for the first 2 years (Fig. 24). In the Carimagua Oxisol with considerably lower phosphorus fixation capacity (400 ppm P to reach 0.2 pprn P in solution as shown in Fig. 22), adapted pasture species such as Brachiaria decumbens require only 50 kg P,05/ha as triple superphosphate to achieve maximum production (Fig. 25).

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

359

A

v)

C

0

c Y

z 0

FIG. 24. Phosphorus response by Panicum maximum and Centrosema hybrid mixtures in an Ultisol from Quilichao during the establishment year. (A) 0, Panicum maximum; 0 , Andropogon gayanus. (B) Centroseinat 0, Panicum maximum; 0 Andropogon gayanus. (Source: CIAT, 1979.)

At such low levels of application, banding is definitely superior to broadcast incorporated application for pasture establishment, particularly if seeding is also done in bands (CIAT, 1978; Fenster and Leon, 1979b). Pasture species have their maximum phosphorus requirements a few weeks after germination, before a deep root system develops (Salinas, 1980). Consequently, it is important to assure that the seedlings have a nearby supply of phosphorus. Band placement also decreases weed growth between rows in these Oxisols (Spain, 1979). After a pasture is well established, maintenance phosphorus applications can be broadcast on the soil surface without incorporation (NCSU, 1976). This

360

PEDRO A. SANCHEZ AND JOSk G. SALINAS

32 0 E

>c

30

-

28-

,O 26-

Y

0 24J

22-

J

400

kg P205/ha FIG. 25. Phosphorus response of Erachiaria decumbens grown on a Carimagua Oxisol (sum of eight harvests). In the annual treatment P was reapplied I year after planting. 0 , TSP residual; 0, TSP annual. (Source: CIAT. 1978.)

permits the use of lower rates as the contact with the high phosphorus fixing soil is minimized. Although the means by which pasture species utilize surfaceplaced phosphorus is not well understood, apparently the superficial roots are able to absorb and utilize it efficiently. B . THENEED TO

IMPROVE SOIL

FERTILITY EVALUATION

PROCEDURES

Another way to increase the efficiency of phosphorus fertilization is to use better methods for determining fertilizer recommendations. The purpose is to identify the initial phosphorus requirement for a particular species or variety either in terms of available soil phosphorus (external critical level) or foliar phosphorus content (internal critical level). These critical levels are those necessary to provide an adequate level of dry matter defined in this review as 80% of the maximum. The use of the Cate-Nelson (1972) diagrams and the linear response and plateau model, described in Section I,B,2, are quite useful for phosphorus, while the use of quadratic response models tend to exaggerate the optimum rates of fertilizer application (Anderson and Nelson, 1975).

36 1

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

Given the phosphorus fixation constraints in these soils, it is tempting to use estimates of phosphorus fixation as guides for the rates of phosphorus to apply. The most common approach is to extrapolate from the phosphorus sorption isotherms the amount of phosphorus that needs to be added in order to achieve a desired level in the soil solution (Fox et a l . , 1971, 1974). The soil solution level extrapolated to the field that produced 95% of the maximum yield was defined by Fox and co-workers as the “external critical phosphorus requirement.” The range of this critical level among species is 0.05-0.6 ppm P (Fox et al., 1974). Table XXVI shows the amounts of broadcast superphosphate required to maintain specific soil solution levels in the field and their equivalence in terms of three common soil test methods. The soil on which the data on Table XXVI were obtained is a clayey Tropeptic Eutrorthox with a high capacity to fix phosphorus (350 ppm P applied to reach 0.2 ppm in solution). When applied to Oxisols and Ultisols of tropical America this approach has been found to exaggerate the phosphorus rate recommendations by a significant amount (Novais and Kamprath, 1979; Smyth and Sanchez, 1980b; Sanchez and Uehara, 1980; Fenster and Leon, 1979a,b). The main reason is found in Table XXVI. The critical soil test levels for grain crops in Latin America, based on the Cate-Nelson approach, is on the order of 8- 15 ppm P by the extractions shown in table (Cano, 1973; Kamprath, 1973; Miranda et al., 1980). Soil solution levels as low as 0.025 ppm P produce soil test values way above the critical soil test levels that have been developed with proper calibration. In addition, it is extremely difficult to establish critical levels of a few parts Table XXVI Solution Phosphorus Levels in Sorption Isotherms Equivalent Soil Test Levels and Amounts of Broadcast Triple Superphosphate Added After 7 Years and 13 Continuous Crops to a Tropeptic Eutrorthox from Hawaii” P maintained in soil solution (ppm P) 0.003 0.006 0.012 0.025 0.05 0.1 0.2 0.4 1.6

P added to soil (kg P,O,/ha)

Soil test P values (ppm P)

Bray I 3 5

14 28 55 72 144

156 339

N.C.

Olsen

In it i a I

6 9 20 35 57 86 I58 209 331

12

80 200 432 682

“Adapted from Yost and Fox (1979).

15

30 44 72 93 164 160 295

I000 1363 1591 1591 3273

Maintenance in 7 yr 1 I4

204 714 1445 2050 2614 369 1 4634 1566

Total I94 404 1 I46 2127 3050 3977 5282 6225 10,839

362

PEDRO A. SANCHEZ AND JOSk G. SALINAS

per billion that often correspond to the agronomically relevant range in such soils. The Langmuir and Freundlich isotherms are difficult to extrapolate at this range. Also, the low concentrations approach the detection limits of conventional spectrophotometers. When considering low levels of phosphorus fertilizer additions (50-150 kg P,O,/ha), the sorption isotherms are of little value (Fenster and Leon, 1979a,b). For example, Fig. 22 shows that the Carimagua Oxisol fixes high amounts of applied phosphorus (400 ppm P or 1818 kg P,O,/ha to reach 0.2 ppm P in solutionj. After 4 years continuous cropping with Brachiaria decumbens, however, an initial application of 50 kg P,O,/ha as triple superphosphate produced 79% of the maximum yield obtained with the 400 kg P,O,/ha rate (Table XXVII). At such low rates, the conventional soil test extraction procedures often do not reflect the amount of fertilizer phosphorus added. Table XXVIII shows the very small increases in Bray I1 available phosphorus when an Oxisol received 0-100 kg P,O,/ha in 20-kg increments. This causes difficulties in making fertilizer phosphorus recommendations based only on soil tests. Some studies have been started to improve the sensitivity of the existing soil tests (CIAT, 1980) Figure 26 shows that increasing the NH,F concentrations in the Bray extractant, which increases the available phosphorus values, reflects the sorbed phosphorus that is available to the plant (CIAT, 1981). Since NH,F is

Table XXVII Relative Agronomic Effectiveness of Several Phosphate Rocks as Determined by Yield of Bruehiunu decumbens Grown in the Field at Carimagua",b Percent relative yield" Phosphorus source TSP Annual TSP Residual" Florida (U.S.) Bayovar (Peru) Gafsa (Tunisia) Huila (Colombia) Pesca (Colombia) Tennessee (U.S. ) Check: 13.6%

25" (32.2)l' 100 (21.1)''

I22 I20 I08 95 110

I04

50"

(34.5) 100 (29.4) 93 80 104 I I3 82 76

Source: Leon and Fenster (1980). "Sum of 13 cuts taken over a 44-month period. "Assumed at 100% for each level of application "P applied in kg P,O,/ha. "Dry matter yields in tonslha.

a

loo"

400"

(35.9)

(43.6)

100 (31.2)

100 (36.8) 104

101 103

I09

104

104

98

110 116 I08

111

96

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

363

Table XXVIII Phosphorus Fractions in the Carimagua Oxisol as a Function of Applied Phosphorus Rates"

0 10 20 40 80 100 I50 200

0 4.4 8.7 17.5 34.9 43.7 65.5 87.3

1.8 I .8 1.9 2.1 2.2 3.5 5.5 6.6

0.9 0.8 1.0 1.1

1.7 1.7 I .9 2.2

0.5 0.6 0.6 0.6 0.9 1.0 1.3 1.5

26 29 32 35 40 42 43 45

29.2 32.2 35.5 38.8 44.8 48.2 51.7 55.3

101

97 97 I08 102 92 101

101

130.2 129.2 132.5 146.8 146.8 140.2 152.7 156.3

"Source: CIAT (1981)

able to extract some of the aluminum- and iron-bonded phosphorus, these fractions might play an important role in releasing phosphorus to the plants, perhaps through root excretion or microbial activity. Table XXVIII shows the phosphorus fractions of the Carimagua Oxisol as a function of phosphorus rates. Increases in calcium- and aluminum-bonded phosphorus contribute to an increase in available phosphorus, but part of the large quantities of iron-bonded phosphorus may be having some influence on the availability of phosphorus. Therefore plants at low rates of applied phosphorus appear to extract phosphorus from these fractions in a way that conventional soil tests are not able to detect. When phosphorus applications are banded, the interpretation of soil tests becomes even more difficult. One possibility is to use tissue analysis as the plant is the ultimate evaluator of soil fertility. Where internal critical levels are available and properly standardized in terms of plant part and age, tissue analysis can be used. Another approach might be to interpret soil test data of samples between the bands in the form outlined in Fig. 27, where soybean yield responses are plotted as a function of soil test values in experiments that involve different broadcast and banding combinations. Where field response data are available, making fertilizer recommendations based on soil tests has the benefit of calibration with known field responses. Table XXIX shows the initial broadcast and annual banding recommendations for clayey Oxisols near Brasilia based on the data shown in Fig. 23. This table shows a decreasing rate of broadcast applications with increasing soil test level.

364

PEDRO A. S h C H E Z AND JOSE G. SALINAS

DM YIELD

P APPLIED

I%) (g/pot)

(kg p205/ha 1

25

0

2-

I

I

1

I

I

FIG. 26. Dry matter production of Brachiaria decurnbens grown in a Carimagua Oxisol as a function of available soil test P levels obtained by different extractant solutions: 0.1 N HCI plus 0.03 (a),0.05 (0).0.10 (A), and 0.20 (A)N NH,F. (Source: CIAT, 1981.)

CROP 2

CROP I

I

I

0 5 BRAY 1 P P P M )

I

10

FIG.27. Relationship between yield response to banding superphosphate applications and available soil test level of a Typic Acrustox from Brasilia, Brazil, grown to two consecutive soybean crops. Band rate (kg P/ha): 0, 22; 0 , 44.(Source:Smyth, 1981.)

365

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

Table XXIX Phosphorus Rate Recommendations for Clayey Typic Haplustox near Brasilia, Brazil, for Continuous Corn Production according to Soil Test Interpretations"

Available P (ppm) (N.C. method) 0.0- 2.0 2.1- 6.0 6.1-10.0 10.1-16.0 >16.0

Soil test interpretation

Relative corn yields (% max.)

Basal broadcast application (kg P,O,/ha)

Banded application per crop (kg P,O,/ha)

Extremely low Very low Low Medium High

0-25 26-50 51-75 76-90 91-100

320 200 80 0 0

so

"Adapted from Miranda

el

80 80 70 60

Total for 9 crops (kg P,O,/ha) I040 920 800 630 540

a / . (1980).

C. USE OF LESSSOLUBLE PHOSPHORUS SOURCES

A third component of the low-input phosphorus management strategy is to utilize the abundant rock phosphate deposits present in tropical South America, shown in Fig. 28. All these deposits, except two, are classified as low-reactivity materials that are considered unsuitable for direct application ( k h r and McClelIan, 1972). The Bayovar rock is considered of high reactivity and the Huila rock is of medium reactivity (Chien and Hammond, 1978; Leon and Fenster, 1979b).

I . Cotnparisons among Sources Table XXX shows the agronomic effectiveness of different phosphate rocks as related to triple superphosphate, using Panicum maximum as the test crop on an Oxisol from the Llanos Orientales of Colombia. High-reactivity phosphate rocks such as North Carolina, Bayovar, and Gafsa performed nearly as well as triple superphosphate. Medium-reactivity phosphate rocks such as Huila and Florida, and even the array of low-reactivity materials from Brazil, Colombia, and Venezuela look promising for direct application in acid soils. The effectiveness of rock phosphates in these soils depends on their solubility, fineness, time of reaction, and soil pH (Khaswahneh and Doll, 1978). On these highly acid soils, even the low-reactivity phosphate rocks are effective with time. Table XXVII shows the results of an experiment conducted on a Carimagua Oxisol with Brachiaria decumbens in which six phosphate rocks with varying agronomic effectiveness ratings were compared to triple superphosphate (Leon and Fenster, 1980). This study included broadcast and incorporated application rates ranging from 0 to 400 kg P,O,/ha. After nearly 4 years, the yields of forage

3 66

PEDRO A. S h C H E Z AND JOS6 G. SALINAS

FIG. 28. Principal rock phosphate deposits in tropical South America. (Sources: Fenster and Leon, 1979a.b. and updated information.)

from the phosphate rock treatments compared favorably with those from triple superphosphate. In many instances the yields with phosphate rocks were considerably higher. For the period of time this experiment has been conducted, a 50 kg P,O,/ha application rate appears to be adequate under field conditions. Similar results have been recorded from a field experiment conducted on Ultisols from Pucallpa and Yurimaguas, Peru (NCSU, 1974; Can0 et al., 1978; Leon and Fenster, 1980), and on an Oxisol of Brasilia, Brazil (NCSU, 1975, 1976; Miranda et al., 1980). In the latter case, the higher phosphorus fixation capacity increased the required rate to about 200 kg P, O,/ha. The use of the very low-reactivity Araxa phosphate in Brasilia had little effect on Brachiaria decumbens growth during the first year of application.

2 . Particle Size of Rock Phosphate Materials The effectiveness of all rock phosphates increases with increasing fineness, in contrast to the opposite effect in water-soluble sources (Terman and Englestad,

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

367

1972). From the practical standpoint, finely ground phosphate rock presents serious problems of handling and spreading that would limit the farmer or the fertilizer dealer in making widespread use of phosphate rock. To solve the problem, the International Fertilizer Development Center initiated a study to determine whether finely ground phosphate rock could be granulated and still retain its agronomic effectiveness. Preliminary greenhouse experiments were carried out using different rates and granule sizes of phosphate rocks, and the results are shown in Fig. 29. The minigranules (-48 + 150 mesh) proved to be as agronomically effective as the finely ground phosphate rock. Apparently, when these "minigranules" came in contact with the soil the KCI binder dissolved. Their effective surface area therefore is similar to that of finely ground materials. Although the larger-sized granules (-6 + 16 mesh) were not as effective initially, they did release increasing amounts of phosphorus with time (CIAT, 1980, 1981). Table XXX Agronomic Effectiveness of Phosphate Rocks as Determined by Yield of Panicurn maximurn Grown on a Las Gaviotas Oxisol in the Llanos of Colombia under Greenhouse Conditions",b Percent age relative yield'' Phosphate rock Brazil Abaete Araxa Catalio Jacupiranga Patos de Minas Tapira Colombia Huila Pesca Sardinata Peru Bayovar Venezuela Lobatera Tunisia Gafsa United States Florida North Carolina

Reactivity rating"

Low Low LOW LOW

LOW LOW

50

I00

200

400

II 30 5

33 33

52 56

6

22

55 58 38

12 27 4

13

19 66 10

42 7

51 72

23

Medium

58

59

LOW

56

61

84 80

83

84

LOW

29

44

68

74

High

99

79

104

91

LOW

56

56

65

76

High

63

72

I14

I05

Medium High

59 70

71 78

86 107

91 108

"Source: Leon and Fenster (1979a,b). "Sum of three cuttings. "Interpreted from Lehr and McClellan (1972) and unpublished sources. "Dry matter yields obtained are with triple superphosphate considered as 100% for each phosphorus rate. Absolute yields: 0.6, 13.3, 19.0, 22.2, and 22.2 g/pot with 0, 50, 100, 200 and 400 mg Plpot as triple superphosphate, respectively.

PEDRO A. SANCHEZ AND JOSE G . SALINAS

368

10

a w

5

4

2 I

0

minigranule;0 , Fic. 29. Effect of rate and granule size of Huila Phosphate Rock (0,ground; 0, regular-size granule) on yield of Pnnicum maximum grown on a Carimagua Oxisol in the greenhouse (2 cuttings). (Source: CIAT, 1980, 1981.)

3 . Applications before Liming f o r Acid-Sensitive Crops

Rock phosphates require an acid soil environment in order to release phosphorus into the soil solution. In some acid soils of tropical America, the effectiveness of high-reactivity rock phosphates decreases if the soil pH increases above 5.0 (Lathwell, 1979). This usually does not pose a problem with most aluminum-tolerant pastures but may inhibit the growth of aluminum-sensitive crop varieties. In terms of crop production, an alternative is to apply the rock phosphate several months ahead of liming in order for it to react at low pH. This procedure is especially advantageous if the first crop to be planted is relatively tolerant to aluminum, as is the case for upland rice. Lime can then be added prior to planting a crop more sensitive to aluminum, such as corn. The time required

369

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

for lime to react in acid soils is less than that needed for the high-solubility rock phosphate sources to react (Sanchez and Uehara, 1980). 4 . Combination with More Soluble Sources

An additional alternative is to broadcast the rock phosphate and apply more soluble phosphorus sources in bands in order to provide phosphorus while the rock phosphate slowly dissolves. Smyth (1981) and CPAC (1980), have shown that broadcasting 200 kg P,O,/ha of low-reactivity Patos de Minas phosphate rock plus annual banded applications of simple superphosphate produce SOYbean yields similar to that of the same rate supplied entirely with simple superphosphate. Table XXXI shows that when' varying ratios of phosphate rock to single or triple superphosphate, the initial growth response of corn in a Colombian Oxisol was proportionate to the amount of soluble phosphorus in the fertilizer mixture (Fenster and Leon, 1979a,b). Comparisons between co-minigranulated phosphate rock and triple superphosphate or simple superphosphate with these soluble phosphorus sources alone show that the granulated materials are superior in every instance. These results indicate that the acid produced from the soluble phosphorus in the granule may be reacting with the phosphate rock, which is releasing additional phosphorus for the plants.

Table XXXI Effect of Ratio of Phosphate Rock to Simple and Triple Superphosphate on Yield of Corn Grown in the Greenhouse on a Carimagua Oxisola,b Percent relative yield" ~~

Phosphorus source

1 :O('

3:1

1:1

1:3

0: 1

Simple superphosphate Triple superphosphate Floriddsimple superphosphate Floriddtriple superphosphate Pescdsimple superphosphate Pescdtriple superphosphate Check: 16%

-

-

-

-

-

1 0 0 (18.9)' 91

71 71 27 27

70 72 53 64

91 92 75 70

99 98 99 89

-

-

"Source: Leon and Fenster (1980). hSum of two harvests.

'All phosphorus rates were averaged. Granule size used: minigranule (-48 "Ratio of phosphate rock to simple and triple superphosphate. "Simple superphosphate assumed at 100%. 'Tissue yield in g/pot.

+ 150 mesh)

370

PEDRO A.

SANCHEZ AND JOSk G.SALINAS 5 . Partial Acidulation

From the aspects discussed previously, it is apparent that many phosphate rocks, although they perform well with time, are initially inferior to the more soluble phosphorus sources for crop production and for certain pastures as well. The work of McLean and Wheeler (1 964) indicates that partially acidulating phosphate rock to levels of 10 or 20% could overcome this problem. The partially acidulated phosphate rock would provide a soluble source of phosphorus initially while still maintaining the residual value of the phosphate rock (Hammond et a f . , 1980). Howeler (CIAT, 1979) has shown very encouraging results with beans. Studies on a Carimagua Oxisol have shown that partial acidulation of low-reactivity Colombian phosphate rocks with H2S04, however, did improve yields when compared with the North Carolina and Florida phosphate rocks (Mokwunye and Chien, 1980). Recent IFDC estimates, however, indicate that cost per unit P of partially acidulated phosphate rocks equals that of superphosphate. 6 . Thermal Alterations

Another group of potentially cheaper souces of phosphorus for acid highfixing soils are basic slag and fused magnesium phosphates, both water-insoluble products of thermal alteration. These types of fertilizers have been used primarily in Europe, but their potential in tropical areas with phosphorus fixation problems is receiving increased attention, particularly since steel industries develop where there are cheaper sources of energy. Basic slag (called “Escorias Thomas” in Latin America) is a by-product of steel manufacture from iron ore high in phosphorus. It has an average content of 4-8% phosphorus and 32% calcium mostly as calcium silicophosphates and calcium silicates. It has been found to be equally or more effective than superphosphates at the same rates of phosphorus application in Oxisols of Brazil and Colombia (Spain, 1979; Sanchez and Uehara, 1980). The Rhenania phosphates are produced by fusing rock phosphates of low citrate solubility with silica and soda ash. When serpentine or magnesium silicates are fused to give calcium or magnesium silicophosphates, the products are called fused magnesium phosphates or thermophosphates in Brazil. These products vary in composition in the ranges 10-12% P, 20-30% Ca, and 0-8% Mg. They have been found to be as effective or more effective than superphosphates in high-fixing Oxisols and Ultisols, particularly if the soils are not limed (NCSU, 1976; CPAC, 1980; Sanchez and Uehara, 1980). Ongoing experiments in Oxisols of Brasilia indicate that an application of 152 kg P/ha as “Termofosfato” decreased aluminum saturation from 70 to 38%,

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

37 1

while no such change was observed with an equal rate of triple superphosphate that produced similar pasture yields. (NCSU, 1976, 1978; CPAC, 1979, 1980). The main disadvantage of the Rhenania phosphates is their high cost of production. In Brazil, for example, the price per kilogram P is almost equal to that of triple superphosphate. Although the liming effect and silicon content may make their use more profitable, the high cost of Rhenania phosphates is a major limiting factor. In areas with ample supply of hydroelectric energy for thermal alteration, the situation may be different. Production of thermally altered phosphates is sometimes suited to small fertilizer plants employing intermediate technology. While fertilizer plants with production capacity as low as 50,000 tons/yr may not be feasible in industrialized nations, developing countries may find it profitable and appropriate to use intermediate technology that depends on utilization of local resources and skills (Sanchez and Uehara, 1980). Unlike superphosphates, thermally altered phosphates do not require sulfur or sulfuric acid plants. Also, phosphate rocks with high silica content can be used for thermal alteration. D. DECREASE OF PHOSPHORUS FIXATION WITH LIMING

The third component of this low-input strategy is to decrease the phosphorus fixation capacity of these acid soils by applying amendments such as lime and silicates. Considerable controversy exists as to whether liming decreases phosphorus fixation or not (Amarasiri and Olsen, 1973; Pearson, 1975). Part of this problem is attributed to reactions of the added phosphorus with freshly precipitated iron and aluminum hydroxides. Therefore the effects of lime on phosphorus availability may depend on the extent to which phosphorus is fixed by the adsorbing surfaces or by reactions with exchangeable aluminum (Smyth and Sanchez, 1980a). Several studies with acid soils in tropical America show that when exchangeable aluminum was neutralized by liming, phosphorus fixation decreased (Mendez and Kamprath, 1978; Leal and Velloso, 1973a,b; Vasconcel10s er al., 1975). Table XXXII shows the results of Smyth and Sanchez (1980a) for Oxisols from Brazil on which lime, silicate, and mixtures of lime and silicate were applied at agronomically relevant rates in an attempt to decrease phosphorus fixation. All amendment treatments decreased phosphorus fixation by about 20-30% in treatments that did not receive phosphorus. These results imply that determination of the amounts of phosphorus required to obtain a given solution concentration should be performed after lime or silicate applications and after sufficient time has been allowed for their reaction; otherwise, the phosphorus requirements may be overestimated (Smyth and Sanchez, 1980a). In the case of

372

PEDRO A. SANCHEZ AND JOSe G. SALINAS

Table XXXII Effects of Soil Amendment and P Applications on the Amount of Sorbed P Needed to Provide 0.1 ppm P in Solution in a Brazilian Oxisol” Percent decrease in P sorption Levelb

Amendment

0

Control CaCO, CaSiO, Combined CaCO, CaSiO, Combined

1

2

Applied P (pprn) : 0

380

0 18 24 18 16 28 32

460

540

44

54

65

59 65 65 62 75 14

68 77 71 77 a2 77

77 84 82 85 91 a5

“Source: Smyth and Sanchez (1980a) *Amendment level is relative to neutralization of exchangeable Al by the factor of 1 and 2, respectively. Initial exchangeable Al 1.45 meq/100 g.

using soil tests as a key to fertilizer recommendations, improvements could be made if samples are taken after lime has reacted. Liming has little or no effect in decreasing phosphorus fixation in soils with pH values of 5-6. Although still acid, they have aluminum saturation levels lower than 45% (Sanchez and Uehara, 1980; Leal and Velloso, 1973b). Furthermore, liming to pH values near or above neutrality may increase, rather than decrease phosphorus fixation because of the formation of relatively insoluble calcium phosphates (Sanchez and Uehara, 1980). Consequently, the effect of lime on phosphorus fixation depends on pH levels. E. SELECTION OF VARIETIES TOLERANT TO Low LEVELS OF AVAILABLE SOILPHOSPHORUS

A fifth component of the low-input phosphorus management strategy is to select plant species or varieties that grow well and produce about 80% of the maximum yields at low levels of available soil phosphorus. Although screening and selection of germplasm for “phosphorus efficiency” or “tolerance to low phosphorus” is less advanced than that for aluminum toxicity, research in that direction is also being conducted in tropical America. 1 . Annual Crops

Salinas (1978) screened a number of commercial varieties of upland rice, corn, and beans for tolerance to low phosphorus availability in the Cerrado of Brazil.

373

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

Figure 30 shows the results with rice expressed as yields relative to a high rate of broadcast superphosphate applications (1363 kg P,O,/ha). This rate provided a high level of available soil phosphorus (26 ppm P by the Mehlich 2 extraction). Most of the rice varieties produced maximum yields at the high soil phosphorus rate, but at different aluminum saturation levels. When aluminum saturation decreased to 63%, by adding 0.5 tons lime/ha, which supplied mainly calcium and magnesium as nutrients, the first three rice varieties (Batatais, Flotante, and IAC-1246) did not approach 80% maximum yield as did IAC-47 and Pratb Precoce. The latter variety had the lowest external phosphorus requirement (10 ppm P) under aluminum stress. These results clearly show differential varietal response to available soil phosphorus under aluminum stress. When aluminum saturation was reduced to 38% by adding 1.5 ton lime/ha, the rice varieties Flotante and IAC-1246 produced 80% of the maximum yield, but with a significant difference in the external phosphorus requirement. The Flotante variety required almost four times more available phosphorus than did IAC-1246. On the other hand, IAC-47 and Prat3o BATATAIS 100%=1303kg/ha 100

FLOTANTE = 758 kg/ha

F

IAC- 1246 =2771kg/ha

r

IAC- 47 = 2984 kg/ha

80 ----------

-a?

-0

17, PRAT~O PRECOCE =3116kg/ha

10

i" '0

10 20 30 40 0 10 20 30 40 0 1020 3040 0 10203040 0 10203040 5 12 8 33 18

AVAILABLE

P - N C EXTRACTION

(ppm)

FIG. 30. Relative yields of rice varieties (percentage of maximum yield of each variety) as a function of soil available P under three levels of Al stress in Brazilian Oxisol. (Source: Salinas, 1978.)

314

PEDRO A. SANCHEZ AND JOSE G. SALWAS

Precoce decreased their external requirements, which indicates a better utilization of phosphorus at low rates when aluminum toxicity is reduced. The economic implications of these results suggest a trade-off between lime and phosphorus. Using 1.5 tons lime/ha could decrease phosphorus requirements. Lime is likely to remain cheaper than phosphorus fertilizers. Under no aluminum stress, all rice varieties approached 80% of maximum yield, but at different available phosphorus levels. The Flotante rice variety always required more available phosphorus to produce well, whereas Pratao Precoce was able to produce over 80% of the maximum yield at one-sixth the phosphorus rate. Figure 31 shows a similar trend with corn varieties, but in all cases with a higher external phosphorus requirement than was the case with rice varieties. These results also confirm the general observation that the recommended rates of

AGROCERES 259 % Al Sat.

loOo/o= 6264 kg/ha

I 00

53

WHITE CARIMAGUA =6542 kg/ha

I[ I - , YELLOW CARIMAGUA

= 6234 kg /ha

AGROCERES 152

CARGILL-I I I

=6403kg/ha

=6842 kg/ha

r

I

-8

20

0

32

33

100 r

5

40

20

I/1 I

'0

1020 3040 0 10 2 0 3 0 4 0 0 1 0 2 0 3 0 4 0 0 10 20 3040 0 10 20 12 27 15

AVAILABLE P - N C EXTRACTION (pprn) FIG. 31. Relative yields of corn varieties (percentage of maximum yield of each variety) as a function of soil available P under three levels of Al stress in a Brazilian savanna Oxisol. (Source: Salinas, 1978.)

375

LOW-INPUT TECHNOLOGY FOR OXlSOLS AND ULTISOLS

%A; Sat,

CARIOCA

RlCO 23

100%=1784kg/ha

=1569kg/ha

lOOr

r

RlCO PARD0

DIACOL

JALO

= 1171 kg/ha

=967kg/ha

=1349kg/ha

r

r

r

AVAILABLE P - N C EXTRACTION ( ppm) FIG. 32. Relative maximum bean yield as affected by available soil P and A1 saturation levels of a Typic Haplustox from Brasilia. Brazil. (Source: Miranda and Lobato, 1978.)

phosphorus for upland rice are much lower than those for corn in Latin America (Kamprath, 1973). Under aluminum stress (63% aluminum saturation), the corn varieties Yellow Carimagua and Agroceres-152 approach 80% of maximum yield. When aluminum saturation was decreased to 38% by adding 1.5 tons lime/ha, the five corn varieties showed lower external phosphorus requirements to approach 80% of maximum yield. This observation underscores the important role that the lime plays in efficiency of phosphorus fertilization. Also, it appears that liming this Oxisol with 1.5 tons/ha enabled the corn plants to utilize both native and fertilizer phosphorus more efficiently (Salinas, 1978). Figures 32 and 33 also show the differential responses of bean and wheat varieties. With the exception of the variety Rico Pardo, the bean varieties had lower external phosphorus requirements because aluminum was neutralized by liming. In addition, varieties differed in their phosphorus requirements under the same level of aluminum stress. In the case of wheat varieties (Fig. 33), the Mexican varieties Sonora and Jupateco, which were developed in calcareous soils, produced significant yields only under no aluminum stress and had higher phosphorus requirements than those of the Brazilian wheat varieties BH- 1 146 and IAC-5. Although IAC-5 had a high external phosphorus requirement, it was

376

PEDRO A. SANCHEZ AND JOSh G.SALINAS BH-lI46

IAC-5

100%=2257kg/ha % d B L IOOr

SONORA

JUPATECO = 1727kgIha

= 1863 kg/ha

L L

-0

i/,

,

‘Ot!, ‘0

7

, ,

,

20 40

0

9

20

40

0

20 40

14

0

9

20 40

AVAILABLE P - N C EXTRACTION ( ppm) FIG. 33. Relative maximum wheat yields as affected by available soil P and Al saturation levels of a Typic Haplustox from Brasilia, Brazil. (Source: Miranda and Lobato, 1978.)

the only wheat variety that produced 80% of its maximum yield under aluminum stress. As aluminum stress decreased, the external phosphorus requirements of all varieties also decreased. 2 . Pastures Similar results are being obtained with tropical grasses and legumes (CIAT, 1977, 1978, 1979, 1980). Tables XXXIII and XXXIV show external and internal

phosphorus requirements for several tropical grasses and legumes. The data indicate substantial differences among ecotypes in internal and external phosphorus requirements. Excellent pasture establishment with low phosphorus fertilizer inputs and the use of grasses and legumes adapted to the acid, infertile soil conditions is taking place in different ecosystems of tropical America (CIAT, 1980).

F. POTENTIAL UTILIZATION OF MOREEFFECTIVE MYCORRHIZAL ASSOCIATIONS

It is well established that several genera and species of vesicular-arbuscular mycorrhizae form symbiotic associations with roots of certain plants and as a

LOW-INPUTTECHNOLOGY FOR OXISOLS AND ULTISOLS

377

result increase the uptake of phosphorus from soils low in this element (Sanders et al., 1975). Many of the plant species considered in this review to be tolerant to acid soil constraints are heavily mycorrhizal in Oxisols and Ultisols: cowpea, cassava, citrus, guava, Brachiaria decumbens, Centrosema pubescens, Pueraria phaseoloides, Stylosanthes guianensis, soybeans, and others (CPAC, 1979, 1980; Waidyanatha et al., 1979; Yost and Fox, 1979). It seems reasonable to speculate that the ability to enter into mycorrhizal associations may be an important characteristic of plant species and varieties adaptable to low-input systems. The advantage of mycorrhizal association lies in the use of fungal hyphae as an extension of the plant root system, which results in a larger surface area for nutrient uptake and the tapping of nutrients that move primarily by diffusion (phosphorus, zinc, and others) from a larger soil volume. There is no evidence that mycorrhizal associations are capable of utilizing forms of soil phosphorus that would be otherwise unavailable (Mosse et al., 1973). Nevertheless, the increase in phosphorus uptake can result not only in increased growth and phos-

Table XXXIII External Critical Phosphorus Levels of Various Tropical Pasture Species"

Species and accession number

Critical level of Bray I1 available P"

Legumes Stylosanthes capitara CIAT 1978 Stylosanthes guianensis ClAT I200 Zornia latifolia ClAT 728 Desmodium ovalifolium CIAT 350 Stylosanrhes cupitata ClAT I3 I5 Stylosanthes capitata ClAT I097 Zornia sp. CIAT 883 Pueraria phaseoloides CIAT 9900 Stylosanrhes capitata CIAT 1019 Stylosanthes capitata CIAT 1338 Stylosanthes guianensis CIAT 1153 Desmodium scorpiurus CIAT 3022 Macroptilium sp. ClAT 536 Desmodium gyroides ClAT 3001

2.5 2.5 2.8 3.0 3.2 3.3 3.4 3.5 3.5 3.6 5.5 8.0 9.5 11.4

Grasses Andropogon gayunus CIAT 62 I Brachiaria decumbens CIAT 606 Panicum maximum ClAT 604

5.0 7.0 10.0

"Sources: CIAT (1978, 1979, 1980). "Soil test level associated with about 80% of maximum yield.

PEDRO A. SANCHEZ AND JOSk G.SALINAS

378

Table XXXIV internal Critical Levelsof Phosphorus Associated with Near-Maximum Yields of Tropical Pastures Species Species

P in tissue (%)

Srylosanthes humilis Centrosema puhesrens Desmodium intortum Glycine wightii Medicago saliva Grasses Andropogon gayanus Brachiaria decumhens Melinis minutijlora Panicum maximum Pennisetum rlandestinuni Chloris gayana Paspalum dilatatum

0.17 0.16 0.22 0.23 0.25

Legumes

0.11" 0.12 0.18" 0.19" 0.22 0.23 0.25

"The source of these values is CIAT (1978). All other values are from Andrew and Robins (1969, 1971).

phorus concentration but increased nodulation and nitrogen fixation in legumes. Table XXXV shows the results of inoculation with and without high-reactivity rock phosphate additions on Pueraria phaseoloides growth in an "acid lateritic soil" or Sri Lanka with pH of 4.5 and 4 ppm Bray I1 available phosphorus. Mycorrhizal infections indeed produced all these favorable effects and, in addition, increased the efficiency of an application of 12 ppm P of Jordanian phosphate rock comparable to that of 60 ppm P without mycorrhizae. In an Oxisol from Hawaii, Yost and Fox (1979) compared the field response of various crops to phosphorus by fumigating part of the plots and leaving the rest in its natural state. Since fumigation killed most of the mycorrhizal population, their importance was evaluated in terms of phosphorus response. They found that mycorrhizae did make a difference in terms of phosphorus uptake, not only at low available phosphorus levels, but up to levels on the order of 0.1 ppm P in solution for soybeans, 0.2 for cowpeas, and 1.6 or greater for Stylosanthes hamata, Leucaena leucocephala, and cassava. At low available phosphorus levels (0.003 ppm P in solution or 3 ppm P Bray I) phosphorus uptake was on the average 25 times greater in mycorrhizal than in nonmycorrhizal plants. Estimates of internal or external critical phosphorus levels in the absence of mycorrhizal associations, such as those based on sand culture, nutrient solution, or fumigated soil, may be grossly exaggerated. Yost and Fox (1979) estimate that the phosphorus requirement of cassava can be exaggerated by a factor of 100 times if estimated in the absence of mycorrhizae.

