Advances in Agronomy, Volume 33

ADVANCES IN

AGRONOMY

VOLUME 33

CONTRIBUTORS TO THIS VOLUME R. W. ARNOLD F. H. BEINROTH R. J . BURESH F. B . CADY M. E. CASSELMAN MARLING . CLINE G . D. FARQUHAR

R. J . HAYNES ERNESTA. KIRKBY T. M. MCCALLA KONRADMENGEL W. H. PATRICK,JR. J . A . SILVA

H . T. STALKER G. UEHARA

P. W. UNGER P. D. WALTON R. WETSELAAR

ADVANCES IN

AGRONOMY Prepared in cooperation with the AMERICAN SOCIETY OF AGRONOMY

VOLUME 33 Edited by N. C . BRADY International Rice Research Institute Manila, Philippines

ADVISORY BOARD H. J . GORZ,CHAIRMAN R.B. GROSSMAN T. M. STARLING I . B. POWELL

J . W . BIGGAR

M. A. TABATABAI M. STELLY, EX

OFFICIO,

ASA Headquarters 1980

ACADEMIC PRESS A Subsidiary of Harcourr Brace Jovanovich, Publishers

New York London Toronto Sydney San Francisco

COPYRIGHT @ 1980, BY ACADEMIC PRESS, INC. ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER.

ACADEMIC PRESS,INC.

I I 1 Fifth Avenue, New York, New York 10003

United Kingdom Edition published by ACADEMIC PRESS, INC. ( L O N D O N ) LTD. 24/ 28 Oval Road, London NWI 7DX

LIBRARY OF

CONGRESS CATALOG CARD

NUMBER:50-5598

ISBN 0-12-000733-9 PRINTED IN THE UNITED STATES OF AMERICA 80 81 82 83

9 8 7 6 5 4 3 2 1

CONTENTS CONTIUBUTORS TO VOLUME 33 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PREFACE .....................................................

ix xi

CONSERVATION TILLAGE SYSTEMS

P. W. Unger and T . M . McCalla I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1 . Historical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III . Tillage Equipment and Use ............................... IV . Crop Yields and Quality ................................. V . Environmental Consideration ............................. VI . Infiltration and Water Conservation ........................ VII . Weed Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII . Insects and Plant Diseases ................................ IX . Soil Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . X . Soil Structure and Other Physical Properties . . . . . . . . . . . . . . . . . XI . Chemical Effects and Microbial Activity .................... XI1. Economics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XI11. Summary and Conclusions ............................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 5 6 13 16 30 35 38 39 43 48 49 51 53

POTASSIUM IN CROP PRODUCTION

Konrad Mengel and Ernest A . Kirkby

I. I1. 111. IV . V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Potassium Availability in the Soil .......................... Potassium in Physiology ................................. Potassium Application and Crop Growth .................... Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References ............................................

59 60 74 91 103 103

UTILIZATION OF WILD SPECIES FOR CROP IMPROVEMENT

H . T . Stalker I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Biosystematics ......................................... V

112 113

vi

CONTENTS

111. The Gap between Hybridization and Utilization . . . . . . . . . . . . . . IV . Approaches for Utilizing Wild Germplasm Resources . . . . . . . . . V . Examples of Species Used in Wild Species Hybridization Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Specific Uses of Wild Species for Crop Improvement . . . . . . . . . VII . Summary and Conclusions ............................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

118 119 126 135 140 141

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS. A REVIEW

R . J . Buresh. M . E . Casselman. and W . H . Patrick. Jr . I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Nitrogen Fixation in the Water Column and on the Soil Surface 111. Nitrogen Fixation in the Aerobic Layer of Flooded Soil . . . . . . . IV . Nitrogen Fixation in the Anaerobic Layer of Flooded Soil . . . . . V . Nitrogen Fixation in the Root Zone of Nonnodulated Plants . . . . VI . Nitrogen Fixation on the Leaf and Stem Surface of Plants . . . . . VII . Environmental Factors Influencing Nitrogen Fixation in Flooded Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII . Comparison of Acetylene Reduction and 15N Methodology . . . . IX * Contribution of Fixed Nitrogen to the Nitrogen Requirements of Plants ............................................ X . Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

150 152 160 161 163 170 174 180 183 185 187

EXPERIENCE WITH SOIL TAXONOMY OF THE UNITED STATES

Marlin G . Cline I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1 . General Reactions to the System .......................... 111. Use of Soil Taxonomy Internationally ...................... IV . Problems for Users of Soil Taxonomy ...................... V . Taxonomic Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Impact of Soil Taxonomy Internationally .................... VII . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

193 195 197 202 207 213 222 223

vii

CONTENTS COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

R . J . Haynes

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Competition in the Pasture Community ..................... 111. Physiological Considerations .............................. IV . Morphological Considerations ............................. V . Competition for Environmental Factors ..................... VI . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

227 229 231 239 248 254 256

NITROGEN LOSSES FROM TOPS OF PLANTS

R . Wetselaar and G . D . Farquhar I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

263 264 111. Possible Pathways of Nitrogen Losses from Tops . . . . . . . . . . . . . 279 IV . Associated Methodology Problems ......................... 293 V . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299

I1 . Record of Observed Decreases of Nitrogen Content of Plant Tops

AGROTECHNOLOGY TRANSFER IN THE TROPICS BASED ON SOIL TAXONOMY

F . H . Beinroth. G . Uehara. J . A . Silva. R . W . Arnold. and F . B . Cady

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

304 304 Soil Classification in Perspective .......................... 309 Agrotechnology Transference Research ..................... 316 Quantitative Verification of Transferability within a Soil Family . 323 Prerequisites for Worldwide Agrotechnology Transfer . . . . . . . . . 332 Conclusion: Implication for Agricultural Development . . . . . . . . . 336 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338

I1 . The Transfer of Agrotechnology ........................... 111.

IV . V. VI . VII .

THE PRODUCTION CHARACTERISTICS OF Bromus inerrnis LEYSS AND THEIR INHERITANCE

P . D . Walton

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. The Nature of the Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

341 343

viii

CONTENTS

III . Seed Production and Establishment ........................ IV . Forage Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V . Forage Yield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Plant Breeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

347 350 353 363 367

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

371

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

R. W. ARNOLD (303), Department of Agronomy, New York State College of Agriculture and Life Sciences, Cornell University, lthaca, New York 14853 F. H. BEINROTH (303), Department of Agronomy and Soils, College of Agricultural Sciences, University of Puerto Rico, Mayaguez, Puerto Rico 00708 R. J . BURESH* (149), Laboratory for Wetland Soils and Sediments, Center for Wetland Resources, Louisiana State University, Baton Rouge, Louisiana 70803 F. B. CADY (303), Biometrics Unit, Department of Plant Breeding and Biometry, New York State College of Agriculture and Life Sciences, Cornell University, lthaca, New York 14853 M. E. CASSELMAN (149), Laboratory for Wetland Soils and Sediments, Center for Wetland Resources, Louisiana State University, Baton Rouge, Louisiana 70803 MARLIN G . CLINE ( 1 93), Department of Agronomy, Cornell University, lthaca, New York 14853 G. D. FARQUHAR (263), Department of Environmental Biology, Research School of Biological Sciences, Australian National University, P.O. Box 47.5, Canberra City, Australia 2601 R . J . HAYNES (227), Department of Soil Science, Lincoln College, Canterbury, New Zealand ERNEST A. KIRKBY (59), Department of Plant Sciences, The University, Leeds LS2 9JT, England T. M. MCCALLA ( l ) , Agricultural Research, Science and Education Administration, USDA, University of Nebraska, Lincoln, Nebraska 68583 KONRAD MENGEL (59), Institute of Plant Nutrition, Justus Liebig University, 0-6300 Giessen, Federal Republic of Germany W. H. PATRICK, JR. (149), Laboratory for Wetland Soils and Sediments, Center for Wetland Resources, Louisiana State University, Baton Rouge, Louisiana 70803 J . A. SILVA (303), Department of Agronomy and Soil Science, College of Tropical Agriculture and Human Resources, University of Hawaii, Honolulu, Hawaii 96822 H . T. STALKER ( 1 1 l), Department of Crop Science, North Carolina State University, Raleigh, North Carolina 27650 *Resent address: International Fertilizer Development Center, P.O. Box 2040, Muscle Shoals, Alabama 35660. iX

X

CONTRIBUTORS

G . UEHARA (303), Department of Agronomy and Soil Science, College of Tropical Agriculture and Human Resources, University of Hawaii, Honolulu, Hawaii 96822 P . W. UNGER ( l ) , Conservation and Production Laboratory, Agricultural Research, Science and Education Administration, USDA, Bushland, Texas 79012 P. D. WALTON (341), Department of Plant Science, The University of Alberta, Edmonton, Alberta T6G 2E3, Canada R. WETSELAAR (263), Division of Land Use Research, Commonwealth Scientific and Industrial Research Organization, P.O. Box 1666, Canberra City, Australia 2601

PREFACE The orderly classification of soils in the field is essential to effective utilization of much production-oriented soil and crop research. Unfortunately, however, there is all too little exchange of information among scientists concerned with soil classification. Many soil classification schemes are sufficiently complicated as to make them not easily understood by most agronomists. This has become increasingly evident with the development in recent years of comprehensive international soil classification schemes. Two of the articles in this volume relate to the classification of soils and to the usefulness to agronomists of classification schemes. Cline reviews agronomists’ experiences with the most comprehensive of the international classification systems, that developed under the leadership of the U.S. Department of Agriculture. The views presented are not only those of other pedologists, but also those of production-oriented scientists and others who have used the new comprehensive soil classification scheme. Beinroth and coauthors relate the findings of soil and crop scientists who have compared the performance of soils that have similar characteristics but are located in different parts of the world. This information is useful in ascertaining the value of soil survey in agrotechnology transfer. Nitrogen continues to be a prominent subject for agronomic research. The fixation of nitrogen in flooded systems is reviewed by Buresh, Casselman, and Patrick. They emphasize the uniqueness of flooded systems in relation both to redox potential and to nitrogen-fixing organisms. High nitrogen losses directly from plants as they mature may account for much of the low rate of utilization of applied nitrogen. Wetselaar and Farquharreview this subject in their contribution. Potassium in crop production has received prominent research attention in recent years, especially in relation to methods of predicting response to this important element. Mengel and Kirkby provide an excellent review of research on potassium availability in the soil and the function of potassium in the plant. The most significant recent change in soil and crop management in the United States has been in tillage and crop residue management. Over wide areas, tillage systems that keep most of the crop residues on or near the soil surface have replaced conventional systems that focused on the moldboard plow. A comprehensive review of the effects of some of these new tillage systems is presented in the article by Unger and McCalla. In recent years, scientists have become increasingly successful in implementing crosses between cultivated plants and wild species. These crosses provide considerable potential to incorporate into cultivated plants such desired characteristics as disease and insect pest resistance and tolerance of drought. Stalker reviews research on this subject. Research related to the production of forage crops is the focus of two of the xi

xii

PREFACE

manuscripts in this volume. Haynes reviews competitive aspects of grass-legume associations and illustrates how this competition accounts for the dominance of either grasses or legumes, depending on the ecological stresses. The characteristics of Bromus inermis Leyss that control its production and inheritance are the subjects of the review by Walton. The authors of the contributions presented in this volume are from six different countries. We express sincere appreciation to them for their efforts. N. C . BRADY

ADVANCES IN

AGRONOMY

VOLUME 33

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY, VOL. 33

CONSERVATION TlLLAGE SYSTEMS P. W. Unger2 and T. M. McCalla3 Agricultural Research, Science and EducationAdministration,U.S. Departmentof Agriculture

I. Introduction . . . . . . , . . . . . . . . . . . . . . . . . . ............................... A. Definition of Conservation Tillage Syst ............................... B. Development and Use of Practices in the United States .... C. Purpose of Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Historical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Tillage Equipment and Use ......................................... A. Machinery Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Seedbed Preparation and Crop Seeding . . . . IV. Crop Yields and Quality . . . . . . . . . . . . . . . . . . . A. Grain Yields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Plant Protein Content and Mineral Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Residue Yields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Root Growth V. Environmental Consideration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Control of Wind Erosion . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Control of Water Erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Runoff Water Quality . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Infiltration and Water Conservation .......... A. Runoff and Infiltration.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Evaporation .................... VII. Weed Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Problem Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Control with Tillage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Control with Herbicides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Control with Rotations ....................................... VIII. Insects and Plant Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Plant Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX. Soil Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Residue Factors Involved . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Biological Effects of Residue . .

2 2 2 5 5 6 6 10 13 13 14 15 16 16 17 21 29 30 30 33 35 35 36 37 38 38 38 39 39 39

40 42

'Contribution from Agricultural Research, Science and Education Administration, U.S. Department of Agriculture, in cooperation with the Texas and Nebraska Agricultural Experiment Stations. zSoil Scientist, Conservation and Production Research Laboratory, Bushland, Texas. 3Microbiologist, University of Nebraska, Lincoln, Nebraska. 1 ISBN 0-12-W733-9

2

P. W . UNGER AND T. M. McCALLA

X. Soil Structure and Other Physical Properties .................................. ........ A. Aggregation .......... B. Porosity and Density .................................................. C. Other Physical Properties . . ................. XI. Chemical Effects and Microbial XII. Economics . . . . . . . . . . . . . . XIII. Summary and Conclusions ................................................ A. Accomplishments ..... ..... B . Needs .............................................................. References . . . .................................................

43 44 46 46

51 51 52 53

1. INTRODUCTION A. DEFINITION OF CONSERVATION TILLAGE SYSTEMS

Conservation tillage systems, as used in this review, are systems of managing crop residue on the soil surface with minimum or no tillage. The systems are frequently referred to as stubble mulching, ecofallow, limited tillage, reduced tillage, minimum tillage, no-tillage, and direct drill. The goal of these systems of plant residue management is threefold: to leave enough plant residue on the soil surface at all times for water and wind erosion control, to reduce energy use, and to conserve soil and water. These systems are used throughout the United States and the world, and can be applied to all kinds of crop residue in many cropping systems. B. DEVELOPMENT A N D USE OF PRACTICES IN THE UNITEDSTATES

Stubble mulching was developed as a result of severe wind erosion in the Great Plains of the United States and Canada during the 1930s. Anchored surface residue kept the soil in place despite the erratic climate of the Great Plains. Crop residue on the soil surface was soon found to be equally effective for controlling water erosion. A surface mulch of plant residue protects the soil against the beating action of raindrops and keeps the surface of the soil open, thus increasing infiltration over that of a bare soil. When enough residue is present, more water is conserved with a mulch system than with the moldboard plow system. Since the 1930s and 1940s, many effective herbicides have come on the market, which reduced the need for tillage to control weeds. Even though the use of crop residue on the soil surface has much merit in controlling soil erosion and conserving water, use of residue by farmers depends in the final analysis upon the effects of surface residue on crop yields. Crop yields are often reduced where plant residue is maintained on the soil surface,

3

CONSERVATION TILLAGE SYSTEMS

particularly on heavier textured soils in the more humid and northern areas of the United States. This apparently is due to factors such as ( 1 ) lack of proper equipment and knowledge of how to manage the residue with the equipment; (2) colder, wetter, and less aerated soil; (3) weed, insect, and disease problems; (4) lower nutrient availability, such as lower nitrate production; and ( 5 ) changes in the microbial status of the soil and the possible production of phytotoxic substances. In many areas, however, crop yields are often increased by plant residue left on the surface. In addition, residue use may keep a crop from being lost by wind erosion. Despite the lower yields and other problems sometimes encountered with the use of crop residue on the soil surface, these systems are useful to U.S. farmers. Demands for improved water quality of the nation’s streams and groundwater have also stimulated the use of crop residue on the soil surface. At present, over 28 million hectares (over 70 million acres) of land are cropped by minimum or no-tillage methods (Table I). Limited tillage, especially sod planting, is one method used in many parts of the United States. In some instances, the use of crop residue on the soil surface alone is not Table I 1978 /1979 No-Till Fanner Survey” No-tillageb State Alabama Alaska Arizona Arkansas California Colorado Connecticut Delaware Florida Georgia Hawaii Idaho Illinois Indiana Iowa Kansas Kentucky Louisiana Maine Maryland Massachusetts‘

Minimum tillage*

1978

I979

66,170 2,670 410 2,020 490 29,340 7,640 8 1,750 1,620 210,320 233,290 139,980 86,600 393,360 8 10

68,110 2,750 870

1.900

96,830 1,340

1978

50

2,140 490 32,120 1 1,530 108,050 2,830 215,300 259,000 146,090 95,100 470,250 810 2.060 112,490 1,380

83,930 1,850 120,190 109,210 390,570 832,980 770 123,030 48,260 8 17,480 1,093,080 2,040,470 698,250 3,001,360 3,862,810 800,650 236,540 158,570 4,330

1979 78,430 2,250 114,930 124.6 I0 410,700 866,370 1,210 133,310 54,230 997,570 1,097,940 2,246,050 696,070 3,338,730 4,091,460 849,980 238,570 182,790 4,130

Conventional tillageb 1978 1,473,130 3,840 425,740 2,715,170 1,603,680 1,571,230 19,790 55,940 490,810 806,560 113,560 1,333,470 6,433,210 3,155,870 5,665,720 4,634,970 410,280 1,458,720 71,310 226,840 18,490

1979 1,510,680 3,840 432,620 2,699,340 1,673,270 1,537,840 18,540 52,210 505,580 728,450 116,960 1,327,400 6,324,560 3,237,560 5,463,380 4,164,310 416,430 1,548,560 76,160 230,230 18,490 (continued)

4

P. W. UNGER AND T. M. McCALLA

Table I (continued) No-tillageb State Michigan Minnesota Mississippi Missouri Montana‘ Nebraska Nevada New Hampshire New Jersey New Mexico New York North Carolina North Dakota Ohio Oklahoma Oregon Pennsylvania Puerto Rico Rhode Island South Carolina South Dakota Tennessee Texas Utah Vermont Virginia Washington West Virginia Wisconsin Wyoming Totals

1978

1979

19,750 47,630 29,950 98,580 8,090 234,320 28,640 400 17,400 2,550 13,150 147,310 3,440 289,360 2,350 12,110 176,040 950 30 8,390 72,240 74,340 59,300 2,430 127,280 3,170 15,580 39,460 -

I 8,050 54,230 35,920 109,470 8,170 225,410 28.940 810 23,470 3,560 13,150 149,330 4,650 297,450 3,160 10,440 176,850 730 30 8,980 62,320 81,950 49,450 2,430 127,840 7,420 17,400 39,660 -

2,890,790 3,092,730

Minimum tillageb 1978

I979

428,980 1,848,640 131,530 1,474,960 263,050 2,613,920 38,360 2,140 8,090 78,310 263,050 255,970 1,297,860 40,470 235,610 127,770 69,130 110

728,250 1,398,220 220,960 607.600 79,930 1,010 157,710 234,690 1,210 359,570 1,210

437,070 1,893,970 148,120 1,445,890 264,470 2,577,090 39,290 2,140 16,190 90,450 265,070 258,800 1,421,690 40,470 239,580 129,830 62,160 I10 734,720 1,643,060 23 1,080 508,040 77,900 1,210 169,970 39 1,460 2,020 361,390 1,210

Conventional tillage* 1978 1,547,470 5,709,230 2,215,900 2,580,760 2,023,470 3,532,580 63,310 9.960 172,800 368,470 443,550 1,391,830 9,531,570 3,914,610 3,681.9 10 576,760 864,020 193,540 2,110 740,150 3,567,990 1,161,390 8.8 17,690 4 15,620 3 11,610 404,530 1,543,830 26,710 1,920,030 220,560

1979 1,541,080 5,679,890 2,199,620 3,831,550 2,058,880 3,146,900 63,770 10,560 157,430 360,180 441,520 1,375,800 9,409,360 3,906,520 3,678,070 566,070 897,610 173,080 2,110 753,540 3,268,310 1,197,090 7,868,000 428,170 31 1,210 402.470 1,382,810 22,260 1,943,100 273,570

27,392,940 28,983,820 90,962,280 89,374,030

“The source of information is the State Agronomists of the Soil Conservation Service. Values have been converted to hectares. Published with permission from No-Till Farmer, Inc. 61 I East Wells Street, Milwaukee, Wisconsin 53202. *Definitions of each of the tillage systems are as follows: No-tillage: Only the intermediate seed zone is prepared. Up to 25% of surface areacould be worked. Includes no-till, till-plant, chisel-plant, rotary strip tillage, and many other forms of conservation tillage and mulch tillage. Minimum tillage: Limited tillage, but the total field surface is still worked by tillage equipment. Conventional tillage: Where 100% of the topsoil is mixed or inverted by plowing, power tiller, or multiple diskings. ‘Estimates provided by Cooperative Extension Service Agronomists.