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

379

The problem with such data is that they only document what is happening in Oxisols and Ultisols under natural conditions, where native mycorrhizae strains are already operating. This information, although highly enlightening, does not produce a new management practice. What is needed is to determine whether inoculation with more effective strains of mycorrhizae can enhance phosphorus uptake. Two questions need to be answered in order to reach a determination: (1) How can mycorrhizae be inoculated on a practical basis? (2) Are there more effective strains that can compete with the native ones and persist in the soil? Unlike rhizobia inoculation, mycorrhizae must be inoculated as fresh hyphae and cannot be mixed with peat and dried. At the experimental level, field inoculation can be carried out by adding soil from mycorrhizal area, but the tonnage required impedes its practical application. Some advances are being made toward answering the second question. Researchers at the Cerrado Center near Brasilia (CPAC, 1980) were able to produce good infection of the mycorrhiza species Anaulospora laevis in the acid-tolerant soybean cultivar UFV-1 in an Oxisol. More work in this direction is needed before mycorrhizae can be a component of low-input soil management technology. G . CONCLUSIONS

Phosphorus is frequently the most expensive purchased input in Oxisols and Ultisols of tropical America. Except for lands recently cleared of rain forests, phosphorus fertilization is almost always essential for sustained crop or pasture production systems. The high phosphorus fixation capacity of loamy and clayey Oxisols and Ultisols has raised fears of huge quantities of phosphorus needed for Table XXXV Effects of Vesicular Arbuscular Mycorrhiza Inoculum in Sterilized "Acid Lateritic Soil" of Sri Lanka on Growth, Phosphorus Uptake, and Nitrogen Fixation by Pueraria phaseoloides under Pot Conditions"

Dry matter Treatment

production (dpot)

Unsterilized check Mycorrhiza only Myconhiza 12 ppm P as PRh Mycorrhiza + 60 ppm P as PRb 12 pprn P as PRb 60 ppm P as PRb

2.4 28.8 31 .O 37.8 3.9 24.6

+

"Adapted from Waidyanatha e t a / . (1979). *Jordan phosphate rock.

Mycorrhizal infection

Plant P

(%)

(YO)

0 76 67 74 I1 0

0.18 0.27 0.28 0.31 0.25 0.25

No. of nodules per pot I 230 241 354 11

96

GH, reduction (pmoYpot/hr) 0. I 55.0 69. I 123.4 1.6 24.8

380

PEDRO A . SANCHEZ AND JOSE G. SALINAS

these vast areas. Five of the major components of low-input soil management technology, either individually or preferably together, can markedly reduce phosphorus requirements and thus increase the efficiency of utilization of this basic resource.

VII. MANAGEMENT OF LOW NATIVE SOIL FERTILITY In addition to aluminum and manganese toxicities, calcium, magnesium, and phosphorus deficiencies, and high phosphorus fixation, many Oxisols and Ultisols of tropical America are also deficient in other essential nutrients, particularly nitrogen, potassium, sulfur, zinc, copper, boron, and molybdenum (Sanchez, 1976; Spain, 1976; Lopes, 1980). This low-fertility syndrome has sometimes caused the least fertile Oxisols to be considered as “fertility deserts” (Spain, 1975). In somewhat less infertile Ultisols of the Peruvian Amazon, Villachica (1978) and Sanchez (1979) recorded deficiencies of all essential plant nutrients except for iron, manganese, and chlorine in continuous crop production systems, now in its 20th consecutive crop. Table I1 shows that 93% of the Oxisol-Ultisol regions suffer from nitrogen deficiency, 77% have low potassium reserves indicative of potassium deficiency, 71% have sulfur deficiency, 62% have zinc deficiency, and 30% have copper deficiency. The areal extent of other micronutrient deficiencies cannot be ascertained with the data available. Although these figures give an indication of the extent of the individual constraint, they are also fairly rough estimates (Sanchez and Cochrane, 1980). The main low-input technologies required to manage low native soil fertility center on (1) maximum use of nitrogen fixation by legumes in acid soils, (2) increasing the efficiency of nitrogen and potassium fertilization, (3) identification and correction of sulfur and micronutrient deficiencies, and (4) promotion of nutrient recycling. A . M A X I M U USE M OF BIOLOGICAL NITROGEN FIXATION

The best known low-input soil management technology is the use of legumerhizobium symbiosis to meet the plant’s nitrogen demand without having to purchase nitrogen fertilizers. Biological nitrogen fixation is limited to legumerhizobium symbiosis in these soils in terms of practical management. Associative symbiosis between nitrogen-fixing bacteria such as Spirillum lipoferum in the rhizosphere of tropical grasses has created widespread expectation about the possibility of nitrogen-fixing grasses, many of which are acid-tolerant (National

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

38 1

Academy of Sciences, 1977a; Neyra and Dobereiner, 1977). Unfortunately, evidence to date indicates that the practical exploitation of such symbiosis in Oxisols and Ultisols appears to be minimal at this time (Hubbell, 1979). This is an example of a low-input component that has not worked to date. Additional basic research, however, may reveal some practical implications in the future and such research should continue. We are fortunate that many of the plant species of economic importance that are adapted to acid soil conditions are legumes. Among the annual food crops, there are three important acid-tolerant legumes, namely cowpeas, peanuts, and pigeon peas, and several less widespread ones, such as lima beans, mung beans, and winged beans. There is also a wealth of very acid-tolerant forage legumes of the genera Stylosanthes, Desmodium, Zornia, Pueraria, Centrosema, and many others. Spontaneous legumes also abound in areas cleared of rain forests. Hecht (1979) recorded 69 tree, shrub, and creeping legume species in pastures of the Eastern Amazon of Brazil. In order for these legumes to fix sufficient nitrogen, it is essential that the nutritional requirements and degree of acid soil tolerance of the associated rhizobium match those of the plant (Munns, 1978). If not, plant growth will be severely hampered because of nitrogen deficiency. Rhizobium strains differ in their tolerance to the various acid soil stresses just as plants do (Munns, 1978; Date and Halliday, 1979; Munns et a l . , 1979; Halliday, 1979; Keyser et al., 1979). Consequently, soil management practices require the matching of nutritional requirements and tolerances of both legume and rhizobia. Until recently, it has been assumed that most tropical pasture legumes growing on acid soils develop effective symbiosis with native “cowpea-type’’ strains of rhizobium, and therefore the selection of specific strains for individual legume species or cultivars is the exception rather than the rule (Noms, 1972). Recent work by Halliday (1 979) and collaborators clearly shows that this is no longer the case. A five-stage screening and matching procedure involving laboratory, greenhouse, and field stages has shown a high degree of strain specificity for obtaining effective symbiosis in the most promising forage legume ecotypes. Recent recommendations, including inoculation technology, are available (CIAT, 1980). Long-term field experiments, however, show that the response to inoculation with selected rhizobium strains generally decreases with time. Protecting the inoculant strain with lime or rock phosphate pelleting often permits an effective infection in an acid soil. The critical point, however, is reached 2-3 months afterward when the primary nodule population decomposes. Then the rhizobia must fend for themselves in an acid soil environment in order to reinfect the plant roots (CIAT, 1979). The selection of effective acid-tolerant strains is therefore highly desirable. Date and Halliday (1979) developed a simple laboratory technique to screen for acid tolerance at the early stages of strain selection, using an

382

PEDRO A.

SANCHEZ AND JOSE G . SALINAS

agar medium buffered at pH 4.2. Rhizobium strains tolerant to acidity grow in such media, whereas those susceptible die. With this approach, specific strains have been identified and recommended for low-input pasture production systems on acid soils for several accessions of Stylosanthes capitata, Desmodium ovalifolium, Desmodium heterophyllum, Zornia spp., Pueraria phaseoloides, Aeschynomene brasiliana, and A . histrix (CIAT, 1980). Differences in acid tolerance among rhizobium strains have also been identified for cowpeas (Keyser et al., 1979) and mung beans (Munns et al., 1979). In both species, the host plant tends to be more tolerant to acidity than many of the rhizobial strains. The opposite is apparently the case with soybeans, for which the current commercial strains of rhizobia appear to be more tolerant than the hosts (Munns, 1980). In terms of nutritional needs, rhizobia require greater amounts of cobalt and molybdenum for symbiotic nitrogen fixation than do the host legume for growth (Robson, 1978). The relative requirements of other nutrients and the interactions between legume nutrition and rhizobium nutrition merit additional research. Nevertheless, it seems clear that the nutritional requirements and acid soil tolerance of legume species should not be determined in the absence of nodulation. This is almost invariably the case with culture solution studies. Screening for acid soil tolerance of legumes should be done with soil and with inoculation. In addition to joint work by soil fertility specialists and plant breeders, the microbiologists must also be involved.

B. INCREASE

OF THE

EFFICIENCY OF NITROGEN AND FERTILIZATION

POTASSIUM

1 . Nitrogen

It appears that no fertilizer nitrogen is likely to be needed for acid-tolerant, legume-based pastures for the acid, infertile soil regions of tropical America. Fertilizer nitrogen applications, however, are essential for cereal or root crop production systems in these regions. Rotating or intercropping grain legumes with cereals may decrease the overall amounts of nitrogen needed, not because of a significant transfer of fixed nitrogen to the cereals, but because the legumes occupy space in the fields. Most of the nitrogen fixed by grain legumes is removed from the field during harvest (Henzell and Vallis, 1977). Consequently, increasing the efficiency of fertilizer nitrogen utilization appears to be the main avenue for decreasing nitrogen fertilizer inputs for nonlegume crops. Exceptions of the above statements are few. Nitrogen responses in these soils

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

383

are almost universal except during the first crop after clearing rain forests or on Oxisols and Ultisols that have been intensively fertilized with nitrogen for many years. Fox et al. (1974) observed no nitrogen responses by corn for six consecutive and relatively high-yielding crops in Ultisols of Puerto Rico, because of a long-term history of intensive fertilization. Extensive nitrogen fertilization research has been conducted with corn, upland rice, sorghum, cassava, and sweet potatoes in Ultisols and Oxisols of tropical America. A review by Grove (1979) shows that these soils typically supply 60-80 kg N/ha to most of these crops and that applications on the order of 80-120 kg N/ha produced about 95% of the maximum yield, which in the case of corn was on the order of 5 tons/ha. When the most efficient rates, sources, and placement methods (urea incorporated right before the period of most rapid plant uptake) were used, apparent nitrogen recovery was about 56% (Grove, 1979). With upland rice recovery is on the order of 30% (Sanchez, 1972). Sulfur-coated urea has failed to produce significant advantages over regular urea or ammonium sulfate on cereal or root crops in Oxisols and Ultisols of tropical America. Higher nitrogen rates than those reported by Grove (1979) are often necessary in high rainfall environments due to leaching. Splitting nitrogen applications in two usually increases nitrogen recovery. The problem with the above summary is that most of the data were collected in experiments in which other fertility constraints had been eliminated. It is not known whether fertilizer nitrogen efficiency would be different when acidtolerant cereal or root crops are grown under low phosphorus and lime inputs. Although corn varieties are known to differ in their ability to utilize fertilizer nitrogen efficiently (Gerloff, 1978), this has not been tested under low-input technology situations. Well-known plant characteristics that increase yield responses to nitrogen, such as short stature and high tillering in upland rice in high-fertility soil, should have a similar effect in acid, infertile soils. Soil testing is of little value for nitrogen fertilization because of the mobility of nitrate in well-drained Oxisols and Ultisols and other factors (Sanchez, 1976). Consequently, fertilizer recommendations are based on field experience and plant uptake data. Nitrogen fertilization for cereal and root crops is therefore one of the weakest components in low-input strategy for these soils. 2. Potassium

The situation for potassium is similar to that for nitrogen. As mentioned before, most of the Oxisols and Ultisols have low potassium reserves in their clay minerals and potassium deficiencies increase with time (Ritchey, 1979). Unlike nitrogen, the identification of potassium deficiency via soil test is straightforward. The established critical levels are in the range 0.15-0.20 meq K/100 g for most crops. Unfortunately, there are no obvious shortcuts for low-input potas-

PEDRO A. SANCHEZ AND JOSB G. SALINAS

384

sium management. There are no major inter- or intraspecific differences in terms of “tolerance to low available soil potassium. ” Potassium fertilizer requirements can reach levels of 100-150 kg K,O/ha/crop. Although not as costly per unit as nitrogen or phosphorus fertilizers, such outlays represent a significant cost to the farmers. The main avenues for increasing the efficiency of potassium fertilization are split applications and avoidance of removal of crop residues, particularly stover, in order to attain some degree of recycling. The efficiency of potassium utilization is becoming an increasingly important concern in Oxisol-Ultisol regions of tropical America, as progress in overcoming acidity, phosphorus, and nitrogen constraints increases yield potential and therefore potassium requirements. A major research thrust on potassium efficiency is badly needed. C.

IDENTIFICATION A N D CORRECTION OF DEFICIENCIES OF SULFUR A N D MICRONUTRIENTS

Oxisols and Ultisols are ofrzn deficient in sulfur and several micronutrients, particularly zinc, copper, boron, and molybdenum (Kamprath, 1973; Cox, 1973; Blair, 1979; Lopes, 1980). Unfortunately, very little is known about the geographical occurrence of these deficiencies, their critical levels in the soil, and the requirements of acid-tolerant species and varieties. Hutton (1979) attributed most of the lack of legume persistence in mixed pastures of Latin America to uncorrected nutrient deficiencies. Many ranchers in tropical America feel that applying triple superphosphate is sufficient fertilization for grass-legume pastures. This fertilizer source provides only phosphorus and some calcium. In tropical Australia, molybdenized simple superphosphate is widely used as the only fertilizer in Alfisols that are very deficient in nitrogen, phosphorus, sulfur, and molybdenum. This source corrects phosphorus, sulfur, and molybdenum deficiencies, allowing the legume to provide nitrogen to the mixture. Given the fundamental differences in soil acidity between soils of tropical Australia where improved pastures are grown (mainly Alfisols) and the Oxisol-Ultisol region of tropical America, it is not possible to extrapolate the Australian fertilization practices (Sanchez and Isbell, 1979). The situation is not much better for crop production because most of the fertilizers available are straight NPK formulations. With the use of higher-analysis sources such as urea, triple superphosphate, and KCl, the sulfur content of such mixtures has decreased and sulfur deficiency has become more widespread. Surveys of the nutritional status of Oxisol-Ultisol regions, such as the one Lopes and Cox (1977a) did in the Cerrado of Brazil, plus on-site field experiments on the nutrients, such as those conducted in Carimagua, Colombia (CIAT, 1977, 1978, 1979, 1980; Spain, 1979) and in Yurimaguas, Peru (Villachica,

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

385

19781, contribute significantly to identifying which nutrients are deficient and which practices are best to correct them. They also aid in identifying possible nutrient imbalances that may be induced by fertilization. Therefore site-specific identification is necessary. These efforts must be related to the nutritional requirements of the main species and varieties. Relatively little is known about the acid-tolerant species mentioned in this article. Table XXXVI shows tentative external and internal critical sulfur levels for important grasses and legume species under Oxisol conditions. When one of these constraints is identified, the results can be extremely positive. Wang et al. (1976) identified sulfur deficiency in rice-growing areas in the lower Amazon of Brazil. By switching from urea to ammonia sulfate applications and thereby applying sulfur, rice production improved dramatically. Similar experiences with micronutrient identification and correction have k e n recorded elsewhere (Cox, 1973; Lopes, 1980). Insufficient knowledge of nutrient deficiencies is probably the weakest component of low-input technology. This gap can be corrected by systematic determination of critical nutrient levels in the soil and in the plants. Fortunately, the application costs are low, and zinc and copper fertilization produce long residual effects.

Table XXXVI Tentative External and Internal Critical Sulfur Levels of Acid-Tolerant Forage Grasses and Legumes Grown in a Carimapa Oxisol in the

Species

Critical soil test levelc (ppm S)

Critical tissue concentration (% S )

Grasses Brachiaria humidicola 679 Andropogon gayanus 621 Brachiaria decumbens 606 Panicum maximum 604

I1 12 13 14

0.14 0.15 0.16 0.15

Legumes Stylosanthes capitara 1315 Desmodium ovalifolium 350 Zornia latifolia 728 Stylosanthes capitata 1019

12 13 14 15

0.15 0.12 0.14 0.17

aSource: CIAT (1981). bEstimated from Cate-Nelson diagrams. “Calcium phosphate extraction.

386

PEDRO A. SANCHEZ AND JOSB G.SALINAS

D.

PROMOTION OF NUTRIENT RECYCLING

Soil management practices in low-fertility soils should encourage nutrient recycling as much as possible. Nutrient recycling is the main reason why acid, infertile Oxisols and Ultisols are able to support exuberant tropical rain forest vegetation in udic environments. The magnitude of this natural recycling is of interest. Two detailed studies conducted on an Oxisol from Manaus, Brazil (Fittkau and Klinge, 1973) and an Oxisol from Carare-Opon, Colombia (Salas, 1978) show that the annual nutrient additions via litter layer ranged as follows (in kg/ha): 106-141 N, 4-8 P,O,, 15-20 K 2 0 , 18-90 Ca, and 13-20 Mg. Nutrient additions through rainwash, wood decomposition, and root decomposition may double the above estimates. In crop production systems, a significant portion of nutrients are removed from the soil at harvest. Simple "maintenance" fertilizer applications aimed at replacing what harvests took away are seldom sufficient for sustained crop yields (NCSU, 1974, 1975). Nutrient recycling therefore offers limited possibilities in crop production systems. One possible application may be leaving crop residues as mulches, particularly in the case of corn stover and rice straw, in order to recycle potassium back into the soil. There is little data on the effect of these or other mulching practices on nutrient recycling. In pasture production systems, there is a natural recycling mechanism by which about 80% of the nitrogen, phosphorus, and potassium consumed by cattle are returned to the soil via excreta (Mott, 1974). This percentage is a very rough estimate and depends considerably on stocking rate, grazing management, and other factors. The limited data available in Oxisol-Ultisol regions show that this is an important mechanism. Figure 34 shows the changes in the top 20 cm of an Orthoxic Palehumult from Quilichao, Colombia, caused by dung deposition in a Brachiaria decumbens pasture under rotational grazing every 15 days. This figure shows that the topsoil inorganic nitrogen content doubled within 15 days within a 1-m radius from the excreta and declined afterward. Available phosphorus, potassium, calcium, and sulfur also showed a similar increase, followed by a more gradual decrease with time than nitrogen. The effects of urine (not shown) indicate a sharper increase in potassium and sulfur than with feces, but a smaller increase in the availability of nitrogen, phosphorus, and calcium (CIAT, 1981). The overall effects of these additions were favorably reflected in increases of all five elements in plant tissue concentration within the first 30 days after excreta deposition. Indirect evidence of nutrient recycling in poorly managed pastures is shown in Fig. 35 in Oxisols of the eastern Amazon of Brazil, where the forest was cut by the slash-and-bum method and Panicum maximum was planted. Serrfio et a l . (1979) sampled soils in unfertilized Panicum maximum pastures of known ages

387

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

101

a 1

I

I I

I

I

30

45

I .8

Exch. K (meq/IOOg)

I .6

0.1 O 0 .

IA---+. 2 p

;

A'

1.4

0.08

I .2

6or

Avail.

0

0

s

15

(ppm)

30

15

DAYS AFTER DEPOSITION

45

DAYS AFTER DEPOSITION

FIG. 34. Nutrient recycling on the top cm of an Orthoxic Palehumult from Quilchao, Colombia, as a result of dung deposition by cattle grazing a Brachiaria decumbens pasture. Distance from dung (cm): 0 , 20; A, 100. (Source: Salinas and Camps, unpublished results.)

in two sites. Soil pH increased from about 4.5 to between 6 and 7 right after burning, and remained constant up to 13 years. Aluminum toxicity was completely eliminated as calcium and magnesium levels were maintained at fairly high levels. Organic matter and nitrogen levels also remained high over the 13-year period. Potassium values remained close to the critical level, while available phosphorus decreased below the critical level (5 ppm P by Mehlich 2) within a few years. These results are from samples of different fields of known

388

PEDRO A.

o

SANCHEZ AND JOSE G. SALINAS

Loomy Oxi~ol,N.Malo Gross

-

0 0.2 ,pd .

'd

Q3

0

b

0.1 ' -

0

too

% At Saturation

80 60 40

20 0 % Total N

0

-

1 3 5 7 9 1 1 1 3

0

1 3 5 7 9 1 1 1 3

Age of Pasture Sampled (years)

FIG. 35. Changes in topsoil properties of Panicum maximum pastures of known age sampled at the same time in two regions of the Eastern Amazon of Brazil (Adapted from SerrHo er al., 1979.)

age after clearing taken at the same time; therefore they confound time and space variability. Nevertheless, it seems clear that many of the chemical properties of thses Oxisols were definitely improved by clearing and grazing. These soil dynamics are in sharp contrast with the rapid fertility decline observed after clearing rain forests and growing annual crops in udic areas of Peru (shown in Fig. 10). The reasons for these differences are not clearly understood and deserve more thorough study. Some factors favoring a less marked decline in eastern Amazonia may be an ustic soil moisture regime that allows for a more

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

389

thorough bum and more ash deposition, and possibly upward movement of cations and anions during the dry season. Also, the periodic burning every few years practiced in these areas and some degree of nutrient recycling by the grazing animal may contribute to the effects shown in Fig. 35. Whatever the reasons, the improvement in the chemical properties of acid, infertile Oxisols is remarkable and shows promise for better managed grass-legume pastures in the Amazon region. Farming systems that include trees are expected to produce better nutrient recycling. Trees of economic importance such as cocoa and oil palm are expected to have a nutrient recycling mechanism similar to that of the rain forest (Alvim, 1981). Actual data to support this hypothesis, however, are very limited. Silva (1978) observed evidence of incipient nutrient recycling of several permanent crops in an Oxic Paleudult of Barrolindia, Bahia, Brazil, in terms of an increase in the exchangeable base content of the top 5 cm of the soil 34 months after burning. The increase is most marked in the young oil palm plantation with a Pueraria phaseoloides ground cover, followed by the pasture, and to a lesser degree in the cassava-banana intercropping that precedes cocoa planting. Similar observations have been made with some planted forestry species with a kudzu understory in an Oxisol of Manaus, Brazil (P. T. Alvim, personal communication). More data covering a longer time span are needed in order to fully ascertain the importance of nutrient recycling in cropping systems of Oxisol-Ultisol regions in tropical America. E. CONCLUS~ONS

The low native fertility of Oxisol-Ultisols cannot be eliminated as a major constraint without significant fertilizer inputs. Several avenues are available for lowering the overall fertilizer requirements. The need for nitrogen fertilization, however, can be essentially eliminated in legume-based pasture systems with the use of acid-tolerant Rhizobium strains in association with acid-tolerant legume species. This is also possible for the acid-tolerant grain legumes, but definitely not for cereal and root crop species. The carryover effect of nitrogen fixed by a legume to a nonlegume crop either intercropped or in rotation appears to be very small since most of the nitrogen is removed in the harvest. Increasing the efficiency of nitrogen fertilization for nonlegumes can be accomplished through improved timing and placement of fertilizers. Little is known about fertilizer nitrogen efficiency of acid-tolerant cereal crops under low-input systems. Potassium and sulfur deficiencies are widespread and in the case of the latter, become more widespread with the use of higher-analysis fertilizers. The identification of deficiencies of these nutrients and the micronutrients is a major gap in tropical America. This can be overcome by effective soil fertility evaluation

390

PEDRO A. SANCHEZ AND JOSE G . SALINAS

services, including the establishment of critical levels and fertilizer recommendations. Nutrient recycling should be promoted, but in crop production systems the possibilities seem largely limited to crop residue utilization. The magnitude of nutrient recycling in pastures and tree systems needs substantial quantification.

VIII. DISCUSSION The previous sections have described the various components for low-input soil management technology that can be used in the acid, infertile soils of the American tropics. Obviously each component is not applicable to all situations or farming systems in the vast target area; some components are mutually exclusive. Also, several components are reasonably well developed and ready for local validation, whereas others are barely more than preliminary observations. As a whole, however, they represent a philosophy of soil management for marginal lands of the tropics. The same philosophy can also be applied to other aspects of agriculture, particularly plant protection. This section of the review examines some of the implications of the use of such technology. A. Low-

VERSUS

HIGH-INPUT APPROACHES

There is considerable ambiguity in the term “low-input technology. How low is low, and relative to what? The terms “zero input” and “minimum input” have also been used. The first one is not appropriate because in most systems zero input results in zero output. Low input as opposed to medium or high input deserves some quantification. In this review, we would like to consider low-input technologyfor acid soils of the tropics as that targeted at obtaining about 80% of the maximum yields of acid-tolerant germplasm with the most eficient use of soils, fertilizers, and lime. This review shows that it is biologically feasible to reach these yield levels with available technology and germplasm at a substantially lower level of input use than by using traditional technology and germplasm. What is wrong with the traditional high-input technology that has been the base of much of our present world food production? There is little wrong with it from an agronomic point of view. If we were farmers in an Oxisol region and the government gave us a choice between overcoming the main soil constraints by financing massive phosphorus applications, sufficient liming, and supplemental irrigation systems, or puttling into practice the components described in this review, we would immediately follow the first alternative. As farmers, we would ”

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

39 1

see the value of our land increasing as it is transformed from marginal to excellent land by the application of inputs. The senior author, in fact, saw his father do exactly that in a 50-ha Oxisol farm, where he grew 3 crops/yr with irrigation and profited handsomely from it. It is difficult to find better soil to manage than an Oxisol once its chemical constraints are eliminated. Such opportunities, however, are the exception rather than the rule in the acid, infertile soil regions of tropical America. The magnitude of investment capital needed to apply high-input technology to these soils is commonly beyond the resources of most governments and private organizations. Political priorities also dictate that farm intensification through high-input use be located where the large concentrations of farmers are, usually in high-base status soils. The increasing costs of petroleum-related inputs and the worldwide emphasis on conserving the earth’s natural resources pose additional restraints to the “maximum input” approach. The development policy goals of many tropical countries require that both producers and consumers with limited resources be the major beneficiaries of improved agricultural technology. Nickel (1 979) observed that if low-income consumers are to benefit, food production increases must be achieved at lower unit costs. These low unit costs can be achieved through biologically based technology that is often scale neutral. To assure that producers with limited resources have access to the benefits of this technology, it should not depend on large amounts of purchased inputs. Consequently, the main justij’ication of low-input soil management technology in Oxisol-Ultisol regions of tropical America is socioeconomic and not agronomic in nature. In the past, farmers adjusted to their lack of purchasing power by applying low amounts of inputs to a farming system designed to operate best at high-input levels. Examples of this abound in Latin America, where nutrient deficiency symptoms are obvious in many fields. Many fanners know that their crops could yield more if more fertilizer were applied to high-yielding varieties, but they either cannot afford to purchase more or do not dare to because of the high risk involved. Another example is the large-scale attempt of beef production in Oxisols and Ultisols of the Amazon of Brazil by planting Panicum maximum without phosphorus fertilization. This clearly is a case of ignoring very obvious soil constraints. As Paulo Alvim has repeatedly mentioned in meetings about the Amazon, “agriculture is different from mining.” Farmers must add fertilizers in order to have sustained production, even in the best soils of the temperate region. Low-input soil management technology for these acid soils is different from the partial adoption of high-input technology. Low-input technology is not less of the same but a different way of managing the soil. The fundamental breakthrough has been the identification of important plant species and varieties that can tolerate significant degrees of acid soil constraints. Then it is a matter of determining how much fertilizer and lime these tolerant species require to produce about 80% of their maximum yield on a sustained basis.

392

PEDRO A. SANCHEZ AND JOSE G . SALINAS

Finally, a better understanding of the favorable attributes of acid, infertile soils converts certain soil constraints into management assets. Four examples follow: 1. By keeping the soil in its acid state, low-reactivity phosphate rocks, abundant in tropical America, can be used directly at a fraction of the cost of superphosphates. In effect, the chemistry of soil acidity replaces the superphosphate factory at considerable energy savings, provided that aluminum-tolerant plants are grown. 2. Extreme acid soil infertility can prevent weed infestations while localized fertilizer applications promote vigorous growth of the desired crop or pasture. 3. Low effective cation-exchange capacity can be considered an asset in many of these soils. Clayey soils with low ECEC generally have better structure and are less erodible than soils with high-activity clays and similar clay content. 4. Low effective cation-exchange capacity permits the gradual increase in the base status of the subsoil through the downward movement of calcium and magnesium. Instead of deterioration, the fertility of these soils actually improves, permitting deeper root development, which, in turn, permits the utilization of hitherto unavailable soil moisture. This is an attractive alternative to the more expensive supplemental irrigation systems.

B. PRODUCTIVITY OF LOW-INPUT SYSTEMS Agronomically sound high-input soil management systems almost invariably produce higher yields than the low-input systems defined here. There are several reasons that account for this observation. When soil constraints are eliminated by fertilization, liming, and irrigation, it is possible to use plant species and varieties that have a higher absolute yield potential and the acid-tolerant varieties presently available. The reason for this difference is very simple. Plant breeders have traditionally concentrated on increasing the yield potential in the absence of soil constraints. Breeding to combine the various high-yielding attributes with acid soil tolerance is in its infancy. As yet, there are no aluminum-tolerant rice varieties with the yield potential of IR8. Andropogon gayanus does not have the production potential or the nutritional quality to match intensively fertilized Pennisetum purpureum, although it has high platibility . Stylosanthes guianensis cannot outproduce alfalfa in terms of quality under optimal conditions. This limitation is probably a matter of time because some tolerances to acid soil stresses are controlled by one or two genes, which are often dominant (Rhue, 1979). Consequently combining acid tolerance with high-yield potential appears feasible from the breeding point of view. Breeding for acid soil tolerance, however, is just beginning. Most of the screening work is based on selecting preexistent germplasm and not segregating populations produced by a breeding

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

393

program for acid tolerance. Joint work of breeders and soil scientists should be intensified. Its payoff could be as important as the successful efforts of plant breeders with pathologists and entomologists in breeding for disease or insect resistance. In fact, the payoff may be even greater because the acid-tolerant varieties may have a longer useful time span than insect- or disease-tolerant varieties. The aluminum ion does not mutate into a more virulent race as many fungi or bacteria strains do. C. SOILMININGOR SOILIMPROVEMENT?

Concerns have been expensed that plant species tolerant to acid soil constraints, particularly those tolerant to lower levels of available phosphorus, may completely deplete the low reserve of nutrients that these soils have and render them totally useless. Low-input technology is sometimes viewed as a last ditch effort to extract the last bit of fertility out of these soils. This argument must be viewed in terms of the total reserves of the soil, the amounts of fertilizers to be added, and total nutrient extraction. With continuous plant growth the supply of certain available nutrients in the soil eventually decreases below the critical level. In Oxisols and Ultisols, this happens rather quickly with nitrogen and potassium, elements that are very mobile in their available form. Nitrogen depletion is very unlikely because of the large reservoir in the organic fraction and its replenishment by root decomposition, nitrogen fixation, and other factors in a farming system. Organic matter contents of these soils are not generally different from the main soils of the temperate region (Sanchez, 1976). The situation with sulfur is similar. The rate of potassium depletion depends on the soil’s reserve in nonexchangeable form, mainly in clay minerals. The potassium reserves of these soils usually provide less than the generally accepted critical level of 0.15 meq/100 g. An equilibrium between available (exchangeable) potassium and nonexchangeable is then established. This level will not support rapid plant growth but will not decrease the soil’s potassium reserves to zero. Since crop residues or mature pastures are usually high in this element, some degree of recycling normally takes place. The “mining” potential for calcium, magnesium, zinc, iron, copper, boron, manganese, and molybdenum appears less likely because the amounts removed by plant harvests are very small in comparison to total soil reserves in Oxisols and Ultisols. Also, the available forms of these elements are less mobile in soils and thus less subject to loss. This leaves phosphorus, the element around which most of the “soil mining” arguments revolve. Total phosphorus contents in the topsoil of Oxisols and Ultisols are on the order of 100-200 ppm P, as compared with about 3000 ppm P in high-base status, high-activity clay soils of the midwestern United States and

394

PEDRO A. SANCHEZ AND JOSE G . SALINAS

Table XXXVII Soil Phosphorus Fractions in the Profile of an Oxisol of Carimagua, Llanos Orientales, Colombia"

Horizon (cm) 0-6 6-15 15-40 40-70 70-100 100-150

Organic Base saturation C pH

(%)

(%)

4.5 4.6 4.6 4.9

2.26

7 7 13

5.1 5.1

1.84

1.13 0.53 0.43 0.24

15

29 21

Percentage of total P Total P Organic Reductant- Occluded AI-P P Ca-P AI-P Fe-P Sol. Fe-P (ppm) 185 1.51

77

126 114 90

73

84

75 55

47 35

0.9 0.6 0.7 0.8 0.6 0.7

0.8 0.9 1.2 1.3

1.0 1.2

10 11

6 7 9 4

9 I1 17 34 41 53

1

I

I 1 1

4

"Source: Benavides (1963).

similar temperate regions (Sanchez, 1976). Some Oxisols, however, have very high total phosphorus contents, such as Eutrustox of the Cerrado of Brazil (Moura er al., 1972), but the limited data base shows that most Oxisols and Ultisols are generally low in total phosphorus. Table XXXVII shows the total phosphorus content of an Oxisol profile from Carimagua, Colombia, representing the least fertile range of the Oxisol-Ultisol regions of tropical America. The total phosphorus reserves of the top 150 cm average 106 ppm P, which is equivalent to 4830 kg P205/haof total phosphorus. Roots of acid-tolerant plants, however, may penetrate deeper than 150 cm. Table XXXVIII shows the total uptake of phosphorus of two acid-tolerant grasses under grazing at Carimagua. Total phosphorus uptake by the forage available to the animals was in the range 3-12 kg P/ha/yr (7.5-28 kg P205/ha). Assuming all the phosphorus is removed from the sward, and thus ignoring recycling, the amounts added as fertilizer (50 kg P205/hdyr)more than compensate for the removal. Therefore there is no soil mining but actually a slow buildup of phosphorus. Table XXVIII confirms that there is a gradual buildup of total phosphorus in these soils of about 16 ppm P/yr on the topsoil with application rates of 50- 100 kg P,O,/ha/yr. In the case of crop production, phosphorus removal rates are higher. Wade (1978) reports that four consecutive harvests of cowpeas, corn, peanuts, and rice, after which the residues were left in place, produced a total removal of up to 68 kg Pz05/hdyr in Yurimaguas. The total amount added was 50 kg P,O,/ha, suggesting a very close balance. An annual application rate of 100 kg PzOs/ha/yr would probably produce a gradual increase in available phosphorus. Long-term soil dynamics data at Yurimaguas show major buildups in available phosphorus,

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

395

calcium, zinc, and copper with continuous cropping in this region (Sanchez, unpublished results). It is well known that plants remove less phosphorus than is applied as fertilizers. Since low-input technologies described in this review do involve fertilization, the soil mining argument appears to have very limited validity.

D. RESEARCHNEEDS This review has shown the feasibility of the low-input approach and presented examples of low-input soil management technology components. Research institutions responsible for developing low-input farming systems for representative soils may want to integrate the components that are relevant to their situation into their farming systems. The authors of this review are not aware of improved low-input farming systems that have all the necessary components sufficiently well developed. Hence, the first research priority in most situations is to fully develop the components of low-input technology for a particular farming system. The items listed in Section I,C of this review could serve as a rudimentary check list, subject to local modification. This review has also identified several major knowledge gaps. A partial list is as follows: Table XXXVIII Phosphorus Content of Andropogon gayanus and Bracharia decumbens Available by Swards under a Stocking Rate of 1.7 Animal Unitdha in a Tropeptic Haplustox of Carimagua, Colombia, Fertilized with 50 kg P,Odha as Triple Superphosphate plus Small Quantities of Ca, Mg, K , and S" ~~~~

Species A . gayanus

( I -yr mean)

B . decwnbens (4-yr mean)

~

~

(%)

Phosphorus uptake (kg Plha)

Annual liveweight gains (kdha)

4.7 5.5

0.16 0.09

1.5 4.9

288 -23

10.2

0.12

12.4

265"

Rainy Dry

0.8 1.6

0.15 0.13

1.2 2.1

I25 4

Annual

2.4

0.14

3.3

I29

Dry matter on offer (tondha)

P content

Rainy Dry Annual

Season

"Adapted from 0. Paladines and P. Hoyos (unpublished data) and CIAT (1980) "Stocking rate of 2.4 animal unitdha.

396

PEDRO A. SANCHEZ AND JOSE G . SALINAS

1. Characterization of main varieties of promising ecotypes of the principal annual crops, pastures, and permanent crop species for their tolerance to the various acid soil constraints in terms of quantitative critical levels. Given the interactions between aluminum, calcium, and available phosphorus levels in the soil, the factors that are held constant should be specified. These constant factors should reflect levels found in the particular soil-farming system, not necessarily eliminating them as constraints. For legume species, plants inoculated with the appropriate rhizobium strain should be used. 2 . Characterization of the critical soil test levels for nutrient deficiency or toxicity in the principal soil types for plant species and varieties used in low-input systems. The main gaps are in the secondary nutrients and the micronutrients. 3. Development of meansfor interpreting land evaluation systems in terms of requirementsfor low-input technology. 4. Study of the changes in soil properties, both chemical and physical, with time in major soil-farming system situations. This monitoring would enable the prediction of changes in nutrient dynamics or soil physical deterioration that could occur, and correction of them before they actually happen. Soil dynamics data are scanty and usually reflect too short a period of time. Long-term monitoring of the changes in soil properties is also needed to establish a better fundamental understanding of what happens to soils managed under low-input systems. Questions about the degree of nutrient recycling, the amount of nitrogen turnover in systems involving legumes, and the efficiency of fertilizer use could be ansered by long-term monitoring of soil properties and their relationships to plant production. 5 . Agroforestry systems must be quantijied. Most of the quantitative data in this article are related to annual food crops and pastures. A data base on farming systems that involve either trees alone or trees in combination with annual crops and pastures needs to be established. 6. Increasing subsoil fertility requires substantial, additional work. A more basic understanding of the chemistry of calcium and magnesium movement is needed, as well as other factors that alleviate subsoil aluminum toxicity through leaching. 7. Tolerance to low available phosphorus needs further understanding. Theories and greenhouse studies on the differential ability of plants to acidify its rhizosphere (Israel and Jackson, 1978; Van Raij and Van Diest, 1979) should be tested and validated in Oxisol-Ultisol conditions. 8 . The various components of low-input phosphorus management technology should be put together as a package. The best source combinations, rates, placement, and the interaction with varieties tolerant to low available phosphorus, rhizobium inoculum, and potential inoculation of improved mycorrhizal strains could be combined for specific soil-farming systems. Improved or less expensive phosphorus fertilizer sources should be developed.

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

397

9. Matching of acid soil tolerance of legume species or varieties with rhizobia strains, in order to make them both compatible to the same degree of acid soil stresses and to favor rhizobia persistence in the soil. 10. Development of novel methods for improving the eflciency of nitrogen fertilizarion in nonlegutne crops and potassium fertilization in all crops. The low recovery of nitrogen and potassium fertilizers is a considerable obstacle preventing decreasing unit costs.

IX. SUMMARY Low-input technology for acid soils of the tropics can be defined as a group of practices that can produce about 80% maximum yields of acid-tolerant plant species and varieties with the most efficient use of soils and chemical inputs. The term ‘‘low’’ is used in relation to “high”-input technology where the application of fertilizers and amendments largely eliminate chemical soil constraints. The identification of plant species and ecotypes tolerant to the main acid soil stresses allows the development of low-input soil management systems for OxisolUltisol regions where socioeconomic constraints prevent the widespread application of large quantities of lime and fertilizers. The basic approach is to use plants adapted to acid soil constraints, to maximize the use of fertilizers and lime needed to produce about 80% of their maximum yield, and to take advantage of favorable attributes of acid, infertile Oxisols and Ultisols. Several technology components are reasonably well identified and could be used as building blocks for specific management systems:

1. Selection of lands dominated by well-drained Oxisols or Ultisols without steep slopes, and identification of the major soil constraints encountered. 2. Selection of species and varieties of annual crops, pastures, or tree crops that can tolerate a reasonable degree of aluminum toxicity, low available phosphorus levels, and/or manganese toxicity, as well as being adapted to climatic, insect, and disease stresses. 3. Land clearing methods in rain forests should include burning in order to take advantage of the fertilizer value of the ash, to minimize soil compaction, and to permit the rapid establishment of a crop or pasture canopy to decrease erosion hazards. Land clearing methods in the savanna are less complicated but should also aim at the quick establishment of a plant canopy. 4. Low-cost pasture establishment techniques include the introduction of improved species into native savanna, its gradual replacement, low-density seeding methods, and crop-pasture relay intercropping. Pasture maintenance techniques must consider the frequency of fertilizer applications.

398

PEDRO A. SANCHEZ AND JOSE G . SALINAS

5. Further soil cover protection can be obtained by mulching annual crops and green manuring, although the results are not always positive. Intercropping and agroforestry combinations are poorly characterized and quantified. 6. Soil acidity constraints can be attenuated without massive lime applications by ( a ) the use of plant species and varieties tolerant to aluminum and manganese toxicities, ( b )the application of sufficient lime to satisfy the calcium and magnesium requirements of plants, (c) the application of sufficient lime to decrease aluminum saturation below toxic levels, if needed, and ( d ) the promotion of the downward movement of calcium and magnesium into the subsoil. 7. Efficient phosphorus management in these soils consists of ( a ) determination of the most appropriate combination of rates and placement methods that enhance initial and residual effects, ( b ) improvement of soil fertility evaluation methods for making fertilizer recommendations, (c) use of less costly sources, such as phosphate rock, ( d ) selection of species and varieties that grow well at lower levels of available soil phosphorus, and ( e ) exploration of the practical possibilities of mycorrhizal inoculations to increase phosphorus uptake by plants. 8 . The main low-input technologies to manage low native soil fertility center on ( a ) the maximum use of nitrogen fixation by legumes using acid-tolerant rhizobia, ( h ) increase of the efficiency of nitrogen and potassium fertilization, (c) identification and correction of sulfur and micronutrient deficiencies, and ( d ) promotion of nutrient recycling. 9. Concerns have been expressed that the use of plants tolerant to acid soil constraints may completely deplete the low nutrient reserves of Oxisols and Ultisols and render them totally useless. An analysis of the total nutrient reserves of such soils, nutrient removal by crops and pastures, and the amounts of fertilizers to be added indicates no evidence of soil reserve depletion, but rather a gradual increase in total soil phosphorus and other nutrients. Since low-input technologies described in this review include fertilization, the soil mining argument appears to have little validity. REFERENCES Abrutia, F. (1980). In “World Soybean Research Conference 11: Proceedings” (F. T. Corbin, ed.), pp. 35-46. Westview Press, Boulder, Colorado. Abruria, F . , and Vicente-Chandler, J . (1967). Agron. J . 59, 539-542. Abrutia, F . , Vicente-Chandler, J . , Becerra, L.. and Bosque-Lugo, R. (1965). J . Agric. Univ. P . R . 39, 413-428. Agboola, A. A . , and Fayemi, A. A . (1972). Agron. J . 64, 6 4 - 6 4 4 . Alvarado, L. (Undated). lnstituto Colombiano Agropecuario (ICA), Bogota, Colombia. Alvim, P. T. (1976). Desarrollo Rurcrl Am. 8, 187-194. Alvim, P. T. (1981). I n ”Amazon Agricultural and Land Use Research.” CIAT, Cali, Colombia (in press). Amarasiri, S . L., and Olsen, S. R. (1973). Soil Sci. Soc. Am. Proc. 37, 716-721. Anderson, R. L., and Nelson, L. A. (1975). Eiomefrics 31, 303-318.

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

399

Andrade, E. B., de. (1979). “Sistemas de Produ@o corn Plantas Perennes em Consorcio Duplo.” EMBRAPA, Belem, Brasil. Andrew, C. S. (1976). I n “Plant Adaptation to Mineral Stress in Problem Soils” (M. J . Wright, ed.), pp. 329-340. Cornell Univ., Ithaca, New York. Andrew, C. S. (1978). In “Mineral Nutrition of Legumes in Tropical and Subtropical Soils” (C. S. Andrew and E. J . Kamprath, eds.), pp. 93-1 12. CSIRO, Melbourne, Australia. Andrew, C. S . , and Hegarty, M. P. (1969). Aust. J . Agrir. Res. 20, 687-696. Andrew, C. S., and Kamprath, E. J . (eds.). (1978). “Mineral Nutrition of Legumes in Tropical and Subtropical Soils. ” CSIRO, Melbourne, Australia. Andrew, C. S., and Robins, M. F. (1969). Aust. J . Agric. Res. 20, 655-674. Andrew, C. S., and Robins, M. F. (1971). Ausf. J . Agric. Res. 22, 693-703. Andrew, C. S . , and Vanden Berg, P. J. (1973). Ausf. J. Agric. Res. 24, 341-351. Bandy, D. E. (1976). Ph.D. Thesis, Cornell University, Ithaca, New York. Bandy, D. E. (1977). “Manejo de Suelos y Cultivos en Sistemas de Agricultura Permanente en la Selva Amazonica del Peru.” Ministerio de Alimentacion, Lima, Peni. Bartholomew, W. V. (1972). lnr. Soil Fert. Eva/. I t p r o v . Prog. Bull. 6, North Carolina State Univ., Raleigh. Benavides, S. T. (1963). M. S. Thesis, Oklahoma State University, Stillwater. Bentley, C. F., Holowaychuck, H., Leskiw, L., and Toogood, J . A. (1980). In “Bonn Conference on Agricultural Production Report. ” Rockefeller Foundation, New York. Bishop, J . P. (1981). In “Amazon Agricultural and Land Use Research.” CIAT, Cali, Colombia (in press). Blair, G . (1979). “Sulfur in the Tropics.” International Fertilizer Development Center, Muscle Shoals, Alabama. Boyd, D. A. (1970). In “Ninth Congress International Potash Institute.” pp. 461-473. International Potash Institute, Berne, Switzerland. Boyd, D. A. (1974). Phosphorus Agric. 7-17. Brinkmann. W. L. F., and de Nascimento, J . C. (1973). Turrialba 23, 284-290. Brown, J. C., and Jones, W. E. (1977a). Agron. J . 69, 410-414. Brown, J . C., and Jones, W. E. (1977b). Commun. Soil Sci. Planr Anal. 8, 1-17. Brown, 1. C., Clark, R. B., and Jones, W. E. (1977). Soil Sci. SOC. Am. Proc. 41, 747-750. Buringh. P., van Heemst, H. D. J., and Staring, C. J . (1975). ‘Computation of the Absolute Maximum Food Production of the World. ” Agricultural Univ., Wageningen, Netherlands. Buol, S. W., Sanchez, P. A,. Cate, R. B., Jr., and Granger, M. A. (1975). In “Soil Management in Tropical America” (E. Bornemisza and A. Alvarado, eds), pp. 126-141. North Carolina State Univ., Raleigh. Calvo, F. A.. Spain, J . M., and Howeler, R. H. (1977). Suelos Ecuaf. 8, 151-159. Camargo, M. N., Freire, E. S., and Venturini, W. R . (1962). Braganria 21, 143-161. Cano, M. (1973). Bol. Tec. 78. Direccion de Investigacion Agraria, Ministerio de Agricultura, Lima, Peru. Cano, M., Davelouis, J . , Valverde, C . , Arca, M., and Mendieta, J . R. (1978). “Efectividad Agronomica y Economica del Fos-Bayovar como Fuente de Fosforo y Perspectivas para su Uso.” INIA, Lima, Perti. Caro-Costas, R . , Abruna, F., and Vicente-Chandler, J . (1964). J . A@. Univ. P . R . 48, 312-317. Cate, R. B., Jr. (1965). Tech. Bull. Ed. Duarte Coelho, Recife, Brazil. Cate, R . B., Jr., and Nelson, L. A. (1971). Soil Sci. Soc. Am. Proc. 35, 658-659. Chien, S. H., and Hammond, L. L. (1978). Soil Sci. Sor. Am. J . 42, 935-939. Chien, S. H . , Leon, L. A,, and Tejeda, H. R. (1980). Soil Sci. SOC. Am. J. 44, 1267-1271. CIAT. (1975). “Annual Report for 1974. ” Centro lnternacional de Agricultura Tropical, Cali, Colombia.

400

PEDRO A. SANCHEZ AND JOSE G. SALINAS

CIAT. (1977). “Annual Report for 1976. ” Centro Intemacional de Agricultura Tropical, Cali, Colombia. CIAT. (1978). “Annual Report for 1977.” Centro Intemacional de Agricultura Tropical, Cali, Colombia. CIAT. (1979). “Annual Report for 1978. ” Centro Intemacional de Agricultura Tropical, Cali, Colombia. CIAT. (1980). “Annual Report for 1979.” Cassava and Tropical Pastures Program. Centro Intemacional de Agricultura Tropical, Cali, Colombia. CIAT. (1981). “Annual Report for 1980: Tropical Pastures Program.” Centro Intemacional de Agricultura Tropical, Cali, Colombia. Cock, J . H. (1981). In “Potential Productivity of Field Crops under Different Environments.” International Rice Research Institute, Los Bafios, Philippines (in press). Cochrane, T. T. (1979). In “Pasture Production in Acid Soils of the Tropics” (P. A. Sanchez and L. E. Tergas, eds.), pp. 1-12. Centro Intemacional de Agricultura Tropical, Cali, Colombia. Cochrane, T. T . , and Sanchez, P. A. (1981). In “Amazon Agricultural and Land Use Research.” Centro Intemacional de Agricultura Tropical, Cali, Colombia (in press). Cochrane, T. T., Porras, J. A,, Azevedo. L. G . , de, Jones, P. G . , and Sanchez, L. F. (1979). “An Explanatory Manual for CIAT’s Computerized Land Resource Study of Tropical America. Centro Intemacional de Agricultura Tropical, Cali, Colombia. Cochrane, T. T., Salinas, J. G . , and Sanchez, P. A. (1980). Trop. Agric. ( T r i n i d a d ) 5 7 , 133-140. Cock, J. H., and Howeler, R. H. (1978). Am. SOC.Agron. Spec. Publ. 32, 145-154. Coimbra, R. 0. (1963). In “Simposio sobre o Cerrado” (M. G. Ferri, Coord.), pp. 359-382. Univ. Sio Paulo, Brazil. Cordero, A. (1964). M. S. Thesis, Univ. of Fiorida, Gainesville. Cox, F. R. (1973). N . C . Agr. Exp. Sra. Tech. Bull. 219, 182-197. CPAC. (1976). “Relatorio Tecnico Anual do Centro de Pesquisa Agropecuaria dos Cerrados. 1975-1976. ” Empresa Brasileria de Pesquisa Agropecuiria, Brasilia, Brasil. CPAC. (1978). “Relatorio Tecnico Anual do Centro de Pesquisa Agropecuiria dos Cerrados. 19761977. ’’ Empresa Brasileria de Pesquisa Agropecuikia, Brasilia, Brasil. CPAC. (1979). “Relatorio Tdcnico Anual do Centro de Pesquisa Agropecuikia dos Cerrados. 1977-1978. ” Empresa Brasileria de Pesquisa Agropecubia, Brasilia, Brasil. CPAC. (1980). “Relat6rio Tecnico Anual do Centro de Pesquisa Agropecubia dos Cerrados. 1978-1979. ” Empresa Brasileria de Pesquisa Agropecubia, Brasilia, Brasil. Date, R. A., and Halliday, J. (1979). Nature (London) 227, 62-64. DeWitt, C. T. (1967). In “Harvesting the Sun” (A. San Pietro, F. A. Greer, and T. J. Army, eds.), pp. 315-320. Academic Press, New York. Donovan, K . (1973). “lnforme de Ensayos con Arroces Chancay y Huallaga 1972-1973.” Agencia de Extension, Zona Agraria IX. Yurimaguas, Peni. Dudal, R. (1980). In “Priorities for Alleviating Soil-Related Constraints to Food Production in the Tropics,” pp. 23-37. International Rice Research Institute, Los Baiios, Philippines. Duke, J. A. (1978). Am. SOC.Agron. Spec. Pub/. 32, 1-61. Duque, F. F., de Melo, J. C., de Sousa, R. L. P., and Gomide, R. L. (1980). In “Cerrado, Us0 e Manejo” (D.Marchetti and A. D. Machado, eds.), pp. 501-519. Ed. Editerra, Brasilia, Brazil. Edwards, D. G.,and Kang, B. T. (1978). Fields Crops Res. 1, 337-346. El-Swaify, S. A. (1977). In “Soil Conservation and Management in the Humid Tropics” (D. J. Greenland and R. Lal, eds.), pp. 71-77. Wiley, Chichester. Evans, C. E., and Kamprath, E. J. (1970). Soil Sci. SOC. Am. Proc. 34, 893-896. FAO-UNESCO. (1971). “Soil Map of the World Volume IV: South America.” UNESCO, Paris. FAO-UNESCO. (1975). “Soil Map of the World Volume 111: Mexico and Central America.” UNESCO. Paris.

”

LOW-INPUT TECHNOLOGY FOR OXlSOLS AND ULTISOLS

401

FAO-UNESCO. (1977). “Soil Map of the World Volume VII: South Asia.” UNESCO, Paris. FAO-UNESCO. (1979). “Soil Map of the World Volume IX: Southeast Asia.” UNESCO, Pans. Fenster, W. E., and Leon, C. A. (1979a). In “Pasture Production in Acid Soils of the Tropics” (P, A . Sanchez and L. E. Tergas, eds.), pp. 109-122. CIAT, Cali, Colombia. Fenster, W. E.. and Leon, L. A. (1979b). I n “Seminar on Phosphate Rock for Direct Application,” pp. 174-210. International Fertilizer Development Center, Muscle Shoals, Alabama. Fittkau, E. J . , and Klinge, H . (1973). Riotropiccr 5 , 2-14. Fox, R. H. (1978). Comtnun. Soil Sci. Plant Anal. 9, 13-37. Fox, R . L., and Kamprath, E. J . (1970). Soil Sci. Soc. Am. Proe. 34, 902-907. Fox. R. H., Hasan, S. M., and Jones, R. C. (1971). Proc. I n r . Symp. SoilFer/. Eval. ( N e w Delhi) 1, 857-864. Fox, R . L., Hashimoto, R. K.. Thompson, J . R . , and de la Pena, R . S. (1974). Inf. Congr. SoilSci.. IOrh, (Moscow) 4, 232-239. Foy, C. D. (1976a). CiPnc. Cuff. (Bruzil) 28, 150-155. Foy, C. D. (1976b). In “Plant Adaptation to Mineral Stress in Problem Soils” (M. J. Wright, ed.), pp. 255-268. Cornell Univ., Ithaca, New York. Foy, C. D., and Brown, J . C. (1964). SoilSci. Soc. Am. Proc. 28, 27-32. Foy, C. D., and Fleming, A. L. (1978). Am. Soc. Agron. Spec. Pub!. 32, 301-328. Foy, C. D., Lafever, H.N., Schwartz, J. W., and Fleming, A. L. (1974). Agron. J. 66, 751-758. Freitas, F. G., and Silveira, C. 0. (1977). In “1V Simposio sobre o Cerrado” (M. G. Ferri, Coord.), pp. 155-209. Univ. Sio Paulo, Brasil. Garcia. A. (1977). M. S. Thesis, Univ. Nac. Colombia, Bogota. Gerloff, G. C. (1978). In “Plant Adaptation to Mineral and Stress in Problem Soils” (M. J. Wright, ed.), pp. 161-174. Cornell Univ. Press, Ithaca, New York. Gomes, J. C., and Howeler, R. H. (1980). In “Cassava Cultural Practices” (E. J. Weber, J. C. Toro, and M. Graham, eds.), pp, 93-102. IDRC-ISIE, Ottawa, Canada. Gonzalez, E. (1976). Ph.D. Thesis, North Carolina State Univ., Raleigh. Gonzalez, E., Kamprath, E. J., Naderman, G. C., and Soares, W. V. (1979). SoilSci. Soc. Am. J . 43, 1155-1 158. Goodland, R. J . A , , and Irwin, H. S. (1975). “Amazon Jungle: Green Hell To Red Desert?” Elsevier, Amsterdam. Grove, T . L. (ed.). (1979). Cornell In/. Agric. Bull. 36. Gualdron, R., and Spain, J. M. (1980). Suelos Ecuar. 10, 131-146. Guimaries, G. A,, and Santos, J. M. S. (1968). Insr. Pesq. Exp. Agropec. Norre Circ. 11, Belern, Brasil. Halliday, J. (1979). In “Pasture Production in Acid Soils of the Tropics” (P. A . Sanchez and L. E. Tergas, eds.), pp. 123-138. CIAT, Cali, Colombia. Hammond, L. L.,Chien, S. H., and Polo, J . R. (1980). Ferf. Res. 1, 37-49. Hancock, J. K., Hill, R. W., and Hargreaves, G. H. (1979). “Potential Evapotraspiration and Precipitation Deficits for Tropical America. CIAT, Cali, Colombia. Hargreaves, G. H. ( I 977). “World Water for Agriculture: Climate, Precipitation Probabilities and Adequacies for Rainfed Agriculture. ” Utah State University, Logan. Hauser, G. F. (1974). In “Fertilizers, Crop Quality and Economy” (V. Hernando, ed.), pp. 343361. Elsevier, Amsterdam. Hecht, S. B. (1979). I n “Pasture Production in Acid Soils of the Tropics” (P. A. Sanchez and L. E. Tergas, eds.), pp. 65-80. CIAT. Cali, Colombia. Hecht, S. B. (1981). Ph.D. Thesis, Univ. of California, Berkeley. Helyar, K. R. (1978). In “Mineral Nutrition of Legumes in Tropical and Subtropical Soils” (C. S . Andrew and E. J . Kamprath, eds.), pp. 207-231. CSIRO, Melbourne, Australia. ”

402

PEDRO A. SANCHEZ AND JOSE G. SALINAS

Hensell, E. F., and Vallis, I. (1977). I n “Biological Nitrogen Fixation in Farming Systems of the Tropics” (A. Ayanaba and P. J . Dart, eds.). pp. 73-88. Wiley, New York. Hewitt, E. J. (1963). I n “Plant Physiology” (F. C. Steward, ed.), Vol. 3, pp. 137-360. Academic Press, New York. Hill, G. D. (1970). Pupua New Guinea Agric. J. 22, 69-71. Howeler, R. H. (1975). Tech. Bull. CIAT, Cali, Colombia. Howeler, R. H., and Cadavid, L. F. (1976). Agron. J. 68, 551-555. Hubbell, D. H. (1979). I n “Pasture Production in Acid Soils of the Tropics” (P. A. Sbchez and L. E. Tergas. eds.), pp. 139-145. CIAT, Cali, Colombia. Hutton, E. M. (1979). I n “Pasture Production in Acid Soils of the Tropics” (P. A. Sanchez and L. E. Tergas, eds.), pp. 8 1-94. CIAT, Cali, Colombia. Hutton, E. M., Williams, W.T., and Andrew, C. S. (1978). Aust. J. Agric. Res. 29, 67-79. IFPRI. (1978). Working Paper 1, International Food Policy Research Institute, Washington, D.C. IITA. (1980). “Research Highlights. ’’ International Institute for Tropical Agriculture, Ibadan, Nigeria. Israel, D. W.. and Jackson, W. A. (1978). I n “Mineral Nutrition of Legumes in Tropical and Subtropical Soils” (C. S. Andrew and E. J. Kamprath, eds.), pp. 113-130. CSIRO, Melbourne, Australia. Jaiyebo, E. O., and Moore, A. W. (1964). Trop. Agric. (Trinidad) 41, 129-139. Jones, M. G . , and Freitas, L. M. (1970). Preq. Agropec. Bras. 5 , 91-99. Jung, G . A. (ed.). (1978). Crop Tolerance to Suboptimal Land Conditions. ASA Spec. Publ. 32, Madison, Wisconsin. Juo, A. S. R., and Lal, R. (1977). Plant Soil 47, 567-584. Kamprath, E. J . (1970). Soil Sci. SOC. Am. Proc. 34, 252-254. Kamprath, E. J . (1971). Soil Crop Sci. SOC. F l a . Proc. 31, 200-203. Kamprath, E. J . (1973). I n “A Review of Soils Research in Tropical Latin America” (P. A. Sanchez, ed.), pp. 138-161, 179-181. North Carolina St. Univ., Raleigh. Kamprath, E. J . , and Foy, C. D. (1971). I n “Fertilizer Technology and Use” (R. A. Olsen et a / . . eds.), 2nd Ed., pp. 105-151. Soil Sci. Kellogg, C. E., and Orvedal, A. C. (1969). Adv. Agron. 21, 109-170. Keyser, H. H., Munns, D. N., and Hohenberg, J . S. (1979). Soil Sci. Sor. A m . J . 43, 719-722. Khasawneh, F. E., and Doll, E. C. (1978). Adv. Agron. 30, 159-206. Kirkby, E. A. (1979). Commun. Soil Sci. Plant Anal. 10, 89-113. Kornelius, E., Sauressig, M. G., and Goedert, W. J. (1979). I n “Pasture Production in Acid Soils of the Tropics” (P. A. Sanchez and L. E. Tergas, eds.), pp. 147-166. CIAT, Cali, Colombia. Lafever, H. N., Campbell, L. G., and Foy, C. D. (1977). Agron. J. 69, 563-568. Lal, R. (1975). IITA Monograph No. I . Ibadan, Nigeria. Lal, R. (1980). I n “Priorities for Alleviating Soil Constraints in the Tropics,” pp. 405-424. International Rice Research Institute, Los Baaos, Philippines. Lal, R., Kang, B. T., Moormann, F. K., Juo, A. F. R., and Moomaw, J. C. (1975). I n “Soil Management in Tropical America” (E. Bornemisza and A. Alvarado, eds.), pp. 372-408. North Carolina State Univ., Raleigh. Lathwell, D. J. (ed.). (1979). Cornell Int. Agric. Bull 35. Leal, J . R., and Velloso, A. C. (1973a). Pesy. Agrop. Bras. 8, 81-88. Leal, J. R., and Velloso, A. C. (1973b). Pesq. Agrop. Bras. 8 , 89-92. Lebdosoekojo, S. (1977). Ph.D. Thesis, University of Florida, Gainesville. Lehr, J . A,, and McClellan, G. H. (1972). Tenn. Val. Aurh. Bull. Y-43. Leihner, D. E. (1979). “New Technology for Cassava Intercropping. ” CIAT, Cali, Colombia. Liener, I. E. (1969). I n “Toxic Constituents of Plant Food Stuffs” (I. E. Liener, ed.), pp. 1-5. Academic Press, New York.

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTISOLS

403

Leon, L. A,, and Fenster, W. E. (1979a). Phosphorus Agric. 76, 57-73. Leon, L. A., and Fenster, W. E. (1979b). “Usos Potenciales de las Rocas Fosforicas Colombianas en Agricultura. IFDC-CIAT, Cali, Colombia. Leon, L. A . , and Fenster, W. E. (1980). “El Us0 de Rocas Fosfatadas como Fuente de Fosforo en Suelos Acidos e Infertiles de Sur America.” IFDC-CIAT. Cali, Colombia. Levitt, J . (1978). Ant. Soc. Agron. Spec. Publ. 32, 161-172. Lohnis, M. P. (1951). Planr Soil 3, 193-222. Loneregan, J . F. (1978). Ant. SOC. Agron. Spec. Publ. 32, 329-343. Lopes, A. S. (1975). M. S. Thesis, North Carolina State University, Raleigh. Lopes, A. S . (1980). I n “Priorities for Alleviating Soil-Related Constraints to Food Production in the Tropics,” pp. 277-298. IRRI, Los Banos, Philippines. Lopes, A. S., and Cox, F. R. (1977a). Soil Sci. SOC. A m . Proc. 41, 742-747. Lopes, A . S., and Cox, F. R. (1977b). Agron. J . 69, 828-831. Lopes. A. S., and Cox, F. R. (1979). Rev. Bras. Cieri. Solo 3, 82-88. McLean, E. O . , and Wheeler, R. W. (1964). Soil Sci. Soc. A m . Proc. 29, 545-550. Marchetti, D., and Machado, A. D. (eds.). (1980). “Cerrado: Us0 e Manejo” Editerra, Brasilia, Brasil. Marin, J . G. (1977). “Fertilidad de Suelos con Enfasis en Colombia.” lnstituto Colombiano Agropecuario, Bogota. Martini, J. A., Kochhann, R . A., Gomes, E. P., and Langer, F. (1977). Agron. J . 69, 612-616. Matsumura, F. (1976). “Toxicology of Insecticides. Plenum, New York. McNeil, M. (1964). Sci. A m . 211, 96-102. Mendez, J . (1973). M. S. Thesis, North Carolina State University, Raleigh. Mendez, J . , and Kamprath, E. J . (1978). Soil Sci. Soc. Am. J . 41, 86-88. Millaway, R. M. (1979). Cominun. Soil Sci. Plant Anal. 10, 1-28. Ministerio das Minas e Energia, Brasil. (1973-1979). “Projeto RADAM-Brasil Vol. I-XI. ” Departamento Nacional da Produqio Mineral, Rio de Janeiro. Miranda, L. N., and Lobato, E. (1978). Rev. Bras. Cien. Solo 2, 44-50. Miranda, L. N . , Mielniczuk, J., and Lobato, E. (1980). I n “Cerrado: Usoe Manejo” (D. Marchetti and A. D. Machado, eds.). pp. 521-578. Editerra, Brasilia, Brasil. Mohr, W. (1960). An. I Congr. N a c . Cons. Solo, pp. 61-76. Campinas, Sio Paulo, Brasil. Mokwunye, A . W., and Chien, S. H. (1980). Soil Sci. Soc. Am. J. 44, 477-482. Mongi, H. 0.. and Huxley, P. A. (eds.). (1979). I C R A F Bull. 1, Nairobi, Kenya. Mosse, B., Hayman, D. S., and Arnold, D. J . (1973). New Phytol. 72, 809-815. Mott, G. 0. (1974). I n “Forage Fertilization” (D. A. Mays, ed.), pp. 323-339. Amer. SOC.Agron., Madison, Wisconsin. Moura, W., Buol, S . W., and Kamprath, E. J . (1972). Experienriae ( B r a z i l ) 13, 235-247. Munn, D. A,, and McCollum, R. E. (1976). Agron. J . 68, 981-991. Munns, D. N . (1978). I n “Mineral Nutrition of Legumes in Tropical and Subtropical Soils” (C. S. Andrew and E. J . Kamprath), pp. 247-264. CSIRO, Melbourne, Australia. Munns, D. N . (1980). I n “World Soybean Conference 11: Proceedings” (F. T. Corbin, ed.), pp. 47-56. Westview Press, Boulder, Colorado. Munns, D. N . , Keyser, H. H., Fogle, V . W., Hohenberg, J . S., Righetti, T. L., Lauter, D. L., Zaroug, M . G . , Clarkin, K . L., and Whitacre, K . A. (1979). Agron. J . 71, 256-260. Mussell, H., and Staples, R . C. (eds). (1979). “Stress Physiology in Crop Plants.” Wiley, New York. Muzilli, O., Santos, D., Palhano, J. B., Manetti, J . , Lantmarr, A. F., Garcia, A,, and Cataneo, A. (1978). Rev. Brus. Cien. Solo 2, 34-40. National Academy of Sciences. (l977a). “World Food and Nutrition Study: The Potential Contributions of Research. ” National Academy of Sciences, Washington, D.C. ”

”

404

PEDRO A. SANCHEZ AND JOSE G. SALINAS

National Academy of Sciences. (1977b). “Leucaena: Promising Forage and Tree Crop for the Tropics.” National Academy of Sciences, Washington, D.C. NCSU. (1972). “Agronomic-Economic Research on Tropical Soils. Annual Report for 1972. North Carolina State Univ., Raleigh. NCSU. (1974). “Agronomic-Economic Research on Tropical Soils.” Annual Report for 1973. North Carolina State Univ., Raleigh. NCSU. (1975). “Agronomic-Economic Research on Tropical Soils.” Annual Report for 1974. North Carolina State Univ., Raleigh. NCSU. (1976). “Agronomic-Economic Research on Tropical Soils.” Annual Report for 1975. North Carolina State Univ., Raleigh. NCSU. (1978). “Agronomic-Economic Research on Tropical Soils.” Annual Report for 1976-77. North Carolina State Univ., Raleigh. NCSU. (1980). “Agronomic-Economic Research on Tropical Soils.” Annual Report for 1978-79. North Carolina State Univ., Raleigh. Neyra, C. A,, and Dobereiner, J. (1977). Adv. Agron. 29, 1-38. Nickel, J . L. (1979). In “CIAT Highlights for 1978,” p. viii. CIAT, Cali, Colombia. Nonis, D. 0. (1972). Trop. Grass/. 6, 159-170. Novais, R. F., and Kamprath, E. J . (1979). Rev. Bras. Cien. Solo 3, 37-41, Oliveira, D., de. (1979). EMBRATER lnf. Mandioq. 30, Brasilia, Brazil. Olmos, 1. L. J . (1971). Equipe de Pedologia e Fertilidade do Solo. Tech. Bull. 16, Rio de Janeiro, Brazil. Olmos, I. L. J., and Camargo, M. N . (1976). Cien. Cult. (Brazil) 28, 171-180. Paladines, 0. P., and Leal, J. (1979). In “Pasture Production in Acid Soils of the Tropics’’ (P. A. Sanchez and L. E. Tergas, eds.), pp. 31 1-322. CIAT, Cali, Colombia. Pearson, R. W. (1975). Cornell Int. Agric. Bull. 30. Pearson, R. W., Abruna, D., and Vicente-Chandler, J. (1962). Soil Sci. 93, 77-82. Pearson, R. W., Perez-Escolar, R., Abmiia, F., Lund, Z. F., and Brenes, E. J. (1977). J. Agric. Univ. P . R . 61, 361-382. Perez-Escolar, R. (1977). J. Agric. Univ. P . R . 61, 82-89. Pinchinat, A. M., Soria, J . , and Bazan, R. (1976). A m . Soc. Agron. Spec. Pub/. 27, 51-62. Plucknett, D. L. (1978). A m . SOC. Agron. Spec. Pub/. 32, 125-144. Polk, E., Konzak, C. F., and Kittrick, J. A. (1978). Crop Sci. 18, 823-827 Ponnamperuma, F. N . (1972). Adv. Agron. 24, 29-96. Ponnamperuma, F. N . (1976). In “Plant Adaptation to Mineral Stress in Problem Soils” (M. J. Wright, ed.), pp. 341-354. Cornell Univ. Press, Ithaca, New York. Ponnampemma, F. N. (1977). IRRl Res. Paper Ser. 6. Pope, R. A . (1976). Ph.D. Thesis, North Carolina State University, Raleigh. Ramalho Filho. A., Pereira, E. G . , and Beek, K . J. (1978). “Sistema de AvaliaSHo da AptidHo Agricola das Terras. Serviso Nacional de Levantamento e Conserva@o de Solos, EMBRAPA, Rio de Janeiro, Brazil. Rhue, R. D. (1979). In “Stress Physiology in Crop Plants” ( H . Mussell and R. C. Staples, ed.), pp. 61-80. Wiley, New York. Ritchey, K . D. (ed.). (1979). Cornell Int. Agric. BUlI. 37. Ritchey, K . D., Djalma, M. G., Lobato, E . , and Coma, 0. (1980). Agron. J. 72, 40-44. Robson, A . D. (1978). In “Mineral Nutrition of Legumes in Tropical and Subtropical Soils” (C. S. Andrew and E. J. Kamprath, eds.). pp. 277-293. CSIRO, Melbourne, Australia. Rodriguez, M .(1975). M. S . Thesis, Cornell University, Ithaca, New York. Salas, G., de las. (1978). Corporacion Nacional de Investigacion y Fomento Forestal, Bogota, Colombia. CONIF Ser. Tec. 8. Salas, G . , de las, and Folster, H. (1976). Turrialba 26, 179-186. Salinas, J . G. (1978). Ph.D. Thesis, North Carolina State University, Raleigh. ”

”

LOW-INPUT TECHNOLOGY FOR OXISOLS AND ULTlSOLS

405

Salinas, J. G. (1980). In “Fertilidad de Suelos-Diagnostico y Control” (F. Silva-Mojica, ed.), pp. 399-420. Sociedad Colombiana Ciencia del Suelo, Bogota. Salinas, J. G . , and Delgadillo, G. (1980). Lat. A m . Soil Sci. Congr., 7rh, Sun Jose, Cosra Rica. June, Salinas, J . G . , and Sanchez, P. A . (1976). Cien. Cult. (Bruzi/)28, 156-168. Salinas, J. G., Gonzalez, E., Kamprath, E. J . , and Sanchez, P. A. (1976). I n “AgronomicEconomic Research on Soils of the Tropics.” pp. 81-98. North Carolina State Univ., Raleigh. Samuels, F., Hernandez, E., and Torres, S. (1975). J . Agric. Univ. P. R . 58, 134-136. Sanchez, C. (1977). “Encalamiento de Ultisoles de Sabana.” Univ. del Oriente, Jusepin, Venezuela. Sanchez, P. A. (1972). N. C. Agric. Exp. Stu. Tech. Bull. 213. Sanchez, P. A . (ed.). (1973). N . C. Agric. Exp. Stu. Tech. Bull. 219. Sanchez, P. A. (1976). “Properties and Management of Soils in the Tropics.” Wiley, New York. Sanchez, P. A . (1977). I n “Proc. Int. Sem. on Soil Environment and Fertility Management in Intensive Agriculture,” pp. 535-566. SOC.Sci. Soil and Manure, Tokyo, Japan. Sanchez, P. A. (1979). I n “Soils Research in Agroforestry” (H. 0. Mongi and P. A. Huxley, eds.), pp. 79-124. ICRAF, Nairobi, Kenya. Sanchez, P. A., and Cochrane, T. T. (1980). I n “Priorities for Alleviating Soil-Related Constraints to Food Production in the Tropics,” pp. 107-140. IRRI, Los Batios, Philippines. Sanchez, P. A., and Isbell, R. F. (1979). In “Pasture Production in Acid Soils of the Tropics” (P. A. Sanchez and L. E. Tergas, eds.), pp. 25-54. CIAT, Cali, Colombia. Sanchez, P. A., and Nurena, M . A . (1972). N. C. Agric. Exp. Sta. Tech. Bull. 210. Sanchez, P. A., Gavidia, A., Ramirez, G. E., Vergara, R., and Minguillo, F. (1973). SoilSci. Soc. Am. Proc. 37, 789-791. Sanchez, P. A., and Uehara, G. (1980). I n “Phosphorus in Agriculture” (F. E. Khaswaneh, E. Sample, and E. J . Kamprath, eds.), pp. 471-514. Amer. SOC.Agron, Madison, Wisconsin. Sanders, F. E., Mosse, B., and Tinker, P. 8. (eds.). (1975). “Endomycorrhizas.” Academic Press, New York. Santana, M. B. M., Cabala-Rosand, F. P., and Vasconcelos Filho, A. P. (1977). Rev. Theobroma 7 , 125-132. Sartain, J. B., and Kamprath, E. J . (1978). Agron. J . 70, 17-20. Schubart, H. 0. R. (1977). Acta Amazon. 7 , 559-567. Schultze-Kraft, R., and Giacometti, D. C. (1979). I n “Pasture Production in Acid Soils of the Tropics” (P. A . Sanchez and L. E. Tergas, eds.), pp. 59-69. CIAT, Cali, Colombia. Serrao, E. A. S., Falesi, I. C., Viega, J. B., and Texeira, J. F. (1979). In “Pasture Production in Acid Soils of the Tropics” (P. A. Sanchez and L. E. Tergas, eds.), pp. 195-226. CIAT, Cali, Colombia. Seubert, C. E., Sanchez, P. A,, and Valverde, C. (1977). Trop. Agric. (Trinidad) 54, 307-321. Silva, A. R., da. (1976). I n “Plant Adaptation to Mineral Stress in Problem Soils” (M. J. Wright, ed.), pp. 223-231. Comell Univ. Press, Ithaca, New York. Silva, L. F. (1978). “InfluCncia do Manejo de um Ecossistema nas Propriedades Edaficas dos Oxisols de Tabuleiro.” Centro de Pesquisas do Cacau, CEPLAC, Itabuna, Bahia, Brazil. Silva, S., and Vicente-Chandler, J . (1974). J . Agric. Univ. P. R . 58, 134-136. Simar, A,. Cradock, F. W., and Hudson, A. W. (1974). Plum Soil 41, 129-140. Smyth, T. J . (1981). Ph.D. Thesis. North Carolina State University, Raleigh. Smyth, T . J . , and Sanchez, P. A. (1980a). SoilSci. Soc. A m . J . 44, 500-505. Smyth, T . J., and Sanchez, P. A. (1980b). Rev. Bras. Cien. Solo. 4, 88-92. Soil Conservation Service. (1975). U . S . Dep. Agric. Handb. 436, Washington, D.C. Souto. C. H., and Dobereiner, J. (1969). Pesq. Agrop. Bras. 4, 129-138. Spain, J. M. (1975). Am. SOC. Agron. Spec. Pub/. 24, 1-8.