CONSERVATION TILLAGE SYSTEMS

5

enough for good soil conservation. In such cases, use of residue in combination with other proven mechanical and conservation cropping practices, such as terracing and contouring, is necessary for more effective soil and water conservation. C. PURFOSEOF REVIEW

McCalla and Army (1961) reviewed the use of crop residue on the soil surface, which at that time was called stubble mulching. Because of research since then and because new equipment and herbicides are available, an inventory of the available information developed since then was needed, along with an appraisal of the merits and faults of managing crop residue on the soil surface for control of water and wind erosion, conservation of water, production of crops, and conservation of energy. Although the use of crop residue on the soil surface has been widely researched, information is still scarce on many aspects of the system’s influence on the physical, chemical, and biological soil environment; on insect, disease, and weed control; on water conservation; and on crop production. Not enough information is available for full implementation of the system across the United States for various soils, climatic conditions, and crops. Improvements are needed in equipment for seeding and applying herbicides. Crop varieties developed specifically for residue management systems are badly needed. Also, the economics of systems involving crop residue need to be fully evaluated. This review is, in general, limited to the salient points that have been researched since 1961. We have not attempted to summarize all the literature pertaining to the use of crop residue on the soil surface. Since 1961, numerous papers have been published, and a number of symposia and reviews have been made concerning the use of crop residue on the surface. Some examples of the latter are American Society of Agronomy (1978), Soil Conservation Society of America ( 1973, 1977, 1979), Great Plains Agricultural Council ( 1962, 1968, 1976), Ohio State University (1972), Plant Protection Limited (1973, 1975), and USDA (1977). Only pertinent and illustrative data are used to describe the areas of application and merits and faults of crop residue on the soil surface. Where essential technical information is lacking, we have called the deficiency to the reader’s attention.

II. HISTORICAL The early work was reviewed by McCalla and Army (1961) and Jacks et al. (1953, and also in a number of the symposia and reviews cited earlier.

6

P. W. UNGER AND T.M. McCALLA

Mulches, to some degree, have been used in agriculture since man first began to cultivate crops. Orchards and garden crops were probably the first to be mulched, and mulches usually were several centimeters thick. The Chinese, many hundreds of years ago, used stone mulches in their agriculture. Also, most primitive agriculture allowed residue to remain on the surface because the simple tools, in most cases, did not cover the plant residue. Crop residue is commonly used as mulches; but paper, plastics, glass wool, cloth, metal foil, sugarcane trash, manures, leaves, peat, litter, stones, and dust mulches have also been used. Natural mulches are snow and volcanic dust. Snow mulch is valuable in protecting a crop such as wheat against winterkill. Snow also is a major source of water for crop production in many cold regions. Dr. F. L. Duley and Professor J. C. Russel conducted the first intensive research in the United States on the use of a mulch for crops. The work was started at Lincoln, Nebraska, in 1937 by the Nebraska Agricultural Experiment Station in cooperation with the Research Division of the Soil Conservation Service, U.S. Department of Agriculture. Since that time, these and many other researchers have studied the management of crop residue on the soil surface. This research has resulted in the development of the crop residue management systems that are now known as conservation tillage systems.

Ill. TILLAGE EQUIPMENT AND USE A . MACHINERY REQUIREMENTS

Regardless of cropping system, a complement of machinery is needed to prepare satisfactorily a seedbed, plant seeds, and control weeds and volunteer crop plants. Because conservation tillage systems rely heavily on surface residue for erosion control and water conservation, it is imperative that the machinery be capable of operating satisfactorily when large amounts of residue are on the soil surface and that most residue be kept on the surface. Tillage systems developed within the last half century are capable of retaining most crop residue on the soil surface. These systems are the stubble mulch system, developed in the late 1930s and early 1940s, and the reduced- or notillage systems, which are essentially still under development. 1 . Stubble Mulch Tillage

Stubble mulch farming is a year-round system of managing plant residue. Stubble mulch tillage is performed with implements that undercut residue, loosen soil, and kill weeds. Because the soil is tilled as often as necessary to control weeds during the period between crops, the stubble mulch system is a tillage-

7

CONSERVATION TILLAGE SYSTEMS

intensive system that requires frequent operations to control weeds. The system was developed primarily for wheat (Triticum aestivum L.) and other small grain crops, but is also adaptable to such crops as sorghum (Sorghum sp.). Good management of a stubble mulch farming system begins with harvest. To minimize tillage problems, crop residue should be spread uniformly by the combine. In the Great Plains, sweeps or blades are generally operated at the 12- to 15-cm depth during the first operation after harvest and shallower during subsequent operations. Weed control generally is best when the soil is dry at the time of tillage. In the dry-farming areas of the Pacific Northwest and at more humid locations where straw production by small grain is usually higher than in the Great Plains, the first operation is similar to that in the Great Plains, but the second operation may be deeper than the first to avoid serious plugging of the equipment by the residue. When unusually large amounts of residue are present, a disk-type implement may be used for the first operation to incorporate some of the residue with soil. This hastens decomposition, but still keeps enough residue on the surface for erosion control. Other implements that may be used in heavy residue are stubble pulverizers or busters (Jacks et al., 1955) and skewtreaders or spike-tooth harrows in conjunction with one-way disk plows (Papendick and Miller, 1977). Tillage implements that maintain surface residue are (1) s w e e p 6 0 cm or wider; (2) rodweeders with semichisels or small sweeps; (3) straight-blade machines; (4) chisel plows; ( 5 ) one-way plows (which generally should be used only when large amounts of residue are present); and (6) rodweeders. The amount of residue remaining on the surface after one operation with several tillage machines is shown in Table 11. The power needed for stubble mulch tillage depends on such factors as the type and size of machine; the depth and speed of operation; and the texture, water content, and slope of soil. Promersberger and Pratt (1958) showed that Table I1 Effect of Tillage Machines on Surface Residue Remaining after Each Operationa Tillage machine Subsurface cultivators Wide-blade cultivator and rodweeder Mixing-type cultivators Heavy-duty cultivator and other type machines Mixing and inverting disk machines One-way flexible disk harrow, one-way disk, tandem disk, offset disk Inverting machines Moldboard and inclined disk plow a From

Anderson ( 1968).

Residue maintained (%)

90 15

50

10

8

P. W . UNGER AND T. M. McCALLA

Table I11 Measured Average Diesel Fuel Consumption for Specific Field Operations on Pullman Clay h m , Bushland, Texas"

Operation Dryland Sweep Sweep Surface-irrigated Moldboard plow Heavy tandem disk Heavy offset disk Lister bedder Disk bedder Rolling cultivator Chisel, 38-cm spacing Chisel, 50-cm spacing Chisel, 100-cm spacing Sweep-rodweed (bed-furrow cultivation) Seeding Grain drill, 25-cm spacing

Tillage depth (cm)

Diesel fuel (liter)

8

6.1 8.4

13

20-25 8-13 8-13

15-20 15-20 15-20

28.1 9.4 11.7 6.5 8.4 5.1 14.0-16.8 12.2 7.5 7.9 3.7

" From Allen er al. (1977). moldboard plowing and field cultivating (with sweeps) a clay soil at a depth of 13 to 18 cm required 21.4 to 23.2 kW hours/ha (1 1.6 to 12.6 hp hourdacre) and 3.3 to 10 kW hours/ha, respectively. Operating a Noble blade 15 to 27 cm deep required 10.7 to 13.3 kW hours/ha. Allen et al. (1977) reported fuel consumption values for performing various field operations, including some that are used in stubble mulch systems, on a clay loam soil (Table 111). Subsurface tillage generally required less power or fuel than disk or chisel tillage, and substantially less than moldboard plowing. Detailed descriptions of implements used in stubble mulch farming systems were given by Fenster (1960), FA0 (1971), and Jacks et al. (1955). Many types and sizes of equipment are available for doing a satisfactory job of stubble mulching. The essential part of any stubble mulching system is to maintain enough residue on the surface to control erosion adequately from harvest to harvest.

2 . Reduced- and No-Tillage Systems One of the major reasons for tilling a soil is to control weeds. Hence, if weeds are controlled by herbicides, the need for tillage is reduced. The development of

CONSERVATION TILLAGE SYSTEMS

9

effective herbicides in recent years has permitted the development of reduced- and no-tillage cropping systems. As with stubble mulch tillage, a major goal of these systems is to maintain crop residue on the surface for soil and water conservation. In some cases, however, the land is moldboard plowed, but the number of secondary operations is greatly reduced. The following are reduced- and no-tillage systems that have been evaluated in research trials and are currently used by some farmers. Additional information pertaining to these systems is given by Fisher and Lane (1973), Lewis (1973), Griffith et al. (1977), Amemiya (19771,Reicosky eral. (1977),Allen etal. (1980), andungerand Wiese(1979). a . Fall Plow, Field Cultivate. In this system, the moldboard plow is used for primary tillage, but secondary tillage is reduced to one shallow cultivation with sweeps at planting. A disk or rotary tiller may be substituted for the field cultivator to produce a finer, firmer seedbed, but it also leaves the soil more erodible. This system is widely used on the dark, nearly level, medium- and fine-textured clay loam soils of the east central Corn Belt. b. Spring Plow, Wheel-Track Plant. This system uses strip seedbed preparation on soil that was initially plowed 12 to 24 hours before planting. By planting soon after plowing, the soil water content is such that wheels break the clods and firm the seedbed. The planted rows may be in the tractor or planter wheel tracks. This system affords greater protection against erosion than fall plowing because crop residue is maintained on the surface until planting. c . Fall Chisel, Field Cultivate. This system is similar to the fall plow, field cultivate system, except that chiseling 20 to 25 cm deep replaces moldboard plowing. Chiseling retains more surface residue than moldboard plowing and, therefore, more effectively controls erosion. d . Disk and Plant. Tillage in this system is performed with standard tandem disks operated 8 to 10 cm deep, heavy disks operated 15 to 20 cm deep, or a combination of the two. The system usually includes one fall disking and one or more diskings before planting in spring. To conserve surface residue, disking should be delayed as long as feasible, and tandem disks rather than heavy disks should be used, because the heavy disks penetrate deeper and incorporate more residue than do tandem disks. e . Till-Plant. In this system, tillage and planting are done in one operation. Normally, tillage is with wide sweeps operated 5 to 8 cm deep on the ridgetop, which moves old stalks and root clumps into the area between rows and provides a trash-free zone for planting. The ridges were made the previous year during cultivation or after harvest with rolling or disk-hiller cultivators, large disk cultivators, or disk bedders (after harvest). Ridges may be re-formed annually in cases where heavy disks are used to cut residue and level old ridges, or they may be permanent, in which case the only tillage needed is for reshaping the ridges in the fall or spring with a rolling or disk-type cultivator. On soils in the southeast United States with compacted subsurface layers, two

10

P. W . UNGER AND T. M. McCALLA

machines have been developed to loosen the layers and plant seeds directly over the loosened zone. The first is the “ripper-hipper,” which subsoils the intended plant row and forms a ridge over the slit with hillers or bedders. Planting can be done in the same operation. The second machine is the subsoiler-planter, which has subsoilers to loosen the compacted layer, treading wheels to firm the loose soil in the slits, and flexible unit planters. Colters can be used with both machines to cut the surface residue. f . Strip Tillage. In a strip tillage system, only a narrow band of soil is tilled. Rotary tillers can be adapted for strip tillage by removing some of the knives. The tillage zone may be 20 cm wide and 5 to 10 cm deep. A standard planter can be attached to the tiller, resulting in a one-pass operation, because stalks can be chopped by the tiller. Another form of strip tillage is the opening of a slot for seed placement in previously untilled ground. The system, referred to as no-tillage, zero tillage, or slot planting, is used to plant in residue of previous crops or in chemically killed sod. Tillage usually is done with nonpowered, fluted colters running ahead of planters that have disk openers. Narrow chisels, angled disks, or straight or slightly rippled colters can also be used to open the soil for seed placement. A press wheel or seed packer wheel is needed for good seed-soil contact after planting. For proper operation of no-tillage planters, crop residue should be uniformly distributed on the soil surface and corn (Zea mays L.) or similar residue should be chopped before planting. The equipment and practices just described can be used in various combinations. Some other possibilities are combinations of stubble mulch tillage and herbicides (Phillips, 1969) and of conventional, reduced, and no-tillage (Allen et al., 1980; Unger and Wiese, 1979). Choice of system must consider the equipment available, soil and climatic conditions, size and type of farming operation, and the producer’s managerial ability and personal preferences (Giffith et a l . , 1977).

B . SEEDBED PREPARATION A N D CROPSEEDING

Regardless of crop or region, a firm, moist seedbed is desirable for rapid seed germination and seedling emergence. In regions with adequate precipitation, the seedbed normally is moist enough at planting for rapid seed germination and seedling emergence and for good plant growth. Under such conditions, continuous cropping is possible. In areas of limited precipitation, leaving land idle for a season to store enough water for a crop may be necessary. This practice is called summer fallowing. During fallow, the land must be kept free of weeds, and enough residue must be kept on the surface for erosion control.

CONSERVATION TILLAGE SYSTEMS

11

I . Small Grains Stubble mulch tillage, either in a summer fallow or continuous cropping system, is widely used for weed control in the Great Plains and the Pacific Northwest where small grains are grown. With stubble mulch tillage, soil may need packing or smoothing to obtain a favorable seedbed. When packing is desirable, it can be done with a treader, which is similar to a rotary hoe with the tongues reversed so that the soil is packed rather than loosened, or by shallow operation of a sweep-rodweeder. The treaders also break heavy stubble and surface crusts, control small weeds, and have a tendency to break surface clods. Because clods are important for controlling wind erosion, treaders must be used with caution where small amounts of residue are on the surface. Successively shallower operation of rodweeders after the initial plowing causes the soil to be firm near the surface and, therefore, no further packing is required. In areas with favorable precipitation, soil packing generally is not necessary because of settling due to precipitation. A major requirement of seeding equipment is trouble-free operation under all surface residue conditions. Small grains can be successfully seeded with shovel-, hoe-, or disk-opener drills. Shovel- and hoe-opener drills work well for placing seeds in moist soil that is overlain by dry surface soil, and they also form ridges that help control wind erosion. These drills generally perform well when large amounts of residue are present because they have high clearance and staggered shanks that support the openers and seed spouts. Drills with disk openers ridge the soil less than those with shovel or hoe openers. They are, therefore, less satisfactory for planting through dry surface soil and for controlling erosion. The disk openers also tend to destroy surface clods, which further decreases their effectiveness for controlling wind erosion. Drills with wide-spaced (25 to 35 cm) disk openers perform satisfactorily in high residue situations and have been used for no-tillage seeding of wheat in wheat (Allen et al., 1976; Unger, 1978a) or corn (Musick et al., 1977) residue, and wheat or barley (Hordeum vulgare L.) in grain sorghum [Sorghum bicolor L. (Moench)] residue (Musick et a l . , 1972, 1977). Because soil penetration with disk openers may be limited under high residue conditions, irrigation or timely precipitation may be required for satisfactory germination and seedling emergence. Small grains can also be seeded with no-tillage drills (Papendick and Miller, 1977) or with one-way disk plows, moldboard plows, or similar implements with attached seeding equipment and press wheels (FAO, 1971; Stonebridge and Fletcher, 1973). Where plows are used, enough residue or clods should remain on the surface for effective wind erosion control. 2 . Row Crops The seedbed requirements for row crops, such as corn, soybeans (Glycine max L.), sorghum, and cotton (Gossypium hirsutum L.), are similar to those for small

12

P. W . UNGER A N D T. M. McCALLA

grains. Because the row spacing for row crops is usually wider than it is for small grains, a wider clean zone can be used for seeding row crops. Seeding can be done with planters equipped with furrow openers or with listers or similar implements operated at a shallow depth. The furrows should be deep enough to produce a clean, moist seedbed, but not deep enough to cover the residue between the rows. Various types of equipment for no-tillage planting of row crops have been developed in recent years. A major requirement of these planters is adequate penetration of the soil for satisfactory seed placement. A colter may be needed ahead of the planter to cut surface residue, but both unit planters with double-disk openers (Unger and Wiese, 1979) and grain drills with some of the seed spouts blocked (Musick et a l . , 1977) have been used successfully without colters. As for small grains, timely precipitation or irrigation after planting assured uniform germination and seedling emergence in cases where large amounts of surface residue restricted the penetration of soil by the planting unit. Where irrigation is not practiced and large amounts of surface residue are present, straight, rippled, or fluted colters are necessary for satisfactory soil penetration for seed placement. Straight colters require less equipment weight for soil penetration than do fluted colters, but fluted colters loosen the soil more and, therefore, have been widely used on no-tillage planters (Griffith et a l . , 1977; Amemiya, 1977; Pitts, 1978). Besides soil penetration, another major requirement of no-tillage planters is a provision for adequate seed-soil contact, which normally is provided by ribbed press wheels or seed packer wheels (Griffith et al., 1977; Pitts, 1978). The planting unit described by Pitts (1978), which has a subsoiler for loosening compacted layers in the soil, also has a spike-toothed wheel to fill the slot made by the subsoiler, which prevents seeds from falling too deeply into the soil.

3 . Grasses and Legumes The small seeds of many grasses and legumes require precise planting for adequate stand establishment. Grasses and legumes have been successfully seeded in stubble mulch systems by scattering seeds from a seeder box mounted ahead of a treader, which covered some of the seed. A preferable method is to use a drill, equipped with double-disk furrow openers and depth bands, that is capable of handling rough seeds and drilling through stubble. This method keeps the soil protected against erosion and maintains more favorable soil water conditions for germination and emergence. The no-tillage method has been successfully used for planting grasses and legumes at numerous locations (Bennett, 1977; Olsen et a l . , 1978; Papendick and Miller, 1977; Reicosky et a l . , 1977). In this system, tilling and seeding are done in one operation in chemically killed sod or residue from previous crops, thus affording excellent protection against erosion.

CONSERVATION TILLAGE SYSTEMS

13

IV. CROP YIELDS AND QUALITY A. GRAINYIELDS

Results from early studies showed that wheat grain yields generally were higher with stubble mulch than with clean tillage in arid and semiarid regions, but lower in subhumid and humid regions (Zingg and Whitfield, 1957). Similar results have been reported for subsequent studies (Johnson and Davis, 1972; Smika, 1976a; Papendick and Miller, 1977). The higher yields resulted from increased water conservation with stubble mulch tillage. Possible reasons for the lower yields in more humid regions with stubble mulch tillage include greater nitrogen immobilization, fertility imbalances, difficulty with stand establishment, reduced seedling vigor, greater weed infestations, and release of phytotoxic decomposition products (Kimber, 1966; McCalla and Norstadt, 1974; Papendick and Miller, 1977). Much research concerning reduced- and no-tillage crop production systems has been conducted in recent years. While better erosion control and lower production costs have caused much of the interest in these systems, producers are not likely to adopt reduced- and no-tillage systems where there is a risk of lower yields, even though production costs are lower (Griffith et al., 1977). As a rule, grain yields were little affected by tillage practices under conditions of adequate soil water, favorable precipitation, and good drainage, provided other factors such as soil fertility, weed control, and plant populations were equal (Bennett, 1977; Reicosky et al., 1977; Amemiya, 1977; Unger et al., 1977; Griffith et al., 1977; Elliott et al., 1977; Rowel1 et al., 1977; La1 et al., 1978). Under conditions of limited soil water and limited precipitation or irrigation, crop yields were equal and often significantly higher with reduced- and no-tillage systems than with conventional tillage (Amemiya, 1977; Fenster, 1977; Unger et al., 1977; Unger and Wiese, 1979; Anderson, 1976; Wicks and Nordquist, 1976; Phillips et al., 1976; Wicks and Smika, 1973; Phillips, 1969). The higher yields with the reduced- and no-tillage systems generally were attributed to increased soil water contents resulting from increased infiltration, decreased runoff, and possibly decreased evaporation. Although workable reduced- and no-tillage systems have been developed for many crops at numerous locations, unfavorable results have been obtained in some cases. In general, yields often are lower with reduced- and no-tillage than with conventional tillage on poorly drained soils (Griffith et al., 1977), with continuous cropping when volunteer crop plants cause excessive plant populations (Allen et al., 1975), and where extremely large amounts of residue are present, as in the Pacific Northwest, where weed control, planting, soil fertility, disease, insect, and rodent problems have been encountered (Papendick and Miller, 1977).

14

P. W . UNGER AND T. M. McCALLA

B . PLANTPROTEINCONTENTA N D MINERALCOMFQSITION

Plant residue contains inorganic nutrients that are potentially available to subsequent crops as the residue decomposes. The rate of decomposition, however, is strongly influenced by tillage methods. Methods that retain residue on the surface result in slow decomposition and, therefore, slow release of nutrients to crops. Residue incorporation with soil generally increases the decomposition rate, but the nutrients may be initially immobilized, then subsequently released in forms available to plants as decomposition progresses. The rates of decomposition and application of fertilizers affect the amounts of nutrients available to plants and, therefore, the mineral composition of plants. Zingg and Whitfield (1 957) summarized the protein contents of wheat grain from seven locations in the Great Plains. At six of the locations, average protein contents for 52 years of data were 13.5 and 14.1% with stubble mulch and clean tillage, respectively. Similar effects of tillage were found by Bennett et al. (1954) in Utah, and Unger (unpublished data) in Texas. Even lower protein contents of grain were obtained for no-tillage wheat. The percentages were 18.4, 18.0, and 17.4 with disk, sweep, and no-tillage, respectively, for dryland wheat and 16.4, 16.1, and 15.9 with the respective tillages for imgated wheat (Unger, unpublished data). The same protein trends were found for sorghum grain in an irrigated wheat-dryland grain sorghum cropping system. Average values were 14.7, 13.9, and 13.5% with disk, sweep, and no-tillage, respectively (Unger and Wiese, 1979). Jacks et al. (1955) reported that the effect of mulches on mineral composition of plants depended on the type and amount of mulch, stage of decomposition of the mulch, soil and climatic conditions, and kind of plant. One of the responses to heavy mulching was a marked increase in the K content of soil, which resulted in increased K contents of plant tissues. Jacks et al. (1955) also showed that Ca, Mg, Mn, and P contents of plant tissues generally were higher on mulched than on bare soil. In contrast, Estes (1972) showed that Ca, Mg, Zn, Mo, B, and A1 concentrations in corn leaf tissue were significantly decreased and K concentrations were significantly increased under no-tillage as compared with conventional tillage at five levels of lime application. Concentrations of P, Fe, and Mn were not affected (Table IV). Changes in element concentration in tissues apparently were related to soil pH and method of lime application. Also, reduced Ca and Mg uptake possibly resulted from greater K uptake (Estes, 1972). Griffith et al. (1977) reported that no-tillage decreased plant uptake of Cu, Zn, B, and Mn, possibly because of reduced root systems, which caused less contact with soil, and because of cooler and wetter soils, which decreased the availability of the micronutrients early in the season. They also reported that plant K was deficient with shallow or no-tillage only under unusually wet conditions on poorly drained soils. Similar results were reported by Bower et al. (1944) and

15

CONSERVATION TlLLAGE SYSTEMS

Table IV Effect of Tillage Method on Nutrient Composition of Corn Ear Leaves' Tillage method Nutrient Phosphorous (%) Potassium (%) Calcium (S) Magnesium (%) Zinc (ppm) Iron (ppm) Molybdenum (ppm) Manganese (ppm) Boron (ppm) Aluminum (ppm) a

No-tillage

0.34 1.66 0.11

0.50 26.0 154.0 2.0 77.0 9.I 56.0

Conventional

0.34 1.53 0.85 0.59 29.0 161.0 3.2 84.0 10.2 60.0

% Change relative to conventional

0.0

+8.56 -9.4b - 15.2b - 10.3b -4.3 -37.5b -8.3 - 10.86 -6.7b

From Estes (1972). Denotes significant difference between tillage methods (P = 0.05).

Lawton and Browning (1948) with stubble mulching in Iowa. Reduced soil aeration was suggested as a possible cause for reduced K absorption. The generally higher soil and plant K concentration with surface residue apparently results from leaching of readily water-soluble K compounds from the residue. C. RESIDUEYIELDS

Zingg and Whitfield (1957) summarized the early data pertaining to straw production by wheat at six locations in the western United States. Wheat grown in rotation with corn, oats (Avena sativa L.), and sweet clover (Melilotus sp.) produced less straw with stubble mulch than with clean tillage. Similar trends occurred for continuous wheat, but straw production was higher with stubble mulch tillage than with plowing in wheat-fallow systems at some locations. The average data for 101 crop-year comparisons showed a 3.9% decrease in straw production with stubble mulch tillage as compared with plowing. On a strawgrain ratio basis, the averages were 1.87 for stubble mulching and 1.98 for plowing. The straw yields and straw-grain ratios varied with varieties, soil fertility, and growing conditions. At similar grain yields, some semidwarf wheat varieties yielded less and others yielded more residue than taller varieties (Bauer and Zubriski, 1978). Relatively few residue yield data as influenced by tillage methods have been reported in recent years, but for studies at Bushland, Texas (Johnson and Davis, 1972; Unger, 1977), wheat straw-grain ratios averaged 2.43 and 2.15 with clean

16

P. W . UNGER AND T. M. McCALLA

and stubble mulch tillage, respectively, for 24 crop-year comparisons. For dryland grain sorghum, the ratios were 1.32, 1.15, and 1.18 with disk, sweep, and no-tillage, respectively (Unger and Wiese, 1979). Limited data for inigated continuous grain sorghum showed opposite trends. The values were 1.70 and 2.42 with clean and no-tillage, respectively (Allen et af., 1975). The high residue yields relative to grain yields with no-tillage resulted from volunteer sorghum, which caused high forage yields, but plant populations that were too high for favorable grain yields. The ratio with clean tillage was also higher than those normally expected (1.0 to 1.2) for irrigated grain sorghum (Eck and Taylor, 1969). D. ROOTGROWTH

Jacks et al. (1955), from their review of the literature, concluded that plant roots tended to accumulate at or near the soil surface when large amounts of mulch were present. Although data concerning the influence of stubble mulch tillage on root development are not available, recent studies showed higher root concentrations near the surface with no-tillage than with clean tillage (Ellis et al., 1977; Griffith et af., 1977). Besides the rooting pattern, weight and size of roots are also affected by type of tillage. No-tillage and strip rotary tillage resulted in lower, and chisel plowing and wheel-track planting resulted in higher corn root weights than conventional tillage. Also, corn roots were larger in diameter with no-tillage, which caused them to have less absorbing surface for water and nutrient uptake per unit weight of roots. However, when the residue resulted in adequate water near the soil surface, there was little correlation between root weight differences due to tillage method and grain yields (Griffith et al., 1977). Apparently, the improved soil water conditions resulting from surface residue with no-tillage compensated for the decreased root weight and penetration depth that were noted.

V. ENVIRONMENTAL CONSIDERATION Much emphasis has been placed on control of air and water pollution in recent years. Because wind and water erosion are major contributors to air and water pollution, respectively, much effort has been directed toward developing improved erosion control techniques. The value of surface residue for erosion control has long been recognized. Recent results show that erosion control can be improved by maintaining crop residue on soil surfaces with reduced- and notillage cropping practices.

CONSERVATION TILLAGE SYSTEMS

17

A. CONTROLOF WINDEROSION

The urgent need to control soil erosion by wind effectively in the late 1930s and early 1940s stimulated the development of the stubble mulch farming system. Because of the success of stubble mulch farming in the drier part of the Great Plains and the continued threat of wind erosion in many years, the stubble mulch system has become the basic tillage method in many dryland farming areas.

I. Factors That Influence Wind Erosion Soil erosion by wind can be a problem wherever the required conditions of soil, vegetation, and climate prevail. These conditions are (1) a soil that is loose, dry, and reasonably finely divided; (2) a smooth soil surface on which vegetative cover is absent or sparse; (3) a large enough field; and (4) wind that is strong enough to move soil (Skidmore and Siddoway, 1978). Sandy soils are extremely susceptible to wind erosion because of little or no coherence between particles, rapid drying after wetting, and small particle sizes. Other soils, however, are also subject to wind erosion when they are dry and loose, and when the soil particles have been finely divided by tillage, raindrop impact, or freezing and thawing. Particles greater than 0.84 mm in diameter are generally considered nonerodible by wind. The smoothness of soils and the amount of residue on the surface are strongly influenced by tillage methods. Operations that leave a rough, cloddy surface or that keep residue on the surface can help to minimize wind erosion. Examples are moldboard, lister, or chisel plowing, which leave the surface rough, and stubble mulch or no-tillage, which keep residue on the surface. The potential for wind erosion is increased by surface smoothing and residue-destroying operations, such as disking, harrowing, cultivating, and land planing. When used for secondary tillage, these operations can also reduce surface roughness initially formed by primary tillage such as moldboard plowing. The impact of raindrops, soil freezing and thawing, and erosion itself can further reduce surface roughness. Field width parallel to the wind direction has a major influence on soil erodibility, which increases as field width increases. Hence, width in the direction of prevailing winds should be kept as narrow as possible. Wind erosion occurs on some highly erodible soils that are only a few meters wide (Skidmore and Siddoway, 1978). With adequate surface residues, the effect of field width is minimized. Soil movement begins at relatively low wind speeds (Chepil and Woodruff, 1963) and progressively increases as wind speed and turbulence increase. Therefore, to minimize wind erosion, the wind speed at the soil-air interface must be reduced to the threshold value below which no significant wind erosion will

18

P. W . UNGER AND T. M. McCALLA

occur (Skidmore and Siddoway, 1978). The influence of wind on soil erosion is extremely complex and includes the processes of soil particle movement, transport, sorting, abrasion, avalanching, and deposition (Woodruff and Siddoway, 1973). Detailed discussions of the mechanics of soil erosion have been established by Bagnold (1943), Chepil and Woodruff (1957), and Zingg et al. (1965).

2 . Wind Erosion Equation A generalized equation expressing the relative quantity of wind erosion from a field was first published by Chepil (1959). As new data have become available, the equation has been modified and is now generally given as

E =f ( f C K L V )

(1)

where E is the potential annual quantity of erosion per unit area and is a function, f, of I, soil erodibility; C, local wind erosion climatic factor; K , soil surface roughness; L , equivalent width of field (maximum unsheltered distance across the field along the prevailing wind erosion direction); and V, equivalent quantity of vegetative cover (Chepil and Woodruff, 1963). The mathematical relationships among the components of the equation are complicated. The relationships, however, have been computed and developed into tables or plotted on graphs, and are useful for estimating annual soil losses by wind erosion and for determining alternate land treatments for wind erosion control. A guide containing this information for the Great Plains states is available (Craig and Turelle, 1964). Tillage has a direct bearing on factors 1, K , and V through its effect on soil cloddiness, soil roughness, and equivalent quantity of vegetative cover. 3. General Principles of Wind Erosion Control

Values for each of the individual primary factors that influence wind erosion must be determined before the potential soil loss can be estimated. The I factor is determined from the percentage of soil particles smaller than 0.84 mm in diameter as determined by dry sieving or from reference tables of known average cloddiness of different soils during the wind erosion season. The C factor for a particular location is estimated from the wind erosion climatic map. Appropriate values for I and C are available in the wind erosion control guide (Craig and Turelle, 1964). Factors K, L, and V are based on field conditions. With I and C determined, appropriate charts and tables in the guide pertaining to K , L, and V are used to determine the erosion potential. The charts and tables can also be used in reverse to determine what conditions or practices are needed to reduce wind erosion to a desired level (Chepil and Woodruff, 1963).

CONSERVATION TILLAGE SYSTEMS

19

4 . Residue and Tillage Effects on Wind Erosion

Surface residue affects wind erosion primarily through influences related to factor V of the wind erosion equation. The effect of tillage is related mainly to factors I and K of the equation. a . Function of Residue. The principal function of surface residue is to decrease the force of wind on the soil itself. Zingg (1954) showed that different amounts, kinds, and arrangements of anchored crop residue in the field eliminated from 5 to 99% of the wind action on the soil surface. The forces that move erodible particles are reduced when the force of the wind is transferred to the residue. Wind erosion is eliminated when the forces at the soil-air interface are reduced to a threshold value below which erosion will not occur. Wind erosion is influenced by the kind, amount, texture, height, and orientation of surface residue. The value of small grain residue for controlling erosion is well known. The wind erosion control guide uses this kind of residue as the basic type and gives equivalents for other types of residue (Craig and Turelle, 1964). On an equal-weight basis, small grain residue is more effective than that of sorghum or corn, which in turn is more effective than that of cotton or soybeans (Woodruff and Siddoway, 1973). The differences result from the different densities of the materials. As residue amounts increase, erosion decreases; therefore, as much residue as possible, up to the limit beyond which no further reduction occurs, should be maintained on the surface for effective erosion control. The amounts required vary with location and soil type. For highly erosive soils, residue production may not be adequate for effective erosion control, especially with such crops as cotton or soybeans, and possibly with corn, sorghum, and even small grains. The texture, or fineness, of residue also influences erosion. For equal amounts, fine residue gives more protection than coarse residue when it is equally distributed and anchored in the soil. Harvesting, tillage, and other crop production operations may remove, flatten, shorten, shred, or redistribute surface residue and, therefore, influence its effectiveness for controlling erosion. In general, standing residue is about twice as effective as flattened residue for controlling erosion. Also, tall residue is more effective than short residue. Sorghum stubble about 30 cm tall is effective under most conditions if it is dense enough to cause wind to flow over it rather than through it. Leaves are important for increasing the density of stubble (McCalla and Army, 1961). Harvesting and tillage operations that retain large amounts of tall, erect residue on the surface are, therefore, most effective for controlling wind erosion. Stubble mulch tillage undercuts surface residue and maintains most of it on the surface. However, even this tillage method destroys surface residue when used repeatedly and when performed at relatively high speeds. The no-tillage system,

20

P. W. UNGER A N D T. M. McCALLA

with no soil disturbance other than that needed to plant the seeds, is especially effective for maintaining surface residue and, therefore, for controlling wind erosion. b. Function of Tillage. In the last section, the effectiveness of stubble mulch and no-tillage were discussed relative to maintaining residue on the surface. In some cases, not enough residue is produced for erosion control, even when all of it is kept on the surface. In other cases, weeds are controlled and seedbeds are prepared with tillage, which reduces or destroys surface residue and, therefore, leaves the soil susceptible to wind erosion. To control erosion under such conditions, soil should be kept in a rough, cloddy condition. Some surface roughness results from normal crop production operations. Roughness for controlling wind erosion is also obtained by tillage or planting operations that cover or replace some of the erodible soil particles with less erodible material, thus reducing wind drag on the remaining particles. There is, however, some evidence that increasing roughness increases wind turbulence and velocity fluctuations at the surface so that some of the benefits gained by roughening the surface are lost (Lyles et al., 1971). The most effective soil roughness height is 5 to 13 cm (Armbrust et al., 1964; Woodruff and Lyles, 1967). Tillage operations that minimize soil pulverization and smoothing are effective for maintaining surface roughness. Special planters, such as the deep furrow and hoe drills for planting small grains in surface residues, produce roughness in the 5- to 13-cm height range and, therefore, are especially effective in providing erosion-resistant surfaces (Woodruff and Siddoway, 1973). Reduced- and no-tillage systems, as compared with conventional tillage, provide greater protection against wind erosion because the surface can be kept rougher due to fewer tillage operations and more residue on the surface. Other factors being equal, soils at locations where the value for the climatic factor is high require more nonerodible clods (>0.84 mm) to control wind erosion effectively than such soils at locations where the climatic factor is low. The degree of cloddiness produced by tillage depends on soil texture, soil water content at tillage, tillage tool, and speed of operation (Woodruff and Siddoway, 1973). Sandy soils have low cohesiveness and, therefore, contain relatively few clods that resist wind erosion. However, if the surface soil contains more than about 8% clay, a cloddy surface that resists wind erosion can be produced by cultivation of sandy soils (Harper and Brensing, 1950). Plowing, cultivating, or planting sandy soils while they are moist also leaves the surface more resistant to wind erosion than when these operations are performed while the soils are dry. For soils with higher clay contents, enough cloddiness can generally be obtained when the operations are performed within a wide range of soil water contents. However, least cloddiness (>0.84 mm) of a silty clay loam soil in Kansas was obtained at an intermediate soil water content (about 20%). Cloddiness increased at lower and higher water contents with moldboard, sweep, and one-way tillage