406

PEDRO A. SANCHEZ AND JOSE G . SALINAS

Spain, J. M. (1976). I n “Plant Adaptation to Mineral Stress in Problem Soils” (M. J. Wright, ed.), pp. 213-222. Cornell Univ. Press, Ithaca, New York. Spain, J . M . (1979). I n “Pasture Production in Acid Soils in the Tropics” (P. A. Sanchez and L. E. Tergas, eds.), pp. 167-175. CIAT, Cali, Colombia. Spain, J . M., Francis, C. A . , Howeler, R. H . , and Calvo, F. (1975). Iri “Soil Management in Tropical America” (E. Bornemisza and A. Alvarado, eds.), pp. 308-329. North Carolina State Univ., Raleigh. Terman, G . L., and Englestad. 0. P. (1972). TVA Bull. Y-21. Thomas, D. (1978). Trop. Agric. (Trinidnd) 55, 39-44. Toledo. J . M., and Morales, V. A. (1979). I n “Pasture Production in Acid Soils of the Tropics” (P. A. Sanchez and L. E. Tergas, eds.). pp. 177-194. CIAT, Cali, Colombia. Toma, N. S. (1978). M. S. Thesis, North Carolina State University, Raleigh. Turenne, J. F. (1969). “Deforestation et preparation du sol bmlis. Modifications des caracteres physico-chimiques de I’horizon superieur du sol. ORSTOM Centre, Cayenne, Guiana Franqaise. Turenne, J. F. (1977). Sirnp. I n / . Ecol. T r o p . , 4 / h , ORSTOM, Fort rle Frcirice. UEPAE de Manaus. (1978). “Relatorio Tecnico Anual para 1977.” EMBRAPA, Manaus, AM, Brazil. UEPAE de Manaus. (1979). “Relatorio Tecnico Anual para 1978.” EMBRAPA, Manaus, AM, Brazil. Ulrich, A. (1952). Annu. Rev. Plun/ Physiol. 3, 207-228. Valverde, C . , and Bandy, D. E. (1981). Irt “Amazon Agricultural and Land Use Research.” CIAT, Cali, Colombia (in press). Van der Ween, R. (1974). Trop. Agric. (Triniclud) 51, 325-331. Van Raij, B., and Van Diest, A. (1979). Plunr Soil 51, 577-589. Vasconcelos, C . A . , Braga, J. M., Novais, R. F.. and Pinto. 0. (1975). Cerrs (Bruzil) 22, 22-49, 62-73. Vicente-Chandler, J . , and Figarella, J . (1962). J . Agric. Univ. P . R . 46, 226-236. Vicente-Chandler. J . , Caro-Costas. R., Pearson, R. W . , Abruria, F . , Figarella, J., and Silva, S. (1964). Univ. P . R . Bull. 187. Vicente-Chandler, J . , Abruria, F., Caro-Costas, R., Figarella, J., Silva, S., and Pearson, R. W. (1974). Univ. P . R . Agric. Exp. S / U . Bull. 223. Villachica, J . H. (1978). Ph.D. Thesis, North Carolina State University, Raleigh. Villagarcia, S. (1973). Ph.D. Thesis, North Carolina State University, Raleigh. Vlamis, J . , and Williams, D. E. (1973). Plurtr Soil 39, 245-251. Wade, M. K. (1978). Ph.D. Thesis, North Carolina State University, Raleigh. Waggoner, P. E . , and Norvell, W . A. (1979). Agron. J. 71, 352-354. Waidyanatha, U. P. S., Yogaratnam, N., and Ariyatne, W. A. (1979). New Ph-yrol. 82, 147-152. Wang, C. H., Liern, T. H., and Mikklesen, D. H. (1976). IRI Res. Bull. 47. Waugh, D. L., Cate, R. B., Jr., Nelson, L. A , , and Manzano. A . (1975). I n “Soil Management in Tropical America” (E. Bornemisza and A. Alvarado, eds.), pp. 484-501. North Carolina State Univ., Raleigh. Whiteaker, G . , Gerloff, G. C . , Gableman, H., and Lindgen, D. (1976). J . Am. SOC. Hort. Sci. 101, 474-479. Wolf, J . M. (1977). Pesq. Agropec. Brus. 12, 141-150. Wright. M. J . (ed.). (1976). “Plant Adaptation to Mineral Stress in Problem Soils.” Cornell Univ. Press, Ithaca, New York. Yost, R. S., and Fox, R. L. (1979). Agron. J. 71, 903-908. Yosr, R. S . , Kamprath, E. J., Lobato, E., and Naderman. G . C. (1979). Soil Sri. Soc. Am. J . 43, 338-343. Zandstra, H. G. (1971). Ph.D. Thesis, Cornell University, Ithaca, New York. ”

ADVANCES IN AGRONOMY, VOL. 34

CYTOGENETICS OF PEARL MILLET* Prem P. Jauhar-t. Department of Botany and Plant Sciences, University of California-Riverside, Riverside, California

.................... I . Introduction . . . . . . . . . . . . 11. Karyotypic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Karyomorphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Meiosis.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........ A. P . typhoides ( B u m . ) Stapf et Hubb. ( 2 n = 2x = 14). . . . . . . . . . . . . . . . . . . . . . . B. P . purpureutn Schumach. (2n = 28, 56). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I V . Abnormal Meiosis and Its Genetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Desynapsis and Its Genetic Basis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Effect of Nutrients on Desynapsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Effect of Ploidy on Desynapsis.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Effect of Desynaptic Gene on B Chromosomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Experimental Induction of Desynapsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Inbreeding and Disruption of Chromosome Pairing: Heterozygosis and Heterosis for Chromosome Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V . Haploidy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Spontaneous and Induced Haploids ................................... ................................... B. Chromosome Pairing in Haploids . . C. Haploids i n Plant Breeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V1. Polyploidy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Spontaneous Occurrence . . . . . . . . . . . . . . . . . B. Factors Favoring Spontaneous Polyploidy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Induction of Polyploidy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Polyploidy and Plant Breedin VII. Aneuploids . . . . . . . . . . . . . . . . . . A. Spontaneous Occurrence . . . . B. Synthesis of Aneuploids of Pe ................................... C. Trisomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Tetrasomics VIII. Structural Changes in Chromosomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Spontaneous Rearrangements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .................. B. Induced Translocations C. Building Up a Complete Interchange Ring . . . . D. Karyomorphology and the Incidence of Interchanges . . . . . . . . . . . . . . . . . . . . . . . .

408

415 415 417 417

422 422 423 424 424 424 427

429 430 432 434 434 435 436 441 442 442 442

444

*This article is dedicated with admiration to Dr. Glenn W. Burton, whose pioneering work has contributed significantly to the genetic improvement of pearl millet. ?Present address: Division of Cytogenetics and Cytology, City of Hope National Medical Center, Duarte. California 91010. 407

Copyright 0 1981 by Academic Press. Inc. All nghts of repduction in any form resewed ISBN 0-12-000734-7

408

PREM P. JAUHAR E. Identification of Translocated Chromosomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Disjunction of Interchange Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G. Interchange Heterozygosity and Plant Breeding . . . . . . . . . . . . . . . .

444 445

IX. B Chromosomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. B Chromosomes as Indicators of the Origin of Pearl Millet. . . . . . . . . . . . . . . . . . . 447

B . Mode of Pollination.. . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . 454

XI. Hybridization and Chromosome Relationships . . . . . . . . . . . . . . . . . .................................... A. lntraspecific Hybrids B. Interspecific Hybrids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Intergeneric Hybridization . . . . . . . . . . . . . . . . . . . . . . . . . . . . XII. Conclusion . ................................ References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

458

472 473

I. INTRODUCTION Pennisetum is one of the most important genera of the tribe Paniceae of the grass family. Pearl millet [Pennisetum ryphoides (Burm.) Stapf et Hubb.] is the most important constituent of this genus. It is a dual-purpose crop: its grain is used for human consumption and its fodder serves as feed for cattle. In Asia and Africa, however, it is grown primarily as a grain crop on an estimated 60 million acres of relatively poor land. It has remarkable ability to grow in areas of low rainfall. In sub-Saharan Africa harvests of pearl millet are obtained with as little as 250 mm of annual rainfall (Brunken, 1977). Its grain is traditionally considered to be nutritious and is put to a variety of uses. As poor man’s bread, it sustains a large proportion of the populace of Africa and Asia. It also contributes to the economy of countries like the United States, where it is grown as a forage crop on an estimated 1 million acres. Pearl millet is also grown as a forage crop in the tropical and warm-temperate regions of Australia and several other countries. Pearl millet originated in West Africa. Selection exercised by the early cultivators under a myriad of cultural contexts led to the development of several morphologically diverse forms. Its protogynous nature facilitated the introgression of characters from other wild and cultivated annual species of the section Penicillaria. It is now widely cultivated in different parts of the world. In terms of annual production, pearl millet is the sixth most important cereal crop in the world, following wheat, rice, maize, barley, and sorghum. Among the millets it is second only to sorghum. In India, it is the fourth most important cereal after rice, wheat, and sorghum.

CYTOGENETICS OF PEARL MILLET

409

Pearl millet is a favorable organism for genetic research. Its chromosome number, 2n = 14, was determined more than 50 years ago by Rau (1929). Several favorable features of the chromosome complement, e.g., small number and large size of chromosomes with one distinctive pair of nucleolar organizers, make pearl millet a suitable organism for cytogenetic studies. Moreover, its protogynous flowers and outbreeding system make it ideal for interspecific hybridization and for breeding work. It is indeed ideally suited for heterosis breeding. Although pearl millet has a remarkable ability to grow on soils of marginal fertility, it responds very well to proper fertilization, which helps in realizing the high yield potential of its hybrids. The hybrids’ greater N use efficiency (biomass production per unit of N in the plant) is probably attributable to the highly efficient (C,) photosynthetic pathway of this crop. Although pearl millet has great agricultural importance, is a very favorable organism for cytogenetic studies and breeding work, and has a low chromosome number that was also determined at about the same time as those of most other crops, the information available on its genetics and cytogenetics is much less than that known for other important crops. There are several reasons why this crop has been largely overlooked as a genetic and cytogenetic tool:

1. It has long been considered to be a crop of secondary importance and, thus, could not compete for research funding with other crops like wheat and corn. 2 . It has a restricted area of use, being a food for the poor only, although it is also an excellent fodder crop. 3. Its potential as a research tool was not appreciated until recently. 4. The existence of long-standing nomenclatural controversies (in the postLinnaean period from 1753 to 1759, pearl millet has been treated as a member of at least six different genera, viz., Panicum, Holcus, Alopecurus, Cenchrus, Penicillaria, and Pennisetum, and has been given different botanical names; see Jauhar, 1981a) could also have had an adverse impact on research. Studies on chromosome pairing in interspecific hybrids-with pearl millet as one of the parents-have contributed to our understanding of phylogenetic relationships between different P ennisetum species and pearl millet. In these studies, the large size of the pearl millet chromosomes has been helpful in ascertaining chromosome relationships. Because of its low chromosome number, pearl millet also offers a particularly favorable material for aneuploid analyses, which should be helpful in the elucidation of its cytogenetic architecture. Primary trisomics constitute a valuable tool for locating genes on different chromosomes and for assigning them to linkage groups. Although considerable progress has been made in developing a set of primary trisomics in pearl millet, the establishment of linkage groups awaits completion. A good deal of information is available on certain other cytogenetic aspects, e.g., polyploidy, interchange heterozygosity , haploidy, and B chromosomes. All these studies should contribute to the improvement programs of pearl millet.

410

PREM P. JAUHAR

The purpose of this article is to summarize and integrate the available information on different aspects of pearl millet cytogenetics. It is hoped that this article will provide useful information to cytogeneticists and breeders engaged in the improvement of pearl millet and other forage species of Penniseturn. This article may also be of interest to a spectrum of other workers engaged in basic research.

II. KARYOTYPIC ANALYSIS Karyotypic analysis includes the study of the number, size, and morphology of chromosomes. Total length and arm ratios of chymosomes are helpful in systematic and phylogenetic investigations. Levitskii (193 1) and Avdulov (1 93 1) pioneered the use of cytological features as aids in establishing taxonomic and phylogenetic relationships among species and genera. Although basic number, size, and morphology of the chromosomes can indeed be useful in taxonomic classification (Hunter, 1934; Constance, 1957), these parameters should be subsidiary to morphological characters in any taxonomic treatment (Pilger, 1954). Modern cytological techniques, e.g., the banding of chromosomes with Giemsa (Vosa and Marchi, 1972; Vosa, 1973, 1975), and staining heterochromatic patterns with fluorochromes like quinacrine mustard (Vosa, 1970) can provide information of phylogenetic value. The occurrence of cytotypes or chromosomal races (intraspecific polyploid series) is a characteristic feature of the perennial species of Penniseturn. However, no such cytotypes exist in the annual cultivated or wild pearl millets, which all have 2 n = 14 chromosomes (Table I); in fact, all these taxa belong to the species P . ryphoides. There is a report of 2 n = 36 chromosomes for a Nigerian collection of “ P . violaceurn (Lam.) L . Rich.” (Olorode, 1975), but this could be an incorrect identification. Since the material classified as P . violaceurn forms fully fertile hybrids with pearl millet (2n = 14), the former must have 2 n = 14 chromosomes (see Section XI,A). KARYOMORPHOLOGY

Chromosomes are generally measured at somatic metaphase after pretreatments that condense and spread them. The main drawback inherent in these studies is that the magnitude of error in the measurements of condensed chromosomes is high. Therefore relatively small size differences among chromosomes of a species, of infraspecific categories, or of different species cannot be resolved accurately. However, karyomorphological studies can be done more precisely on pachytene chromosomes in taxa with low chromosome numbers, e.g., P .

41 1

CYTOGENETICS OF PEARL MILLET

TABLE I Chromosome Numbers of Different Taxa in the Section Penicillaria of Genus Penniserurn Taxa Cultivated pearl millet P. typhoides (Burm.) Stapf et Hubb. [Syn. P. tvphoideutn Rich. P. spicarurn (L.) Koern. P. glaucum (L.) R. Br. P. amerironum ( L . ) K. Schum.] Annual relatives of pearl millet" P. alhicauda Stapf et Hubb. P. rmcvhchaele Stapf et Hubb. P. rinereum Stapf et Hubb. P. dalzielii Stapf et Hubb. P. echinurus (K. Schum.) Stapf et Hubb. P. fullax (Fig. & De Not.) Stapf et Hubb. P. gambiense Stapf et Hubb. P. leanis Stapf et Hubb. P. tnuiwa Stapf et Hubb. P. nigrilarurn Schlecht. P. mollissimum Hochst. P. perrottetii (Klotzch ex A.Br.)K. Schum P. pyrnostachyum (Steud.) Stapf et Hubb. P. pynostachvunr var. gambia P. versicolor Schrad. P. violareum (Lam.) L. Rich Perennial relative of pearl millet P. purpureutn Schum.

References

2n

14

14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 I4? 28

56

Rau (1929)

Thevenin (1952) Krishnaswamy (1951) Krishnaswamy (1951) Krishnaswamy (1951);Thevenin (1952) Krishnaswamy (1951) Thevenin (1952) Thevenin (1952)

Thevenin (1952) Krishnaswamy (1951) Bilquez and Lecomte (1969) Mehra et ul. (1968) Burton (1942); Nishiyama and Kondo ( 1942) Krishnaswamy and Raman (1948); Gadella and Kliphuis (1964)

"These and other annual, cultivated, or wild relatives of pearl millet have 2n = 14 chromosomes. They are not reproductively isolated from the cultivated species-P. fyphoides-and in fact do not deserve specific ranks.

ramosum ( 2 n = 10) and P . typhoides ( 2 n = 14). For critical comparisons, the DNA content of chromosomes can also be measured. The genus Pennisetum is a heterogeneous assemblage of species with chromosome numbers ranging from 2 n = 10 to 2 n = 7 2 , being multiples of 5, 7, 8, and 9. Their chromosome morphology is also very diverse, with tremendous size differences; a noteworthy feature is that the species with lower numbers have the larger sizes. Thus, pearl millet ( P . ryphoides) has only 2 n = 14, but relatively very large chromosomes. Avdulov (1931) noted that pearl millet had 14 large chromosomes, larger than those of any other member of the tribe Paniceae.

412

PREM P. JAUHAR

However, I think that the annual (or rarely biennial) species P . ramosum (2n = 10) has the largest chromosomes in the genus Pennisetum and probably in the entire tribe Paniceae. The chromosomes of P . ramosum are approximately 5% larger than those of P . typhoides. Thus, in the genus Pennisetum, the species with the lowest chromosome number (2n = 10) has the largest chromosomes. In contrast, the species with higher chromosome numbers (e.g., P . orientale, 2n = 18, 36, 54) have strikingly smaller chromosomes than those of P . ramosum or P . typhoides. The trend of species with low chromosome numbers to have much larger chromosomes is evident in several other plant groups. In Sorghum, for example, the average lengths of chromosomes of S. versicolor (2n = lo), S . vulgare (2n = 20), and S . halepense (2n = 40) were 4.86, 2.24, and 1.98 p m , respectively (Karper and Chisholm, 1936). 1 . P . typhoides (2n = 1 4 )

Rau (1929) determined from root tips the chromosome number of pearl millet as 2n = 14. Moreover, he mentioned that “the chromosomes are very large” and that the homologous pairs could be easily distinguished. Avdulov (1931) studied the chromosomes of pearl millet, which was at that time classified as Penicillaria spicata Willd. His drawing shows 14 chromosomes with median to submedian centromeres, the shortest chromosome being the satellited one. It is interesting to note that as early as 1931 when cytological techniques were not perfected, Avdulov noticed one pair of satellited chromosomes; this observation has been confirmed by numerous workers. The small nucleolar bivalent is clearly observed to be associated with the nucleolus (Fig. 1 ) . Pantulu (1 958) examined the chromosomes at pachytene and grouped them into four classes on the basis of relative length and position of centromere: ( 1 ) two large pairs (chromosomes 1 and 2) with median centromeres; ( 2 ) two somewhat shorter pairs (chromosomes 3 and 4) with median to submedian centromeres; (3) two medium-sized pairs (chromosomes 5 and 6) with submedian centromeres; and (4)the shortest pair (chromosome 7) with the nucleolus organizer. Later, essentially similar results were obtained on the analysis of karyotype at pachytene (Venkateswarlu and Pantulu, 1968; and Lobana and Gill, 1973), at pollen mitosis (Krishnaswamy and Raman, 1953a), and at somatic metaphase (Burton and Powell, 1968). Virmani and Gill (1972) and Tyagi (1975a) karyotyped the somatic chromosomes and classified them as follows: chromosomes 1 , 2, 3, and 5 as metacentric; chromosomes 4 and 6 as submetacentric; and chromosome 7 as subterminal. Thus, there are minor disagreements among different workers as to the position of centromere. Looking at a condensed chromosome at somatic metaphase, it is not unexpected that one worker locates the centromere as median, whereas another classifies it as submedian. The same workers, looking at pachytene and somatic chromosomes, can also arrive at different conclusions. For example,

CYTOGENETICS OF PEARL MILLET

413

FIG. 1. Diakinesis in pearl millet ( 2 n = 14) showing one nucleolar bivalent; note small rod associated with the nucleolus. Also note chiasma terminahation. [ X 12601

Virmani and Gill (1972) studied somatic chromosomes and classified chromosome 1 as metacentric; whereas, based on pachytene analysis, Lobana and Gill (1973) considered it to be submetacentric. There is no doubt that pearl millet has a fairly symmetrical karyotype. It is certainly not very easy to identify all of the seven chromosomes by the techniques currently used; therefore, Giemsa banding (see Vosa, 1973, 1975; Zelleret al., 1977; Filion and Blakey, 1979) of somatic prometaphase chromosomes must be tried to identify individual members of the complement. It has mostly metacentric or submetacentric chromosomes, the longest being approximately 1.5 times the shortest; both these features are indices of symmetry of karyotype. Under Stebbins’ (1958) classification of types of asymmetry, pearl millet will fit best in the class la, i.e., the most symmetrical of the 12 karyotypes described. The shortest chromosome pair is somewhat subterminal with the satellite on its short arm. It can be identified in somatic plates as well as at pachytene and diakinesis, where, as a small bivalent, it is associated with the nucleolus (Fig. 1). Chromosomes of some diploid taxa of the section Penicillaria, which are annual relatives of pearl millet, have been observed. The materials classified as Pennisetum cinereum, P. echinurus, P . gumbiense, P. leonis, and P. pycnosruchyum had 2n = 14 chromosomes, as in cultivated pearl millet (Krishnaswamy, 1951). Veyret (1957) found that P. ancylochaete, P . gambiense, P . maiwa, and P. nigritarum had 2n = 14 chromosomes, and their chromosome morphologies were similar to one another and also to that of cultivated pearl millet. Genetic studies by Bilquez and Lecomte (1969) and Brunken (1977)

414

PREM P. JAUHAR

have shown that P . violuceum and P . fullux-two of the important wild, annual relatives of pearl millet-are not reproductively isolated from it; their hybrids with pearl millet were highly fertile. Although these workers have not mentioned the chromosome number of these wild taxa, they obviously have 2n = 14 chromosomes in order to form fertile hybrids with pearl millet.

2. P . purpureum (2n

= 4x =

28)

Napier grass, an allotetraploid and a relative of pearl millet, has a somewhat asymmetrical karyotype consisting of chromosomes with median, submedian, and subterminal centromeres. On the basis of pachytene studies, Pantulu and Venkateswarlu (1968) reported that the longest chromosome of the complement (chromosome 1) was 2.7 times the length of the shortest (chromosome 14), thus making the karyotype asymmetrical. Based on these observations, the karyotype of P . purpureum will fall in the category 2b in Stebbins’ (1958) classification of types of asymmetry. Pantulu and Venkateswarlu (1968) reported that chromosomes 1 and 14 have nucleolus organizers. The largest chromosome of the complement (chromosome 1) is certainly satellited, as evidenced by the association of the largest bivalent with the nucleolus (Fig. 2 ) . The other nucleolar bivalent is one of the smallest, if not the smallest, in the complement. If chromosome 14 is indeed satellited, then

FIG.2. Diakinesis in napier grass ( 2 n = 4x = 28) showing 14 bivalents (9 rings and 5 rods) with terminalized chiasmata. Note one large bivalent (the largest in the complement) and one relatively small bivalent associated with the nucleolus. Also note an additional small nucleolus (marked with arrow). [ x 12601

CYTOGENETICS OF PEARL MILLET

415

P. purpureum shares an important karyotypic feature with P. typhoides, i.e., the shortest chromosomes of both the species are satellited. Moreover, during meiotic prophase both typhoides and purpureurn show rapid terminalization of chiasmata (see Section 111). They also seem to have similar patterns of centromeric heterochromatin. Thus, P. typhoides and P. purpureum seem to share some important karyological features of phyletic value.

Ill. MEIOSIS All penicillarias fall into the x = 7 group (see Table I). They have conspicuously penicillate anther tips (see Fig. 14a,b). Of these, only one species ( P . purpureum) is a perennial tetraploid. All other taxa are annual and diploid with 2n = 2x = 14 chromosomes. The annual, semiwild taxa are not reproductively isolated from the cultivated pearl millet and must be considered as infraspecific categories within P . typhoides. They have regular meiosis with 7,,, as in P. typhoides. A . P . ryphoides ( B u R M . )STAPFET HUBB.( 2 n = 2x = 14)

Rau (1929) determined the chromosome number of pearl millet as 2 n = 14. Rangaswamy (1935) studied meiosis and found at diakinesis mostly seven ringshaped bivalents having two terminalized chiasmata each. In different populations of pearl millet, mostly ring bivalents with two chiasmata each are observed at diakinesis, but the nucleolar bivalent is generally a small rod with one chiasma (Figs. 1 and 3b). The rapid terminalization of chiasmata seems to be a characteristic feature, so that at diakinesis the bivalents generally appear loose and dissociated (Figs. 1 and 3b,c). At metaphase, both ring and rod bivalents are observed. In some cultivated varieties in India, the mean chiasma frequency at metaphase was found to be 12.10 per cell and 0.86 per paired chromosome; this means that ring bivalents are preponderant (see Fig. 3d). Some populations of pearl millet show secondary associations of bivalents. Two groups of two bivalents each were clearly observed (Fig. 3c) in some cells. Although the phyletic significance of secondary associations in diploid species remains controversial, such associations cannot be entirely meaningless. In hexaploid wheat, such associations are known to take place between genetically and evolutionarily related chromosomes (Riley, 1960; Kempanna and Riley, 1964). In pearl millet, the secondarily associated bivalents look very similar to each other, although their genetic and phyletic relatedness cannot be determined. In the haploid complement when their homologous partners are missing, the chromosomes involved in these secondary associations probably form bivalents.

416

PREM P. JAUHAR

FIG. 3. Meiotic stages in pearl millet. (a) Late diplotene showing 7 bivalents (711).(b) Diakinesis with 711with teminalized chiasmata. Note 6 ring bivalents and the small, nucleolar rod bivalent. (c) Diakinesis with 711.Note the secondary associations of two pairs of bivalents; the associated bivalents look similar in size and shape. (d) Metaphase 1 with 611.[(a, d) X ca. 2050; (b, c) x ca. 21501

CYTOGENETICS OF PEARL MILLET

417

It is interesting to note that two bivalents have been reported in haploids studied by different workers (see Section V,B, Table 11; Fig. 6c). These observations lend favor to the suggestion that the complement of typhoides has been derived from a basic set of x = 5 chromosomes (see Jauhar, 1968, 1970b; Sections V,B and XI,B,l,c). B . P . purpureum SCHUMACH. ( 2 n = 28, 56)

Elephant or napier grass ( P . purpureum) is a perennial relative of pearl millet and is native to Africa. Burton (1942) and Nishiyama and Kondo (1942) determined its somatic chromosome number as 2n = 28, which is tetraploid based on x = 7. It shows diploid-like meiosis, 14,, being regularly formed at diplotene, diakinesis, and metaphase (Fig. 4a-c). No multivalents or univalents are generally formed. The occasional occurrence of a quadrivalent (Olorode, 1974) can be attributed to a floating interchange in certain populations. At diakinesis, there is a rapid terminalization of chiasmata (Figs. 2 and 4b)-a feature also characteristic of pearl millet. Two bivalents are generally associated with the nucleolus (Fig. 4a,b). One of the nucleolar bivalents is the largest in the complement, whereas the other is a small one (see also Section 11,2). Occasionally, additional nucleolar material is organized (see Fig. 2). At metaphase, there are noticeable size differences among bivalents; the smaller ones are generally rod-shaped with one chiasma each, whereas the majority of the large ones are ring-shaped with mostly two chiasmata each (Fig. 4c). Chiasma frequency per cell and per paired chromosome was found to be 18.9 and 0.68, respectively, in some collections. Several factors speak for the allotetraploid nature of elephant grass: ( I ) its 2n = 28 chromosome number; (2) the regular bivalent formation and high pollen fertility; and (3) the noticeable size differences among bivalents. This is further borne out by studies on chromosome pairing in its hybrids with pearl millet (see Section XI,B,2,d). This allotetraploid can be genomically represented as A'A'BB. The A' genome is homoeologous with the A genome of pearl millet (which is genomically AA). The large bivalents observed at metaphase in P . purpureum are evidently formed by the A'A' genome, whereas the small ones belong to the BB genome, the donor of which is not yet known.

IV. ABNORMAL MEIOSIS AND ITS GENETICS The nature of events that lead to synapsis and crossing over during the meiotic prophase remains one of the most intriguing problems in cytogenetics today. It is

418

PREM P. JAUHAR

FIG. 4. Meiotic stages in napier grass ( 2 n = 4x = 28). (a) Early diakinesis with 14 bivalents (l4,,). Note that two bivalents (one the largest in the complement and one much smaller) are associated with the nucleolus. (b) Diakinesis with 14,: Due to terminalization of chiasmata, most bivalents appear to be loose and dissociated. Note that 2,, are associated with the nucleolus; the large bivalent-largest in the complement-is lying on the nucleolus. (c) Metaphase I with 14,,; I,, is separated. [ x 14801

CYTOGENETICS OF PEARL MILLET

419

known, however, that meiosis is an integrated process consisting of a series of sequential events that are under the control of specific genes. Genotypes that deviate from the normal course of meiosis have been described in numerous organisms, including some species of Pennisetum, and their study has led to a better understanding of this complex process. Several major genes that, in homozygous recessive condition, bring about failure or disruption of chromosome pairing are now known (see Riley and Law, 1965; Jauhar and Singh, 1969a; Gottschalk and Klein, 1976; Golubovskaya, 1979). Genes that influence the very initiation and, hence, bring about complete failure of pairing at meiotic prophase are referred to as asynaptic ( a s ) , whereas those that disrupt the maintenance of pairing between initially synapsed chromosomes are termed desynaptic ( d s ) . It is, however, very difficult to distinguish between the phenomena of asynapsis and desynapsis, distinction being all the more difficult between partial asynapsis and partial desynapsis. Studies on the pairing variants have nevertheless helped in the elucidation of the genetic control of synapsis and its concomitant process of recombination. A . DESYNAPSIS A N D ITS GENETIC BASIS

Generally, desynapsis is of more common occurrence than asynapsis. Although both these phenomena have been described in pearl millet, the former is certainly much more common. Krishnaswamy et a f . (1949) reported on a desynaptic plant derived from X-ray-treated seeds. At pachytene, synapsis was noted in one or two pairs, but at diplotene-diakinesis there was almost complete disruption of pairing, resulting in 14 univalents. Irregular and random anaphasic separation of univalents resulted in 98.5% sterility. Later, Patil and Vohra (1962), Minocha et al. (1968), Jauhar (1969), Dhesi et al. (1973), and Singh et a / . (1977a) reported cases of partial desynapsis. Figure 5b, for example, shows a cell with 211 10,; two of the newly desynapsed chromosomes lie juxtaposed. More recently, Rao and Koduru (1978a) observed that failure of pairing (Fig. 5c,d) was associated with syncyte formation. The desynaptic condition was attributed to the double recessive condition of the gene ds (Minocha et al., 1975; Pantulu and Rao, 1976). In the same variety (T55) of pearl millet, Patil and Vohra (1962) and Jauhar (1969) reported different degrees of desynapsis at diplotene, diakinesis, and metaphase I. At metaphase I , for example, Jauhar (1969) observed a mean of 3.46,,, whereas Patil and Vohra (1962) had recorded only 0.3 1 in their desynaptic. Although the genetics of desynapsis was not determined, it was probably under monogenic recessive control as in other desynaptic stocks of pearl millet. Different degrees of desynapsis could be due to different environmental conditions. Like other mutant genes that are generally less buffered against environmental fluctuations (Darlington, 1958), the degree of manifestation of a desynaptic gene can be affected

+

,,

420

PREM P. JAUHAR

FIG.5. Meiotic stages i n normal, partially desynaptic, and desynaptic plants of pearl millet. (a) Diakinesis in a normal plant showing 7 bivalents (7,,). (b) Diakinesis in a partial desynaptic showing 2 + 10,. Note that two of the desynapsed homologs lie juxtaposed. (c) A PMC (at a stage comparable to diakinesis) in a desynaptic showing 14,. (d). Meta-anaphase I in the desynaptic showing 14,. Note that some univalents are arranged on the metaphase plate, while others are randomly scattered. [(a, b) x ca. 1860; (c) X ca. 2120; (d) X ca. 16201 Negatives of c and d were kindly supplied by Dr. P. R. K . Koduru.