CONSERVATION TILLAGE SYSTEMS

21

(Lyles and Woodruff, 1962). Emergency chiseling of high-clay soils effectively reduces wind erosion by increasing surface cloddiness, even when these soils are quite dry. In stubble mulch studies, cloddiness generally was greatest with 5-cm-wide chisels and 80-cm-wide sweeps, followed in order by the one-way disk, rodweeder with shovels, and 2.75-m-wide sweeps. In a wheat-fallow rotation in Montana, the nonerodible fraction (>0.84 mm) of a sandy loam soil also increased as the amount of residue on the surface increased (Black, 1973). Similar results were obtained for a silt loam soil in Idaho (Woodruff and Siddoway, 1973) and a clay loam soil in Texas (Unger, unpublished data). For no-tillage, Unger et al. (1979) reported fewer large dry aggregates (B0.84 mm) with no-tillage than with disk or sweep tillage on Pullman clay loam (Torrertic Paleustoll) at Bushland. The soil, however, was protected against wind erosion by surface residue. B . CONTROL OF WATEREROSION

Water erosion is the dominant problem on about 72 million of the 172 million hectares of cropland in the 48 contiguous states of the United States (Hayes and Kimberlin, 1978). With clean tillage systems, water erosion may occur at any time on most soils, but the potential is generally greatest while the surface is bare after plowing, during seedbed preparation, and at seedling establishment. Conservation systems, which involve surface residue, are especially effective for controlling water erosion (Wischmeier, 1973; Hayes and Kimberlin, 1978). Soil erosion by water is a process of particle detachment and transport that requires energy. Both rainfall and flowing water (runoff) have detachment potential, but transport is mainly by runoff. At upslope positions, the energy is supplied mainly by rainfall and the slope gradient. On bare soil, most of the kinetic energy of raindrops is dissipated at the surface where the impacting drops detach soil particles. Splash action and shallow sheet flow transport many of the detached particles to runoff concentrations. Drop impact also disperses soil aggregates, reduces surface roughness, and promotes surface sealing and crusting, thereby increasing runoff (Wischmeier, 1973). As runoff increases, rill and finally gully erosion may occur. Gully erosion is the most obvious type, but sheet and rill erosion are responsible for the major part of water erosion on cropland (Hayes and Kimberlin, 1978). I . Factors That Influence Water Erosion

For effective water erosion control, seeding and tillage practices should decrease raindrop impact on the soil, increase water infiltration, decrease runoff velocity, and decrease soil detachability (Wischmeier, 1973). These factors are

22

P. W. UNGER AND T. M. McCALLA

influenced by intensity and duration of rainfall; steepness and length of soil slope; texture, organic matter content, roughness, and ridging of soil; amount, type, and distribution of surface residue; and type of erosion control practice (e.g., contouring, strip cropping, terracing). The factors influencing erosion have been studied extensively and reviews and guidelines pertaining to erosion control have been published by Hayes and Kimberlin ( 1978), Kimberlin ( 1976), Stewart et al. (1975), Wischmeier (1973), and Wischmeier and Smith (1978). The guidelines generally involve use of the universal soil loss equation, which helps to establish relationships between the amount of erosion and the factors influencing erosion. 2 . The Universal Soil Loss Equation (USLE) The universal soil loss equation is A

= RKLSCP

(2)

where A is computed soil loss per hectare; R , rainfall factor based on the number of erosion-index units in a normal year’s rainfall at a specific location; K, soil erodibility factor; L, length of slope factor; S, slope gradient factor; C, crop management factor; and P, erosion control practice factor. All factors are unitless, except A and K . Units for A are metric tondhectare (or tondacre) per year and those for K are metric todhectare (or todacre) per erosion index unit (Hayes and Kimberlin, 1978). Values for the factors of the equation are available for many conditions at numerous locations (Stewart et al., 1975).

3 . General Principles of Water Erosion Control To determine potential erosion at a specific location, a value for each factor of the USLE must be determined. The R factor is based on rainfall records at a location and, therefore, is fixed. The K , L, and S factors at a given location are also based on prevailing conditions and, therefore, are not subject to change without major soil alterations. Changing the value of the C factor, however, offers a major potential for reducing erosion. Crop management practices involved include tillage, rotations, and residue management practices. When potential erosion at a given location cannot be reduced to acceptable levels by crop management, then engineering-type erosion control practices, which affect factor P, must be used. In the remainder of this section, we discuss mainly practices that affect the C factor of the USLE. 4 . Residue and Tillage Effects on Water Erosion

The influence of crops, rotations, and management on factor C of the USLE is illustrated in Table V. In general, the value of C decreases as increasing amounts

23

CONSERVATION TILLAGE SYSTEMS

Table V Generalized Values of the Crop Management Factor, C, in the 37 States East of the Rocky Mountains" C value Line no.

Crop, rotation, and management*

Base value: continuous fallow, tilled up and down slope

Productivity levelr HigL Moderate

.oo

1.oo

0.54

0.62 0.59 0.52 0.49 0.48 0.44 0.35 0.30 0.24 0.28 0.26 0.23 0.24 0.20 0.17

1

Corn C, RdR, fall TP, conv (1) C, RdR, spring TP, conv ( I ) C, RdL, fall TP, conv (1) C, RdR, wc seeding, spring TP, conv ( I ) C, RdL, standing, spring TP, conv (1) C, fall shred stalks, spring TP, conv (1) C (silage)-W(RDL, fall TP) (2) C, RdL, fall chisel, spring disk, 40-30% rc ( I ) C (silage), W wc seeding, no-till pl in c-k W (1) C (RDL)-W(RDL, spring TP) (2) C, fall shred stalks, chisel pl, 40-30% rc (1) C-C-C-W-M, RdL, TP for C, disk for W (5) C, RdL, strip till row zones, 55-408 rc (1) C-C-C-W-M-M, RdL, TP for C, disk for W (6) C-C-W-M, RdL, TP for C, disk for W (4) C, fall shred, no-till pl, 70-50% rc (1) C-C-W-M-M, RdL. TP for C, disk for W (5) C-C-C-W-M, RdL, no-till pl 2d & 3rd C (5) C-C-W-M, RdL, no-till pl 2d C (4) C, no-till pl in c-k wheat, 90-708 rc ( I ) C-C-C-W-M-M, no-till pl2d & 3rd C (6) C-W-M, RdL, TP for C, disk for W (3) C-C-W-M-M, RdL, no-till pl 2d C (5) C-W-M-M, RdL, TP for C, disk for W (4) C-W-M-M-M, RdL, TP for C, disk for W (5) C, no-till pl in c-k sod, 95-80% rc (1)

0.50 0.42 0.40 0.38 0.35 0.31 0.24 0.20 0.20 0.19 0.17 0.16 0.14 0.12 0.11 0.087 0.076 0.068 0.062 0.061 0.055 0.051 0.039 0.032 0.017

Cottond 27 28

Cot, conv (Western Plains) (1) Cot, conv (South) (1)

0.42 0.34

0.49 0.40

Meadow 29 30 31

Grass & legume mix Alfalfa, lespedeza, or Sericia Sweet clover

0.004 0.020 0.025

0.01

1

2 3 4 5 6 7 8 9 10 11

12 13 14 15

16 17 18 19 20 21 22 23 24 25 26

0.18

0.14 0.13 0.11 0.14 0.11 0.095 0.094 0.074 0.061 0.053

(continued)

24 Table V

P. W. UNGER AND T. M. McCALLA (continued)

C value

Line no.

Crop, rotation, and managementb

Sorghum, Grain (Western Plains)d 32 RdL, spring TP, conv (1) 33 No-till pl in shredded 70-50% rc

Productivity levelC High Moderate

0.43

0.11

0.53 0.18

B. RdL, spring TP, conv (1) C-B, TP annually, conv (2) B, no-till pl C-B, no-till pl. fall shred C stalks (2)

0.48 0.43 0.22 0.18

0.54 0.51 0.28 0.22

W-F, fall TP after W (2) W-F. stubble mulch, 560 kg rc (2) W-F, stubble mulch, 1120 kg rc (2) Spring W, RdL, Sept TP, conv (N & S Dak)( I ) Winter W, RdL, Aug TP, conv (Kans) (1) Spring W, stubble mulch, 840 kg rc (1) Spring W, stubble mulch, 1400 kg rc (1) Winter W, stubble mulch, 840 kg rc (1) Winter W, stubble mulch, 1400 kg rc (1) W-M, conv (2) W-M-M, conv (3) W-M-M-M, conv (4)

0.38 0.32

Soybeansd 34 35 36 37

Wheat 38 39 40 41 42 43 44 45 46 47 48 49

0.21 0.23 0.19 0.15 0.12

0.11 0.10 0.05 0.026

0.021

From Stewart et al. (1975). This table is for illustrative purposes only and is not a complete list of cropping systems or potential practices. Values of C differ with rainfall pattern and planting dates. These generalized values show approximately the relative erosion-reducing effectiveness of various crop systems, but locationally derived C values should be used for conservation planning at the field level. Tables of local values are available from the Soil Conservation Service. Abbreviations used are defined as follows: B-soybeans; C--corn; c-k--chemically killed; convconventional; cot--cotton; F-fallow; M-grass-and-legume hay; pl-plant; W-wheat; wc-winter cover; kg rc-kilograms of crop residue per hectare remaining on surface after new crop seeding; % rc-percentage of soil surface covered by residue mulch after new crop seeding; 70-50% rc70% cover for C values in first column, 50% for second column; RdR-residues (corn stover, straw, etc.) removed or burned; RdL-all residues left on field (on surface or incorporated); TP-turn plowed (upper 13 or more cm of soil inverted, covering residues). bNumbers in parentheses indicate number of years in the rotation cycle. No. (1) designates a continuous one-crop system. High level is exemplified by long-term yield averages greater than 4700 kg corn or 6.7 metric tons grass-and-legume hay; or cotton management that regularly provides good stands and growth. dGrain sorghum, soybeans, or cotton may be substituted for corn in lines 12. 14, 15, 17-19, 21-25 to estimate C values for sod-based rotations.

CONSERVATION TILLAGE SYSTEMS

25

of residue are maintained on the soil surface for increasing amounts of time during the crop production cycle. a . Function of Residue. The amount of crop residue available and how it is managed influences the erosion control effectiveness of tillage systems. Surface residue dissipates raindrop energy, thus decreasing soil detachment, surface sealing, crusting, and, in turn, runoff. Soil detachment and subsequent erosion are further decreased by surface residue because the residue decreases the shear stress exerted on the soil by runoff (Wischmeier, 1973). The residue serves as dams that slow the flow rate of water across the surface. 1. Residue rate. For maximum effectiveness, the soil surface should be completely covered with residue. Light-weight, hollow-stemmed materials, such as small grain residue, provide greater coverage per unit of weight than materials such as corn or sorghum stubble. Greater than 95% coverage is provided by about 5.6 metric tons/ha of small grain straw or about 8.2 metric tons/ha of chopped cornstalks (Wischmeier, 1973). The effect of wheat straw mulch rate and soil slope on water and soil losses during simulated rainfall is given in Table VI. Water and soil losses decreased as mulch rates increased. Soil slopes had little or no effect on water losses, but soil losses increased as the percent slope increased (Lattanzi et al., 1974). Similar results, reported by Wischmeier (1973), indicated that even the low rates of mulch kept infiltration rates relatively high. Runoff velocities with 1 . 1 and 2.2 metric tons/ha of wheat straw were about 50 and 33% of those with no mulch. The decreased runoff velocities apparently greatly lowered soil losses as mulch rates increased. 2 . Residue type. On Miami silt loam (Typic Hapludalf) with 5% slope in Indiana, soybean and wheat residue left on plots at harvest were equally effective in controlling erosion when compared on a dry-weight basis in late autumn. In contrast, 4.5 metric tons/ha of corn stover was only half as effective as an equal weight of wheat straw. At higher mulch rates, the differences among the materials decreased (Wischmeier, 1973). 3. Mulch distribution. Wischmeier (1973) also measured the influence of residue distribution on the silt loam soil in Indiana. Mulch in rows across the slope, with two-thirds of the area bare in alternate strips, was as effective in controlling erosion as uniform distribution of the same amount of residue over the entire area. Much of the soil detached from bare strips was deposited in the mulched strips, which showed that strip tillage can control erosion if the residue strips are on the contour. b. Function of Tillage. Conservation tillage systems, such as no-tillage and till planting, retain all or most crop residue on the surface. They are especially effective in controlling erosion by water. Numerous studies concerning the effect of these systems on erosion have been conducted, but we will show only two

Table Vi Water Loss by Runoff and Soil Loss in Runoff during Initial, Wet, and Very Wet Runs"

Rainapplied

Initial run, 60 minutes

Wet runr 30 minutes

Very wet runC 30 minutes

Mulch rate (metric todha)

0 0.5 2 8 0 0.5

2 8 0 0.5 L

8

Soil loss (g/m*)

Water loss (kg/mz)

S = 6%

S = 12%

S = 20%

S =2%

49.3 50.9 43.5 2.5 28.6 29.9 27.7 2.5 28.4 30.8 28.9 4.0

52.7 52.4 43.7 4.3 28.9 29.8 28.1 5.0 29.3 30.8 28.1 6.4

53.0

49.2 52.4

950 600 240 7 410 260 90 2

1230 750 260 10 550 370 130

400

560 380 120 5

51.1 44.0 4.1 28.2 29.5 27.2 9.3 29.1 30.1 28.2 12.5 ~~~

44.1

5.7 27.9 30.0 27.3 7.1 28.1 30.6 29.3 9.1

260 80 1 ~~~

Rain intensity was 6.4 c d h o u r . From Lattanzi er al. (1974). b S = slope. 'Wet run made 1 day after initial run; very wet run made 15 minutes after wet run. a

S =6%

Sb = 2%

5

S = 12%

S =20%

1870 970 3 10 6 770 480 150 3 720 480 150 3

2140 1250 490 1

L

820 520 210 8 810 540 220 5

27

CONSERVATION TILLAGE SYSTEMS

examples, which show the tremendous potential of these systems for controlling erosion. Harrold and Edwards (1972) measured rainfall, runoff, and sediment yield on three watersheds for a storm near Coshocton, Ohio, having an expected recurrence frequency of over 100 years. More than 12.7 cm of rain fell in 7 hours. Corn was grown on all watersheds. Rainfall was identical and the slopes were similar for clean-tilled watersheds with sloping or contour rows (Table VII), but runoff and sediment yield from the contoured watershed were only 52 and 14%, respectively, of that from the sloping-row watershed. No-tillage (corn planted in sod) with contour rows resulted in 57 and 0.1% runoff and sediment yield, respectively, of that from the sloping-row watershed, even though the slope was much greater on the no-tillage watershed. Onstad ( 1972) in South Dakota, measured runoff and soil losses from plots on Egan and Wentworth silty clay loams (Udic Haplustolls) having about 6% slopes. Although average runoff, based on either the actual amount or percentage of rainfall, was lower than at locations with more rainfall, the results showed the value of surface residue and improved tillage practices in decreasing runoff and soil losses (Table VIII). Surface conditions other than residue also influence the effectiveness of tillage practices in controlling erosion. These include the amount of residue incorporated, surface roughness, surface ridging, and the portion of surface disturbed (Wischmeier, 1973). Residue mixed with the surface soil by chiseling or disking is less effective than residue on the surface, but incorporation is better than removal because the incorporated residue tends to increase infiltration and decrease runoff and, hence, erosion. Wischmeier and Smith (1965) showed 40% less runoff for conservation tillage corn systems where the residue was incorporated by plowing than where it was removed at harvest. Soil loss was reduced about 12% for each 2.2 metric tons/ha (1 todacre) of corn residue Table VII Runoff and Sediment Yield from Corn Watersheds at Coshocton, Ohio, during a Severe Rainstorm on 5 July 1%9"

Tillage

(a)

Rainfall (cm)

Runoff (cm)

Sediment yield (kdha)

Plowed, clean-tilled sloping rows Plowed, clean-tilled contour rows No-tillage, contour rows

6.6

14.0

11.2

50,700

5.8

14.0

5.8

7,200

20.7

12.9

6.4

70

Slope

(I

From H m l d and Edwards (1972).

28

P. W . UNGER AND T. M. McCALLA

Table VIII Average Rainfall, Runoff, and Soil Loss on Tillage Plots at Madison, South Dakota" Average values (1965 to 1970)

Tillage practice

Rainfall (cm)

Fallow Conventional Mulch Till-plant (with slope) Till-plant (on contour)

42.0 42.0 42.0 42.0 42.0

Runoff (cm)

Soil loss (todha)

5.0 2.9 2.4 2.1

17.5 6.0 3.7 3.5 0.9

ab b c c 1.0 d

ab b c c d

From Onstad (1972). bValues within a column followed by the same letter are not significantly different (Duncan Multiple Range Test, 5% level).

incorporated (Wischmeier and Smith, 1978); therefore, the value of incorporated residue as compared with removed residue for erosion control is obvious. On soils where the amount of surface residue is limited, tillage-induced roughness and cloddiness can increase infiltration, reduce runoff velocity, and thereby reduce the potential for soil loss. Because high-intensity rains early in the season rapidly decrease the roughness and cloddiness of bare soils, the resulting soil crust must be broken by cultivation to enhance subsequent infiltration (Wischmeier, 1973). As shown in Table VII, contour ridges or furrows greatly reduce runoff and soil losses as compared with ridges with the slope. Similar decreases are possible with graded furrows (Richardson et al., 1969). A major factor involved is runoff velocity. On sloping rows, the rapidly flowing water readily transports detached soil particles down the slope, whereas on contour or graded rows, the runoff velocity is much lower. Runoff velocity and erosion can be decreased, even on sloping rows, when residue is maintained in the furrows (Kramer and Meyer, 1969; Meyer and Mannering, 1963; Mannering and Meyer, 1961; Taylor et al., 1964). Any tillage operation, even stubble mulch tillage which undercuts the surface, disturbs the surface and reduces the amount of surface covered by residue. Because soil detachability is inversely related to surface cover, it presumably increases with tillage, thus increasing the potential for soil loss. Whereas soil loosening by tillage increases infiltration as compared with no-tillage (Wischmeier, 1973), erosion can be kept to a minimum if contour plowing is used in a strip tillage system. This allows the strips of undisturbed residue to trap soil detached from the tilled zone.

29

CONSERVATION TILLAGE SYSTEMS

C. RUNOFFWATERQUALITY

The present concern for clean water demands that pollutants be kept from entering surface and ground waters. Sediment, which is the end product of soil erosion, is by volume the largest single pollutant of surface waters. It is also the principal carrier of some chemical pollutants (Stewart et al., 1975). Hence, decreases in erosion by water also decrease the pollution of surface waters by sediment and some chemicals. Water erosion control was discussed in Section V, B. The same principles that apply to controlling erosion also apply to controlling pollution. However, all soil lost from a given location does not necessarily enter downstream water because much of it may be deposited before it enters the streams. Stewart et al. (1975) roughly estimated sediment delivery ratios for drainage areas of different sizes (Table IX), but they recognized that soil texture, relief, type of erosion, sediment transport system, and areas of deposition within the watershed would all influence the amount of sediment delivered to the downstream waters. The inability to predict accurately the transport of pollutants from fields to downstream bodies of water is recognized as one of the greatest problems in recommending specific control practices for a given site (Frere et al., 1977). Accurate prediction or even measurement of pollutant transport from fields is not possible; therefore, we will present only qualitative data on runoff water quality. The transport of sediment depends primarily on the volume and velocity of water flow. When the velocity is reduced, the transport capacity is also reduced. Any sediment in excess of the reduced capacity settles out. Because larger and heavier particles settle out first, the remaining sediment has a higher percentage of fine particles. The finer material has a higher capacity per unit of sediment to absorb such chemicals as phosphorus and pesticides. Also, light-weight organic Table IX Influence of Drainage Area on Sediment Delivery Ration ~~

~~

Drainage area Square kilometers I .3 2.6 13.0 26.0 130.0 260.0 518.0

Square miles 0.5 1 .o

5.0 10.0

50.0 100.0 200.0

From Stewart et al. (1975).

Sediment delivery ratio 0.33 0.30 0.22 0.18 0.12 0.10 0.08

30

P. W . UNGER AND T. M. McCALLA

materials tend to be associated with the fine particles. The transported sediment that reaches downstream bodies of water, therefore, contains more clays, organic matter, nutrients, and pesticides than the original field soil (Frere, 1976; Frere et al., 1977).

VI. INFILTRATION AND WATER CONSERVATION A major dryland crop production goal in subhumid and semiarid regions is to store enough water in the soil between crops so that the subsequent crop will not be too severely stressed for water and, therefore, will produce a favorable yield. Even in more humid regions, storage of additional water in soil is beneficial for alleviating the adverse effects of short-term droughts. To conserve water, the water that would normally be lost by runoff must infiltrate into the soil and it must then be protected against loss by evaporation or use by weeds. Weed control is discussed in Section VII. At some locations, drainage of excess water is an important production practice, but drainage will not be discussed in this report. A. RUNOFFAND INFILTRATION

The processes of runoff and infiltration are closely related in that water that infiltrates a soil is prevented from leaving it as runoff. However, runoff reduction or even prevention does not necessarily mean that the water will infiltrate into the soil, because the water may be lost by evaporation. Some results pertaining to runoff were included in the discussion of water erosion control (Section V, B) (Tables VI, VII, VIII). The large decreases in runoff with conservation tillage result from surface residue, which dissipates the energy of falling raindrops, decreases dispersion and surface sealing, and increases infiltration. Surface mulches decreased runoff under such widely different conditions as found in the high rainfall areas of the tropics (Barnett et al., 1972; Rockwood and Lal, 1974) and the southeastern United States (Batchelder and Jones, 1972), and the low rainfall area of the U.S. Great Plains (Onstad, 1972). Some effects of residue on runoff were shown in Section V, B. Hence, only limited additional data will be used to illustrate the effects of residue on runoff and infiltration. In the tropics (Nigeria), Rockwood and La1 (1974) measured much lower runoff and soil losses from no-tillage areas than from bare-fallow and plowed areas. In Puerto Rico, Barnett et al. (1972) applied artificial rainfall and measured runoff from three soils with different slopes (Table X). Cropping treatment

31

CONSERVATION TILLAGE SYSTEMS

Table X Cropping Effects on Rainfall Runoff fkom Three Soils in Puerto R i d Rainfall' (cm) Cropping treatmentb Fallow

Tobacco Conventional tillage

Mulch tillage

Grass strips

Pangola grass Full sod Tops removed

" From Barnett ef

Runoff (cm)

Slope Soil

(%)

1

2

3

1

2

3

Humatas clay Juncos silty clay Pandura sandy loam

35 32 26

6.4 6.5 6.6

13.0 12.4 13.0

19.4 18.9 19.7

0.0 4.4 0.5

7.2 13.0 5.8

7.2 17.4 6.3

Humatas clay Juncos silty clay Pandura sandy loam Humatas clay Juncos silty clay Pandura sandy loam Humatas clay Juncos silty clay Pandura sandy loam

38 33 28 38 34 26 37 32 29

6.0 6.6 7.2 6.0 6.2 6.5 6.4 6.2 6.8

11.9 12.6 12.9 11.9 12.3 13.1 13.0 12.6 12.5

18.0 19.2 20.1 18.0 18.6 19.6 19.4 18.8 19.3

1.4 4.0 0.2 1.4 3.6 0.0 1.1 3.8 0.1

8.2 11.6 4.8 9.9 11.5 2.9 10.6 11.0 3.2

15.6 5.1 11.2 15.0 2.9 11.7 14.8 3.2

Humatas clay Humatas clay

39 46

6.0 6.0

12.7 11.9

18.7 17.9

0.1 1.7

0.7 9.1

0.7 10.8

9.6

al. (1972).

Conventional tillage was plowing, smoothing, and planting on contour; mulch tillage was planting in sod; grass strips were one row wide between three clean-tilled rows; and Pangola grass conditions were as shown. Rainfall was applied at 6.35 cm/hour for 60 minutes (Storm 1); 12.7 c d h o u r for 60 minutes (Storm 2) beginning 10 minutes after Storm 1. Results for Storm 3 are the sum for Storms I and 2.

did not greatly affect runoff on Humatas clay (Typic Tropohumult), apparently because water conductivity through this fine-textured soil was lower than the potential infiltration rate. On Pandura sandy loam (Typic Eutropept), mulch tillage and grass strips greatly reduced runoff. According to Barnett et al. (1972), the high runoff from Juncos silty clay (Vertic Eutropept) is misleading because the rainfall entered the soil, but returned as interflow at the lower end of the plots. All soils were highly aggregated and remained so throughout all storms. The imposed rainstorms had occurrence frequencies of 2 years (Storm l), 50 to 100 years (Storm 2), and 150 to 500 years (Storm 3). Batchelder and Jones (1972) measured rainfall and irrigation-water runoff from topsoil, exposed subsoil, and mulched, exposed, subsoil plots on Groseclose clay loam (Typic Hapludult) in Virginia from 1966 to 1968 (Table XI). Except in 1966, runoff always was lowest from the mulched subsoil. In 1966, irrigation runoff was higher from mulched than from unmulched subsoil. This indicated that the mulched subsoil plots were excessively irrigated. In subsequent

32

P. W. UNGER AND T. M. McCALLA Table XI Soil Management Effect on Runoff from Topsoil and Exposed Subsoil"

Topsoilb Dates

Factor

1,

I0

9 May-20 Oct. 1966

Rainfall Rainfall runoff Irrigation Irrigation runoff Rainfall Rainfall runoff Imgation Irrigation runoff Rainfall Rainfall runoff Irrigation Irrigation runoff

48.1 11.8 34.1

48.7

24 May-3 1 Oct. 1961

10 May-25 Oct. 1968

5.5

31.5 9.5 19.8 5.9 49.9 12.4 17.1

3.1

12.6 0 0 31.5 11.5

0 0 49.9 16.6 0 0

Subsoilb 1,

48.7 11.3 34.1 2.2 31.5 1.9 11.5 2.2 49.9 10.9 14.2 2.0

I0

48.7 15.8 0 0 31.5 9.1 0 0 49.9 13.7 0 0

Mulched subsoilb 1,

I,

48.7 10.4 34.1 3.4 31.5 1.9 12.5 0.3 49.9 1.2 6.2 0.1

48.1 2.7 0 0 31.5 0.3 0 0 49.9 0.5 0 0

From Batchelder and Jones (1912). Values given in centimeters. bI, and I, refer to imgated and nonirrigated plots, respectively.

years when irrigation amounts were adjusted, the mulch effectively controlled runoff. Mulches greatly decreased runoff in several seedbed preparation studies by Hays (1961) and Taylor et al. (1964) in Wisconsin. Wheel-track planting or mulch planting (field cultivator) for corn after hay decreased runoff about 50% as compared with normal seedbed preparation. A stover mulch in corn after corn decreased runoff to less than it was from corn in a rotation on soils at La Crosse and Madison. Runoff from small grain was lower with field cultivation than with fall plowing. Even less runoff occurred with mulching and field cultivation (Table XII). In all studies, yields were not greatly affected by seedbed preparation techniques, but runoff control greatly reduced soil loss. Under furrow-irrigatedconditions on Pullman clay loam at Bushland, Allen et al. (1975) evaluated rainfall and irrigation water runoff from clean- and notillage areas cropped continusouly to grain sorghum. For four irrigations in 1971 totaling 36.4 cm, runoff totaled 9.7 and 4.1 cm from clean- and no-tillage areas, respectively. For the fifth irrigation, when 6.3 cm of rain fell immediately after application of 8.2 cm of water, runoff from the respective areas totaled 5.8 and 4.6 cm. Less runoff from the no-tillage areas was attributed to slower water advance due to residue in the furrows, which allowed deeper penetration of the water than in clean-tillage furrows. The additional water caused greater plant growth, but yields were not increased because uncontrolled volunteer sorghum plants raised plant populations to levels above the optimum for grain production.

33

CONSERVATION TILLAGE SYSTEMS

Table XI1 Tillage Effects on Runoff and Crop Yields" Soil, period, crop seedbed preparation Fayette silt loam, 15% slope, 1955-1959 average (corn) Seedbed-Corn after hay-Normal Seedbed-Corn after hay-Wheel-track Seedbed-Corn after hay-Mulch Fayette silt loam, Lacrosse, 1954- 1959 average (Corn) Corn in corn-grain-hay rotation Corn after corn and corn stover mulch Miami silt loam, Madison, 1955-1959 average (Corn) Corn in corn-grain-hay rotation Corn after corn and corn stover mulch Fayette silt loam, 15% slope, 1957-1959 average (Small grain) Fall plowed Field cultivated Corn stover mulched plus field cultivated

Runoff (cm) .

Yield (kg/ha)

2.1 0.9 1.1

61 10 5440 5950

1.5 0.3

7530 7880

2.0 0.2

4930 4770

9.0 7.2 4.8

5720 5240 5540

"From Hays (1961).

B. EVAFQRATION

In many areas, evaporation accounts for the major loss of water from agricultural soils. In the Great Plains, for example, about 60% of the 50 cm of average annual precipitation is lost directly from soil by evaporation (Bertrand, 1966). Water evaporates from the soil surfaces after precipitation, but before it enters the soil; as the surface dries; and from within the soil, especially before plant canopies completely cover the surface. As plant canopies develop, evaporation decreases and transpiration increases. Soil water evaporation occurs in three stages (Lemon, 1956). In the first stage, evaporation is rapid and steady, and depends on the net effects of water transmission to the surface and the aboveground conditions, such as wind speed, temperature, relative humidity, and radiant energy. During the second stage, evaporation decreases rapidly as the soil water supply decreases. Soil factors control the rate of water movement to the surface, and aboveground factors have little influence. Evaporation during the third stage is extremely slow and is controlled by adsorptive forces at the liquid-solid interface. According to Lemon (1956), the greatest potential for decreasing soil water evaporation lies within the first two stages. Potential methods include (1) decreasing turbulent transfer of water vapor to the atmosphere; (2) decreasing capillary continuity; and (3) decreasing capillary flow and water-holding capacity of surface soil layers. Crop residue has been studied as a potential mulch for decreasing evaporation. Mulches greatly affect first-stage evaporation (Bond and Willis, 1969; Unger,

34

P. W. UNGER AND T. M. McCALLA

1976; Unger and Parker, 1976), but the long-term effect of mulches on evaporation is difficult to establish because of their interacting influences on water infiltration, distribution in soil, and subsequent evaporation. Consequently, higher soil water contents resulting from surface mulches may be due to lower evaporation, but water infiltration and distribution in soil may be involved also, especially in field studies where the researcher has little control over soil wetting by precipitation. In field studies in Colorado, Montana, and Nebraska, precipitation storage as soil water during fallow was 16% with no residue and 34% with 11 metric tons/ha of wheat straw on the surface (Greb ef af., 1967). When Unger (1978a) placed wheat straw on Pullman clay loam at Bushland, precipitation storage during fallow ranged from 23% with no mulch to 46% with 12 metric tons of mulch/ha. Dryland grain sorghum grown after fallow yielded 1780 and 3990 kg/ha with the 0 and 12 metric todha mulch treatments, respectively. Unger and Wiese (1979) used no-, sweep, and disk tillage for residue management and weed control from wheat harvest until sorghum planting in an irrigated wheatdryland grain sorghum cropping system. Precipitation storage, sorghum grain yields, and water-use efficiency were highest with no-tillage and lowest with disk tillage (Table XIII). Even though factors besides evaporation control undoubtedly were involved, these studies showed that residue from high-residue producing crops, when maintained on soil as a mulch, can greatly increase precipitation storage and crop yields in areas where water for crop production is limited. The thicker the mulch, the more effectively it decreases evaporation (Bond and Willis, 1969; Hanks and Woodruff, 1958; Unger and Parker, 1976). Because material density greatly influences the thickness obtained with a given weight of material, low-density materials, such as wheat straw, more effectively decrease evaporation than sorghum stubble or cotton stalks, which are more dense. For Table XI11 Effect of Tillage Method on Average Precipitation Storage, Sorghum Grain Yield, and Water-Use Emciency for the Sorghum Crop in an Irrigated Wheat-Dryland Grain Sorghum Cropping Systema ~~~~~~

Tillage method Factor

No-tiliage

Sweep

Disk

Precipitationb storage (%) Grain yield (kg/ha) Water-use efficiency (kdha-cm)

35 3140 89

23 2500 77

15 1930 66

From Unger and Wiese (1979). Precipitation averaged 34.8 cm during fallow and 26.4 cm during the growing season.

CONSERVATION TILLAGE SYSTEMS

35

similar evaporation decreases, about two and four times as much sorghum stubble and cotton stalks, respectively, were needed as compared with wheat straw on a weight basis (Unger and Parker, 1976). Crop residue mulches effectively decrease evaporation during the first stage (Bond and Willis, 1969; Unger, 1976; Unger and Parker, 1976). However, for maximum water conservation over long periods, either enough water must be added to penetrate deeply into the soil profile or large amounts of residue must be present (Bond and Willis, 1971; Gardner and Gardner, 1969; Unger, 1976). Because the water content near the surface of mulched soil often is higher than that of bare soil, especially soon after water additions, mulches are very useful for improving seedling establishment (Army et al., 1961; Bertrand, 1966; Smika, 1976b).

VII. WEED CONTROL Weeds compete with crops for water, nutrients, and light; therefore, effective weed control is essential if crops are to produce maximum yields under the prevailing environmental conditions. The technology of weed control uses cultural methods and herbicides, either singly or in various combinations, to prevent weed seedling establishment and to eliminate those seedlings or plants that have survived the initial control measures. Best weed control is obtained when differences in the biological characteristics of crops and competing weeds are exploited (Wiese and Staniforth, 1973). A. PROBLEM AREAS

The life history of weedy plants has a major effect on methods used to control them in a given crop production system. Weeds may be annuals, biennials, or perennials. Annual weeds can be either summer or winter annuals. Summer annuals, whose seeds germinate during warm weather, often have growth habits similar to summer crops. Troublesome summer annuals include barnyard grass [Echinochloa crusgalfi (L.) Beauv.], crabgrass [Digitaria sanguinalis (L.) Scop.], sandbur (Cenchrus), pigweed (Amaranthus sp.), and green foxtail (Setaria viridis L . ) . Volunteer crop plants of summer and winter annuals may cause problems in succeeding crops when the crops are grown continuously. Winter annual weeds have life cycles similar to fall-planted small grains. Especially troublesome in winter wheat fields are cheatgrass (Bromus secalinus L.), hairy chess (Bromus cornmutatus Schrad.), downy brome (Bromus tectorum

36

P. W . UNGER A N D T. M. McCALLA

L.), tansy mustard [Descurainia pinnata (Walt.) Britt.], and henbit (Larniurn urnplexicaule L.). Biennial weeds live more than 1 year but less than 2. Some common biennials are wild carrot (Daucus carota L.), burdock (Arctiurn sp.), common mullein (Verbascum thapsus L.), and bull thistle [Cirsiurn vulgare (Sair) Tenore], but they are not considered to be especially troublesome (Wiese and Staniforth, 1973). Perennial weeds may spread by seeds, rhizomes, bulbs, tubers, or stolons, and once established, they compete vigorously with most annual crops and may be difficult to control. Weed species differ in different sections of the country, but some of the most troublesome perennial weeds in the United States are Johnson grass [Sorghum halepense (L.) Pers.], quack grass [Agropyron repens (L.) Beauv.], nutsedge (Cypersus sp.), field bindweed (Convolvulus arvensis L.), leafy spurge (Euphorbia esula L.), perennial sow thistle (Sonchus arvensis L.), Bermuda grass [Cynodon dactylon (L.) Pers.], Canada thistle [ Cirsium arvense (L.) Scop.], horse nettle (Solanurn carolinense L.), silverleaf nightshade (Solanum elaeagnifolium Cav.), Russian knapweed (Centaurea repens L.), and woollyleaf bursage (Franseria tomentosa Gray) (Wiese and Staniforth, 1973). B . CONTROL WITH TILLAGE