,,

CYTOGENETICS OF PEARL MILLET

42 1

by temperature, e.g., in wheat (Li et al., 1945) or by other environmental conditions (Gottschalk, 1976). Another possibility that the degree of expression of the desynaptic gene in pearl millet is controlled by a set of modifiers (or perhaps even polygenes), which tend to make desynapsis a sort of quantitative trait, cannot be excluded (Jauhar, 1969). Rao’s (1980) work also suggests that some modifying genes may be responsible for the differential expressivity of the ds gene. With the failure of pairing, the activity of the gene responsible for normal disjunction is also probably hampered. Anaphasic separation is therefore very irregular, the univalents moving to the poles like “unguided missiles’’ (see Fig. 5d). Thus, desynapsis generally results in high to complete male sterility.

B. EFFECTOF NUTRIENTS O N DESYNAPSIS It is well known that chiasma frequency is under genetic control and that it can be influenced easily by various environmental factors like temperature. It has been claimed that different nutrients can alter chiasma frequency. Thus, Dhesi er al. (1975) reported that the application of phosphate and potash resulted in an increase in chiasma frequency of pearl millet desynaptics. In other words, these treatments reduced the strength of desynapsis. Phosphate treatments were reported to increase chiasma frequency in desynaptic strains of rye (Bennett and Rees, 1970) and barley (Fedak, 1973). In desynaptic barley, Fedak (1973) observed “a strong positive relationship between phosphate treatment and chiasma frequency. ” It is difficult to explain how nutritional status of the soil significantly alters chiasma frequency of normal or desynaptic plants. If a higher mineral status of soil does increase chiasma frequency in desynaptic plants, mineral starvation should accentuate the effect of the desynaptic genes. Lakshmi er al. (1979) recently estimated the phosphate and potassium contents in the flag leaves of normal pearl millet plants, and in desynaptics with 2-6 univalents, 2-10 univalents, and 10-14 univalents per cell, but they did not detect any significant differences in mean values. More studies are needed therefore before any definite conclusions can be derived regarding the effect of mineral nutrition on chiasma frequency. C. EFFECTO F PLOIDY O N DESYNAPSIS

A desynaptic tetraploid was observed by Rao (1978) in the open-pollinated progeny of some desynaptic diploids. It showed a mean chromosome configura-

422

PREM P. JAUHAR

+

tion 1.02,, + 1.30,,, + 7.37,1 5.28, per pollen mother cell (PMC), suggesting an improvement in pairing apparently resulting from chromosome doubling. Its mean chiasma frequency was 15.12, whereas the parental desynaptic diploids had mean frequencies of 0.2-8.4. Thus, four doses of the desynaptic ( d s ) gene seemed to increase chiasma frequency in the desynaptic tetraploid over that in the desynaptic diploid, which had two doses of the ds gene. Since desynapsis in pearl millet is due to the double recessive (dsds) condition of the ds gene, the desynaptic tetraploid should have the nulliplex condition (dsdsdsds) for this gene in order for this character to be expressed. D . EFFECTO F DESYNAPTIC GENEO N B CHROMOSOMES

Pantulu and Rao (1976) found that in a desynaptic stock with B chromosomes, the desynaptic gene brought about a reduction in synapsis and chiasma frequency of the A as well as B chromosomes. This is the only report of the effect of a ds gene on B chromosomes. However, it has been observed that the diploidizing genes controlling diploid-like pairing of A chromosomes in hexaploid tall fescue (Jauhar, 1975a,c) seem to regulate pairing of B chromosomes also to bivalent level even when up to 10 B chromosomes are present (Jauhar, 1980). That a particular gene(s) should coordinately control the behavior of A and B chromosomes is quite understandable. E. EXPERIMENTAL INDUCTION OF DESYNAPSIS

Since desynapsis and asynapsis are mostly under genetic control, they can be artificially induced like any other mutation. The first case of desynapsis in pearl millet recorded by Krishnaswamy et a / . (1949) was probably a mutant induced as a result of X irradiation. Later, some desynaptic mutants of pearl millet were produced by both physical and chemical mutagens (Singh e t a / . , 1977a; Lakshmi and Yacob, 1978). Colchicine, which is primarily an effective polyploidizing agent, has been reported to induce a desynaptic mutant in pearl millet (Rao, 1980). It is now becoming evident that the whole meiotic sequence from premeiotic mitosis to spore formation is controlled by different genes that, acting in a coherent fashion, produce normal meiosis and, hence, male and female fertility. Mutation of any of these genetic determinants can disrupt, distort, or even block a particular stage and with it the subsequent stages in this integrated and ordered sequence of events.

CYTOGENETICS OF PEARL MILLET

423

F. INBREEDINGA N D DISRUPTION O F CHROMOSOME PAIRING: HETEROZYGOSIS A N D HETEROSIS FOR CHROMOSOME BEHAVIOR

Pearl millet is an allogamous crop. When subjected to enforced selfing, it shows inbreeding depression, which is associated with several meiotic abnormalities, like reduction in chiasma frequency and desynapsis (Pantulu and Manga, 1972). In an inbred line that was maintained by selfing, some desynaptics were obtained (Koduru and Rao, 1978). In another inbred line, partial “asynapsis” was accompanied by chromosome breakages that seemed to be localized near the centromeric regions (Rao and Koduru, 1978b). Hybrids between inbred lines have been reported to show heterosis for chiasma frequency and meiotic behavior. Thus, Mahadevappa and Ponnaiya (1 967) observed that a hybrid had higher chiasma frequency than that of the parental inbreds. Pantulu and Manga (1972) found that all the F, hybrids obtained by intercrossing inbred lines had mean chiasma frequencies higher than those of the parents involved. When tested under stress and nonstress conditions, the heterozygotes showed consistently higher chiasma frequencies and less variation in their means compared to the parental inbreds (Manga and Pantulu, 1974). Harinarayana and Murty (1971), however, reported that pearl millet inbreds had higher chiasma frequencies than did the parental outbred populations. This increase in chiasma frequency under enforced selfing was attributed to a “buffering mechanism of chiasma frequency under inbreeding, referred to as cytological homeostasis. ” Other studies have not confirmed the presence of the so-called cytological homeostasis in pearl millet. Recently, Srivastava and Balyan (1977) also reported heterosis for chiasma frequency and chromosome behavior. The pearl millet hybrids had 15-47% higher chiasma frequencies and also showed fewer meiotic abnormalities than the parental inbreds. Some of the deleterious genes, which otherwise remain masked in the heterozygous condition, are uncovered by inbreeding, and they adversely affect meiotic pairing and chiasma frequency in pearl millet inbreds. A similar decline in chiasma frequency also results from enforced inbreeding of other outbreeders, e.g., rye (Secale cereale), but on intercrossing of inbreds a positive heterosis for chiasma frequency is obtained (see Rees, 1961a). The results obtained by Srivastava and Balyan (1977) suggested a positive association of increased chiasma frequency with desirable quantitative characters like grain yields. They noted that the highest-yielding pearl millet hybrid, NHB-5, also had the highest chiasma frequency. Since the hybrids show heterosis for chiasma frequency, and heterosis for yields also closely parallels the heterozygosis of the breeding material (Burton, 1968), it is not surprising that some hybrids with high chiasma frequencies should also have high grain yields. Further studies along these lines may prove rewarding.

424

PREM P. JAUHAR

V. HAPLOIDY Haploid sporophytes, with half the chromosome complement, are useful cytogenetic tools. They offer an opportunity for the study of chromosome relationships within species because the internal homologies, which generally remain masked in the diploid, can be revealed in the complement in haploid state. Haploids can be usefully employed in several other areas of fundamental research, such as studies on induced mutagenesis, gene dosage effects, interaction of genes, and linkage analyses. Haploids can also help accelerate a breeding program. Haploids of polyploids are more appropriately termed polyhaploids, whereas haploids of diploid species are referred to as rnonoploids, rnonohaploids, or simply haploids. However, I feel that the term “monoploid” should be used for the haploids of true diploids with no history of polyploidy, whereas haploids of such doubtful cases as corn (see Section VI1,B) should, preferably, not be called monoploids, but simply haploids. A. SPONTANEOUS A N D INDUCED HAPLOIDS

Haploids of pearl millet (2n = x = 7) have been reported to occur spontaneously (Pantulu and Manga, 1969; Jauhar, 1970b; Powell et al., 1975). They have also been isolated in the progeny of trisomics, as well as induced artificially by treating pearl millet seeds with methyl methane sulfonate (Gill et al., 1973). B. CHROMOSOME PAIRINGI N HAPLOIDS

In the pearl millet haploids studied by different workers, mostly univalents were observed at stages comparable to diplotene (Fig. 6a), diakinesis (Fig. 6b), and metaphase I (Fig. 6c), and the subsequent meiotic stages were highly disorganized. Anaphase I disjunction led to 4 : 3, 5 : 2, 6 : I , or 7 : 0 distribution of chromosomes to the poles (Fig. 6d-g). Thus, as a result of 7 : 0 distribution, some balanced male (and also perhaps female) gametes were formed that fused to produce some viable seeds on the haploid earhead studied by Powell et al. (1975). This provided direct evidence of the formation of unreduced gametes in haploid pearl millet. The frequencies of bivalents observed at diakinesis and metaphase were very low (Table 11). A significant feature, however, was the occasional formation of some chiasmate bivalents. Jauhar (1970b) observed both ring and rod bivalents,

FIG.6 . Meiotic stages in haploid pearl millet (2n = x = 7). (a) Late diplotene with 7 univalents (71). (b) Diakinesis with 111 + 51. Note that the chiasma in the bivalent is completely terminalized and two univalents are showing end-to-side association. (c) Metaphase I with 2" 3,. (d-g) Anaphase I stages showing 4 : 3,5 : 2 , 6 : l , and 7 : 0 distributions to poles. [(a) X ca. 1970; (b) x ca. 1836; (c-g) x ca. 14221. Negatives of a, c-g were kindly supplied by Dr. J. B. Powell.

+

426

PREM P. JAUHAR

Table I1 Range and Mean of Bivalents Observed in Different Haploids of Pearl Millet References 0-2 0-2 0-2 0-2 0-2 0-2 0-2

0.57 0.40 0.20 1.13 0.46 0.58 0.87

Pantulu and Manga ( 1 969) Manga and Pantulu (1971) Jauhar (1970b) Gill et a / . (1973) Powell el a / . (1975) Powell et a / . ( I 975) Powell el a / . (1975)

but the chiasma terminalization was very rapid, which is a characteristic feature of diploid pearl millet (see Section III,A and Figs. 1 and 3b,c). Aside from chiasmate pairing, secondary associations of univalents were also observed. More significant among these were the s-s (side-to-side) associations, which are generally considered to result from residual homology (Person, 1955; Riley and Chapman, 1957). The realization of a maximum of two bivalents per cell, as also a maximum of two s-s pairs, was attributed to homologies between four members of the complement that may have resulted from duplication (Jauhar, 1970b). It was inferred therefore that the complement (x = 7) of pearl millet was derived from a basic set of x = 5 chromosomes. Intrahaploid pairing in the typhoides complement in P . typhoides x P . purpureutn and P . ryphoides x P . orientale hybrids lends support to this view (see Section XI,B.) If the x = 7 complement of pearl millet has indeed arisen from a basic set of x = 5 chromosomes, the duplicated chromosomes must have undergone sufficient differentiation during the course of evolution so that they pair as bivalents in the diploid pearl millet, although they occasionally show conspicuous secondary associations (see Section III,A and Fig. 3c). Gill et al. (1973) studied different haploids of pearl millet and reached the same conclusions as Jauhar (1970b). Manga and Pantulu (1971) also made essentially similar observations, but they did not support the view that the complement of pearl millet has been derived from a lower basic number. Haploid meiosis may not provide unequivocal evidence in regard to the phyletically basic chromosome number of a species. It is, nevertheless, interesting to note that different workers, studying different haploids of pearl millet from diverse sources, observed a maximum of two bivalents per cell (Table 11; Fig. 6c). The mean bivalent frequency, however, varied from 0.20 to 1.13. Haploids from different genetic backgrounds are expected to show such a variation in chromosome pairing.

CYTOGENETICS OF PEARL MILLET

427

It is noteworthy that apparently normal synaptinemal complexes (SCs) have been observed in haploids of maize (Ting, 1971, 1973) and barley (Gillies, 1974). Maize does not seem to be a true diploid (see Section VII,B) and some pairing accompanied by the formation of SCs is not unexpected. Also, extensive pachytene pairing in haploid barley (Sadasivaiah and Kasha, 1971) coupled with the formation of SCs-which are structurally similar to those in diploid barley (Gillies, 1974)-may not be without phylogenetic significance. However, it is not known whether pairing in pearl millet haploids is accompanied by the formation of SCs. Such studies should of course yield valuable information. C . HAPLOIDS I N PLANTBREEDING

If haploids of an open-pollinated crop like pearl millet could be produced on a large scale, they would help produce homozygous lines in one step and, thus, accelerate a breeding program. Such a haploidy method of breeding diploid species has great potential. It involves the following main steps: (1) production of haploids in the desired material of diverse origin; (2) selection, if any, at the haploid level; (3) induction of chromosome doubling in selected haploids to produce homozygous, diploid lines; and (4) the usual testing and selection in the homozygous lines before using them in heterosis breeding or production of synthetics. It is interesting to note that doubled haploids of maize have been utilized in the production of commercial hybrids (Chase, 1974). However, the frequency of the spontaneous occurrence of haploids in nature is very low. For example, the frequency of one haploid per 10,000 plants found in the pearl millet inbred Tift 23A (Powell et al., 1975) is far too low to be of any use in a breeding program. Therefore various methods of producing haploids by anther culture, including the anther-panicle culture technique of Kasperbauer et al. (1980), should be tried. It is noteworthy that ability to produce haploids is genetically controlled. The production of haploids in corn, for example, is under genetic control, different strains apparently having distinct rates of haploid frequency (see Carlson, 1977). Coe (1959) identified a high haploidy line, Stock 6, which on self-pollination produces more than 3% haploids. More recently, the high haploidy-inducing potential of some inducer lines was reported to be a highly heritable trait not appreciably influenced by environment (Aman and Sarkar, 1978). This trait may therefore be transferred rather easily to other desirable lines. If similar genetic mechanisms of inducing haploidy are discovered in pearl millet, its breeding program could be accelerated, at least in some cases. In this context, it is interesting to note that haploid-derived hornozygous maize lines are being used profitably in breeding programs in the Soviet Union, where some superior single hybrids have been developed from these lines (Golovin, 1979).

428

PREM P. JAUHAR

VI. POLYPLOIDY The fact that a large proportion of the existing plant species-including many of our crop plants-are polyploids clearly shows the importance of polyploidy in evolution. Since plant breeding is, in essence, man-made evolution, induced polyploidy has been used as a tool in improving diverse types of plant species. Most of the artificially induced polyploids of grain crops have not proved to be of direct practical value, although a small number of species have responded favorably to chromosome doubling (Stebbins, 1956a). Thus, autotetraploid rye, further improved upon through hybridization and selection among the tetraploid strains, has shown some promise (Muntzing, 1951, 1954). Like rye, pearl millet has certain features considered desirable for a favorable response to induced polyploidy. It is a diploid with low chromosome number (2n = 14) and is typically outbreeding; hence, it offers promise for polyploidy breeding. A. SPONTANEOUS OCCURRENCE

Krishnaswamy and Ayyangar (1941b) reported a spontaneous autotriploid (2n = 3x = 21) in the selfed progeny of a partially sterile (and presumably partially desynaptic) diploid plant (2n = 2x = 14) of pearl millet. Since a partially desynaptic plant can produce a small proportion of both normally reduced and unreduced gametes, the triploid must have resulted from the fusion of an unreduced gamete with a normal gamete. During meiosis the triploid formed an average of 4.95111per cell, 6% of the PMCs having 7111.Thus, the triploid showed meiotic pairing typical of an autotriploid. Burton and Powell (1968) also observed spontaneously occurring autotriploids that were highly sterile, but were not studied meiotically . Recently, Koduru and Rao (1978) reported spontaneous triploid and tetraploid plants in the progeny of an inbred line. While the triploid showed chromosome pairing typical of an autotriploid, forming up to 7111in 2% of the cells with a mean frequency of 4.37111per cell, the tetraploids fell into two categories. Four of the tetraploids behaved more or less like autotetraploids, but one nonhairy tetraploid was reported to show reduced quadrivalent frequency and higher univalent frequency. The low chromosome pairing in the nonhairy tetraploid was attributed to partial desynapsis. However, differences in quadrivalent frequency can also be caused by different genotypes. In the selfed progeny of a doubletelotrisomic (2n = 13 2 telocentrics) that was also a translocation heterozygote, Pantulu and Rao (1977b) obtained a normal triploid (2n = 3x = 21) that showed meiotic pairing typical of an autotriploid.

+

CYTOGENETICS OF PEARL MILLET

429

Powell and Burton (1968) observed tetraploids among twin and triplet seedlings arising from polyembryonic caryopses of pearl millet. A later study by Hanna et al. (1976) showed that autotetraploids occurred at frequencies as high as 1 per 13 plants in polyembryonic and I per 292 plants in nonpolyembryonic material. At metaphase, the mean chromosome configurations per cell were 1.491v 0.38,,, + 8.97,, + 2.64,, which are far less than those in thecolchicineinduced tetraploids (see Section VI,C,2).

+

B . FACTORSFAVORING SPONTANEOUS

POLYPLOIDY

Spontaneous polyploidy appears to be common in the higher plants. The grass family is particularly characterized by the preponderance of polyploid taxa. Of the different Pennisetum species whose chromosome numbers are known, including the intraspecific cytotypes, nearly 76% are polyploids. There is no reason to believe that the proportion would be much different in the rest of the family. Spontaneous chromosome doubling seems to be common in grass species as well as hybrids. The functioning of unreduced gametes of either one or, rarely, both parents of interspecific, intergeneric, and even intervarietal hybrids has contributed significantly towards the creation of spontaneous polyploids in nature (see Harlan and de Wet, 1975; de Wet, 1980). In a synthetic interspecific hybrid (2n = 3x = 21) between P . typhoides and P . purpureum, Jauhar and Singh (1969b) recorded a case of spontaneous chromosome doubling resulting in amphidiploidy apparently induced by decapitation or severe pruning. This was a case of somatic doubling. Meiotic nonreduction, however, is more common in hybrids. Stebbins ( 1956b) has listed different factors responsible for the high frequency of amphiploidy in grasses: (1) the occurrence of several species together gives ample opportunities for hybridization; (2) the production of abundant wind-borne pollen and the existence of self-incompatibility systems promote crosspollination; and (3) most of the species and their hybrids are perennials and have efficient means of vegetative propagation. It may be pointed out that in the genus Pennisetum, polyploids occur only in perennial, vegetatively propagated species of Pennisetum. All these factors of course help in the production, survival, and successful establishment of new hybrid combinations, many of which colonize new ecological niches. In addition to these factors, I feel that through natural decapitation (for example, severe grazing by cattle or wild animals), amphiploids of natural interspecific hybrids may be produced, thus further contributing towards the abundance of allopolyploidy in grasses.

430

PREM P. JAUHAR C. INDUCTION OF POLYPLOIDY

1 . Triploidy

Autotriploids (2n = 3x = 21) of pearl millet have been synthesized by crossing colchicine-induced autotetraploids with diploids (Gill et al., 1969, 1970a; Jauhar, 1970a). In addition to triploids, some hypotriploids (2n = 20, 19) are also obtained by this method (see, for example, Fig. 7). Triploids have also been produced through gamma irradiation (Pantulu, 1968) as well as by combined treatments with gamma rays and ethyl methane sulfonate (Singh et al., 1977b). At diakinesis in the spontaneous autotriploids, mean trivalent frequencies of 4.95 (Krishnaswamy and Ayyangar, 1941b) and 4.37 (Koduru and Rao, 1978) were observed; however, most of the synthetic triploids had ironically lower trivalent frequencies: 2.62 at diakinesis and metaphase (Gill et ul., 1970a), 2.76 at metaphase (Jauhar, 1970a), 3.57 at diakinesis and 3.41 at metaphase (Singh et al., 1977b). Such a variation in trivalent frequencies is not unexpected in synthetic triploids that have been produced by different means and have different genetic backgrounds. While Jauhar (1970a) used nearly raw autotetraploids for making triploids, Gill et al. (1969) used relatively diploidized material. An unusually high trivalent frequency was recorded in a radiation-induced

FIG. 7. Metaphase I in a hypotriploid cell (2ri = 20) showing 6 trivalents and I bivalent. Note chain-, V-, and frying pan-shaped trivalents. [ x 18201

CYTOGENETICS OF PEARL MILLET

43 1

triploid, in which more than 50% of the PMCs were reported to have 7111 each, with a mean of 6.35111and 6.1811, per cell at diakinesis and metaphase, respectively (Pantulu, 1968). The autotriploids of pearl millet have proved very useful in the synthesis of a series of aneuploids (see Section VI1,B). 2 . Tetraploidy The first experimental induction of tetraploidy in pearl millet was accomplished by Krishnaswamy et al. (1950) by injecting aqueous solution of colchicine into shoots of young seedlings. Studies on the C, generation of this allotetraploid showed up to 71vin about 0.8% of the cells, the mean per cell for 15 plants being 2.61v (Raman et al., 1962). Later, several workers produced synthetic tetraploids by colchicine treatment (Gill et al., 1966, 1969, 1970a; Jauhar, 1970a; Minocha et al., 1972) and by X irradiation (Singh et al., 1977b). Varying numbers of multivalents were reported in the C, generation colchi-tetraploids produced in different materials, the means being 2.6,, (Raman el al., 1962), 3.011~+111 (Gill et al., 1969), and 4.311v+111 (Jauhar, 1970a). Raman et al. (1962) also observed some trivalents, but they did not report the mean trivalent frequency. Different multivalent frequencies in these synthetic tetraploids may be attributed to their different genotypic backgrounds. Minocha et al. (1972) studied meiotic pairing in the autotetraploids of the open-pollinated variety, T55, and its inbred, BIL-4. Whereas the autotetraploids of T55 showed a mean frequency of 2.541~+111 per cell, the autotetraploids derived from the inbred BIL-4 had a mean of 3.011~+111 per cell. This indicates that the degree of heterozygosity of parental material can affect the multivalent frequency of the derived autotetraploids, with homozygosity favoring higher multivalent frequency.

3 . Reversion to Diploidy The colchi-tetraploids showed considerable meiotic instability as evidenced by the recovery of diploid revertants in their progeny. In the C, population raised by selfing the C, plants, Raman et al. (1962) observed about 18% diploids, which they attributed to haploid parthenogenesis in the parental tetraploids. A similar reversion from tetraploidy to diploidy was observed by Gill et al. (1 969). In other synthetic tetraploids (maintained in the first two generations by sib-mating and later by open pollination among tetraploids) diploid revertants were recovered even up to the C6 generation (Jauhar, 1970a; Jauhar et al., 1976). Since the diploid revertants were rnatromorphic, they may have arisen from the unfertilized (and perhaps pseudogamous) egg through parthenogenesis. The

432

PREM P. JAUHAR

intra- andor extracellular factors contributing to stability or instability of synthetic polyploids are not well understood today. The occurrence of hypotetraploid aneuploids in the progeny of pearl millet tetraploids showed that imbalanced gametes were functional.

4 . Cytological Diploidization Gill et a / . (1969) and Jauhar (1970a) studied chromosome pairing in the raw (C,) and advanced generation tetraploids and noted a marked cytological diploidization in successive generations, i.e., a gradual shift from multivalent to bivalent type of association. This was probably due to natural selection of genes that condition regular meiosis with bivalent formation (Jauhar, 1970a). Equally interesting was the fact that an increase in bivalents was not associated with an increase in univalent frequency; hence, the fertility of the tetraploids improved in later generations. However, even in the evolved tetraploids with 1.841v+1,1 per cell, the seed set was much lower than in the parental material. It was inferred that meiotic abnormalities coupled with genetic factors were jointly responsible for sterility of the advanced generation tetraploids (Jauhar, 1970a) (see also Section VI,D). Synthetic pearl millet tetraploids had bolder grains compared to the parental diploids. However, for them to be of any practical use, their fertility must be improved by a program of intercrossing among the tetraploid strains and selecting for high seed fertility. Providing this kind of evolution would probably improve their meiotic behavior by bringing about cytological diploidization. D. POLYPLOIDY A N D PLANTBREEDING

Many of our crop plants, e.g., wheat, oats, cotton, tobacco, potato, peanut, alfalfa, tall fescue, and napier grass owe their success to polyploidy. Their diploid progenitors are either extinct or, if living, cannot compete with them. It was envisaged therefore that the existing diploid species could be improved by inducing polyploidy . Although initial expectations of the success of polyploidy as a tool in plant improvement have not been realized, it has contributed somewhat to the improvement of ornamental plants and forage crops. Other areas of its usefulness include the following: (1) bridging the ploidy gap between taxa to permit their hybridization; (2) overcoming interspecific sterility barriers and thereby facilitating gene transfers; and (3) applications in special breeding techniques like the haploidy method of breeding diploids. The genomic makeup of an organism is delicately balanced. Artificial chromosome doubling upsets this balance and produces sterility. At the chromosomal

CYTOGENETICS OF PEARL MILLET

433

level, the main problem is irregular meiotic behavior that results in imbalanced gametes and sterility. However, genetic factors can also contribute to sterility. Doggett (1 964) found that fertility of autotetraploid sorghums was genotypically controlled with apparently “fairly simple” inheritance. Moreover, high fertility proved to be dominant to low fertility and, hence, was rather easily transferable from one taxon to another. Doggett suggested therefore that there was no “barrier to the successful development of cultivated tetraploid sorghum as a grain crop. ’ ’ Since trivalents and univalents are the main causes of aneuploidy, a reduction in their frequency should improve fertility. In rye and Lolium tetraploids, Hazarika and Rees (1 967) and Crowley and Rees (1968) found that an increase in chiasma frequency was accompanied by a reduction in univalents and trivalents, an increase in quadrivalents, and an improvement in fertility. These workers (Hazarika and Rees, 1967; Rees and Jones, 1977, p. 63) have suggested, therefore, that improvement in fertility can be achieved by increasing the frequency of quadrivalents, although they feel that increase in bivalent frequency (cytological diploidization) would be an “equally acceptable’ ’ alternative. The question now arises as to which of the two alternatives provides a more realistic approach to improving fertility of synthetic tetraploids. In this connection, it is important to note that quadrivalent formation cannot be achieved with the same efficiency as bivalent formation, because the former can be easily affected by the availability and redistribution of chiasmata. Chiasma frequency, in turn, is easily influenced by environmental stress. In the event of a slight reduction in chiasmata, some quadrivalents (especially the chain-type) could easily be converted into trivalents plus univalents. It has indeed been observed that the frequency of trivalents and univalents increases with decrease in chiasma frequency (see Rees and Jones, 1977). Moreover, chiasmata can be redistributed in favor of bivalents (Hossain, 1978). Thus, complete quadrivalent formation cannot be achieved. In that case, a combination of quadrivalents and bivalents would be a good compromise. Even so, the regular disjunction of all quadrivalents cannot be precisely achieved. On the other hand, complete bivalent formation ensures regular disjunction and balanced gamete formation. There are several reports that suggest that bivalents, but not quadrivalents, are associated with increased fertility (Miintzing, 1954; Hilpert, 1957; see McCollum, 1958; Aastveit, 1968; Gill er al., 1969; Jauhar 1970a). All these factors weaken the argument of Rees and Jones (1977) that increasing quadrivalent frequency by increasing chiasma frequency is an effective means of improving fertility of autotetraploids. It may be concluded therefore that cytological diploidization offers the best means of improving fertility of synthetic tetraploids. After all, this is the route nature has taken by genetically diploidizing polyploid crop plants such as wheat and oats.

PREM P. JAUHAR

434

VII. ANEUPLOIDS The aneuploids of a crop plant are helpful in assigning genes to particular chromosomes and for establishing linkage groups. By virtue of its low chromosome number, pearl millet offers a particularly favorable material for aneuploid analyses. Some progress has been made in building up a trisomic series in pearl millet, but the establishment of linkage groups still awaits completion. In describing different aneuploids in this article, the terminology used by Khush (1973, pp. 3-5) is adopted with minor modifications. An individual with an extra chromosome is referred to as a trisomic (2n + l), whereas the extra chromosome itself is designated as a trisome. Similarly, whereas an individual deficient for a chromosome is termed a monosomic (2n - I ) , the chromosome involved is referred to as a monosome. A telocentric chromosome will be referred to as a telosome or simply as a telo and will be symbolized by the letter t. A few monorelodisomic (2n - 1 + I t ) plants which are deficient for one chromosome arm are also reported in pearl millet; thus, they have 2n = 13 It. Some new terms have been coined, e.g., doubletelotetrasomic (2n = 14 2t), tripletelotetrasomic (2n = 13 + 3t), and quadruplerelotetrasomic ( 2 n = 12 4t) to describe adequately the newly isolated aneuploids of pearl millet (see Section VI1,D).

+ +

+

A. SFONTANEOUS OCCURRENCE

In pearl millet some spontaneous aneuploids, mostly trisomics, have been reported. Li and Li (1943) described four dwarf, weak plants that presumably arose in the progeny of a triploid. Two of the plants were primary trisomics (2n 1 = 15) and formed one trivalent each. The third plant was a double trisomic (2n + 1 + 1 = 16) involving one large and one small chromosomes, which formed two trivalents. This plant was very weak with no tillers and had indehiscent anthers. The fourth plant was extremely weak and sterile and had three extra chromosomes (2n 3 = 17), two medium-sized and one large. From the figures given by Li and Li (1943, p. 141), it appears that this plant formed a chain quadrivalent and a chain trivalent, indicating that it was probably tetrasomic for a medium-sized chromosome and trisomic for a large chromosome. This plant can, therefore, be represented as (2n + 2 + 1 = 17). The fact that the 17-chromosome plant was extremely weak and completely sterile shows that there is a limit to the tolerance of extra chromosomes in pearl millet, as in most other diploids. However, corn is an exception in this respect and has a tremendous tolerance for extra chromosomes. From a triploid x

+

+

435

CYTOGENETICS OF PEARL MILLET

diploid cross, Punyasingh (1947) obtained plants having up to eight extra chromosomes (2n 8 = 28). Of the trisomics of pearl millet observed by other workers, one was from the progeny of a desynaptic plant (Krishnaswamy et al., 1949), another was found in the progeny of a spontaneous triploid (Koduru and Rao, 1978), whereas yet another occurred spontaneously (Burton and Powell, 1968). Meiotic studies in these trisomics were, however, not reported.

+

B. SYNTHESIS OF ANEUPLOIDS OF PEARL MILLET

Of all the aneuploids, the trisomic condition for different chromosomes is best tolerated in pearl millet. Moreover, trisomics, particularly primary trisomics, are the easiest to obtain. However, there is one isolated report of the occurrence of a monosomic (2n - 1 = 13) in the progeny of triploid X diploid crosses (Jauhar, 1970a). The monosomic that survived until maturity was a diminutive plant with thin, grassy leaves and a highly sterile, globose head. It generally formed 6,, lI; rarely, 511 3, were observed. Detailed studies on the meiotic behavior and transmission of monosome could not be done. It is nevertheless interesting that, unlike most diploids, a monosomic condition in pearl millet is tolerated so that it grows until maturity. Some monotelodisomic plants with 2n = 13 1 telocentric, i.e., monosomic for one chromosome arm, have also been produced (Pantulu et al., 1976) (see Section VII,C,S). Recently, Koduru et al. (1980) have reported a case of interchange monosomy in pearl millet. In corn (2n = 20), however, 9 of the possible 10 monosomics (2n - 1 = 19) have been produced along with some occasional double monosomics (2n - 1 1 = 18) and even triple monosomics (2n - 1 - 1 - 1 = 17) (Weber, 1970, 1973). Corn is again an exception in this regard, because no other diploid species is known to withstand a triply rnonosomic condition. In theory, of course, haploids of diploid species are monosomic for n chromosomes. Haploids of corn and pearl millet, for example, are essentially monosomics for 10 and 7 chromosomes, respectively. But in haploids the complement is in a balanced state, even though in half a dose, and is better tolerated. Such an ability to tolerate extensive whole chromosome deficiencies as well as whole extra chromosomes would cast doubt on corn being a truly diploid species. I would imagine that corn is an ancient, secondarily balanced species with an extensive duplication (and probably redundance) of genetic information in the form of whole chromosomes as well as large segments of chromosomes. There is some evidence in favor of pearl millet also being a secondarily balanced species (Jauhar, 1968, 1970b; Minocha and Singh, 1971a,b; Gill et al., 1973). (See also Sections III,A, V,B, and XI,B,l,c.)

+

+

+

PREM P. JAUHAR

436

C. TRISOMICS

+

Since trisomics (2n 1 = IS) carry an extra chromosome, the phenotypic ratios for the genes located on this chromosome are modified in segregating progenies, thus helping in assigning genes to particular chromosomes. Khush (1973) has listed numerous types of trisomics reported in various plant groups. Several types of trisomics are reported in pearl millet.

I . Primary Trisomics A trisomic in which the extra chromosome is one of the normal chromosomes of the complement is referred to as primary single trisomic, or primary trisomic, or simply trisomic (2n + 1); such a trisomic may form one trivalent per cell. Similarly, a plant having an extra dose each of two different chromosomes of the complement can be referred to as primary double trisomic, or simply double trisomic (2n + 1 + l), and may form a maximum of two trivalents per cell. Likewise, a plant with one extra dose of three different chromosomes of the complement, i.e., having three different chromosomes in triple dose, can be referred to as primary triple trisomic or triple trisomic (2n + 1 + 1 + 1). Such a trisomic should ordinarily form a maximum of three trivalents per cell. a . Single Trisomics. Primary single trisomics constitute a particularly valuable tool for locating genes on different chromosomes and for establishing linkage groups. Jauhar (1970a) isolated two trisomics in the progeny of triploid X diploid crosses. These trisomics showed 611 + I,,, (Fig. 8a,b) or 711 1, (Fig. 8c). A systematic attempt at synthesizing a series of trisomics in pearl millet was, however, made by Gill et al. (1970b,c), Gill and Minocha (1971), Virmani and Gill (1971), and Minocha and Gill (1974), who isolated a set of primary trisomics from the progenies of triploid plants as well as from triploid x diploid crosses. Manga (1976) isolated seven trisomics from the selfed progeny of a triploid. Later, some primary trisomics were obtained from the progeny of a desynaptic (Tyagi, 1976c) and from the selfed progeny of primary interchange heterozygotes (Tyagi, 1977). i . Assigning morphological types to trisomes. The primary trisomics isolated by Gill et al. (1970~)were distinguishable from related diploids by their relatively poor vigor, shorter height, narrower leaves, and later flowering. They were also distinguishable from one another and were classified into seven morphological types designated as Tiny, Dark Green, Lax, Slender, Spindle, Broad, and Pseudonormal. The extra chromosome involved in the trisomics was identified by Gill et al. (1970~)at somatic metaphase on the basis of total length, arm ratios, and presence or absence of satellite. Thus, they found that Tiny was trisomic for chromo-

+

437

CYTOGENETICS OF PEARL MILLET

FIG. 8. Chromosome pairing in some primary trisornics ( 2 n + 1 = 15) of pearl millet. (a) Metaphase 1 with I,,, 6,,;the trivalent i s marked with arrow. (b) Metaphase 1 with l,,, 6,,;note the frying-pan trivalent marked with an arrow. ( c ) Metaphase I showing 7,, I,; the univalent is marked with an arrow. [ X 2 3 9 0 ]

+

+

+

438

PREM P. JAUHAR

some 1, Dark Green for chromosome 2 , Lax for chromosome 3, Slender for chromosome 4, Spindle for chromosome 5 , Broad for chromosome 6; the chromosome involved in Pseudonormal was identified at diakinesis as (nucleolar) chromosome 7 (Gill et al., 1970c; Gill and Minocha, 1971). Manga (1976) also described the morphology of seven primary trisomics developed by her and designated them as Dwarf, Bush, Slender, Semidwarf, Purple, Robust, and Pseudonormal. Thus, some of the primary trisomics described by her differ in their morphological features from those reported by Gill et a l . (1970~). As pointed out in Section II,l , it is difficult to identify some chromosomes of pearl millet at somatic metaphase. Extreme caution must be exercised therefore before assigning morphological types to trisomes. The information on somatic karyotype should be supplemented by karyomorphological studies at pachytene. Differential Giemsa banding patterns, when established (see Section 11, I ) , may prove extremely helpful in the identification of individual trisomes. ii. Translocatioii testers and identification of trisomes. Tyagi ( 1 9 7 6 ~1977) employed translocation-tester stocks for the identification of extra chromosomes involved in primary trisomics of pearl millet. This identification is based on the assumption that, in a trisomic F, hybrid between a primary trisomic type and a translocation tester, a configuration of l v + 511 indicates that the extra chromosome in the trisomic is one of those involved in the translocation, whereas the formation of lIv + 1111+ 411 implies its noninvolvement. iii. Trivalent frequency. Virmani and Gill (1 97 1) found a positive correlation between trivalent frequency and the length of the extra chromosome in different pearl millet trisomics. Thus, the trisomic for chromosome 1 showed 84% of the PMCs with a trivalent, whereas in the trisomic for chromosome 7 only 24% of the PMCs had a trivalent. Manga (1976), on the other hand, did not observe any correlation between chromosome length and trivalent frequency, except in the trisomic for chromosome 7. It is somewhat surprising that, barring the trisomic for chromosome 7, she observed higher trivalent frequencies in the trisomics involving shorter chromosomes than in the trisomic for the longest chromosome. Since one of the important factors governing chiasma formation is the length of the chromosome (or the length of a particular arm) in which chiasma formation takes place, the trisomics for the longer chromosomes generally form trivalents at a higher frequency, for example in maize (Einset, 1943) and tomato (Rick and Barton, 1954). In the trisomics for the shorter chromosomes, the opportunity for chiasma formation to bind the extra member is lower probably because of spatial limitation and, hence, the trivalent frequency should be lower. In any case, trivalent frequency is a variable feature easily influenced by genetic and environmental factors, and must never be used for identification of trisomics for different chromosomes as has been done by Virmani and Gill (1971) and Gill and Minocha (1971). Trivalent frequency probably would also be affected by chiasma frequency.