Tillage helps control weeds by ( 1 ) killing emerging seedlings; (2) burying weed seeds and delaying growth of perennial weeds; (3) leaving a rough surface to hinder weed seed germination; (4) providing enough loose soil at the surface to permit effective cultivation; ( 5 ) leaving a clean uniform surface for efficient action of herbicides; and (6) incorporating herbicides when necessary (Richey et al., 1977). Different tillage methods affect soils differently; therefore, they also affect the degree of weed control obtained with their use. Some weeds can be controlled with clean tillage methods involving moldboard plowing or disking (one-way, tandem, offset), but these tillage methods either bury or mix crop residue with soil, thus increasing the potential for erosion and decreasing the potential for water conservation. Alternate tillage systems that retain varying amounts of residue on the surface include chisel, sweep or blade, till-plant, and no-tillage. Normally, cultivation is required for growing season weed control unless herbicides are used, as they are in the no-tillage system. Chisels are widely used to loosen soils while retaining surface residue, but weed control is difficult with chisels because only narrow bands in soil are disturbed. Better weed control is generally obtained with stubble mulch tillage (sweeps, blades, etc.) than with chisels because the entire surface is undercut, thereby severing the deep roots of weeds. Stubble mulch tillage, however, is less effective than plowing because seeds are not buried and the soil and weeds are not inverted as with plowing. With stubble mulch tillage, soil often remains in contact with roots, and plants may continue to grow if rain occurs soon after

CONSERVATION TILLAGE SYSTEMS

37

tillage or if the soil is wet at tillage. Small weeds may be especially hard to kill with stubble mulch tillage. Skewtreaders and rodweeders can improve weed control in a stubble mulch system. The till-plant system provides essentially a clean seedbed, but only partially covers weeds and residue (Williams and Wicks, 1978). Weeds in the seedrow can be controlled with herbicides or by a rolling cultivator, which can operate in thc interrows without clogging (Richey et al., 1977). The till-plant system, even with some bare surfaces, can effectively control erosion when tillage is done on the contour (see Section V, B). With the no-tillage system, soil disturbance or loosening is limited to that required to place seeds in soil, and weeds are controlled with herbicides. Also, the system leaves the weed seeds on the surface in a poor environment for germination. When herbicides fail, tillage may be necessary to save a crop. However, effective mechanical cultivation may be difficult in an established no-tillage crop because the soil may be too firm (Richey et al., 1977). C . CONTROL WITH HERBICIDES

Herbicides are relied upon to control some weeds in many cropping systems involving tillage and to control all weeds in no-tillage systems. The mode of action of herbicides and the type of plants to be controlled largely determine which herbicides can be used in a particular cropping system. Herbicides must be compatible with present and future crops to avoid crop damage. Some herbicides must be incorporated with soil and, therefore, tillage is required. Disks, rotary tillers, and rolling cultivators are generally satisfactory for incorporating herbicides. Plows, even though they invert the soil, are less effective because they mix herbicides with soil only slightly. Chisels and sweep plows also cause little mixing of herbicides with soil. Tillage methods also strongly influence the effectiveness of surface-applied herbicides because they affect the amount of residue that remains on the soil surface. For maximum effectiveness, surface-applied herbicides should be uniformly placed on the entire soil surface. With some tillage methods, residue may intercept the herbicides, leaving some areas of soil untreated. In Indiana, corn plant residue covering 85% of the soil surface intercepted 30% of the applied atrazine [2-chloro-4-(ethylamino)-6-(isopropylamino)-~-triazine]. Many areas under the residue remained untreated (Richey et al., 1977). In Texas, however, Unger et al. ( I97 I ) and Unger and Wiese (1979) completely controlled volunteer wheat with surface-applied atrazine when the atrazine was applied to areas having up to 1 1 metric tons/ha of wheat straw on the surface. Contact herbicides also controlled volunteer wheat under high residue conditions (Unger, 1977, unpublished data). Use of herbicides in no-tillage systems often has resulted in a shift in weed

38

P. W.UNGER AND T. M . McCALLA

species. In Indiana, herbicide-resistant weeds, such as milkweed (Asclepias syriaca L.), fall panicum (Panicum dichotomij7orurn Michx.), and briars, increased with continuous no-tillage (Richey et al., 1977). In Kansas, the weed population in a wheat-sorghum-fallow system shifted from broadleaf species, which were susceptible to atrazine, to sandbur, which was resistant. As a consequence, yields decreased unless sandbur was controlled with tillage. The herbicide-tillage combination resulted in yields of 3700 kg/ha as compared with 2300 kg/ha with sweep tillage alone (Phillips, 1969). D. CONTROL WITH ROTATIONS

Some weed species in some crops are difficult to control with tillage or herbicides because of the similar biological characteristics of the weeds and crops, and the vigorous growth habits of some perennial weeds. Species most difficult to control vary, depending on locations (Southern Weed Sci. SOC., 1979), but some examples are cheat, wild oats (Avena fatua L.), and downy brome in winter wheat; barnyard grass, foxtail (Setaria sp.), sandbur, and fall panicum in corn or grain sorghum; cocklebur (Xanthium sp.) and velvetleaf (Abutilon theophrasti Medic.) in soybeans; cocklebur, field bindweed, prickly sida (Sisa spinosa L.), spurred anoda [Anoda cristata (L.) Schlecht.], sicklepod (Cassia obtusifolia L.), and silverleaf nightshade in cotton; and Johnson grass, Bermuda grass, and nutsedge in all crops. Because problem weeds may be difficult to control with tillage or herbicides in continuous cropping systems, a crop rotation may be the most effective and economical control method available. Fields with summer annual weed problems can be rotated to winter grain crops. Then the problem weed can be controlled with tillage or herbicides during the period between crops. Conversely, fields with troublesome winter annual weeds can be rotated to spring- or summerplanted crops. Crop rotation also makes it possible to select crops that are most competitive against perennial weeds (Wiese and Staniforth, 1973). A crop rotation, or even skipping a crop, and using intensive weed control measures during the noncropped period may be necessary to reduce or eliminate heavy infestations of troublesome perennial weeds.

VIII. INSECTS AND PLANT DISEASES A. INSECTS

In recent years, considerable attention has been given to insect problems with no-tillage systems (Gregory and Musick, 1976). Many of the insect problems are not affected by tillage system, but others may be greater with no-tillage. Accord-

CONSERVATION TILLAGE SYSTEMS

39

ing to Phillips and Young (1973), insects causing slightly greater problems with no-tillage included sod webworms, cutworms, armyworms, and root aphids (various species of each). Phillips and Young (1973) also reported that slugs (Deroceras laeve Muller) caused greater problems with no-tillage. Other reports, however, have shown different results for some insects. For example, Phillips and Young (1973) showed that wireworm (Melanotus cribulosus LeConte, and others) problems were similar for the two systems, but Musick and Beasley (1978) showed that wireworms caused more damage with no-tillage. The reported differences possibly resulted from differences in location, previous crops, and overall management of the systems. B . PLANTDISEASES

Considerable work has been done in recent years concerning the influence of crop residue on plant disease, as indicated by Cook et al. (1978). As for insects, disease problems generally were similar for conventional and no-tillage systems (Phillips and Young, 1973). Exceptions included anthracnose and yellow leaf blight for corn; and bacterial blight, bacterial pustule, wildfire, anthracnose, and sclerotial blight for soybean for which the disease incidence was greater with no-tillage. Other exceptions included “take-all” for small grain and Pytophtora, Rhizoctonia, Fusarium root rot, and stem rot for soybean for which the disease incidence was lower with no-tillage. According to Boosalis (1979), conservation tillage may affect plant disease by providing a habitat for plant pathogens; serving as foci for the dissemination of inoculum; affecting the growth and multiplication of pathogens associated with residue; altering the physical and chemical components of the soil; creating an environment hostile or beneficial to the pathogen; through residue decomposition, producing compounds that adversely or beneficially affect the pathogen; using pesticides that may affect the pathogen; and in altering the physical and chemical environment of the soil surface with residue that may affect the growth and configuration of roots to make them more resistant or vulnerable to soilborne diseases. An example of an indirect beneficial influence of crop residue was the reduction of stalk rot of sorghum in a reduced tillage system in Nebraska (Boosalis and Cook, 1973).

IX. SOIL TEMPERATURE A. EFFECTOF SURFACE RESIDUE

Plant residue placed on soil can significantly influence soil temperature. The primary mechanism of this effect is the change in radiant energy balance (Van

40

P. W . UNGER A N D T. M. McCALLA

Doren and Allmaras, 1978), but an insulating effect may be involved also, because mulched soil normally is warmer than bare soil during cold weather, even during daylight hours (Unger, 1978b). The radiation balance is influenced by heating of air and soil, evaporating of soil water, and reflecting of incoming radiation by surface residue (Van Doren and Allmaras, 1978). As reflectance increases, soil temperature generally decreases. The insulating effect increases as residue mulch thickness increases (Unger, 1978b). B. RESIDUEFACTORS INVOLVED

Residue characteristics involved in the reflectance of incoming radiation include residue age, color, geometry (whether standing or matted on the surface), distribution, and amount. The general condition of the plant residue must be considered because plant aging, which causes yellowing and bleaching of the plants, increases reflectance in the visible wavelengths, but decreases it in the near infrared wavelengths. Further aging and decomposing of residue (after crop harvest) usually produces gray and darker shades, which reduce reflectance over the visible range (Van Doren and Allmaras, 1978). Because reflectance is highest with bright colored residue, the temperature difference between mulched and bare soil is greatest with bright-colored residue, then decreases as the residue ages. Gausman et al. (1975) compared the reflectance from bare soil with that from standing or surface-matted sugar cane (Saccharum oficinarum L.) after the cane was frozen and had bleached and turned yellow. Reflectance was greater from matted residue than from bare soil, but less from standing sugar cane residue than from bare soil or matted residue at all measured wavelengths, apparently because shadows occurred in the standing stubble. Gausman et al. (1975), using dried and bleached avocado (Persea americana L.) leaves, showed also that maximum reflectance is reached asymptotically as the thickness of target increases. For avocado, two leaves (leaf area index of 2) gave near maximum reflectance. For randomly placed residue, such as wheat straw or corn stover after harvest or initial subsurface tillage, complete surface coverage would be needed to obtain maximum reflectance. Under field situations, residue cover often is incomplete because of natural distribution or incorporation by tillage; therefore, some reflectance value less than maximum would be expected for a given residue condition (Van Doren and Allmaras, 1978). The greater reflectance from surface residue than from bare soil causes less soil heating and potentially less evaporation of soil water. However, Hanks et al. (1961), under conditions of their research, found no direct relationship between net radiation and evaporation. Factors within the soil, especially after the surface dried, apparently were more important than incoming radiation in determining water losses by evaporation during the fallow season.

CONSERVATION TILLAGE SYSTEMS

41

40

P 35 I

W

30 I-

4

a 25

a W I-

20 I

1

22 23 24 25 DATES-AUQUST 1973

FIG. 1. Temperatures (10-cm depth) of Pullman clay loam at Bushland, Texas, during a hot period (21 through 25 August 1973) as affected by wheat straw surface mulch rates (metric tons/ha, numbers at lines). Air temperature is also given (from Unger. 1978b).

Radiation reflectance approaches a maximum as surface coverage by residue approaches 100% (Van Doren and Allmaras, 1978). Thus, residue in excess of those required to obtain 100% coverage should have no effect on soil temperature if radiation reflectance alone is involved. However, residue mulches at rates greater than those required for complete surface coverage do affect soil tempera-

10

P I

5

W

a 3

I - 0 4

a W

a

I -5 W

I-

a

;-lo

-I 5

FIG. 2. Temperature (10-cm depth) of Pullman clay loam at Bushland, Texas, during a cold period ( I through 7 January 1974) as affected by wheat straw surface mulch rates (metric tons/ha. numbers at lines). Air temperature is also given (from Unger, 1978b).

42

P. W . UNGER AND T. M. McCALLA

21

y

2c

W

a

2 I1 4

a

W

g

IC

W

k

c Y

C 0

1

2

-

0.084 X

+ 0.104

4 MULCH RATE

!*

' 0.943

8

- tonrlha

.

COLD PERIOD

12

FIG.3. Relationships between mean soil temperature (Y)and wheat straw mulch rates (X metric tons/ha) for seasons of fallow (spring, fall, winter, summer), and a 5-day hot, a 7-day cold, and a 5-day near-sorghum-planting period at Bushland, Texas, on Pullman clay loam (from Unger, 1978b).

tures (McCalla and Duley, 1946; Unger, 1978b), apparently because the mulches have an insulating effect. Effects of wheat straw mulches at rates from 0 to 12 metric tons/ha on temperature of Pullman clay loam at Bushland during a hot and cold period are shown in Figs. 1 and 2, respectively. The 4-metric todha rate covered almost 100% of the surface (Unger, 1978b). Relationships between mean soil temperature and mulch rates during different seasons and periods are shown in Fig. 3. The mulch rate effect on temperature was greatest for the hot period and least during winter. The trends (Figs. 1, 2, and 3), in general, were similar to those reported for residue-mulched soils by Allmaras et al. (1973) and Van Doren and Allmaras (1978).

c.

BIOLOGICAL

EFFECTS OF RESIDUE

In cooler parts of the United States, such as in the Corn Belt or the northern Great Plains, cool soil temperatures under a mulch in the spring of the year may adversely affect seed germination or plant growth. Favorable soil temperatures for germination and seedling emergence may occur up tn 7 days later in notillage than in conventional tillage seedbeds in the northern United States, but no temperature-induced delay of planting is expected at southern U.S. locations (Unger and Stewart, 1976).

CONSERVATION TILLAGE SYSTEMS

43

Early planting of warm-season crops favors higher yields when the frost-free period is limited or when high temperatures, droughts, or insect and disease problems at later growth stages adversely affect yields. Because surface residue with stubble mulch and conservation tillage decreases soil temperature, considerable research has been conducted on how soil temperature affects plant growth. In Iowa, chopped corn stalks applied at a rate of 0 to 9 metric tons/ha lowered soil temperatures at a 10-cm depth by an average of 0.4'C/ton of mulch during May and June (Burrows and Larson, 1962). Straw mulches caused similar temperature decreases at other northern U.S. locations (Allmaras et al., 1964; Willis et al. 1957; Van Wijk et al., 1959), which decreased early corn growth. In South Carolina, where soil temperatures were considerably higher than those in Iowa, Ohio, and Minnesota, a mulch did not appreciably affect corn growth rate (Van Wijk et al., 1959). In central Texas, a straw mulch decreased soil temperature, but had little or no effect on early grain sorghum growth as compared with that on bare soil (Adams, 1962, 1965, 1967, 1970). At a much higher elevation in northwest Texas, increasing the straw mulch rate delayed the time that soil reached a favorable temperature for sorghum germination and growth. However, the temperature was near optimum before normal sorghum planting dates for the region, and thus the mulches affected the time of sorghum emergence only slightly. The sorghum on plots with large amounts of mulch grew more slowly, but yielded more than that on plots with little or no mulch. Yields were higher because more water had been stored in the plots with large amounts of mulch on the surface during the fallow period that preceded the sorghum crop (Unger, 1978b). Because most biological, chemical, and physical reactions depend on temperature, mulch-induced temperature differences in soil undoubtedly affect factors other than seed germination, seedling emergence, and plant growth. Interactions of soil temperature and soil water content resulting from a surface mulch greatly altered the distribution of corn roots in Minnesota (Allmaras and Nelson, 1971). Also, there is evidence that root zone temperature influences the uptake of water and nutrients and the distribution of products of photosynthesis within the plant. However, the final effect of these processes on plant growth and yield is not fully understood (Nielsen, 1974).

X. SOIL STRUCTURE AND OTHER PHYSICAL PROPERTIES

Water conservation and water and wind erosion control are major goals of conservation tillage systems. To achieve these goals, the conditions at the soil surface and within the profile must allow water to enter the soil readily and still keep the soil resistant to erosion. Soil physical factors that influence water infiltration and soil erodibility include soil aggregation, porosity, and density.

44

P. W . UNGER AND T. M. McCALLA

A . AGGREGATION

Soil aggregation refers to the cementing or binding together of several soil particles into secondary units. Water-stable aggregates, which do not disperse, are of special importance for high water infiltration, good soil structure, and good plant growth. Large stable aggregates at the soil surface are important also for controlling erosion by wind and water. The binding substances for natural soil aggregates have mineral or organic origins. Soils of humid tropic and subtropic regions have generally high infiltration rates, even on steep slopes (Table X). Many particles in tropic soils are the size of sand grains and consist mainly of altered minerals cemented by iron (Donahue et al., 1977). Because of the stable particles, the infiltration rate into tropic and subtropic soils often is similar to that for deep sands, and the rate remains high with prolonged rainfall. Organic substances that contribute to soil aggregation are derived from plant materials, either after alteration by soil animals, bacteria, and fungi, or directly from the plants. Earthworms beneficially affect soil structure by increasing infiltration (Hopp and Slater, 1961). While feeding on organic materials and burrowing in soils, earthworms secrete gelatinous substances that coat and stabilize soil aggregates. Water-stable aggregates are formed also with water-insoluble gummy substances secreted by bacteria, fungi, and actinomycetes (Donahue et al., 1977). Earthworm activity and intensive tillage are highly incompatible. Hence, there are few earthworms in most cultivated soils. Bacteria and other microorganisms, however, feed on decaying plant roots and other plant parts returned to the soil by tillage. Besides enhancing aggregation and thus water infiltration, soil microorganisms influence soil productivity through their effect on plant nutrients. For maximum earthworm activity, no-tillage is desirable. Where the soil is tilled, enough crop residue should be kept on or returned to the soil to provide an abundant food source for soil organisms, thus providing the potential for increased soil aggregation and water infiltration. The direct influence of plants on soil aggregation is manifested through exudates from roots, leaves, and stems, and leachates from weathering and decaying plant materials, which bind soil particles together; plant canopies and surface residue, which protect surface aggregates against breakdown due to raindrop impact, abrasion by wind-blown soil, and dispersion in flowing water; and root action in soil, which promotes the formation of aggregates. If aggregates formed through these processes are subsequently maintained on the surface, water infiltration will be higher than it is in intensively cultivated, poorly aggregated soils (Donahue et al., 1977). Crop rotations involving grasses and legumes have long been known to increase soil aggregation and maintain organic matter contents at higher levels than

CONSERVATION TILLAGE SYSTEMS

45

do continuous row crops (Johnston et al., 1943; Mazurak et al., 1955; Van Bavel and Schaller, 1951; Wilson and Browning, 1946). On Marshall silt loam (Typic Hapludoll) in Iowa, aggregates were largest with continuous bluegrass (Poa pratensis) and successively smaller after red clover (Trifolium prateme), oats, and corn in a 10-year rotation, and after continuous corn. The clover maintained a loose, granular structure, whereas continuous corn resulted in a cloddy soil that was difficult to manage. With continuous corn, organic matter content decreased from 3.39% in 1931 to 2.86% in 1942. Organic matter contents with the rotation and with continuous bluegrass were similar. Less runoff and soil erosion were associated with the larger aggregates and higher organic matter contents. With limited water, yields of rotation and continuous corn were similar, but with adequate water, yields were higher with rotation corn (Johnston er al., 1943). Similar results were reported by Van Bavel and Schaller (195 1) and Wilson and Browning (1946). When row crops replaced sod crops, aggregation and infiltration decreased and soil losses generally increased (Adams, 1974; Jensen and Sletten, 1965; Mazurak and Ramig, 1963; Van Bavel and Schaller, 1951). The residual effect on aggregation increased with the age of the sod before plowing. Replacing grain crops with grasses increased aggregation and water infiltration, which generally improved with the age of the sod (Mazurak and Conard, 1959; Mazurak and Ramig, 1962; Mazurak et al., 1960). Cool-season grasses, as a group, more favorably affected aggregation and water infiltration than warm-season grasses (Mazurak and Conard, 1959). About 4 years in sod was needed before substantial increases in water infiltration were measured (Mazurak et al., 1960). In addition to retaining water on soils, thus providing more time for infiltration, crop residue and growing crops protect soil surfaces from dispersion due to raindrop impact and flowing water. The protection from flowing water may be important when the soil is irrigated or when precipitation causes runoff. The protection against dispersion maintains the favorable surface structure, which decreases surface sealing and, therefore, permits more rapid infiltration than would occur through a dispersed, sealed surface layer. Surface protection against dispersion is largely a function of the amount of coverage that the surface materials provide. Because the residue density and diameter differ among different crops, equal weights of different residue provide different amounts of surface coverage and, thus, differently influence infiltration (Van Doren and Allmaras, 1978). For example, wheat straw is about three times more effective than German millet (Setaria italica) straw and seven times more effective than grain sorghum stalk per unit weight of material for preserving infiltration. The protection provided by growing crops, like that of residue, is related to the extent of surface coverage. The importance of large (>0.84 mm), stable, dry aggregates or clods in the control of wind erosion was discussed in Section V, A.

46

P. W . UNGER AND T. M. McCALLA

B . POROSITY AND DENSITY

Soil porosity and bulk density are inversely related; therefore, any practice that affects one also affects the other. The factor generally determined is bulk density. Except in unusual soils, the bulk density of the tillage layer was normally lower in plowed soil than in unplowed soil, such as areas in grass or soil horizons not recently plowed (Unger, 1970, 1972). However, the method of plowing had little effect on soil density when relatively small amounts of crop residue were involved (Blevins et ul., 1977; Johnson, 1950; Unger, 1969; McCalla, 1959). With increasing amounts of residue, soil bulk density normally decreases (Black, 1973; Juo and Lal, 1977; Koshi and Fryrear, 1973). Bulk densities in the 0- to 15-cm layer of minimum-tillage soil in Nigeria were 1.38, 1.37, 1.49, 1.46, and 1.59 g/cm3 for bush fallow (bush regrowth), Guinea grass (Panicum maximum), pigeon pea (Cujunus cujun Millsp.), corn with residue returned, and corn with residue removed, respectively (Juo and Lal, 1977). Except for the last treatment, all residue was returned to the plots as a surface mulch each year. The low density in the fallow plots (bush fallow, Guinea grass, and pigeon pea) was attributed to high biological activity, which resulted in a porous surface horizon. Greater density in the corn plot without residue than with residue resulted from compaction of the surface layer (Juo and Lal, 1977), apparently because of no protection by surface residue against raindrop impact and soil dispersion. Under lower rainfall conditions, compaction of bare soil due to raindrop impact seems to be slight, because a surface mulch mainly decreased compaction due to tractor traffic (Koshi and Fryrear, 1973). C. OTHERPHYSICAL PROPERTIES

In addition to the previously mentioned physical properties, others affected by tillage include soil texture, crusting, hydraulic conductivity, and water storage capacity. Tillage-induced texture changes result primarily from the mode of action and depth of tillage. Plows that do not invert soil have little effect on soil texture. Soil-inverting plows can cause major changes in texture at the surface, especially when the texture changes rapidly with depth and plowing is deeper than previous plowing. Erosion on bare soil plots and earthworm activity in residue-covered plots also contribute to texture changes. In Nigeria, the amount of gravel increased by 5 to 7% and silt and clay decreased by 4 to 6% in corn plots without residue as compared with that in residue-covered plots (Juo and Lal, 1977). The fine materials apparently were eroded from the plots by runoff water. Wind erosion causes similar changes in texture. A sandy soil initially deep-plowed to bring clods to the surface to aid in wind erosion control increased the clay content

CONSERVATION TILLAGE SYSTEMS

47

from 4 to 14%. Within 5 years, the clay content was at about 4% again because wind erosion during that period had removed or buried most of the clay initially brought to the surface (Chepil et al., 1962). Tillage methods that leave bare soil in the planted row may cause severe crusting and seedling emergence problems if heavy rainfall occurs before emergence. Crusting at other times may decrease infiltration and reduce plant growth. Any tillage method that keeps residue on the surface and, thereby, protects the soil against dispersion by raindrop impact and ponded or flowing water decreases crusting. Surface residue resulting from stubble mulch and especially no-tillage practices greatly reduce soil crusting (Johnson, 1950; Juo and Lal, 1977; Lal, 1976 Unger, unpublished data). When a crust has formed, it must often be broken to obtain satisfactory plant populations. Breaking the crust also may enhance subsequent infiltration of water. The saturated hydraulic conductivity of soil increases as soil porosity increases and density decreases. Conductivity was 46 and 65% lower in corn plots with and without surface residue, respectively, than in fallowed plots with residue (bush fallow, Guinea grass, pigeon pea). Juo and La1 (1977) reported that the amount of surface mulch, the type of vegetative cover, and the differences in rooting pattern probably influenced the saturated hydraulic conductivity of the soil. When Koshi and Fryrear ( 1 973) placed a cotton bur mulch on Acuff loam (Aridic Paleustoll) at rates from 1 1.2 or 22.4 metric tons/ha, hydraulic conductivity in crop rows was eight times greater with mulch than it was in bare soil. The higher infiltration rate and hydraulic conductivity that result from tillage practices keeping residue on the surface allow soil profiles to be more readily refilled with water. Consequently, soil water content often is higher in reducedor no-tillage cropping systems. Although higher water content may be detrimental to plant growth and yields on some soils at humid locations (Boone et al., 1976; Van Doren and Triplett, 1969), the additional water generally improves crop growth and yields, especially during short-term droughts and at subhumid and semiarid locations. Besides allowing soil profiles to be refilled more readily with water, reducedor no-tillage systems may also change the amount of plant-available water held in a soil. The water-holding capacities (difference between retention at -0.1 and - 15 bar matric potentials) were 13.6, 11.2, 13.7, 14.9, and 9.1% by weight for soil from bush fallow, Guinea grass, pigeon pea, corn with residue, and corn without residue plots, respectively (Juo and Lal, 1977). La1 (1976), who found higher water-holding capacities on no-tillage plots than on plowed plot%, attributed the differences to changes in organic matter content and texture in the surface horizon of plowed plots. The type of change was not specified, but apparently involved decreases in organic matter and fine soil particles in plowed plots because, when these remained constant, the available water-holding capacities of coarse-textured core and sieved soil samples were similar. As clay

48

P. W.UNGER AND T. M. McCALLA

content increased, the water-holding capacity, based on - 1/3 and - 15 bar matric potentials, increased slightly (Unger, 1975).

XI. CHEMICAL EFFECTS AND MICROBIAL ACTIVITY Some chemical and microbial effects of leaving crop residues on the surface are shown in Table XIV. Doran (1980) provided some additional information. Soil beneath the plant residue cover is generally cooler, wetter, and less aerated than where residue is plowed under. Doran, in his study of surface soil from long-term tillage plots at seven locations around the United States, found that the surface layer (0 to 7.5 cm) of most reduced-tillage soils had higher microbial populations, higher phosphatase and dehydrogenase enzyme activity, and higher levels of total nitrogen and potentially mineralizable nitrogen than conventionally tilled soils. Aerobic microorganism counts increased 10 to 80%, and anaerobic bacteria, including denitrifiers, increased 60 to 300% in the surface of reducedtillage plots compared to conventional tillage. At the depth adjacent to the plow layer (7.5 to 15 cm), populations of aerobic organisms (especially nitrifiers) with Table XIV Effect of Stubble Mulching on Some Chemical and Biological Properties of the Soil as Compared with Plowing" Type of determination ChemicalSoil Ammonia loss Nitrites Nitrates Nitrification rate Organic matter PH HCI-soluble phosphorus Biological Crop yields Bacteria Actinomycetes Fungi Earthworms Nematodes Denitrifiers Azotobacter Legume bacteria

Stubble mulch compared to plowing

Slightly higher with legume residues No difference, and low amount present About 5 to 10% less No difference

1

May be slightly higher in surface 2.5 cm of soil; in 2.5- to 15.2cm depth, no difference

Variable-may

be higher in dry years and lower in wet years

Greater number of organisms in surface 2.5 cm with residues on surface

May be higher number in surface layer of soil

1

From McCalla (1958).

No difference in numbers or effect on nodulation

CONSERVATION TILLAGE SYSTEMS

49

plowing were significantly higher than those with no-tillage; however, populations of facultative anaerobes and denitrifiers were higher with no-tillage. This indicates that the biological environment of no-tillage soils is less oxidative than that for conventional tillage. Under such conditions, organic matter and total nitrogen would tend to increase. The significance of higher microbial populations in surface soil and more denitrifiers has not yet been identified, but these results indicated that soils with residue on the surface may need more nitrogen than those with plowed surfaces. However, corn used nitrogen more efficiently in no-tillage systems than in conventional systems (Moschler and Martens, 1975; Moschler et al., 1972) and wheat gave variable results with respect to nitrogen fertilizer and tillage system. More nitrogen was needed with no-tillage than with conventional tillage on heavy clay soils (Davies and Cannell, 1975). Phosphorus tends to increase in the surface soil of mulched soil, and is available to plants (Fink and Wesley, 1974). However, only fragmentary evidence is available in regard to transformation and availability of other nutrients needed by the crop. Phytotoxic substances that occur in plant residue or are produced by microorganisms may, in some instances, inhibit plant growth and may be related to reduced yield (McCalla and Norstadt, 1974; Elliott et al., 1978).

XII. ECONOMICS To be economically advantageous over an existing system, a new crop production system must be either less expensive or more efficient or both. A new production system is less expensive if it requires less labor, fuel, and equipment. A system is more efficient if it increases the quantity or improves the quality of products to be sold or used in relation to the production inputs. Because of rapidly changing prices, assigning dollar values to various cropping systems has little meaning. We will, therefore, discuss factors that affect expenses and income, and give only limited data regarding the economics of different systems. Labor and equipment (tractor, plows, fuel, etc.) expenses for crop production can be reduced by eliminating field operations, by reducing the number of time-intensive operations, or by using larger equipment. A lower labor and equipment requirement is a major advantage of the reduced- and no-tillage cropping systems because three, and sometimes more, operations can be eliminated (Allen et al., 1977; Hough, 1979). The labor and equipment savings are offset to some degree by higher expenses for herbicides. The high cost of herbicides was a major deterrent to adoption of early no-tillage systems. However, weed control with herbicides rather than with tillage is now more economical in some cropping systems (Unger and Wiese, 1979; Wiese et al., 1979). With the same size tractor, the labor and equipment requirement per unit area is greatly influenced by the type of tillage operation. As tillage depth and inten-

50

P. W.UNGER AND T. M. McCALLA

sity increase, the time required to perform the operation increases. With factors such as soil type and water content unchanged, the time required to perform different operations is related to the amount of fuel expended. Allen et al. (1977) reported fuel consumption values for performing different operations on Pullman clay loam at Bushland (Table 111). Some of the differences were related to depth of tillage. However, moldboard plowing required the most fuel and was followed in decreasing order by chiseling (narrow spacing), disking, and sweep plowing. The values would be different for other soils, plowing depths, and soil water contents, but for all conditions, eliminating the fuel-intensive operations reduces the labor and equipment requirement for tillage. The differences among systems would be minimized where two or more operations with disk or sweep plows are needed to obtain the weed control provided by one moldboard plowing. A further labor savings is possible by using larger equipment. However, larger equipment may require a more skilled operator and is also more costly. Thus, when considering the purchase of larger equipment, all advantages (labor savings, timeliness of operations, etc .) must be weighed against possible disadvantages (higher tractor and equipment costs, need for higher-skilled labor, alternate use of unused labor). When production expenses remain unchanged, then crop values must be increased to obtain higher returns from a new crop production system. Because of higher yields, stubble mulch tillage was more economical than one-way tillage for wheat production at Bushland, even though fuel use was the same (Allen and Fryrear, 1979). When production expenses are decreased and yields are inTable XV British National Average Expenses for Establishing a Cereal Crop into Stubble Based on the Cost of Owning and Operating New Equipment" Number of operations or expense items Operation or expense Ploughing Disking Herbicide Herbicide application Harrowing Seeding Tine cultivation

Conventional tillage I 2 1 pint 1

1 1

Totals From ICI-Plant

Minimum tillage

1 pint 1 1 1 3

Expenses (British pounddacre) Notillage

2 pints 1 1 1

Conventional tillage 6.75 5.20 1.87 1S O 1.60 3.30

20.22

Protection (1976).

Minimum tillage

NOtillage

-

-

1.87 1.50 1.60 3.30 8.50

3.74 1.50 1.60 3.80

16.77

10.64

51

CONSERVATION TILLAGE SYSTEMS

Table XVI Cost of Tillage and Herbicides for Various Cropping Sequences with Surface Irrigation on the Southern High Plains" Operations and total expenses Cropping sequence Wheat to sorghum, double-cropped Wheat to wheat

Sorghum to sorghum

Wheat-sorghum-fallow

a

Clean tillage

Limited tillage

Disk, disk, bed, apply atrazine ( I .8 kg/ha) $52/ha Disk, disk, bed, cultivate $44/ha Disk, disk, chisel, bed, cultivate $49/ha Disk, disk, disk, bed, cultivate, cultivate $44/ha

Apply atrazine (1.8 kg/ha) $17/ha Disk-bed, cultivate $26/ha Shred, split beds, cultivate $30/ha Apply atrazine (3.4 kg/ha) and 2.4-D ( 1 . 1 kg/ha) $30/ha

From Wiese er a / . (1979).

creased, remain constant, or even slightly decreased, the reduced-tillage systems are more economical than tillage-intensive systems. In irrigated areas, additional benefits from reduced-tillage systems are derived from greater water conservation, which results in higher yields with reduced expenses for energy, equipment, and labor for pumping irrigation water (Allen and Fryrear, 1979; Section IV, A). Based on equipment and herbicide expenses, production expenses for some crops are about twice as high with conventional tillage as with limited- and no-tillage (Tables XV and XVI). In other systems, the expenses may not differ much for different production systems (Unger and Wiese, 1979). However, if yields are increased or if the reduced-tillage system permits better utilization of all farm resources, economic returns to the crop production enterprise are increased by using the reduced-tillage system (Brown and White, 1973; Unger and Wiese, 1979).

XIII. SUMMARY AND CONCLUSIONS

Much has been accomplished since 1961 in regard to the use of crop residue on the surface in conservation tillage systems. The rapid technological advances in the use of herbicides have done much to reduce the need for tillage. The direct-

52

P. W.UNGER AND T. M . McCALLA

drill system with no subsequent cultivation is well established in the United States, Europe, Asia (particularly in Japan), and Africa under certain conditions. Use of direct drilling or limited tillage controls water and wind erosion, conserves from 5 to 15 cm additional water from rain-fed agriculture, allows more timely planting of crops, and improves the quality of surface water. The amount of energy required for tillage is reduced by direct drilling and no-tillage, but additional energy may be needed for fertilizers and herbicides. The net energy balance may still be slightly in favor of direct drilling. In the United States, use of direct drilling or limited tillage will be necessary to reduce soil erosion losses to levels that meet the requirements of Section 208 of Public Law 92-500, as amended by the Congress in 1972. The wide interest in and adaptation of this practice in this country and throughout the world has been an important achievement.

B. NEEDS While great advances have been made in the use of direct drill or limited tillage, there are still considerable gaps in our knowledge of how to use the system most effectively. For optimizing crop yield and quality in conservation tillage systems, an in-depth understanding of the influence of tillage and residue management systems on the physical, chemical, and biological components of the plant-soil environment is essential. However, little effort has been made to obtain this information through an integrated approach. In most past research involving plant residue systems, the main factor determined was crop yield. Only minor emphasis was placed on other biological considerations. We know that tillage alone can have a dramatic effect on the biological equilibria and microbial population in the soil. The reductions in crop yield that have occurred with some conservation tillage systems have shown the need for a complete understanding of the biological changes in the soil environment as a result of the physical changes imposed by the system (Doran and McCalla, 1977). To determine the biological effects of minimum tillage and crop residue management, multidisciplinary research is needed to study the microbiology, biochemistry, chemistry, and physics of the plant-soil environment, as well as the physiology and disease and insect vulnerability of the crop itself. Because the soil environment with residue systems is generally cooler, wetter, and less aerated than with the moldboard plow system, microbial activity is slower and also results in different transformations of nutrients. Not until we more clearly understand the physical, chemical, and biological changes brought about by the use of crop residue on the surface in conservation tillage systems can we more effectively use the additional stored water and conserve the organic matter and nutrients in the soil. This information will help us to apply more wisely fertilizer and

CONSERVATION TILLAGE SYSTEMS

53

other treatments to maximize yield, improve the quality of the food produced, and minimize some present harmful effects on crop yields. It will also enable us to reduce energy input, control water and wind erosion, improve crop quality by managing nutrient composition of the crop, and curtail the physiological disorders that result from insects and diseases. Only through a complete understanding of the physical, chemical, and biological effects of conservation tillage systems on the soil and plant environment can predictions be made of how the information obtained in one area can be applied in other areas. A basic need is the development of a model, with appropriate baseline parameters (lowest category or variable in an equation or model), that is applicable to interregional as well as interdisciplinary research. How is a model developed that will use both physical and biological data collected in different regions? One answer to this question may be in the choice of dependent variables for which data are available in the literature and which are most frequently used in agricultural research endeavors. One such universal “dependent variable” is crop yield. Of course, within research disciplines, equations will be developed that are defined by the parameters important to that particular area of study. However, baseline parameters should be chosen so that data measurements can be applied and used by other disciplines. Examples of a few such parameters would be soil temperature, soil air content, and soil water content. These parameters are affected by the physical properties of the soil and have a direct influence on the chemical and biological components of the soil-plant environment.