439

CYTOGENETICS OF PEARL MILLET

Manga (1976) reported that the trivalent frequency and chiasma frequency were “correlated” in the trisomics studied by her. She recorded a higher chiasma frequency in the trisomic for chromosome 7 than in the trisomic for chromosome 1; the latter also had the lowest frequency of all the trisomics as well as the diploid sib. The differences in chiasma frequencies among trisomics, and between the trisomics and their diploid sibs can probably be attributed to genotypic differences. A primary trisomic of tall fescue was also found to have significantly higher chiasma frequency per cell and per paired chromosome than its sister disomics (Jauhar, 1978). Minocha and Gill (1974) and Minocha et al. (1976) studied the transmission frequency of the extra chromosomes involved in the seven primary trisomics and found that it was significantly higher through the egg than through the pollen. In the selfed progeny of trisomics, only 3.4% were trisomics and such a low recovery was a major impediment in the maintenance of trisomics. b. Double and Triple Trisomics. Some double and triple trisomics have been isolated in pearl millet (Gill et al., 1970b). In the selfed progenies of triploids and from triploids x diploid crosses, a total of 169 aneuploids were isolated of which 96% were primary trisomics (2n + 1 = 15), while the remainder consisted of double trisomics (2n 1 + 1 = 16), triple trisomics (2n 1 + 1 + I = 17) and tetrasomics (2n + 2 = 16). The double trisomics, like the primary trisomics, had a reduced vigor and were sterile. They showed at diakinesis 2,,, + 511, or 1111 + 611 + l,, or 711 + 21,in 21, 5 5 , and 24% of the cells, respectively (Gill and Minocha, 1971).

+

+

2 . Secondary Trisomics In pearl millet there is one isolated example of gamma-ray-induced trisomic plant (Pantulu, 1967a). The extra chromosome involved in this trisomic was probably an “isochromosome,” as it formed a “ring trivalent” in two cells (Pantulu, 1967a), and apparently gave an indication of secondary trisomic condition. However, since this plant was a translocation heterozygote, a ring trivalent could also have formed in other ways. There is therefore no clear report of secondary trisomy in pearl millet. Rajhathy and Fedak (1970) described a clear case of secondary trisomy (2n = 14 1 isochromosome) in the diploid species Avena strigosa, in which the isochromosome formed a ring trivalent, or an isoring by interarm pairing. This plant was male- and female-sterile presumably because of the imbalance caused by the tetrasomic condition of the genes on the quadruplicated arm. It would be interesting to know the effect of secondary trisomy in pearl millet.

+

3 . Tertiary Trisomics

Since a tertiary trisomic has only one extra a m , or a part of an arm, of a particular chromosome, the genetic ratios are modified only for the genes located

PREM P.JAUHAR

440

on that arm or segment of the arm.Tertiary trisomics are therefore very useful for assigning genes to a particular chromosome arm and for the location of the position of centromere. Tertiary trisomics of pearl millet have been produced from selfed interchange heterozygotes and their crosses (Minocha et al., 1974; Tyagi, 1975b; Venkateswarlu and Mani, 1978), selfed interchange trisomics (Murthy et al., 1979), and from the selfed progeny of a triploid (Venkateswarlu and Mani, 1978). The expected pentavalent configuration was observed in 15% of the cells, and it was either chain-shaped or dumbbell-shaped (Minocha et al., 1974). Recently, Murthy et al. (1979) observed that of all the different pentavalent configurations formed in two tertiary trisomics, only 10% were dumbbell-shaped. Such differences in chromosome pairing can be caused by different factors, genotype being an important one. 4 . Interchange Trisomics

In pearl millet, interchange trisomics were reported from crosses between two interchange heterozygotes (Minocha and Brar, 1976), in the progeny of interchange heterozygotes (Manga, 1977; Venkateswarlu and Mani, 1978), in the selfed progeny of a triploid (Venkateswarlu and Mani, 1978), and in the progeny of primary trisomics (Rao and Rao, 1977). Tertiary trisomics are also formed as a result of translocation, but they can be distinguished from interchange trisomics. In the tertiaries, the trisome is normally involved in the dumbbell-shaped pentavalent configuration, or in other multivalent associations, like trivalents. Moreover, when the trisome remains as a univalent, the other chromosomes should generally form bivalents, but no multivalents. In the interchange trisomics, however, the trisome may be present as a univalent in addition to an interchange configuration. For example, the occurrence of a ring quadrivalent and a univalent may provide a crucial test of the presence of interchange heterozygosity coupled with trisomy . Thus, in an interchange trisomic a ring hexavalent configuration plus 41, l1 were observed in the majority of the cells (Minocha and Brar, 1976). Manga (1977), Rao and Rao (1977), and Venkateswarlu and Mani (1978) also found a ring quadrivalent plus a univalent in the interchange trisomics studied by them.

+

5 . Telotrisomics and Other Aneuploids with Telocentrics

Some telotrisomics and other aneuploids with telocentric chromosomes have been reported in pearl millet. Pantulu et al. (1976) described a plant with 2 n = 13 2 telocentrics. The telocentrics (t) had arisen from one of the chromosomes as a result of misdivision of its centromere, and they paired with the normal homolog to form a trivalent. Such a trisomic can be technically described as doubletelotrisomic, but is essentially a pseudotrisomic because no new genetic

+

441

CYTOGENETICS OF PEARL MILLET

material is added. In the progeny of a doubletelotrisomic, Pantulu et al. (1976) observed monotelodisomic (2n = 13 I t ) and telotrisomic (2n = 14 It) plants. From crosses between doubletelotrisomic (2n = 13 + 2t) and normal disomic plants, the authors isolated two monotelodisomics (2n = 13 + I t ) that were deficient for one chromosome arm and were also heterozygous for an interchange. Doubletelotrisomic plants ( 2 = ~ 13 + 2t) for different chromosomes of pearl millet have been isolated: for chromosome 1 , in the progeny of an autotriploid (Pantulu and Rao, 1977a); for one of the metacentric chromosomes, in triploid X diploid crosses (Rao et al., 1977); and for chromosome 7, in the progeny of a primary trisomic for chromosome 3 (Rao, 1977).

+

D.

+

TETRASOMICS

By selfing triploids and from triploid x diploid crosses, Gill et a/. (1970b) isolated a few tetrasomic (2n 2 = 16) plants of pearl millet. However, the cytology of these plants was not reported. Some aneuploids with tetrasomic number (2n = 16) but having varying number of telocentrics (t) have also been described. Thus, in the progeny of a doubletelotrisomic (2n = 13 2t), Pantulu et al. (1976) isolated some plants with 2n = 14 2t that can be termed doubletelotetrasomics. From the progeny of another doubletelotrisomic (2n = 13 2t), Rao e t a / . (1977) observed a plant with 2n = 13 3t. This plant with three telocentrics, thus making up a tetrasomic chromosome number (2n = 16), can be described as tripletelotetrasomic. Similarly, on selfing a doubletelotrisomic for chromosome 1, plants with 2n = 12 4t were obtained (Pantulu and Rao, 1977a). These plants with 4 telos making a tetrasomic chromosome number (2n = 16) can be most appropriately described as quadrupletelotetrasomics.

+

+

+

+

+

+

VIII. STRUCTURAL CHANGES IN CHROMOSOMES The preponderance of interchange heterozygotes in several natural populations is a clear indication of the adaptive value of certain chromosomal rearrangements (Bumham, 1956; Darlington, 1956; Rees, 1961b). Chromosomal hybridity confers adaptive advantage because it helps conserve adapted gene complexes. Interchanges, being good cytogenetic markers, are useful in basic studies in cytology and genetics, and offer potential for chromosome engineering (see Bumham, 1962; Ramage, 1964, 1970; Ramage and Wiebe, 1969). Thus, the study of the course and consequences of structural alterations is important to our understanding of many cytogenetic and evolutionary phenomena.

442

PREM P. JAUHAR A. SWNTANEOUS REARRANGEMENTS

There are a few reports of spontaneously occurring chromosomal rearrangements in pearl millet. Pantulu (1958) observed a reciprocal translocation in a first-generation synthetic, Gahi-1, evolved by Burton in the United States. During meiosis it showed, in addition to five bivalents, a ring or chain of four chromosomes that resulted in 40% pollen sterility. Krishnaswamy (1962) noted some partially sterile plants in selfed lines that formed a configuration of four chromosomes, thus indicating heterozygosity for an interchange. Later, Powell and Burton (1 969) reported on some plants that were heterozygous for one or two interchanges. Heterozygosity for a single interchange in the inbred lines Tift 13 and Tift 239 formed a configuration of four chromosomes, whereas an African introduction had a rearranged chromosome common to both interchanges, as evidenced by the formation of a hexavalent configuration. Plants heterozygous for one and two interchanges had, respectively, about 67 and 48% pollen fertility. The present author has observed spontaneous chromosomal interchanges in several populations of pearl millet. B. INDUCEDTRANSLOCATIONS

Chromosomal interchanges have been induced in pearl millet by a variety of physical and chemical mutagens. On the whole, the physical mutagens seem to induce more translocations than the chemical mutagens. Thus, the 90-min treatment with thermal neutrons produced four times as many plants (28%) with chromosome interchanges as with the highest concentration (0.6% aqueous solution) of ethyl methane sulfonate (EMS) (Burton and Powell, 1966). Similar results were reported by La1 and Srinivasachar (1979). Using X-ray treatments, Krishnaswamy and Ayyangar (1941a) induced interchange heterozygosity for one, two, or three reciprocal translocations, which resulted in the formation of quadrivalent, hexavalent, and octavalent configurations, and semisterility of the plants. Later, Pantulu (1967a) and Jauhar (1972, 1974) induced several multiple interchanges by treating seeds with gamma rays. Figure 9b,c shows PMCs of interchange heterozygotes for one reciprocal translocation. C. BUILDING UP A COMPLETE INTERCHANGERING

Through recurrent irradiation coupled with intercrossing of cytologically established interchange heterozygotes, Jauhar (1972, 1974) obtained several com-

CYTOGENETICS OF PEARL MILLET

443

FIG.9. Diakinesis in normal plants (a) and interchange heterozygotes (b-d) of pearl millet produced by gamma irradiation. (a) Seven bivalents (711)in a normal plant. (b,c) Early and late diakinesis with one ring configuration of 4 chromosomes (0') 5". (d) A multiple interchange of 10 chromosomes (partly broken) + 211in a complex interchange heterozygote. [(a) X 1460. (b-d) x ca. 19451

+

444

PREM P. JAUHAR

plex stocks (see, for example, Fig. 9d). However, interchange heterozygotes involving more than 12 chromosomes could not be obtained even by further cycles of irradiation. Irradiations probably broke higher interchange multiples in some cases. Moreover, sterility problems in the interchange stocks involving several chromosomes impeded their successful intercrossing. Interchange heterozygotes involving several chromosomes are difficult to obtain probably because of the limitations imposed on viability of the gametes and/or seeds. Jauhar (1974) believed that the somatic and gametic sieves operated more rigorously with the involvement of more chromosomes in interchanges. Brar et al. (1973), Tyagi and Singh (1974), and Tyagi (1976a), however, were successful in synthesizing interchange stocks involving all 14 chromosomes of pearl millet; these stocks were, of course, almost completely sterile. D. KARYOMORPHOLOGY A N D T H E INCIDENCE O F ~ N T E R C H A N G E S

The response of a species to induced interchange heterozygosity is probably conditioned by certain chromosomal attributes. James (1965, 1970) concluded that lsotoma petraea is cytologically preadapted to the occurrence and maintenance of interchange hybndity. It is interesting to note that karyomorphology and meiotic features, including the mode of chiasma formation, in pearl millet are essentially similar to those of Isotoma petraea. As described in Section I I , l , the chromosomes of pearl millet are mostly isobrachial with median and submedian centromeres, making the karyotype symmetrical. Mean chiasma frequency per bivalent (mean taken of several populations) is 1.84 ? 0.085, so that most chromosomes associate as ring bivalents (Fig. 9a). Moreover, the chiasmata are mostly terminally localized, and at diakinesis their terminalization is nearly complete (see Figs. 1 , 3b,c, and 5a). The favorable response of pearl millet to initial induction of interchanges may be, at least partly, a function of its symmetrical karyotype (Jauhar, 1974). The heterochromatic regions flanking the centromeres may also be helpful in the induction of breakages, because the break points are reported to be localized in the heterochromatic regions (see Evans and Bigger, 1961). E. IDENTIFICATION

OF

TRANSLOCATED CHROMOSOMES

The availability of translocation testers can greatly help in the identification of the chromosomes involved in different interchange stocks. In a cross between a stock and an established tester with known chromosome, the formation of 2,"

+

CYTOGENETICS OF PEARL MILLET

445

311 would suggest that the translocated chromosomes in the unknown stock and the tester are different, whereas l,, + 41, would show that the unknown stock has one chromosome in common with the tester stock. Thus, Tyagi (1976b) was able to identify the chromosomes involved in some translocation stocks by crossing them with tester stocks. F.

DISJUNCTION O F I N T E R C H A N G E CONFIGURATIONS

In a structural heterozygote, the shape and orientation of the quadrivalent configuration depend upon (1) the site of the occurrence of crossing over, and ( 2 ) the co-orientation of the four centromeres involved in the quadrivalent. Thus, two basic types of orientation may be obtained: (1) adjacent, and ( 2 ) “zigzug ” or alternate. Only combinations of chromosomes resulting from alternate orientation, however, result in genetically balanced gametes. Cytological and genetic studies by Burnham (1934) proved the existence of two types of adjacent orientation: adjacent I , in which homologous centromeres co-orientate and thus move to opposite poles; and adjacent 2 , in which nonhomologous centromeres co-orientate and thereby move to opposite poles. Using interchanges T ( 1 , 3) and T (3, 6) of pearl millet, Koduru (1979) studied at metaphase I the orientation types-alternate 1 and 2 , and adjacent 1 and 2 . He found that the relative frequencies of various orientation types were influenced by the genetic background. Frequency of alternate orientation was earlier found to be genotypically controlled in rye heterozygotes (Thompson, 1956), and selection for high and low frequencies of alternate co-orientations gave rise to distinct lines with high and low frequencies of such configurations (Sun and Rees, 1967). These studies provide evidence for the presence of a genetically regulated potential mechanism in meiosis that could be used to offset sterility due to translocations. By exercising selection for high seed fertility in pearl millet, it may be possible to increase the frequency of alternate orientations or vice versa. Such studies may prove useful from the breeding standpoint (see the following section). G. INTERCHANGEHETEROZYGOSITY A N D PLANT BREEDING

The advantage associated with chromosomal repatterning is evident from the fact that interchanges occur in high frequencies in several plant populations. Thus, their contribution to fitness must more than compensate for the infertility of the heterozygotes (Bailey er al., 1978). To explain the benefits of chromosomal hybridity, Darlington (1963) introduced the concept of hybridity

446

PREM P. JAUHAR

optimum, which implies that each species has an optimal level of heterozygosity to which it is accustomed, and that any departure from this level results in deleterious effects. The occurrence of interchanges in several populations of pearl millet suggests that they confer some adaptive advantage. It is likely that heterozygosity for certain interchanges locks up favorable gene arrays (“super genes” of Darlington and Mather, 1952). Interchange hybridity probably helps preserve heterosis associated with certain segments. Recent observations on the possible association of heterosis for yield with a chromosome segment in some hybrids between radiation-induced mutants and the parental line Tift 23 (Burton and Hanna, 1977) suggest the possibility of inducing translocation heterozygotes or certain duplications of value in plant breeding. Since chromosomal hybridity does confer a certain amount of heterosis in the outbreeder Isotoma petraea (James, 1970), it is logical to deduce correspondence between interchange heterozygosity and heterosis in pearl millet. By irradiating interspecific hybrids of Pennisetum, chromosome segments of one species can be transferred into the chromosome complement of the other. Induced translocations may thus be helpful in the interspecific gene transfers (see also Section XI,B,S).

IX. B CHROMOSOMES The occurrence of a special category of chromosomes, in addition to the normal complement, has been reported in numerous plant and animal species. Such extra chromosomes-which are usually smaller than normal chromosomes, do not have synaptic homology with the normal chromosomes, are mostly heterochromatic and genetically inert, and generally vary in number from organism to organism or even in different meiocytes within an organism-are referred to as B, accessory, supernumerary, diminutive, inert, parasitic, or even ghost chromosomes. The nuclear-genetic makeup of an organism is delicately balanced. Whereas each normal chromosome and, indeed, every segment is essential for normal development, B chromosomes are nonessential-if not detrimental-to organisms possessing them. Rhoades and Dempsey (1972) stated that B-chromosome DNA is specialized and shows little transcription or translation. Miintzing (1974) considers the term “accessory” to be more appropriate to describe these chromosomes. However, for the sake of simplicity and brevity, it is preferable to call them uniformly as B chromosomes (the normal chromosomes being A chromosomes). Considerable controversy hangs over the question of the origin of B chromosomes. Although no definite answer has been found so far,

CYTOGENETICS OF PEARL MILLET

447

there is substantial-albeit circumstantial-evidence to support the theory that they originated from normal chromosomes and that they are the by-products of karyotypic stabilization (see Jones, 1975). B chromosomes have been reported in only a few species of the genus Pennisetum, including pearl millet. When chromosome numbers of other species are known, Bs would probably be found in several of them. A. B CHROMOSOMES AS INDICATORS OF T H E ORIGIN OF P E A R L

MILLET

Based on the more frequent occurrence of B chromosomes in primitive varieties than in selected, commercially bred cultivars, Muntzing (1958) suggested that their occurrence might be used as an indicator of a crop’s center of origin. On the basis of the occurrence of B chromosomes in the pearl millet collections, the Sudan (Pantulu, 1960) and Nigeria (Powell and Burton, 1966a; Burton and Powell, 1968) were considered as this crop’s centers of origin. Drawing conclusions on this basis, however, is fraught with danger, because there are several different ecological and edaphic factors that influence the occurrence of B chromosomes. The number of Bs, for example, is reported to be higher in meadow fescue (Fesruca pratensis) growing in areas with clay soil than in areas with lighter soil (Bosemark, 1956); and in rye (Secale cereale) the frequency of Bs is higher in the material growing on acidic than on basic soils (Lee, 1966). Still more striking are the observations of Kishikawa (1970), who worked on clonal plants of rye grown under different regimes of temperature, soil, and moisture. He found that the frequency of Bs was lower in the progeny populations derived from plants grown under higher temperature or drier soil conditions. Such a preferential occurrence of B chromosomes in certain ecological niches would suggest a better selective advantage of Bs in certain conditions than in others. In view of these considerations, great caution should be exercised before drawing conclusions about the center of origin of a crop plant based on the presence or absence of Bs. Moreover, before using any such criteria several collections of a crop plant from different environments should be analyzed. The occurrence of B chromosomes in certain pearl millet collections from the Sudan and Nigeria does not indicate that pearl millet originated at these places. Recent evidence suggests that the Sahel region of West Africa is, in fact, the center of origin of pearl millet (Brunken er al., 1977). B. MEIOTICBEHAVIOR A N D POLYMORPHISM

Pantulu (1 960) reported 1-3 B chromosomes in a population of pearl millet from Sudan. The Bs were predominantly heterochromatic with subterminal cen-

448

PREM P. JAUHAR

tromeres, and there were no morphological differences between plants with Bs and those without. When 3 Bs were present, they paired either as a trivalent or as 111 + 11,or they remained unpaired in a small proportion of PMCs. Similarly, the presence of 4 Bs resulted in the formation 211,1111+ l,, 4,,or occasionally a quadrivalent (Venkateswarlu and Pantulu, 1970). Recently, Rao et al. (1979) studied the effect of 3 and 5 B chromosomes on the duration of mitotic cycle of pearl millet. They found that 3 Bs had little effect, whereas 5 Bs extended the duration by about 39%. Powell and Burton (1966a) reported 1-5 B chromosomes, which they considered euchromatic, in an inbred line derived from seed collected in Nigeria. The Bs either remained unpaired or formed bivalents (Fig. 10). The authors reported a somewhat novel property of the Bs in the Nigerian material, i.e., their ability to organize nucleoli. There is no clear report of the presence of a nucleolusorganizing region in B chromosomes of any species, although Flavell and Rimpau (1975) found evidence of genes coding for ribosomal RNA in the B chromosomes of rye. Additional nucleolar material, as observed by Powell and Burton (1966a), can also be organized by the normal chromosomes. The Bs in the Nigerian material (Powell and Burton, 1966a) showed a size variation among themselves; they were small, intermediate, or large (Fig. 10.). The interarm pairing in the large B coupled with its median centromere suggested that it arose as an isochromosome from the large arm of the intermediate-sized B. The small Bs must also have arisen as telocentrics and/or as isochromosomes from other Bs as a result of misdivision of the centromere. In any case, the Bs in pearl millet show polymorphism with respect to their morphology, if not function. Polymorphic Bs have been reported, among others, in maize (see Jones, 1975). C. EFFECTS O N NORMALCHROMOSOME PAIRING

In the course of this discussion, the terms “association,” “pairing,” and “synapsis” have been used interchangeably. There are varied reports of the effect of B chromosomes on pairing of normal or A chromosomes. Powell and Burton (1966a) observed that the number of B chromosomes per cell had no effect on pairing behavior of A chromosomes. However, the association of Bs among themselves and the A chromosome pairing seemed to be related. Thus, the microsporocytes with the least amount of B chromosome association also seemed to have the least A chromosome pairing. The pattern of pairing in the B chromosomes, however, had no relation to the pairing behavior of the A chromosomes in the material of a different genetic background (Venkateswarlu and Pantulu, 1970). Pantulu and Manga (1975) found that up to 4 Bs did not affect the mean chiasrna frequency of A chromo-

CYTOGENETICS OF PEARL MILLET

449

FIG. 10. Meiotic behavior and polymorphism of B chromosomes. Note that B bivalents (B,,)are orientated on the metaphase plate. (a) Metaphase I with I B,,+ I B,(marked with mows) in addition to 711 of normal chromosomes. (b) Metaphase I with 2B,, (marked with mows). (c) Metaphase I showing 4 B, lying off the metaphase plate. Note the size differences of the B univalents. [ x ca. 20001

450

PREM P. JAUHAR

somes; however, when the number of Bs exceeded 4 and went up to 8, they appeared to have a depressing effect on mean chiasma frequency, although the variance increased with increasing number of Bs. In rye (Secale cereale), Jones and Rees (1967) also reported that 0-8 Bs did not have a significant effect on the mean A-chromosome chiasma frequency; but they did find increase in variance between PMCs within plants, as well as an increase in variance between bivalents within cells. Thus, Bs could generate additional variability by increasing the amount of recombination. More recently, Rao and Pantulu (1978) studied the effects of the standard Bs and their derivatives-standard, deficient, and iso-Bs-on meiosis. With regard to their effect on A-chromosome chiasma frequency, the deficient Bs apparently had a depressing effect and the iso-Bs appeared to have an enhancing effect, whereas the standard Bs had no influence. Thus, the extra euchromatin present in the form of iso-Bs was interpreted to have an enhancing effect, whereas the extra heterochromatin in the form of the deficient Bs was considered to have a depressing effect. However, it is difficult to understand why the standard Bs and the deficient Bs apparently having the same amount of heterochromatin would have differential effect on chiasma frequency. That extra euchromatin in the form of trisomes can have an enhancing effect on chiasma frequency is known in pearl millet (Manga, 1976) and tall fescue (Jauhar, 1978). But the so-called depressing effects of heterochromatin and the enhancing effects of euchromatin (Pantulu and Manga, 1975, p. 244; Rao and Pantulu, 1978, pp. 125, 127) are much too low to be of any interest to a cytogeneticist or a breeder. Thus, in the report by Rao and Pantulu (1978, p. 125), for example, the means of the standard, deficient, and is0 1-4 B classes are 12.74, 12.55, and 13.41, respectively. These differences observed by Rao and Pantulu (1978) are so low that they could be resolved only by scoring a very large number of cells, i.e., 500 in each class. It is debatable whether the Bs, by lowering or promoting chiasma frequencies-if at all-are really regulating the release of variability, and whether or not such an additional variability supposedly released is of any essence to an outbreeder like pearl millet. D. B CHROMOSOMES-A BIOLOGIST'S DILEMMA

As mentioned earlier, B chromosomes generally have erratic meiotic behavior and do not seem to obey the normal laws of cytology and genetics, yet they persist in populations by some devious mechanisms, such as nondisjunction. In this respect, they constitute a cytologist's dilemma. Their genetic organization is not known. No major genes have been located on the B chromosomes, yet they are well known to influence several gene-controlled phenomena, e.g., chiasma

CYTOGENETICS OF PEARL MILLET

45 1

frequency of normal chromosomes. Thus, they also constitute a geneticist’s dilemma. One of the phenomena-nondisjunction-controlled by the B chromosome of maize has nevertheless proved to be a useful genetic tool (Carlson, 1978). Recently, they have been implicated in the suppression of homoeologous chromosome pairing. A report by Evans and Macefield (1972) that B chromosomes strongly suppress homoeologous pairing in a Lolium hybrid and their suggestion that ‘‘this would give, in terms of chromosome behavior, wheatlike (stable) amphidiploids . . . . has aroused considerable interest in breeding circles. Later work has shown, however, that the diploidizing effect of Bs, if any, is not consistent in several Lolium-Festuca hybrids and amphiploids and, hence, not dependable from the breeding standpoint (Jauhar, 1976, 1977a). After all, in nature the well-established diploidizing mechanisms operative in polyploid crop plants (see Jauhar, 1977b) are not related to any B chromosome activity. B chromosomes do not even occur in commercial varieties of crop plants. Moreover, as stated above, B chromosomes have arisen from normal chromosomes and as such must owe their ability to suppress homoeologous pairing, if any, to normal chromosomes. It is doubtful that the incorporation of B chromosomes into the complement of interspecific hybrids of Pennisetum, Lolium-Festuca complex, or any other plant group would help stabilize their meiosis. Moreover, the B chromosome DNA is specialized with no functional cistrons and may not even be transcribable (Rhoades and Dempsey, 1972). Rhoades and Dempsey (1972) further report that B chromosomes normally exist as “an unwanted guest who feasts on the cell’s resources. Thus, depending on B chromosomes for such an important function as stabilization of meiosis would seem fruitless. ”

”

X. FLORAL BIOLOGY AND HYBRIDIZATION

Information on a crop plant’s floral biology is directly useful in its breeding program. Knowledge of floral characteristics, anthesis, mode of pollination and pollen viability is necessary for controlled matings. The inflorescence of pearl millet is a compound, cylindrical, unbranched, tapering spike (Fig. 11). The spikelets are commonly borne in clusters of two on rachillae that are seriately arranged on the main axis. Each spikelet has two florets, one bisexual and the other staminate (Fig. 12). The bisexual or hermaphrodite floret is pedicellate and consists of a single pistil with two feathery stigmas (see Fig. 13a), and three anthers. Borne below the hermaphrodite floret is a sessile, staminate (male) floret with three anthers.

452

PREM P. JAUHAR

FIG. 11. A spike of pearl millet showing protogyny. Note stigmas have emerged all over the earhead and anthers are emerging toward the tip.

CYTOGENETICS OF PEARL MILLET

453

FIG. 12. A scanning electron micrograph of a pearl millet spikelet with two florets; the larger one with protruding feathery stigma is hermaphrodite and the smaller one is staminate. [ ~ 2 4 1

454

PREM P.JAUHAR

A . PROTOGYNY A N D ANTHESIS

An important feature of Pennisetum species is their protogynous (profos = before or ahead) nature, which means that the carpels emerge and mature before the stamens. Consequently, anthesis starts only after most or all stigmas have emerged. Protogyny is particularly conspicuous in pearl millet (Fig. 11). It is, of course, a welcome feature from the evolutionary and breeding standpoints. It facilitates the introgression of characters from wild or semiwild, annual penicillarias into pearl millet and, thus, has helped in the genetic enrichment of this species. The emergence of stigmas generally starts near the tip of the partially exserted spikes and proceeds downward. Sometimes the spikes still inside the boot leaf have fully exserted stigmas. Pearl millet, like most other species of Pennisktum, has bifid, feathery stigmas (Fig. 13a). The two stigmatic branches provide enough surface for effective pollinations. The fully emerged stigmatic branches of pearl millet are glistening white with a bluish tinge. They usually remain receptive for three days. Since pollination is accomplished by wind (anemophily), the stigmatic surface-a reticulum of feathery hairs (Fig. 13b)-plays an important role in bringing about effective trapping of pollen. Generally, one day after the process of the emergence of stigmas is completed on a particular head, the anthers start emerging toward the tip (Fig. 11) and work their way down the head. The time lag between the two processes depends primarily on the temperature conditions. Warmer temperatures are conducive to earlier emergence of anthers and their dehiscence. The species of the section Penicillaria (all cultivated and semiwild or wild pearl millets, and napier grass) are characterized by the presence of conspicuously penicillate anther tips, i.e., they end in a tuft of fine hairs (Fig. 14), whereas most other species of Pennisetum have glabrous anther tips. However, the function of these cilia or trichomes is not known to date. B . MODEOF POLLINATION

Although anemophily-or wind pollination-is the rule in pearl millet, insects are reported to effect occasional cross-pollination (Leuck and Burton, 1966). Using a marker gene, Rao ef al. (1949) estimated 77.8% natural crossing in pearl millet. Of course, the extent of cross-pollination varies with different factors, such as the time of flowering of the parental plants, spacing between plants, and wind velocity and direction. In a uniformly flowering germplasm pool, Burton (1974) estimated natural crossing of 69 and 82% in two consecutive years. In controlled matings, glassine bags must be used.

CYTOGENETICS OF PEARL MILLET

455

FIG. 13. Scanning electron micrographs of feathery stigmas of pearl millet. (a) A bifid stigma; note numerous stigmatic hairs. (b) A portion of stigma enlarged to show a reticulum of feathery hairs that help in effectively trapping wind-borne pollen. [(a) ~ 2 1 (b) ; ~451

C. POLLENMORPHOLOGY AND VIABILITY

The pearl millet pollen grains are generally spheroidal (Fig. 15a). They are uniaperturate (monoporate) with nearly isodiametric pore (porus) (Fig. 15b), 2.5-4 k m in diameter. The nexine (the inner unsculptured layer of exine) is thickened around the porus to form a costa. There is slight variation in the costae of different cultivars; in some it is more pronounced than in others. Under humid conditions, the fresh pollen grains are generally inflated to give a spheroidal appearance. Under dry and hot conditions, however, they shrink to varying degrees. Shrinkage probably makes them lighter and more buoyant so that they can be carried long distances. Thus, this feature seems to have an adaptive value in wind-pollinated species. With the moisture from the stigmatic surface, the shrunken pollen grains swell up and this is probably the first step toward germination. The pearl millet pollen normally remains viable for a few hours, although it

456

PREM P. JAUHAR

FIG. 14. Scanning electron micrographs of pearl millet anthers showing penicillate anther tips. (a) Complete anther. (b) Anther tip magnified to show the tuft o f hairs. [(a) X28; (b) X2401

can be preserved under suitable conditions for future crossing. Cooper and Burton (1965) reported that pollen stored at 4.5"C for 3 weeks gave 80% as good seed set as fresh pollen, whereas that stored for 4 weeks was inviable. Cryopreservation of pollen should be tried with a view to preserving pearl millet germplasm for future use. The hybridization work is best carried out in the mornings between 7 A . M . and 9 A . M . under the Indian conditions. However, Cooper and Burton (1965) have found that hybrids may be made at any time of the day, but those made at midday generally set the least amount of seed per centimeter of spike.

XI. HYBRIDIZATION AND CHROMOSOME RELATIONSHIPS Although mutations have played a significant role in bringing about diversity in the biological world and, hence, are a major force in organic evolution, the catalytic effect of hybridization upon evolution should not be underestimated.

FIG. 15. Scanning electron micrographs of a pollen grain of pearl millet. (a) A spheroidal grain; note a single germ pore (potus) with cmfa around it. (b) A portion of pollen grain magnified to show the porus with costa (C)around i t . [(a) X2600; (b) x5400l

458

PREM P. JAUHAR

Natural hybridization between related species followed by natural polyploidy has in fact given rise to our most important food, fiber, oilseed, and fodder crops. The catalytic effect of hybridization on evolution lies primarily in enlarging the size of the gene pool that would facilitate a favorable response of a population to a changing environment (Grant, 1963; Stebbins, 1969, 1974). The study of chromosome pairing in hybrids has facilitated genome analyses and, thus, helped in the elucidation of phylogenetic relationships between different taxa. Meiotic pairing in hybrids, of course, depends upon the nature and extent of differentiation among the parental genomes. Moreover, different types of genetic control of chromosome pairing can complicate the pairing patterns in intergeneric, interspecific, and even intervarietal hybrids between geographically diverse ecotypes, especially of polyploid species (see Jauhar, 1975c, 1977b). In view of these restrictions, de Wet and Harlan (1972) have questioned the value of chromosome pairing data in interpreting phylogenetic affinities. However, I do feel that, in spite of the aforesaid limitations, chiasmate pairing has provided and will continue to provide useful information on the nature of ploidy of a species and on phylogenetic affinities between species. Chiasmate pairing is a specific process generally confined to chromosomes (or segments) of corresponding genetic similarities. Homologous or homoeologous segments are probably able to recognize each other based on congruence of pairing sites (recognition is probably based on similarity in nucleotide sequences). Information on chromosome pairing in hybrids coupled with that in their amphiploids should therefore give a realistic picture of genomic relationships of the parental species. In the genus Pennisetum, some interspecific, intergeneric, and even intertribal (Penniserum X Oryza) hybrids have been reported. The interspecific hybrids involved mostly pearl millet as one of the parents, and have helped in the study of chromosome relationships between pearl millet and several other species of Pennisetum. A . INTRASPECIFIC HYBRIDS

There are several semiwild, annual, diploid races in the section Penicillaria (Table I) with which the cultivated pearl millet is interfertile and essentially forms a single, composite reproductive unit. So the hybridization between pearl millet and other annual penicillarias will, in effect, be intraspecific hybridization. Pennisetum typhoides is a polymorphic species. Its polymorphism could be due primarily to its hybridization with the taxa in the section Penicillaria. The wild, annual relatives of pearl millet have been treated as separate species by Stapf and Hubbard (1933, 1934) and Clayton (1972), although the latter suggested their merger into single species with pearl millet. They are interfertile with pearl millet and have contributed to its genetic enrichment. Such a process

CYTOGENETICS OF PEARL MILLET

459

of hybridization and intefflow of genes is facilitated by the protogynous nature of the species. Chromosome studies by Krishnaswamy (1951), Thevenin (1952), and Veyret (1957) have shown that the annual relatives of pearl millet-viz., P . ancylochaete, P . cinereum, P . echinurus, P . gambiense, P . leonis, P . maiwa, P . nigritarum, and P . pycnostachyum'-are diploids with 2n = 14 chromosomes like the cultivated pearl millet (see Table I). Moreover, the fact that the chromosome morphology of several of these taxa was similar to one another and also to that of pearl millet (Veyret, 1957) further indicated their affinity with pearl millet. P . violaceurn and P . fallax are two of the important wild, annual relatives of pearl millet. Genetic studies by Bilquez and Lecomte (1969) and by Brunken (1977) have shown that these taxa form fully fertile hybrids with pearl millet and, hence, are not reproductively isolated from it. They must obviously have 2 n = 14 chromosomes in order to form fertile hybrids with pearl millet. The chromosomal, hybridization and genetic studies showed conclusively that the wild or semiwild annual relatives of pearl millet are not sufficiently isolated from it to deserve specific ranks. Brunken (1977) has therefore merged all annual penicillarias with pearl millet (Pennisefum typhoides), which he calls P . americanum. For the sake of convenience, he has subdivided the species into three subspecies: (1) subsp. americanum, which includes all the cultivated races of pearl millet; (2) subsp. monodii, which encompasses all the wild and semiwild races of pearl millet; and (3) subsp. stenostachyum, which is morphologically intermediate between subspecies americanum and monodii, and includes all mimetic weeds associated with the cultivation of pearl millet. B.

INTERSPECIFIC

HYBRIDS

Several hybrids between pearl millet and other species of Pennisetum have been made by different workers. In some cases amphidiploids have also been produced by colchicine-induced chromosome doubling of the interspecific hybrids or by natural meiotic nonreduction.

1 . P . typhoides x P . purpureum Hybrids Interspecific hybrids between P. typhoides ( 2 n = 14) and P . purpureum (2n = 4x = 28) are the most widely studied in the genus Pennisetum and probably

also in the entire tribe Paniceae. These species of the section Penicillaria are reported to cross in nature to produce spontaneous hybrids (Stapf and Hubbard, 'These are all infraspecific categories within the species P . ryphoides

460

PREM P. JAUHAR

1934). They have also been hybridized by numerous workers using either one of them as the female parent. However, using pearl millet as the female parent has several advantages: (1) its protogynous nature generally eliminates the need for emasculation; (2) seed shattering is absent or minimal; and (3) it is easier to identify hybrids at seedling stage. Burton (1944), working in the United States, was the first to make interspecific hybrids, which were later produced in India (Krishnaswamy and Raman, 1949, 1953a,b; Krishnaswamy, 1951), South Africa (Gildenhuys, 1950; Gildenhuys and Brix, 1958, 1964), Pakistan (Khan and Rahman, 1963), Australia (Pritchard, 1971; Muldoon and Pearson, 1977), Sri Lanka (Dhanapala et a l . , 1972), Nigeria (Aken’Ova and Chheda, 1973), and other countries. The main objective of hybridizing these species was to produce a high-quality, highyielding, perennial fodder plant that would inherit pearl millet’s forage quality, nonshattering nature, and capacity to establish readily, and also have the perennial, aggressive nature and rust resistance of napier grass. In South Africa, the main objective was to breed a large-seeded perennial for use in ley farming (Gildenhuys, 1950). The hybrids produced in different countries are generally high-yielding and more acceptable as fodder plants than napier grass. They exhibit considerable heterosis for both fodder yield and quality (Burton, 1944; Krishnaswamy and Raman, 1949; Patil, 1963; Khan and Rahman, 1963; Hussain et al., 1968; Pritchard, 1971; Aken’Ova et a f . , 1974; Gupta, 1974; Muldoon and Pearson, 1977). Powell and Burton (1966b) described a commercial method of producing interspecific hybrids by using a male-sterile line of pearl millet (Tift 23A) as the female parent; the hybrid thus produced was described as the highest-yielding forage millet grown in the United States. However, the hybrids’ complete seed sterility (they can be propagated only by vegetative means) has restricted their adoption on a large scale. While these hybrids are widely grown in some countries, particularly in the Indian subcontinent, they are being grown in trials in several other countries (see Muldoon and Pearson, 1979; see also Section XI ,B,2 ,e) . a. Gross Morphology. With respect to several vegetative characters such as panicle morphology, internode length, leaf and ligule size, the hybrid is either intermediate between the parents or more often approaches the purpureum parent. This is to be expected in view of the greater contribution of genetic material from the purpureum parent. However, depending on the genotype of the parental species used, there is a considerable variation in expression of heterosis for different vegetative characters like height, stem thickness, tillering, and leaf size. b. Chromosome Pairing. The hybrid is a triploid with 2n = 3x = 21 chromosomes (Fig. 16a), 7 contributed by diploid (2x) typhoides and 14 by tetraploid (4x) purpureum. As discussed in Section I1,1, pearl millet has very large chromosomes, larger than those known for any other species of Pen-

CYTOGENETICS OF PEARL MILLET

46 1

nisetum, except P . ramosum. Burton (1944) could identify the 7 large chromosomes of pearl millet ir. the interspecific hybrids, but he did not study chromosome pairing. Krishnaswamy (1951) and Krishnaswamy and Raman (1953a, 1954) studied chromosome pairing in the triploid hybrids. On the basis of the formation of 711 + 7, in most of the cells and the absence of trivalents, these authors concluded that the genome of typhoides was homologous to one of the genomes of purpureum. Khan and Rahman (1963), Ramulu (1968), Sethi et al. (1970), and Rangaswamy (1972) made essentially similar observations on chromosome pairing in the hybrids and reached the same conclusions regarding the genomic makeup of the parental species. Raman (196% reviewing the earlier work of himself, Krishnaswamy, and co-workers, designated the genomic constitution of P . ryphoides as AA, and that of P . purpureum as A'A'BB. The formation of 711in the triploid hybrid (AA'B) was attributed to synapsis between the A genome of typhoides and A' of purpureum. It was evident, however, that A and A' were not completely homologous. c . Analysis of Inter- and Intragenomal Chromosome Pairing. In the hybrids studied by Jauhar (1968), the easily recognizable size differences of the parental chromosomes (Fig. 16a-c) permitted a detailed analysis of inter- and intragenomal chromosome pairing. A range of 0-9 bivalents was observed, the mean per cell being 5.3. Whereas the majority of bivalents were formed between A and A' genomes and, thus, were clearly heteromorphic (Fig. 16b,c), some bivalents resulted from intragenomal pairing. Up to five heteromorphic AA' bivalents were observed per cell, which suggested that the two genomes are evolutionarily related, and that they probably arose from a common progenitor with x = 5 chromosomes. All three genomes-A, A', and B-showed intragenomal (autosyndetic) chromosome pairing. Such bivalents were almost homomorphic and hence easily distinguishable from the heteromorphic, intergenomal (A-A') bivalents described above. The bivalents formed by the A genome chromosomes were the largest (Fig. 16b), whereas those of A' and B genomes were, respectively, intermediate and the smallest in size. The intragenomal pairing appeared to be limited to a maximum of two bivalents, further suggesting x = 5 as the phyletically basic number from which, probably, the genomes with x = 7 were subsequently derived. Thus, it was inferred that x = 5 is the original basic number of the genus Pennisetum and that P . typhoides is a secondarily balanced species (Jauhar, 1968). Later studies by Jauhar (1970b), Minocha and M . Singh (1971b), and Minocha and A. Singh (1971a) provided corroborative evidence favoring these conclusions. It may be pointed out that as early as 1951, Krishnaswamy had suggested that intragenomal pairing occurred within the B genome. On the basis of pachytene pairing in the hybrid, Pantulu (1967b) reported that the chromosomes 1-5 of typhoides were homologous with chromosomes 1-5 of

PREM P. JAUHAR

462

Flc. 16. Meiotic stages in the Penniserurn ryphoides X P . purpureum ( 2 n = 3 x = 21). Note clear size differences of parental chromosomes. [Small arrow, wphoides univalents; medium arrows, intragenomal bivalents within ryphoides complement; thick arrows, intergenomal bivalents formed by A genome of typhoides and A ' of purpureurn.] (a) Metaphase I showing 21 chromosomes, 7 large ryphoides chromosomes ( A genome) and 14 small purpureum chromosomes (A'B genomes). (b) 7, that comprise 1 heteromorphic intergenomal bivalent (thick m o w ) , 2 large Metaphase I with 7,,

+

CYTOGENETICS OF PEARL MILLET

463

purpureum, and that chromosomes 6 and 7 of typhoides were homologous with chromosomes 8 and 14 of purpureum, respectively. However, in view of the markedly larger chromosomes of typhoides, a part of its chromosome should remain unpaired with the corresponding purpureum chromosome. Consequently, terminal forks, terminal unpaired regions, or intercalary loops should be seen at pachytene. Pantulu (1967b) did not, however, report any such structures, as his drawings of different bivalents show almost perfect pairing.

2 . P . typhoides x P . purpureum Amphidiploids a . Chromosome Pairing and Fertility. Krishnaswamy and Raman (1949) produced amphidiploids by treating the P. typhoides X P . purpureum hybrid seedlings with 0.4% colchicine. Chromosome doubling largely restored the fertility in the synthetic amphidiploids (AAA'A'BB; 2n = 6x = 42). During meiosis they generally formed 2 I,,, multivalents being either absent or infrequent (Krishnaswamy, 1951; Krishnaswamy and Raman, 1954; Ramulu, 1968, 1971; Jauhar and Singh, 1969b; and Jauhar, unpublished results). Some univalents observed at metaphase resulted from precocious separation of bivalents and caused disjunctional abnormalities. In view of the formation of 711 + 71 in the triploid hybrid (AA'B), some quadrivalents or trivalents should be expected in the derived amphidiploids, but they are observed rarely, if at all. This is probably due to preferential pairing between the A-A, A'-A', and B-B genome chromosomes, resulting in 21,1. Although the A and A' genomes are somewhat differentiated, the corresponding chromosomes of these genomes can pair in the absence of their own homologous partners. On chromosome doubling, however, bivalent formation is probably brought about by strong preferential pairing. However, the possibility of some sort of genetic control of pairing cannot be ruled out. The diploidizing genes in a double dose probably bring about diploid-like (genetically enforced preferential) pairing in the allohexapioid. [Such diploidizing genes that are effective only in a double dose are known in polyploid species of Festuca (Jauhar, 1975a,c).] Thus, the synthetic amphidiploid behaves meiotically like a typical allohexaploid. b. Chromosomal Instability. Gildenhuys and Brix (1961) also produced an amphidiploid by colchicine treatment of the cuttings of triploid hybrid. Although largely fertile, the amphidiploid showed marked instability in somatic chromointrahaploid bivalents within typhoides complement (medium m o w ) , and 4 intragenomal bivalents within A ' and B genomes; 2 large univalents belong to A genome and the remaining 5 univalents belong to A' and B genomes. (c) Metaphase I with 6,, + 9,. Note clearly heteromorphic bivalents (thick arrows) and distinct size differences among univalents. (d) A cell at anaphase with 22 chromosomes showing 2 typhoides and 5 purpureum chromosomes going to one pole, 3 fyphoides and 9 purpureum chromosomes going to other pole, and 3 Wphoides chromosomes lagging. The ryphoides chromosomes are marked with m o w s . [(a-c) x ca. 1800; (d) x ca. 13501

464

PREM P. JAUHAR

some number within the plant; its number was in the range 2 n = 36-49, with 2 n = 42 occurring most frequently (Gildenhuys and Brix, 1964). The backcross progeny obtained from the cross 2x P. typhoides X 6x amphidiploid also showed intraplant variation in chromosome number. Gildenhuys and Brix ( 1 964) concluded that this intraplant numerical mosaicism was under genetic control and that the genetic determinants expressed themselves only when present in a double or higher dose. Thus, it appears that these genes were hemizygous ineffective. Conversely, Krishnaswamy and Raman (1956) and Ramulu (1968, 1971) did not report any intraplant or even interplant numerical mosaicism in the amphidiploids or their derivatives. Because of the formation of univalents, however, some variation in chromosome numbers in the progenies of amphidiploids is inevitable and may be detected if a large population is scored. c . Phenotypic Manifestation of Different Genomes. The amphidiploids are vigorous in growth and show gigantic features typical of several polyploids. Generally, they have thicker stems, broader leaves, and larger panicles compared to the parental triploids. They show greater morphological resemblance to P. purpureum than to P . typhoides. This is expected in view of the greater genomic contribution from the tetraploid purpureum. Contrarily, if it is considered that A genome of typhoides and A‘ of purpureum are similar in genic content, then the amphidiploids have, in effect, four A genomes and should resemble the typhoides parent more closely. However, their greater resemblance to purpureum would show that either the A and A’ genomes are sufficiently differentiated and have different phenotypic expressions, or else it is more likely that the B genome dominates the A as well as A‘ genomes and masks their phenotypic manifestation. That the B genome is indeed “dominant” is borne out by the studies of Krishnaswamy and Raman (1949, 1953a, 1954, 1956), Raman and Krishnaswami (1961), Raman et al. (1963), and Raman and Nair (1964). All these workers have demonstrated that even when the ratio of A to B genomes is altered from 2 : 1 to 5 : 1, the phenotypic manifestation of the B genome is noticeably greater than that of A genomes combined. In an amphiploid derivative with AAAAA’B constitution (Raman and Krishnaswami, 1961), for example, there was only one dose of B genome, but the characters of the wild parent, P. purpureum, were still expressed. This indicated that one dose of B genome was dominant (or perhaps epistatic) over five doses of the A genome. Studying the morphology of selfed progenies of the amphidiploids by metroglyph analysis, Ramulu and Ponnaiya (1967) and Ramulu (1971) also found a distinct skewedness towards P . purpureum in the expression of several morphological features. d . Amphidiploid Derivatives. Synthetic amphidiploids have been successfully backcrossed to pearl millet to produce derivatives with different genomic constitutions. Thus, Krishnaswamy and Raman (1956) found that hybrids could be easily secured whichever way the cross was attempted, e.g., 2x x 4x, 2x x

CYTOGENETICS OF PEARL MILLET

465

6x, 4x x 2x,or 6x x 2x. All nine tetraploids produced from the cross 2n P. typhoides x 6x amphidiploid had 2n = 28 chromosomes, showing that the amphidiploids generally formed 2 1 -chromosome gametes or else such gametes were at a competitive advantage over the aneuploid gametes. This is in sharp contrast to the observations of Gildenhuys and Brix (1964), who reported that the compatibility of the cross improved when subhaploid pollen of the hexaploid fused with the haploid egg cell of P . typhoides. Gildenhuys and Brix ( 1 964) also found a high degree of incompatibility when amphidiploid was used as pollen parent in backcrosses to P . typhoides. Of the 31 offsprings thus obtained, only 4 had the expected number of 2n = 28 or less. The remainder had 2n = 35 or thereabout, and hence arose from the unreduced egg cells of P . typhoides. It appeared that the pollen from the hexaploid was more compatible with diploid rather than haploid eggs of typhoides. Conversely, the compatibility also seemed to be enhanced when the pollen from hexaploid contained less than the haploid set of chromosomes (Gildenhuys and Brix, 1964). From the standpoint of plant breeding, however, this is not a welcome situation. It was further inferred from these results as well as from the cross [2x P. typhoides x 6x (P.typhoides x P . purpureum)] x 2x P . typhoides, that incompatibility probably resided in the hybrid embryo itself and that its normal development (e.g., the success of the cross) was not dependent on the normal endosperm acting as a nurse tissue (Gildenhuys and Brix, 1964, 1965). Chromosome pairing was studied in several other amphiploid derivatives, with various genomic constitutions, obtained after a series of backcrossings of 6x amphidiploid to 2x and 4x P. typhoides, followed by selfing and further crossing of the derivatives (Raman and Krishnaswami, 1960, 1961; Raman et af., 1963; Nair et al., 1964; Raman and Nair, 1964). As expected, these derivatives formed different frequencies of multivalents, bivalents, and univalents, depending upon their genomic composition. They were sterile by varying degrees. From the studies on triploid hybrid (AA'B), the amphidiploids (AAA'A'BB), autoallotetraploids (AAA'B) and numerous other amphiploid derivatives with different genomic constitutions, the following conclusions can be made:

I . When present in duplicate, the genomes A, A', and B maintain their meiotic integrity, forming mostly bivalents between homologous partners; this is probably due to genetically enforced preferential pairing. 2. When present in a single dose or in more than two doses, there is intergenomal pairing between A and A'; however, A-B and A'-B synapsis is seemingly absent, showing thereby that A and A' are largely homologous to each other, whereas they both are nonhomologous to the B genome. 3. There is some amount of intragenomal pairing probably limited to a maximum of two bivalents within each of the three genomes. These studies have confirmed the allotetraploid nature of P . purpureum. It is a

466

PREM P. JAUHAR

genomic allotetraploid (A’A’BB) with one genome largely homologous to the typhoides (AA) genome at least in terms of its pairing behavior. The donor of B genome may be one of the diploid members of the section Penicillaria, which may exist somewhere in Africa or is probably extinct. e . Selection of Superior Forage Hybrids. As described above, several superior triploid hybrids between pearl millet and napier grass have been produced in different countries. However, they have not been adopted on a wide scale primarily because they are completely sterile and can be propagated only by vegetative means. From the point of view of easy distribution to farmers, superior, fertile amphidiploid hybrids or derivatives would need to be developed, so that seed can be supplied to farmers. Since the colchicine-induced amphidiploids are largely regular meiotically and their progenies show a wide range of pollen and seed fertility, it may be possible to select fertile, superior forage amphiploids which produce large amounts of good seed. f. Possible Synthesis of Perennial Pearl Millet. From the point of view of incorporating the desirable features of P . purpureum into P . vphoides, the triploid plant obtained by Raman and Krishnaswami (1960) is interesting. This plant was derived by crossing 2x typhoides (AA) with an autoallotetraploid (AAA’B). The derived triploid had four nucleolar chromosomes, and the pattern of pairing showed that it had mostly chromosomes of A and A’ genomes and probably none of the B genome. It is interesting that most, if not all, B genome chromosomes were selectively eliminated. Raman and Krishnaswami (1960) suggested that from such triploid plants it might be possible to secure fertile allodiploids with A and A’ genomes that would probably combine the desirable features of ryphoides and purpureum. It was observed earlier at the University of Hawaii Agricultural Experiment Station (Anonymous, 1947) that the 34 F , hybrids, obtained by crossing a selfed line of pearl millet (2n = 14) with an East African strain of napier grass (2n = 28), included plants with 2 n = 28,21, 18, 16, and 14. The occurrence of plants with 2 n = 14 suggests the possibility of selecting the desired “allodiploid” plants. In this way perhaps a perennial pearl millet can be produced. The selfed pearl millet line used in the above cross was probably desynaptic and produced imbalanced gametes and, hence, aneuploid hybrids. If a desynaptic purpureum line can be used as a male parent and crossed with a diploid typhoides with desirable features, different plants with a whole array of chromosome numbers can be obtained to exercise selection for the desired 2 n = 14, meiotically regular, fertile plants. In successive backcrosses of the 6x amphiploids to P . typhoides, Gildenhuys and Brix (1969) observed a selective elimination of imbalanced gametes and zygotes, particularly those with chromosomes of the B genome, and to a lesser extent, those of the A‘ genome (of purpureum). After three generations of backcrossing, all plants expressed the annual pearl millet habit. However, Gildenhuys and Brix (1969) could not combine the desirable features of typhoides

CYTOGENETICS OF PEARL MILLET

467

(viz., fertility, date of flowering, and grain size) with the perennial habit of purpureurn. If further attempts are made using gamma-ray treatments also, it may be possible to incorporate the desirable segments from purpureum into the typhoides complement. If a perennial pearl millet is indeed produced through the techniques described above, it would be useful in the drought-stricken, semiarid regions of Africa, tropical India, and other tropical and subtropical regions.

3 . P . typhoides x P . squamulatum These species have been hybridized with a view to evolving a grass combining the forage quality of pearl millet with the frost resistance and perenniality of P . squamulatum (2n = 6x = 54). Patil et al. (1961) and the present author (Jauhar, unpublished results) successfully obtained the hybrids. Taking advantage of the protogynous nature of pearl millet, we dusted profusely the pollen of squamulatum on a number of typhoides ears that had freshly emerged, receptive stigmas. From the progeny, Patil et al. (1961) picked up a single hybrid that had 2n = 41 chromosomes. The hybrid obviously arose from the union of an unreduced 14-chromosome egg of pearl millet with a haploid 27-chromosome male gamete of squamulatum. Morphologically, the hybrid resembled the squamulatum parent very closely although it had the leafiness of pearl millet. Patil et al. (1961) observed mostly 1611 9, in the hybrid. It appeared to the authors that seven bivalents (which must have been much larger than the rest) were formed by the typhoides complement, whereas the remaining nine bivalents (the small ones) and nine univalents resulted from pairing within the squamulatum complement. The small bivalents in the hybrid could have formed as a result of intergenomal pairing between the three genomes of squamulatum and some by intragenomal pairing.

+

4 . TrispeciJic Hybrids: ( P . typhoides x P . purpureum) x P . squamulatum

Several hybrids involving three distinct species have been made in grasses, e.g., Lolium-Festuca complex (Jauhar, 1975b, 1976). In the genus Pennisetum, there is one report of a trispecific hybrid obtained by crossing ( P . typhoides X P . purpureum) hybrid with P . squamulatum (Rangaswamy and Ponnaiya, 1963). The two hybrid plants thus obtained resembled the squamulutum parent in several features, including the panicle and spikelet characters. They also had penicillate anther tips, a typical character of typhoides (see Fig. 14) and purpureum. Rangaswamy and Ponnaiya (1963) found that the hybrid had 2n = 48 chromosomes, which showed that it had arisen from an unreduced 21-chromosome egg of the female parent (typhoides X purpureum hybrid) fertilized by the normal 27-chromosome male gamete from squamulatum. The formation of quadrivalents and pentavalents at meiosis suggested relationship of the genomes of the

468

PREM P. JAUHAR

three species (Rangaswamy and Ponnaiya, 1963; Menon and Devasahayam, 1964). 5 . P . typhoides

X

P . orientale

P . orientale is a valuable, perennial forage species consisting of different cytotypes with 2 n = 18, 27, 36,45, 54. It belongs to the section Heterostachya. With a view to evolving a perennial forage grass, we hybridized pearl millet and diploid P . orientale (2n = 18). The interspecific hybrids were first produced by Patil and Singh (1964) and later by Jauhar in 1966 and 1967 (unpublished results), using pearl millet as the female parent (Fig. 17). a. Gross Morphology and Chromosome Pairing. The hybrids showed considerable hybrid vigor. In gross morphology, they were intermediate between the parental species and easily recognizable even at seedling stage. All the hybrids were completely sterile and could be propagated only by vegetative means. However, some variation was noted in the clonal progeny raised from a single hybrid. For example, some clones produced spikes of varying compactness. It appears that there was “somatic segregation” for this character.

FIG. 17. Spikes of P ryphoides (a), P. orientale ( c ) , and their hybrid (b). Note intermediate morphology of the hybrid spike and its heterosis for size.

CYTOGENETICS OF PEARL MILLET

469

They hybrid had 2n = 16 chromosomes, 7 contributed by the female typhoides parent and 9 by orientale. There were very marked size differences among the parental chromosomes (see Fig. 18). Preliminary studies by Patil and Singh (1964) revealed some pairing between the typhoides and orientale chromosomes, suggesting their ancestral relationship. They also observed one or two bivalents as a result of pairing (autosyndesis) within the orientale complement. Taking advantage of the easily recognizable size differences (Fig. 18a), Jauhar (1973, 1981b) made a detailed analysis of the inter- and intragenomal chromosome pairing relationships. Association between typhoides and orientale chromosomes resulted in the formation of distinctly heteromorphic bivalents (Fig. 18b) or heteromorphic multivalents. Generally one or two heteromorphic bivalents were noticed, but a maximum of five was observed in one cell. Apart from intergenomal pairing, intracomplement associations within the typhoides and the orientale complements were also observed. Within the typhoides complement a maximum of 2 bivalents were noted in 6 cells, and 1 bivalent in 11 cells, out of the total of 50 analyzed. Within the orientale complement, autosyndetic bivalents (Fig. 18b,c) were observed; rarely, a loose trivalent was noticed in some cells (Fig. 18b). Of the 50 cells studied, 5 had an autosyndetic trivalent and 30 had 1-3 autosyndetic bivalents. A maximum of 4 autosyndetic bivalents were observed in 2 cells. From this trend of chromosome pairing, it was inferred that a part of the orientale genome is partially homololous (or homoeologous) to a part of the typhoides genome. The formation of a maximum of 5 intergenomal bivalents, 2 intragenomal bivalents within the typhoides complement, and 4 intragenomal bivalents within the orientale complement suggests that x = 5 is the phyletically basic number for the genus Pennisetum and that x = 7 and x = 9 have been derived from it subsequently during the course of evolution. This hypothesis is necessarily speculative and difficult to test experimentally. b. Developing Superior Hybrids. As mentioned, the typhoides x orientale hybrids are completely sterile. Colchicine-induced amphidiploids should be produced to restore fertility and seed set. These characters are important not only from the point of view of further breeding but also for easy distribution to farmers. Pairing between the parental chromosomes should facilitate interspecific transfer of desirable genes to produce superior hybrids combining the aggressiveness and tillering ability of P . orientale with the leafiness and forage quality of P. typhoides. Gamma irradiation may be tried to accelerate interspecific gene transfers. Following gamma irradiation of the vegetative tillers, desired somatic segregants (having the typhoides complement with segments of orientale chromosomes) may be picked up. If the vegetative segregants have 7 typhoides chromosomes, stable, homozygous diploids may be produced by colchicine treatment.

470

PREM P. JAUHAR

FIG. 18. Meiotic stages in P . ryphoides X P. orientale hybrids ( 2 n = 16). Note striking size differences between parental chromosomes. [Thin arrows, typhoides univalents; medium arrows, autosyndetic bivalents within the orientale complement; thick arrows, heteromorphic bivalents formed by parental chromosomes.] (a) Diakinesis with 16 univalents (2 associated with the nucleolus). Note 7 large typhoides chromosomes and 9 small orientale chromosomes. (b) A PMC with 2 heteromorphic bivalents + I autosyndetic bivalent within orientale complement. (c) Metaphase I showing 2 autosyndetic bivalents within orientale complement + 7 typhoides univalents + 5 orienlale univalents. (d) Early anaphase 1 showing asynchrony of behavior of parental chromosomes. Note the large typhoides chromosomes still on the metaphase plate, while the orientale chromosomes are moving to poles. [(a,b,d) X ca. 2360; (c) X ca. 19801

CYTOGENETICS OF PEARL MILLET

6. P . typhoides

X

47 1

P . setaceutn (Syn. P . ruppellii)

Fountain grass ( P . setaceurn) is a triploid ( 2 n = 3x = 27). Since it is an apomict, it must be used as male parent in any hybridization program. Using a pearl millet male-sterile line Tift 23 DA as female parent, Hanna (1979) produced an apomictic interspecific hybrid P . typhoides x P . setaceum. The hybrid closely resembled the polyploid seraceum parent. This is expected in view of greater contribution of chromatin from setaceum. Of the three hybrids produced, Hanna (1979) found that two had 2n = 25 (7 from pearl millet and 18 from fountain grass) and the third had 2n = 24 (7 from pearl millet and 17 from fountain grass). On the basis of size differences, the chromosomes of the parental species were readily distinguishable. During meiosis, up to 3 typhoides chromosomes associated with 3 setaceum chromosomes, and at least one such association was observed in 50% of the PMCs. The hybrids were male-sterile. Interestingly enough, Hanna (1979) found that in somatic divisions some of the setaceum chromosomes were eliminated and somatic segregation for parental characters occurred in the clonal progeny. Based on the somatic elimination of setaceum chromosomes, Hanna (1979) attempted to recover from the clonal progeny some plants with the typhoides genome incorporating some segments of setaceum chromosomes. If successful, this should be an elegant technique (see also Section XI,B,5) that may be applied to several other interspecific and intergeneric hybrids. C . INTERGENERIC HYBRIDIZATION

Numerous intergeneric hybrids have been reported in the grass family. Up until 1972, over 800 intergeneric hybrids had been reported in the Gramineae (Knobloch, 1972), and at least one (Tritium X Secale) has led to the evolution of a new man-made cereal-Triticale. A few intergeneric hybrids have been synthesized in Pennisetum. P . typhoides x Cenchrus ciliaris

Buffelgrass, Cenchrus ciliaris [Syn. Pennisetutn ciliare (L.) Link], is an excellent fodder grass that is highly nutritious before flowering. It is also recognized as one of the most important perennial pasture species in the semiarid regions of northern India (Whyte, 1957). It is an apomict (Bashaw, 1962; Taliaferro and Bashaw, 1966). Read and Bashaw (1974) made extensive efforts to cross a male-sterile line (Tift 23A) of P . typhoides ( 2 n = 14) with an apomictic cultivar of C.ciliaris (2n = 36) and

472

PREM P. JAUHAR

produced one F, hybrid. Morphologically, the hybrid resembled the polyploid male parent, C. ciliaris. This hybrid, in a way, is an intergeneric hybrid, although until recently C . ciliaris has been considered as a species of Pennisetum. In the somatic cells of the hybrid, Read and Bashaw (1974) observed 2 n = 25 chromosomes (7 large ones from pearl millet and 18 small ones from buffelgrass). However, chromosome pairing relationships could not be studied. The hybrid was completely sterile and did not produce any seed after pollinations with pollen of either parent. It appeared to have inherited the aposporous mechanism from the apomictic parent ( P . ciliaris), but did not produce any seed even by apomixis. It is nevertheless interesting from the breeder’s standpoint that apomixis can be transferred from the apomictic parent to its hybrid.

X I I . CONCLUSION

Meeting the ever-expanding demand for food for the ever-increasing world population is the biggest challenge confronting agricultural scientists. One way to meet this demand is to bring additional areas-for example, the dry and relatively infertile lands in the tropical and subtropical regions of the worldunder cultivation. Pearl millet has a remarkable ability to grow in some of the driest agricultural conditions. It already provides food for millions of poor people in Africa and Asia. In terms of annual production, it is the sixth most important cereal crop in the world. Because of its ability to provide feed for cattle, pearl millet acquires added importance. Therefore the need for the genetic improvement of this crop cannot be overstated. Fortunately, pearl millet is favorable for both basic studies as well as applied work. Pearl millet is a favorable organism for basic research in cytogenetics. Because of its small number but large size of chromosomes, it provides a suitable tool for studying chromosome pairing and chiasma frequencies and for understanding the factors controlling these intriguing phenomena. Some of the pairing variantsdesynaptics and partial desynaptics-can facilitate these studies. Pearl millet also lends itself for aneuploid analyses that should elucidate its cytogenetic architecture. Although considerable progress has been made in producing a set of trisomics, the establishment of linkage groups awaits completion. Basic studies on chromosomal rearrangements and induced mutagenesis can also be done in this crop. Pearl millet has an efficient photosynthetic ( C , ) pathway and responds very well to fertilizers. It also responds very well to heterosis breeding. Dwarf hybrids should therefore be evolved for maximum grain yields. The development of cytoplasmic male-sterile (cms) lines by Burton (1958, 1965) in the United States and later by Athwal(l965, 1966) in India has greatly facilitated the production of commercial

CYTOGENETICS OF PEARL MILLET

413

hybrids. The speed with which the Indian breeders accomplished the development of high-yielding grain hybrids using cms lines, particularly Burton's Tift 23A, is considered to be one of the most remarkable plant breeding success stories of all time. This should serve as a model for emulation by other Asian and African plant breeders. Although several commercial hybrids in India yield nearly twice as much as the best standard varieties, undoubtedly there is scope for further improvement. Genetic enrichment of their nutritional status-particularly the protein content and amino acid balance-also deserves greater attention, so that pearl millet can better feed the underprivileged man. Superior, high-yielding forage hybrids of pearl millet, pearl millet x napier grass, and pearl millet X other species should also be evolved to better feed our cattle. Cytogenetic studies could help stabilize the interspecific hybrids. Because of its distinctly protogynous nature, pearl millet is well suited for hybridization work. Distinct size differences between the chromosomes of pearl millet and other species permit a study of inter- and intragenomal pairing relationships and should help in elucidating phylogenetic trends in the polybasic, fascinating genus Pennisetum. Several heterotic hybrids combining the desirable characters of pearl millet and other forage species should be produced. Another area that merits special attention is the development of a perennial pearl millet that can yield some grain as well as forage for several years (see Section XI,B,2,f). A perennial strain of pearl millet should be very useful in arid and semiarid regions of Africa and Asia. A knowledge of different aspects of cytogenetics of pearl millet and related species should also help formulate further rational breeding programs. Evidently, pearl millet provides excellent opportunities for both fundamental and applied research. Such studies have already produced enough dividends to encourage further work. With more concerted research, pearl millet should emerge as a leading, economically viable crop that will play an ever-increasing role in the welfare of man. REFERENCES Aastveit, K . (1968). Hereditas 60, 294-315. Aken'Ova, M. E., and Chheda, H. R. (1973). Niger. Agric. J . 10, 82-90. Aken'Ova, M. E . , Crowder, L. V . , and Chheda, H. R . (1974). Agron. Abstr. p. 49. Madison, Wisconsin. Aman, M. A , , and Sarkar, K . R. (1978). Indian J . Genet. Plant Breed. 38, 452-457. Anonymous (1947). Forage crops. In Biennial Rep. Univ. Hawaii Agric. Exp. Sta. for the Biennium ending June 30, 1947. Athwal, D. S. (1965). lndian Farming 15(5), 6-7. Athwal, D. S . (1966). Indian J . Genet. Plant Breed. Symp. Suppl. 26A, 73-85. Avdulov, N. P. (1931). Bull. Appl. Eot. Genet. Plant Breed. Suppl. 43. [Russian]

474

PREM P. JAUHAR

Bailey, R. J., Rees. H., and Adena, M. A. (1978). Heredity 41, 1-12. Bashaw, E. C. (1962). Crop Sci. 2, 412-415. Bennett, M. D., and Rees, H. (1970). Gener. Res. 16, 325-331. Bilquez, A,-F., and Lecomte, J. (1969). Agron. Trop. 24, 249-257. [French]. Bosemark, N. 0. (1956). Heredifas 42, 189-210. Brar, D. S., Minocha, J. L.,and Gill, B. S. (1973). Curr. Sci. 42, 653-654. Brunken, J. N. (1977). Am. J. Bot. 64, 161-176. Brunken, J. N., de Wet, J. M. J., and Harlan, J . R. (1977). Econ. Eot. 31, 163-174. Burnham, C. R. (1934). Generics 19, 430-447. Burnham, C. R. (1956). Bot. Rev. 22, 419-552. Burnham, C. R. (1962). “Discussions in Cytogenetics. ” Burgess, Minneapolis, Minnesota. Burton, G . W. (1942). Am. J. Bor. 29, 355-359. Burton, G. W. (1944). J. Hered. 35, 227-232. Burton, G. W. (1958). Agron. J. 50, 230-231. Burton, G. W. (1965). Crops Soils 17(5), 19. Burton, G. W. (1968). Crop Sci. 8, 229-230. Burton, G. W. (1974). Crop Sci. 14, 802-805. Burton, G. W., and Hanna, W . W. (1977). Mutar. Breed. Newsleft. 9, 3. Bunon, G. W., and Powell, J. 9 . (1966). Crop Sci. 6 , 180-182. Burton, G. W., and Powell, J. 9 . (1968). Adv. Agron. 20, 49-89. Carlson, W. R. (1977). In “Corn and Corn Improvement” (G. F. Sprague, ed.), pp. 223-274. Amer. SOC.Agron., Madison, Wisconsin. Carlson, W. R. (1978). Annu. Rev. Genet. 16, 5-23. Chase, S . S. (1974). In “Haploids in Higher Plants: Advances and Potential’’ (K. J . Kasha, ed.), pp. 211-230. Univ. of Guelph, Canada. Clayton, W. D. (1972). Gramineae. In “Flora of West Tropical Africa” (F. N. Hepper, ed.), 2nd Ed., Vol. 3, Pt. 2., pp. 349-512. Crown, London. Coe, E. H. (1959). Am. Nut. 93, 381-382. Constance, L. (1957). Am. J. Bot. 44, 88-92. Cooper, R. B., and Burton, G . W. (1965). Crop Sci. 5 , 18-20. Crowley, J. G . , and Rees, H. (1968). Chromosoma 24, 300-308. Darlington, C. D. (1956). Proc. R. SOC. Ser. B 145, 350-364. Darlington, C. D. (1958). “Evolution of Genetic Systems.’’ Oliver & Boyd, Edinburgh. Darlington, C. D. (1963). “Chromosome Botany and the Origin of Cultivated Plants,’’ rev. 2nd Ed. Allen & Unwin, London. Darlington, C. D., and Mather, K . (1952). “The Elements of Genetics.” Allen & Unwin, London. de Wet, J . M. J. (1980). If1 “Polyploidy: Biological Relevance” (W. H. Lewis, ed.), pp. 3-15. Plenum, New York. de Wet, J. M. J., and Harlan, J . R. (1972). Taxon 21, 67-70. Dhanapala, S. B., Siriwardene, J. A. de S., and Pathirana, K. K. (1972). Ceylon Vet. J . 20, 77. Dhesi, J. S., Gill, 9 . S., and Sharma, H.L. (1973). Cytofogia 38, 311-316. Dhesi, J. S., Minocha, J. L., and Sidhu, J. S. (1975). Curr. Sci. 44, 862-863. Doggett, H. (1964). Heredify 19, 543-558. Einset, J. (1943). Genetics 28, 349-364. Evans, G. M.,and Macefield, A. J. (1972). Nature (London), New Biof. 236, 110-ill. Evans, H. J . , and Bigger, T. R. L. (1961). Genetics 46, 277-289. Fedak, G. (1973). Can. J. Gener. Cytol. 15, 647-649. Filion, W. G., and Blakey, D. H. (1979). Can. J. Genet. Cyfol.21, 373-378. Flavell, R. B., and Rimpau, J. (1975). Heredity 35, 127-131. Gadella, T. W.J., and Kliphuis, E. (1964). Acta Bot. Neerl. 13, 432-433. Gildenhuys, P. J. (1950). Farming S. Afr. 15, 189-191.

CYTOGENETICS OF PEARL MILLET

475

Gildenhuys, P. J . , and Brix. K . (1958). Heredity 12, 441-452. Gildenhuys, P. J., and Brix, K . (1961). Heredity 16, 358-363. Gildenhuys. P. J . , and Brix, K . (1964). Heredity 19, 533-542. Gildenhuys, P. J., and Brix, K . (1965). Ann. Bor. 29, 709-715. Gildenhuys. P. J . , and Brix, K . (1969). Agroplantae 1, 77-83. Gill, B. S . , and Minocha, J. L. (1971). J. Cytol. Genet. Congr. Suppl. pp. 33-36. Gill, B. S . , Sraon, H., and Minocha, J. L. (1966). J. Res. Punjab Agric. Univ. 3, 260-263. Gill, B. S., Minocha, J. L., Gupta, D . , and Kumar, D. (1969). Indian J. Genet. Plant Breed. 29, 462-467. Gill, B. S., Minocha, J . L., George, M. K., and Brar, D. S . (l970a). J . Res. Punjab Agric. Univ. 7 , 41 7-422. Gill, B. S., Virmani, S. S., and Minocha. J. L. (1970b). Experientia 26, 1021. Gill, B. S., Virmani, S. S., and Minocha, J. L. (1970~).Can. J. Genet. Cyrol 12, 474-483. Gill, B. S . , Sharma, H. L., and Dhesi, J. S. (1973). Cytologia 38, 411-416. Gillies. C. B . (1974). Chromosoma 48, 441-453. Golovin. V. P. (1979). Sel. Semenovod. 3, 13-14. [Russian]. Golubovskaya, I. N. (1979). Int. Rev. Cytol. 58, 247-290. Gottschalk, W . (1976). In “Induced Mutations in Cross-Breeding,’’ pp. 37-53. IAEA, Vienna. Gottschalk, W., and Klein, H. D. (1976). Theor. Appl. Genet. 48, 23-34. Grant, V. (1963). “The Origin of Adaptations.” Columbia Univ. Press, New York. Gupta, V. P. (1974). In “Breeding Researches in Asia and Oceania” (S. Rumanujam and R. D. Iyer, eds.). pp. 162-172. Indian SOC.Genet. & Plant Breeding, New Delhi. Hanna, W. W . (1979). J . Hered. 70, 425-427. Hanna, W . W . , Powell, J. B. and Burton, G . W. (1976). Can. J . Genet. Cytol. 18, 529-536. Harinarayana, G., and Mufly, B. R. (1971). Cytologia 36, 435-448. Harlan, J . R . , and de Wet, J. M. J. (1975). Dot. Rev. 41, 361-390. Hazarika, M. H., and Rees, H. (1967). Herediry 22, 317-332. Hilpen, G. (1957). Hereditas 43, 318-321. Hossain, M. G . (1978). Cytologia 43, 21-34. Hunter, A. W. S. (1934). Can. J . Res. Ser. C 11, 213-241. Hussain, A , , Ullah, M., and Ahmad, B. (1968). W. Pak. J . Agric. Res. 6(2), 69-72. James, S. H. (1965). Heredity 20, 341-353. James, S. H. (1970). Heredity 25, 53-77. Jauhar, P. P. (1968). Generica 39, 360-370. Jauhar, P. P. (1969). Natunvissenschafren 56, 571-572. Jauhar, P. P. (1970a). Generica 41, 407-424. Jauhar, P. P. (1970b). Genetica 41, 532-540. Jauhar, P. P. (1972). Experientia 28, 710. Jauhar, P. P. (1973). Proc. Inr. Congr. Genet., 13th, Gal$. (in Genetics 74, 126-127.) Jauhar, P. P. (1974). Theor. Appl. Genet. 44, 58-62. Jauhar, P. P. (1975a). Nature (London) 254, 595-597. Jauhar, P. P. (1975b). Chromosoma 52, 103-121. Jauhar, P. P. ( 1 9 7 5 ~ )Chromosoma . 52, 363-382. Jauhar, P. P. (1976). In “Chromosomes Today,” Vol. 5 (P. L. Pearson and K.R.Lewis, eds.), pp. 165-177. Wiley, New York. Jauhar, P. P. (1977a). Proc. Int. Congr. SABRAO. 3rd, (SOC.for the Adv. of Breed. Res. in Asia and Oceania) Canberra 14(b), 34-39. Jauhar, P. P. (1977b). Theor. Appl. Genet. 49, 287-295. Jauhar, P. P. (1978). J. Hered. 69, 217-223. Jauhar, P. P. (1980). Genetics 94 (Suppl.), 49. (Abstr.) Jauhar, P. P. (1981a). Indian J. Bot. 4, 1-4.

476

PREM P. JAUHAR

Jauhar, P. P. (1981b). (in press). Jauhar, P. P., and Singh, U. (1969a). Theor. Appl. Genet. 39, 315-319. Jauhar, P. P., and Singh, U. (1969b). Curr. Sci. 38, 420-421. Jauhar, P. P., Singh, U.,and Alice, C. J . (1976). Genet. Iber. 28, 15-20. Jones, R. N. (1975). f n t . Rev. Cytol. 40, 1-100. Jones, R. N., and Rees, H. (1967). Heredity 22, 333-347. Karper, R. E., and Chisholm, A. T. (1936). Am. J . Bot. 23, 369-374. Kasperbauer, M. J., Buckner, R. C., and Springer, W . D. (1980). Crop Sci. 20, 103-107. Kempanna, C., and Riley, R. (1964). Herediry 19, 289-299. Khan, M.-D., and Rahman, H. (1963). W . Pa&. J . Agric. Res. 1, 61-65. Khush, G. S. (1973). “Cytogenetics of Aneuploids.” Academic Press, New York. Kishikawa, H. (1970). Jpn. J . Breed. 20, 269-274. Knobloch, I. W . (1972). Taxon 21, 97-103. Koduru, P. R. K. (1979). Genet. Res. 34, 69-76. Koduru, P. R. K., and Rao, M. K. (1978). Cytologia 43, 445-452. Koduru, P. R. K., Murthy, T. G. K., and Lakshmi, K. V. (1980). Chromosoma 78, 365-370. Krishnaswamy, N. (1951). fndian J. Genet. Plant Breed. 11, 67-74. Krishnaswamy, N. (1962). “Bajra, Pennisetum typhoides S. Kt H.” Indian Council Agricultural Research, New Delhi. Krishnaswamy, N., and Ayyangar, G . N. R. (1941a). J . fndian Bot. SOC.20, I 1 1-1 17. Krishnaswamy, N., and Ayyangar, G. N. R. (1941b). Proc. fndian Acad. Sci. Ser. B 13, 9-23. Krishnaswamy, N., and Raman, V. S. (1948). Curr. Sci. 17, 153-154. Krishnaswamy, N., and Raman, V. S. (1949). Curr. Sci. 18, 15-16. Krishnaswamy, N., and Raman, V. S. (1953a). Proc. Sci. Workers Conf, f s t . Agric. Coll. Res. Inst., Coimbatore. pp. 43-71. Krishnaswamy, N., and Raman, V. S. (1953b). Cytologia 18, 305-321. Krishnaswamy, N., and Raman, V. S. (1954). Genetica 27, 253-272. Krishnaswamy, N., and Raman, V. S . (1956). Genetica 28, 345-360. Krishnaswamy, N., Raman, V. S., and Madhavamenon, P. (1949). Proc. fndian Acad. Sci. Ser. B 30, 195-206. Krishnaswamy, N., Raman, V. S . , and Nair, N. H. (1950). Curr. Sci. 19, 252-253. Lakshmi, K. V . , Murthy, T. G. K., and Koduru, P. R. K. (1979). Theor. Appl. Genet. 55, 189-190. Lakshmi, N., and Yacob, Z. (1978). Microbios Lett. 7, 75-82. Lal, J . , and Srinivasachar, D. (1979). Theor. Appl. Genet. 54, 27-32. Lee, W . J. (1966). Korean J . Eot. 9(3-4), 1-6. Leuck, ,D. B., and Burton, G. W . (1966). J . Econ. Entomol. 59, 1308-1309. Levitskii, G. A. (1931). Bull. Appl. Bot. Genet. Plant Breed. 27, 220-240. Li, C. H., and Li, H . W. (1943). Chin. J . Sci. Agric. 1, 139-141. Li, H . W . , Pao, W.K., and Li, C. H. (1945). Am. J . Bot. 32, 92-101. Lobana, K. S., and Gill, B. S. (1973). Cytologia 38, 713-717. Mahadevappa, M., and Ponnaiya, B. W . X. (1967). Madras Agric. J . 54, 85-88. McCoUum, G. D. (1958). Chromosoma 9, 571-605. Manga, V. (1976). Can. J . Genet. Cytol. 18, 11-15. Manga, V. (1977). Experientia 33, 1581-1582. Manga, V . , and Pantulu, J. V. (1971). Genetica 42, 319-328. Manga, V . , and Pantulu, J. V. (1974). Proc. fndian Acad. Sci. Ser. B 80, 105-120. Mehra, P. N., Khosla, P. K., Kohli, B. L., and Koonar, J . S. (1968). Res. Bull. Punjab Univ. 19, 157-230. Menon, P. M., and Devasahayam, P. (1964). Madras Agric. J . 51, 70. (Abstr.) Minocha, 1. L., and Brar, D. S . (1976). Indian J . Genet. Plant Breed. 36. 153-155.

CYTOGENETICS OF PEARL MILLET

477

Minocha, J. L., and Gill, B. S. (1974). In “Advancing Frontiers in Cytogenetics” (P. Kachroo, ed.), pp. 51-53. Hindustan Publ., Delhi. Minocha, J. L., and Singh, A. (1971a). Sci. Cult. 37, 198-199. Minocha, J. L., and Singh, kl. (1971b). J . Cvtol. Genet. Congr. Suppl. pp. 270-273. Minocha, J. L . , Gill, B. S., and Singh, S . (1968). J . Res. Punjab Agric. Univ. 5, 32-36. Minocha, J. L., Sharma, H. L., and Gill, B. S . (1972). Indian J . Genet. Plant Breed. 32,211-214. Minocha, J. L . , Brar, D. S . , and Gill, B. S. (1974). Experientia 30, 623-624. Minocha, J. L . , Dhesi, J. S., and Sidhu, J. S. (1975). Indian J . Genet. Plant Breed. 35, 470-471. Minocha, J. L . , Sharma, H. L., Sidhu, J. S . , and Gill, B. S. (1976). Indian J . Genet. Plant Breed. 36, 38-43. Muldoon, D. K . , and Pearson, C. J. (1977). Aust. J . Exp. Agric. Anim. Husb. 17, 277-283. Muldoon, D. K . , and Pearson, C. J. (1979). Herb. Absrr. 49, 189-199. Miintzing, A. (1951). Hereditas 27, 273-308. Miintzing, A. (1954). Proc. Indian Acad. Sci. Ser. E 34, 227-241. Miintzing, A. (1958). Trans. Bose. Res. Inst. (Calcuttu) 22, 1-15. Miintzing, A. (1974). Annu. Rev. Genet. 8, 243-266. Murthy, J. V. V. S. N., Sukhadev, P.,Rao, M. V . S., and Manga, V. (1979). Indian J . Eot. 2, 70-72. Nair, M. K., Raman, V. S . , and Ponnaiya, B. W. X . (1964). Madras Agric. J . 51, 356-357. Nishiyama, I . , and Kondo, N. (1942). Seiken Ziho 1, 26-28. [Japanese] Olorode, 0. (1974). Cytologia 39, 429-435. Olorode, 0. (1975). Brirtonia 27, 63-68. Pantulu, J. V. (1958). Curr. Sci. 27, 497-498. Pantulu, J. V. (1960). Curr. Sci. 29, 28-29. Pantulu, J. V. (1967a). Nature (London) 213, 101-102. Pantulu, J. V. (1967b). C.vtologia 32, 532-541. Pantulu, J. V. (1968). Curyologia 21, 11-15. Pantulu, J. V., and Manga, V. (1969). Curr. Sci. 38, 143-144. Pantulu, J. V., and Manga, V. (1972). Theor. Appl. Genet. 42, 68-74. Pantulu, J. V., and Manga, V. (1975). Genetica 45, 237-251. Pantulu, J. V., and Rao, M. V. S. (1976). Curr. Sci. 45, 418-420. Pantulu, J. V., and Rao, G. J. N. (1977a). Curr. Sci. 46, 390-392. Pantulu, J. V., and Rao, G. J . N. (1977b). Proc. Indian Acad. Sci. Ser. E 86, 15-22. Pantulu, J. V., and Venkateswarlu, J . (1968). Genetica 39, 41-44. Pantulu, J. V., Manga, V . , and Rao, M. V. S. (1976). Theor. Appl. Genet. 47, 85-86. Patil, B. D. (1963). Indian Farming 12(11), 20, 23. Patil, B. D., and Singh, A . (1964). Curr. Sci. 33, 255. Patil, B. D., and Vohra, S . K. (1962). Curr. Sci. 31, 345-346. Patil, B. D., Hardas, M. W., and Joshi, A. B. (1961). Nature (London) 189, 419-420. Person, C. (1955). Can. J . Eot. 33, 11-30. Pilger, R. (1954). Eor. Jahrb. 76, 281-384. Powell, J. B., and Burton, G . W. (1966a). Crop Sci. 6, 131-134. Powell, J. B., and Burton, G. W. (1966b). Crop Sci. 6, 378-379. Powell, J. B., and Burton, G. W. (1968). Crop Sci. 8, 771-773. Powell, J. B . , and Burton, G. W. (1969). Crop Sci. 9, 252-253. Powell, J. B., Hanna, W. W., and Burton, G . W. (1975). Crop Sci. 15, 389-392. Pritchard, A. J. (1971). Trop. Grassl. 6, 35-39. Punyasingh, K. (1947). Genetics 32, 541-554. Rajhathy, T., and Fedak, C. (1970). Can. J . Genet. Cytol. 12, 358-360. Ramage, R. T. (1964). Int. Barley Genet. Symp., Ist, Wageningen, pp. 99-115.

478

PREM P. JAUHAR

Ramage, R. T. (1970). Proc. Inr. Barley Genet. Symp.. 2nd. Pullman, Wash. pp. 89-93. Ramage. R. T . , and Wiebe, G. A. (1969). Proc. IAEA F A 0 S v r p . , Pullman, Wash. pp. 655-659. Raman, V. S. (1965). I n “Advances in Agricultural Sciences and Their Applications” (S. Krishnamurthi, ed.), pp. 122-143. Agric. Coll. Res. Inst., Coimbatore, India. Raman, V. S., and Krishnaswami, D. (1960). J . Indian Bot. Soc. 39, 382-385. Raman, V . S . , and Krishnaswami, D. (1961). J . Indian Bot. Soc. 40, 61-65. Raman, V . S., and Nair, M. K. (1964). J . Indian Bor. SOC. 63, 508-514. Raman, V . S., Krishnaswami, D.. and Nair, M. K. (1962). J . Indian Bor. SOC. 41, 268-270. Raman, V . S . , Nair, M. K . , and Krishnaswami, D. (1963). J . Indian Bor. SOC. 62, 469-473. Ramulu, K. S. (1968). Caryologia 21, 147-156. Ramulu, K. S. (1971). Cytologia 36, 652-668. Ramulu, K. S . , and Ponnaiya, B. W . X . (1967). Madras Agric. J . 54, 503-51 I . Rangaswamy, K. (1935). J . Indian Bar. Soc. 14, 125-131. Rangaswamy, S. R . S. (1972). Generica 43, 257-273. Rangaswamy, S. R. S., and Ponnaiya, B. W. X. (1963). Madras Agric. J . 50, 239-241. Rao, M. K., and Koduru, P. R. K. (1978a). Theor. Appl. Genet. 53, 1-7. Rao, M. K., and Koduru, P. R. K. (1978b). Proc. Indian Arad. Sci. Ser. B 87, 29-35. Rao, M. K., Pantulu, J. V., and Jayalakshmi, K. (1979). Experierzria 35, 1154. Rao, M. V. S. (1978). Proc. Indian Acad. Sci. Ser. B 87, 17-22. Rao, M. V. S. (1980). Theor. Appl. Genet. 56, 85-89. Rao, M. V . S . , and Pantulu, J. V. (1978). Chromosoma 69, 121-130. Rao, P. K., Nambiar, A. K., and Munhy, 1. V. G. K. (1949). Madras Agric. J . 36, 526-529. Rao, P. S. R . L. N. (1977). Curr. Sci. 46, 464. Rao, P. S. R. L. N., and Rao, I. N. (1977). Curr. Sci. 46, 314-315. Rao, P. S. R. L. N., Rao, M. V. S . . and Rao, 1. N. (1977). Experientia 33, 308-309. Rau, N . S. (1929). J . Indian Bor. SOC. 8, 126-128. Read, J. C . , and Bashaw, E. C. (1974). Crop Sci. 14, 401-403. Rees, H. (1961a). Bor. Rev. 27, 288-318. Rees, H . (1961b). Evolurion 15, 145-152. Rees. H.. and Jones, R. N . (1977). “Chromosome Genetics.” Univ. Park Press, Baltimore, Maryland. Rhoades, M. M., and Dempsey, E. (1972). Generics 71, 73-96. Rick, C. M., and Barton, D. W. (1954). Generics 39, 640-666. Riley, R. (1960). Nature (London) 185, 751-752. Riley, R., and Chapman, V. (1957). Heredify 11, 195-207. Riley, R., and Law, C. N. (1965). Adv. Genet. 13, 57-1 14. Sadasivaiah, R . S . , and Kasha, K. J. (1971). Chromosorna 35, 247-263. Sethi. G. S . , Kalia, H. R., and Ghai, B. S. (1970). Cyrologia 35, 96-101. Singh, R. B., Singh, B. D., Laxmi, V., and Singh, R . M. (1977a). Cyrologia 42, 41-47. Singh, R. B.. Singh, B. D., Singh, R. M., and Laxmi, V . (1977b). Cyrologia 42, 633-637. Srivastava, H . K., and Balyan, H. S. (1977). J . Hered. 68, 338-340. Stapf. O., and Hubbard, C. E. (1933). Kew Bull. pp. 269-302. Stapf, O., and Hubbard, C. E. (1934). I n “Flora of Tropical Africa” (D. Prain, ed.), Vol. 9, Part 6, pp. 954-1070, Reeve, Ashford, Kent, England. Stebbins, G . L. (1956a). Brookhaven Symp. Biol. 9, 27-52. Stebbins, G. L. (1956b). Am. J . Bot. 43, 890-905. Stebbins, G. L. (1958). Cold Spring Harbor Symp. Quant. Biol. 23, 365-378. Stebbins, G. L. (1969). Taxon 18, 26-35. Stebbins, G. L. (1974). “Flowering Plants: Evolution above the Species Level.” The Belknap Press of Harvard Univ. Press, Cambridge, Massachusetts.

CYTOGENETICS OF PEARL MILLET

479

Sun, S., and Rees, H. (1967). Heredity 22, 249-254. Taliaferro, C. M., and Bashaw, E. C. (1966). Crop Sci. 6, 473-476. Thevenin. L. (1952). Ann. C. R . A . Bambey (Senegal) 8, 102-105. [French] Thompson, J. B. (1956). Heredity 10, 99-108. Ting. Y . C. (1971). Genetics 68, Suppl., 67 (Abstr.) Ting, Y. C. (1973). Cytologia 38, 497-500. Tyagi, B. R. ( I 975a). Proc. Indian Natl. Sci. Acad. Ser. B 41, 462-465. Tyagi, 9.R. (1975b). Proc. Indian Natl. Sci. Acad. Ser. B 41, 545-549. Tyagi, 9.R. (1976a). Nucleus 19, 58-63. Tyagi, B. R. (1976b). Proc. Indian Natl. Sci. Acad. Ser. B 42, 117-121. Tyagi, 9.R. (1976~).Nucleus 19, 112-114. Tyagi, B. R. (1977). Proc. Indian Natl. Sci. Acad. Ser. B 43, 66-69. Tyagi, 9.R., and Singh, R. B. (1974). Proc. Indian Natl. Sci. Acad. Ser. B 40, 359-362. Venkateswarlu, J.. and Mani, J . N. R. (1978). Genetica 48, 145-149. Venkateswarlu, J., and Pantulu, J. V. (1968). J . Hered. 59, 69-70. Venkateswarlu, J., and Pantulu, J. V. (1970). Cytologia 35, 444-448. Veyret, Y. (1957). Agron. Trop. 12, 595-598. [French] Virmani, S. S., and Gill, B. S. (1971). Caryologia 24, 427-433. Virmani, S. S., and Gill, B. S. (1972). Cytologia 37, 257-260. Vosa, C. G. (1970). Chromosoma, 30, 366-372. Vosa, C. G. (1973). Chromosoma, 43, 269-278. Vosa, C. G. (1975). Curr. Adv. Plant Sci. 6, 495-510. Vosa, C. G., and Marchi, P. (1972). Nature (London), New Biol. 237, 191-192. Weber, D. F. (1970). Maize Genet. Coop. Newslett. 44, 203. Weber, D. F. (1973). Theor. Appl. Genet. 43, 167-173. Whyte, R. 0. (1957). "The Grassland and Fodder Resources of India." Scientific Monograph No. 22, Indian Council of Agricultural Research, New Delhi. Zeller, F. J., Kimber, G., and Gill, 9.S. (1977). Chromosoma 62, 279-289.

This Page Intentionally Left Blank

Index A

Betulu. 205

Blight, sclerotina, 48 Blotch, Webb, 48 Bluegrass, Kentucky, 27 Borer, lesser cornstalk, 49 Bruchiuria decumbens, 302-307.3 18.322-325, 341, 350, 360, 365-366, 377, 378, 385, 387, 395 Bruchiarim humidicola, 305, 307, 333, 350, 385 Brassicu chinesis, 3 Brassica juncea, 97 Bmssica nupobrussica. 98 Brassicu oleracea var. Botrytis, 98, 102 var. cupitatu, 102 var. gemmiferu, 77, 98, 102 var. itulicu. 98, 102 Brassicu pekinesis, 3 Brazil nut, 301 Broccoli, 98, 102 Brussels sprout, 77, 97, 98, 102 Buffelgrass, 47 I

Acidity management, 334-354 Aegilops squurrosu. 146 Aflatoxin, 47 Agropyron gluucum. 3 Alfalfa, 3, 27, 28. 97. 98, 102, 205, 208 Aluminum management, 334-35 I tolerance, 7-96-304 Anunus comosus, 300 Andropogon gayunus, 303-307,323, 341, 350, 358, 377. 378, 385. 395 Anther culture, 2-7 Aphis craccivoru, 47 Arachis cardenusii, 46, 47 Arachis chacoense. 46-47 Arachis correntina, 46 Arachis duranensis. 46 Arachis hypogea, 3 I , 33-34, 46. 47, 297 Armyworm, fall, 49 Artocarpus heteroplyllus, 300 Ash, nutrient, 308-310 Asparagus, 132 Aspergillus glavus. 47 Aspergillus niger, 85 Avena sativa, 86, 205 Averrhou carumbolu. 300

C Cabbage, 102 Chinese, 3 Cujanus cajan, 298, 33 I

Calcium, 338-346 B

Calopogonium mucunoides, 305

Canary grass, reed, 30 Banana, 300 Barley, 95, 102, 104, 134. 136, 140, 205 Bean, 76, 102, 324, 375 fava, 33 field, 226 h a , 298 mung, 298 navy, 263 winged, 298 Beet, 102 sugar, 3, 103, 134, 205 Berseem, 97 Bertholletiu excelsu. 30 I Beta vulguris. 3, 102. 103, 134, 205

Capiscum unnun. 3

Carambola, 300 Curicu. 101

Cashew, 300, 341 Cassava, 296, 310, 318, 324, 330-332, 336, 341, 378, 389 Cauliflower, 96, 98 Cenchrus ciliaris. 302, 47 1 Centrosemu hybrid, 350 Centrosema plumier;, 303 Centrosemapubescens, 305. 352-353,358,378 Cercosporu aruchidicola, 46-47 Cercosporidium personatum, 46-47 Chick-pea, 31 48 I

482

INDEX

Chicorium intybus. 29 Chicory, 29 Chloris gayunu. 378 Citrus uuruntiifoliu. 300 Citrus microcurpa, 3 Cirrus purudisi. 300 Citrus sinensis. 1 I , 300 Clover, 22 alsike, 28, 29 crimson, 28, 29 Egyptian, 28, 29 ladino, 28, 29 red, 27, 28, 29, 98, 102 subterranean, 16, 28, 29, 85, 86 white, 27, 28, 208 Cocoa. 301, 389 Coconut, 300 Cocos nuciferu. 300 Coffea urubica. 301 Coffee, 300-301 Collectotrichum gloesporoides. 306 Colocusiu esculentu. 33 1 Copper, 105-107, 205-206 Cordiu uleodora, 333 Corn.4, 1 1 , 102,299,310,311,327-332,341, 356-358, 365, 370, 374-375 see also Maize Cotton, 136, 140 Cowpea, 17, 34, 297, 328, 331, 332, 341. 382 Cucurbita maxima, 136 Cylindrocludium crotulariue, 47 Cytology, regenerated plants, 6-7

E Elueis quineensis. 132, 136, 301 Elusmopalpus lignosellus. 49 Empouscu fubae, 48 Eucalyptus grandiflora, 30 I

F

Flax, 3, 81, 214 Forest, clearing, 308-3 19

G Galuctiu striatu, 305 Genetics, chloroplast, 119-130, 139-142, 149-15 I intergenomic interaction, 1 17- 195 mitochondria, 119-1 39 peanut breeding, 39-72 pearl millet cytogenetics, 407-479 Glycine mux. 3, 30, 136, 140, 299 Glycine wightii, 352, 378 Gmelina arboreu, 301, 333 Gossypium hirsutum, 136, 140 Granadilla, 300 Grapefruit, 300 Grass. pasture, 103 Griselinia littoralis, 2 1 I Guarani, 301, 332 Guava, 300 Guilielma gusipaes. 301

D Dalbergia nigru, 301, 333 Deois incompleta. 307 Desmodium gyroides, 305, 377 Desmodium heterophyllum, 305, 350 Desmodium ovulifolium, 303, 305, 322-326, 333, 341, 350, 377, 385 Desmodium scorpiurus. 377 Desmodium uncinutum, 352 Diabrotica undecimpunctutu howurdi, 48 Digitariu decumbens, 303 Dioscorea, 33 1 Diplodia gossypina, 48

H Heterosis, 131-142, 164-174 Heveu brusiliensis, 3, 4, 301 Hordeurn vulgure, 95, 102, 136, 205 Hypurrheniu rufu, 302-305

I Intercropping, 330-333 lpomoeu bututus. 299

INDEX J

Jacaranda, 301 Jackfruit, 300

K Kudzu, 329

L Lac.tucu sutivu. I03 Laurel, 333 Lead, 215 Leafhopper, potato. 42, 48 Leafspot, cercospora, resistance, 42, 46-47 Legume, nitrogen fixation, 15-38 pasture, 103 Lentil, 31 Lespedeza, Korean, 27, 28 Lespedeza stipulaceu. 21, 28 Lettuce, 103 Leuruenu Ieucocephulu. 30 I, 350, 352, 378 Lime, 287-289. 300 Liming, 95-96, 303, 334-346, 368, 372 Linum usitutissimum, 3, 8 1 Lolium multiflorum, 205 Lolium perenne. 85 Lotononis buinesii. 352 Lupine, 3 I , 34, 265 Lupinus ulbus. 265 Lycopersicon c’scutentum. 103, 140, 205

M Mucroptilium. 350, 377 Macropriliurn utropurpureum. 352-353 Macroptilium lurhvroides. 352 Magnesium, 208-209, 338-346 Maize, 3, 76. 80, 86, 132, 133, 136, 140 see also Corn Malanga, 331 Manganese tolerance, 351 -553 Mango, 300, 341 Manguiferu indita. 300 Munihot esculentu. 296

483

Manure, green, 329 Medicugo, 205 Medicugo donriculutcr. 3 Medicago sutivu, 27, 28, 98, 102, 352, 378 Melilotus ulbu. 27, 28 Me/inis minutiflora. 305, 327, 350, 378 Meloidogyne urenuriu, 48 Meloidogyne hapla. 48 Millet, pearl, cytogenetics, 407-479 Mite, two-spotted, 49 Molybdeqosis, 105 Molybdenum, 206, 210 determination of, 8 1-85 soil, plant, and animal, 73-115 Mulch, 326-329 Muso purudisiacu. 297 Musu supiensis, 300 Mutant selection, 9-1 1 Mycorrhizae, 2 10-2 13. 376-379

N Napier grass. 414, 417 Nematode, lesion, 48 root-knot, 48 Nicotiunu ulutu. 8 Nicotiunu rusricu. 8 Nicotiunu tubacum, 8, 9, 103 Nitrate reductase, 78-80 Nitrogen fertilizer, 382-383 fixation, 15-38, 380-382 Nutrient recycling, 386 Nutrition studies, dilution effect, 197-224

0 Oat, 86, 205 Orange, 300 sweet, 3 Oryzu sutivcr. 2, 80, 87, 298 Oxisol, management, 279-406

P Palm, oil, 132, 136, 300, 301, 332, 389 peach, 301

484

INDEX

Panax notoginseng, I0

Poplar, 3

Pangola grass, 304

Populus nigra. 3

Panicum maximum, 103, 302-304, 310, 318.

Potassium, 208, 383-384 323-329, 341, 350, 358-359, 377, 378, Potato, I I , 136, 292, 297, 298 385-388 sweet, 299, 332 Papaya, 101 Pratylenrhus braehyurus, 48 Paspalum dilatutum. 378 Protoplast fusion, 154- 157 Paspalum notatum, 305 isolation, 7-9 Passijora edulis. 300 Psidium guajava. 300 Pasture, grass, 103 Psophocarpus tetragonolobus, 298 legume, 103 Puccinia arachidis. 45. 46 savanna, 32 1-325 Pueraria phaseoloides, 303, 305, 322-323. tropical, 301-307 329, 333, 341, 377-379, 389 Paullinia cupana. 301 Punica granutum. 300 Pea, 17, 77, 136, 140 Pythium, 48 leafless plant, 225-277 pigeon, 298, 331 Peanut, 33-34, 297, 328, 332. 341 R breeding, 39-72 Peat scours, 99, 105 Raya, 97 Pennisetum clandestinum. 378 Rehmannia glutinosa, 3 Pennisetum orientale, 426, 468-470 Rhizobium. 85, 86 Pennisetum purpureum, 285, 305, 341, 350, Rice, 2-3, 9, 80, 87, 298, 310, 319, 328-332, 41 I , 414-417, 426, 459-467 341, 347, 348, 369, 373-374 Pennisetum ramosum, 4 10 Rootworm, southern corn, 48 Pennisetum setaceum, 47 I Rot, collar, 48 Pennisetum syuamulatum, 467 cylindrocladium black, 47-48 Pennisetum typhoides, 408-411, 426, 459 southern stem, 48 Pepper, 3 Rubber, 300, 301. 332-333 black, 301 Rubber tree, 3, 4-5 Phalaris aquatica, 87 Rust, peanut, 45 Phularis arundinacea, 30 Rutabaga, 98 Phalaris tuberosa. 85, 87 Rye, 136, 140 Phaseolus, 17, 76 Ryegrass, 205 Phaseolus lunatus, 298 perennial, 29, 85 Phaseolus vulgaris. 31, 34, 91, 102, 263, 265, 286, 297, 299, 325 Phleum pratense, I03 S Phoma arachidicola. 48 Phosphorus. 96, 208-210, 215, 304-305 Saccharum officinarum, 301 management, 354-380 Sclerotinia sclerotiorum, 48 Photosynthesis, efficiency, 140-142, 179-1 82 Sclerotium rovsii, 47 Pineapple, 300, 341 Secale cercale. 136, 140 Pinus caribea, 301 Seopolia ocutagula, I0 Piper nigrum, 301 Slash-and-burn, 308-3 I7 Pisum sativum. 3 I , 34, 77. 136, 140, 225, 265 Soil acidity, 209-210 Plantain, 297, 330-331, 341 tropical management, 279-406 P o a pratensis, 27 Solanum tuberosum. 136, 298 Pod breakdown, 48 Sorghum, 207 Pomegranate, 300 grain, 299, 303

485

INDEX Sorghum hicolor, 299 Sorghum vulgurr. 3 Soybean, 3.9, 17,21,30-33,77,97, 132, 136. 140, 263, 299, 310, 328-332, 338, 348-349, 382 Spodopteru frugiperdu. 49 Sterility, cytoplasmic male, 142-145 Streptomyces scorhies, 298 Stylosunthes cupitutu. 303-305, 341, 350. 377 385 Stylosunthes guiunensis. 305, 307, 324, 350, 353, 377 Stylosunthes humutu, 378 Stylosanthes humilis. 352. 378 Stylosunthes scuhru. 305 Stylosunthes r~iscosu.305 Sugar cane, I I , 300, 301 Sulfur, 97, 208, 384-385 Sweetclover. 27-29

U

Ultisol, management, 279-406

V Verticillium wilt, 48 Vetch, 27, 28, 29 Vicia fubu. 8, 31, 33, 226 Vicia villosa, 27, 28, 29 Vignu rudiuru, 298 Vignu unguiculutu. Virus, rosette, 47

w Water use, 213-214 Wheat, 2, 3, 29, 86-87, 103, 104, 134, 136, 140, 145-147, 213, 299-300, 347, 375 Wheatgrass. crested, 263

T Temperature, nutrient uptake, 214-215 Tetrunychus urticue. 49 Theohromu cacao, 30 I Thrip, tobacco, 49 Timothy, 103 Tissue culture advances, 1 - 13 Tobacco, 97, 103, 341 Tomato, 103, 140, 205, 263 Trifolium ulexundrinum, 28, 29, 97 Trifolium hyhridum. 28, 29 Trifolium incurnarum, 29 Trifolium indicu. 28 Trifolium prutense, 27, 28, 98, 102 Trifolium repens. 27. 28 Triticule. 3, 146 Triticum uestivurn. 3, 29, 86, 103. 136, 140, 299 Triticum dicoccoides. 146 Triticum machu, 146

X Xunthosomu, 331

Y

Yam, 331 Yautia, 331 Yield, economic, 252-264 improvement, 1 74- I82 maximum biological, 239-251, 262

2 Zeu mays, 3, 76, 102. 136, 238, 299 Zinc, 208 Zorniu lutifoliu. 303-306, 341, 377, 385

This Page Intentionally Left Blank