REFERENCES Adams, J. E. 1962. Agron. J. 54, 257-261. Adams, J . E. 1965. Agron. J. 57, 471-474. Adams, J . E. 1967. Agron. J. 59, 595-599. Adams, 1. E. 1970. Agron. J . 62, 785-790. Adams. J . E. 1974. Agron. J. 66, 229-304. Allen, R. R., and Fryrear, D . W. 1980. In “Conservation Tillage in Texas” (B. L.Harris and A . E. Coburn, eds.), pp. 31-45. Tex. Agr. Ext. Service B-1290. Allen, R. R., Musick, J . T . , and Wiese, A. F. 1975. Tex. Agr. Exp. Sin. PR-3332 C, pp. 66-78. Allen, R. R., Musick, 1. T . , and Wiese, A. F. 1976. Trans. Am. SOC.Agr. Eng. 19,234-236.241. Allen. R. R., Stewart, B. A., and Unger, P. W. 1977. J. Soil Wafer Conserv. 32, 84-87. Allen, R. R., Musick, J. T.. and Dusek, D . A. 1980. Trans. Am. SOC. Agr. Eng. 23, 346-350. Allmaras, R. R., and Nelson, W. W. 1971. Soil Sci. SOC.Am. Proc. 35, 974-980. Allmaras, R. R., Burrows, W. C., and Larson, W. E. 1964. SoilSci. SOC.Am. Proc. 28,271-275. Allmaras, R. R., Black, A. L., and Rickman. R. W. 1973. I n “Conservation Tillage: The Proceedings of a National Conference,” pp. 62-86. Soil Conserv. SOC. Am., Ankeny, Iowa. Amemiya, M. 1977. J. Soil Wafer Conserv. 32, 29-36. American Society of Agronomy. 1978. Prw. Symp., Crop Residue Management Systems, Houston, Texas. Spec. Pub. No. 31.

54

P. W. UNGER AND T. M. McCALLA

Anderson, C. H. 1976. In “Conservation Tillage,” pp. 33-45. Great Plains Agr. Council Pub. No. 77. Anderson, D. T. 1968. In “Conservation Tillage in the Great Plains,” pp. 83-91. Great Plains Agr. Council Pub. No. 32. Armbrust, D. W., Chepil, W. S., and Siddoway, F. H. 1964. Soil Sci. SOC.Am. Proc. 28,557-560. Army, T. J., Wiese, A. F., and Hanks, R. J. 1961. Soil Sci. SOC.Am. Proc. 25, 410-413. Bagnold, R. A. 1943. “The Physics of Blown Sand and Desert Dunes.” William & Morrow, New York. Barnett, A. P.,Carrecker, J. R., Abruna, F.,Jackson, W. A., Dooley, A. E., and Holladay, J. H. 1972. Agron. J. 64, 391-395. Batchelder, A. R., and Jones, J. N., Jr. 1972. Agron. J . 64, 648-652. Bauer, A.. and Zubriski, J. C. 1978. Soil Sci. SOC. Am. J . 42, 777-781. Bennett, 0. L. 1977. J. Soil Water Consew. 32, 9-12. Bennett, W. H., Pittman, D. W., Tingey, D. C., McAllister, D. R., Peterson, H. B., and Sampson, I. G. 1954. Utah Agr. Exp. Stn. Bull. No. 371, 1-81. Bertrand, A. R. 1966. I n “Plant Environment and Efficient Water Use”(W. H. Pierre, D. Kirkham, J. Pesek, and R. Shaw, eds.), pp. 207-235. Am. SOC.Agron., Madison, Wisconsin. Black, A. L. 1973. Soil Sci. SOC. Am. Proc. 37, 943-946. Blevins, R. L., Thomas, G. W., and Cornelius, P. L. 1977. Agron. J . 69, 383-386. Bond, J. J . , and Willis, W. 0. 1969. Soil Sci. SOC. Am. Proc. 33, 445-448. Bond, J. J . , and Willis, W. 0. 1971. Soil Sci. Soc. Am. Proc. 35, 984-987. Boone, F. R . , Slager, S., Miedema, R., and Eleveld, R. 1976. Neth. J. Agr. Sci. 24, 105-1 19. Boosalis, M. G. 1979. Proc. Seminar, Conservation-Production Systems, Research and Extension Approaches, SEA-AR, USDA, and Univ. of Nebraska, Lincoln. Boosalis, M. G., and Cook, G. E. 1973. In “Conservation Tillage: The Proceedings of a National Conference,” pp. 114-120. Soil Conserv. SOC.Am., Ankeny, Iowa. Bower, C. A., Browning, G. M., and Norton, R. A. 1944. Soil Sci. SOC.A m . Proc. 9, 142-146. Brown, R. E., Jr., and White, T. K. 1973. Indiana Agr. Exp. Srn. Bull. No. SB 19, 1-30. Burrows, W. C., and Larson, W. E. 1962. Agron. J. 54, 19-23. Chepil, W. S . 1959. J . Soil Water Conserv. 14, 214-219. Chepil, W. S . , and Woodruff, N. P. 1957. Am. J . Sci. 255, 206-213. Chepil, W. S., and Woodruff, N. P. 1963. Adv. Agron. 15, 211-302. Chepil, W. S . , Moldenhauer, W. C., Hobbs, J. A , , Nossaman, N. L., andTaylor, H. M. 1962. U.S. Dep. Agr. ARS Prod. Res. Rep. No. 64, 1-14. U.S. Govt. Printing Office, Washington, D.C. Cook, R. J.. Boosalis, M. G., and Doupnik, B. 1978. In “Crop Residue Management Systems” (W. R. Oschwald, ed.), pp. 147-163. Am. SOC.Agron. Spec. Pub. No. 31. Craig, D. G., and Turelle, I . W . 1964. “Guide for Wind Erosion Control on Cropland in the Great Plains States. ” U.S. Dep. Agr.-Soil Conservation Service. Davies, D. B., and Cannell, R. Q. 1975. Outlook Agr. 8, 216-220. Donahue, R. L., Miller, R. W., and Shickluna, J. C. 1977. “Soils-An Introduction to Soils and Plant Growth,’’ 4th ed., pp. 155-171, 302-320, 463-490. Prentice-Hall, New York. Doran, J. W. 1980. Soil Sci. SOC.Am. J . 44, 518-524. Doran, J. W., and McCalla, T. M. 1977. In “Research Progress and Needs, Conservation Tillage,” pp. 1-7. U.S. Dept. Agr., ARS-NC-57. U.S. Govt. Printing Office, Washington, D.C. Eck, H. V.,and Taylor, H. M. 1969. Soil Sci. Soc. Am. Proc. 33, 779-783. Elliott, J. G., Ellis, F. B., and Pollard, F. 1977. J . Agr. Sci. (Cambridge) 89, 621-629. Elliott, L. F., McCalla, T. M., and Waiss. A., Jr. 1978. In “Crop Residue Management Systems” (W. R. Oschwald, ed.), pp. 131-146. Am. SOC.Agron. Spec. Pub. No. 31. Ellis, F. B., Elliott, J. G., Barnes, B. T., and Howse, K. R. 1977. J. Agr. Sci. (Cambridge) 89, 631-642.

CONSERVATION TILLAGE SYSTEMS

55

Estes, G. 0. 1972. Agron. J . 64, 733-735. FAO. 1971. Food and Agr. Org. of the United Nations, Rome. Agr. Dev. Paper No. 92, 1-53. Fenster. C. R. 1960. Soil Sci. SOC.Am. Proc. 24, 518-523. Fenster, C. R. 1977. J . Soil Woter Conserv. 32, 37-42. Fink, R. J., and Wesley, D. 1974. Agron. J . 66, 70-71. Fisher, W. F., and Lane, D. E. 1973. In “Conservation Tillage: The Proceedings of a National Conference,” pp. 187-194. Soil Conserv. SOC.Am., Ankeny, Iowa. Frere, M. H. 1976. In “Control of Water Pollution from Cropland Vol. 11-An Overview” (B. A. Stewart, D. A. Woolhiser, W. H. Wischmeier, J . H. Caro, and M. H. Frere, eds.), pp. 59-90. U.S. Dept. Agr. Rep. No. ARS-H-5-2. U.S. Govt. Printing Office, Washington, D.C. Frere, M. H., Woolhiser. D. A , , Caro, J. H.,Stewart, B. A,, and Wischmeier, W. H. 1977. J. Soil Water Conserv. 32, 260-264. Gardner, H. R., and Gardner, W. R. 1969. Soil Sci. SOC.Am. Proc. 33, 192-196. Gausman, H. W., Gerbermann, A. H., Wiegand, C. L.. L a m e r , R. W., Rodriguez, R. R., and Noriega, J . R. 1975. Soil Sci. Soc. Am. Proc. 39, 752-755. Great Plains Agricultural Council. 1962. Proc. Workshop, Stubble-mulch Farming in the Great Plains, Lincoln, Nebraska. Great Plains Agricultural Council. 1968. Roc. Workshop, Conservation Tillage in the Great Plains, Lincoln, Nebraska, Pub. No. 32. Great Plains Agricultural Council. 1976. Proc. Great Plains Workshop, Conservation Tillage, Fort Collins, Colorado, Pub. No. 77. Greb, B. W., Smika, D. E., and Black, A. L. 1967. Soil Sci. SOC.Am. Proc. 31, 556-559. Gregory, W. W.. and Musick, G. J. 1976. Bull. Entomol. SOC.Am. 22, 302-304. Griffith, D. R., Mannering, J. V., and Moldenhauer, W. C. 1977. J . Soil Water Conserv. 32, 20-28. Hanks, R. J., and Woodruff, N. P. 1958. Soil Sci. Soc. Am. Proc. 86, 160-164. Hanks, R. J., Bowers, S. A,, and Bark, L. D. 1961. Soil Sci. 91, 233-238. Harper, J., and Brensing, 0. H. 1950. Okla. Agr. Exp. S m . Bull. No. B-362, 1-28. Harrold, L. L., and Edwards, W. M. 1972. J . Soil Water Conserv. 27, 30. Hayes, W. A , , and Kimberlin, L. W. 1978. In “Crop Residue Management Systems” (W. R. Oschwald, ed.), pp. 35-48. Am. SOC.Agron. Spec. Pub. No. 31. Hays, 0. E. 1961. J . Soil Warer Conserv. 16, 172-175. Hopp, H., and Slater, C. S. 1961. In “The Challenge of Earthworm Research” (R. Rodale, ed.), pp. 67-83. The Soil and Health Foundation, Emmaus, Pennsylvania. Hough, C. 1979. In “What’s New in Water Conservation,” pp. K 1-3. Crop Prod. and Util. Symp. Proc., Amarillo, Texas. 1CI-Plant Protection. 1976. “Handbook of Direct Drilling,” 1977 Addendum, pp. 24-25. Kynoch Press, Birmingham, England. Jacks, G. V., Brind, W. D., and Smith, R. 1955. Commonw. Bur. Soil Sci. ( G . B . ) Tech. Commun. No. 49, 1-87. Jensen, M. E., and Sletten, W. H. 1965. U S . Dep. Agr.-ARS Conserv. Res. Rep. No. I , 1-25. U.S. Govt. Printing Office, Washington, D.C. Johnson, W. C . 1950. U . S . Dep. Agr. Circ. No. 860, 1-18. Johnson, W. C., and Davis, R. G. 1972. U.S. Dep. Agr. Conserv. Res. Rep. No. 16, 1-32. U.S. Govt. Printing Office, Washington, D.C. Johnston, J . R., Browning, G. M., and Russell, M. B. 1943. Soil Sci. SOC.Am. Proc. 7 , 105-107. Juo, A. S. R., and Lal, R. 1977. Plant Soil 41, 567-584. Kimber, R. W. L. 1966. Ausr. 1. Agr. Res. 18, 361-374. Kimbedin, L. W. 1976. U.S. Dep. Agr.-ARS Spec. Pub. No. 21, 339-346. U.S. Govt. Printing Office, Washington, D.C.

56

P. W. UNGER AND T. M. McCALLA

Koshi, P. T., and Fryrear, D. W. 1973. Soil Sci. Soc. Am. Proc. 37, 758-762. Kramer, L. A,, and Meyer, L. D. 1969. Trans. Am. Soc. Agr. Eng. 12, 638-645. Lal, R. 1976. Soil Sci. Soc. Am. J . 40, 762-768. Lal, R., Maurya, P. R., and Osei-Yeboah, S. 1978. Exp. Agr. ( G . B . ) 14, 113-120. Lattanzi, A. R., Meyer, L. D., and Baumgardner, M. F. 1974. Soil Sci. Soc. Am. Proc. 38, 946-950. Lawton, K., and Browning, G. M. 1948. Soil Sci. Soc. Am. Proc. 13, 31 1-317. Lemon, E. R. 1956. Soil Sci. Soc. Am. Proc. 20, 120-125. Lewis, W. M. 1973. In “Conservation Tillage: The Proceedings of a National Conference,” pp. 182-187. Soil Conserv. SOC.Am., Ankeny, Iowa. Lyles, L., and Woodruff, N. P. 1962. Agr. Eng. 43, 150-153. Lyles, L., Disrud, L. A., and Krauss, R. K. 1971. Trans. Am. Soc. Agr. Eng. 14, 285-289. McCalla. T. M. 1958. J. Soil Water Consent. 13, 255-258. McCalla, T. M. 1959. Nebrk. Agr. Exp. Sfn. Bull. No. 453, 1-31. McCalla, T. M., and Army, T. J. 1961. Adv. Agron. 13, 125-196. McCalla, T . M., and Duley, F. L. 1946. Agron. J. 38, 75-89. McCalla, T. M., and Norstadt, F. A. 1974. Agr. Environ. 1, 153-174. Mannering, J. B., and Meyer, L. D. 1961. Soil Sci. SOC.Am. Proc. 25, 506-510. Mazurak, A. P., and Conard, E. C. 1959. Agron J . 51, 264-267. Mazurak, A. P., and Ramig, R. E. 1962. Soil Sci. 94, 151-157. Mazurak, A. P.. and Ramig, R. E. 1963. Soil Sci. Soc. Am. Proc. 27, 592-595. Mazurak, A. P., Cosper. H. R., and Rhoades, H. F. 1955. Agron. J. 47, 490-493. Mazurak, A . P., Kriz, W., and Ramig, R. E. 1960. Agron. J . 52, 35-37. Meyer, L. D., and Mannering, J. V. 1963. Trans. Am. Soc. Agr. Eng. 6, 322-327. Moschler, W. W . , and Martens, D. C. 1975. Soil Sci. SOC.Am. Proc. 39, 886-891. Moschler, W. W., Shear, G . M., Martens, D. C., Jones, G. D., and Wilmouth, R. R. 1972. Agron. J . 64, 229-231. Musick, G. J., and Beasley, L. E. 1978. In “Crop Residue Management Systems’’ (W. R. Oschwald, ed.), pp. 173-186. Am. SOC.Agron. Spec. Pub. No. 31. Musick, J. T., Allen, R. R., Dusek, D. A., and Wood, F. 0. 1972. Tex. Agr. Exp. Stn. PR-3043, 1-7. Musick, J. T., Wiese, A. F., and Allen, R. R. 1977. Trans. Am. Soc. Agr. Eng. 20, 666-612. Nielsen, K . F. 1974. In “The Plant Root and Its Environment” (E. W. Carson, ed.), pp. 293-333. Univ. of Virginia Press, Charlottesville. Ohio State University. 1972. Roc. No-Tillage Systems Symp., Columbus. Olsen, F. J . , Jones, J. H., and Faix, J. 1. 1978. Ill. Agr. Exp. Stn. DSAC 6, 127-131. Dixon Springs Agr. Center, Simpson, Illinois. Onstad, C. A. 1972. Trans. Am. Soc. Agr. Eng. 15, 287-289. Papendick, R. I . , and Miller, D. E. 1977. J. Soil Water Conserv. 32, 49-56. Phillips, S. H . , and Young, H. M., Jr. 1973. “No-Tillage Farming.” Reiman, Milwaukee, Wisconsin. Phillips, W. M. 1969. Weed Sci. 17, 451-454. Phillips, W. M., Unger, P. W., and Greer, H. A. 1976. In “Conservation Tillage,” pp. 56-63. Great Plains Agr. Council Pub. No. 77. Pitts, N. E. 1978. Soil Conserv. 43(9). 9- 1 1 . Plant Protection Limited. 1973. Minimum Tillage Issue. Ourlook Agr. 7, 142-200. Jealott’s Hill Res. Stn., Bracknell, Berkshire, England. Plant Protection Limited. 1975. Reduced Cultivation and Direct Drilling Issue. Outlook Agr. 8, 21 1-260. Jealott’s Hill Res. Stn., Bracknell, Berkshire, England. Promersberger, W. J., and Pratt, G. L. 1958. N . Dak. Agr. Exp. Stn. Tech. Bull. No. 415, 1-33.

CONSERVATION TILLAGE SYSTEMS

57

Reicosky, D. C., Cassel. D. K.,Blevins, R. L., Gill, W. R., and Naderman, G . C. 1977. J. Soil Water Conserv. 32, 13- 19. Richardson, C. W., Baird, R. W., and Fryrear, D. W. 1969. J. Soil Water Conserv. 24, 60-63. Richey, C. B., Griffith, D. R., and Parsons, S. D. 1977. Adv. Agron. 29, 141-182. Rockwood, W. G., and Lal. R. 1974. Span 17(2), 77-79. Rowell, D. L., Osborne, G. J., Matthews, P. G., Stonebridge, W. C.. and McNeill, A. A. 1977. Aust. J . Exp. Agr. Anim. Husb. 17, 802-811. Skidmore, E. L., and Siddoway, F. H. 1978. I n “Crop Residue Management Systems’’ (W. R. Oschwald, ed.), pp. 17-33. Am. SOC.Agron. Spec. Pub. No. 31. Smika. D. E. 1976a. I n “Conservation Tillage,” pp. 78-92. Great Plains Agr. Council Pub. No. 77. Smika, D. E. 1976b. The 7th Conf. Int. Soil Tillage Res. Org., Sweden. 37, pp. 1-6. Soil Conservation Society of America. 1973. Proc. Natl. Conf., Conservation Tillage, Des Moines, Iowa. Soil Conservation Society of America. 1977. “Conservation Tillage: Problems and Potentials. ” Spec. Pub. No. 20. Soil Conservation Society of America. 1979. “Effect of Tillage and Crop Residue Removal on Erosion, Runoff, and Plant Nutrients.” Spec. Pub. No. 25. Southern Weed Science Society. 1979. “Research Report, 32nd Annual Meeting.” Auburn Univ. Printing Serv., Auburn, Alabama. Stewart, 8 . A., Woolhiser, D. A,, Wischmeier, W. H.,Caro, J. H., and Frere, M. H. 1975. “Control of Water Pollution from Cropland, Vol. L A Manual for Guideline Development. ” U.S. Dep. Agr. Rep No. ARS-H-5-1. U.S. Govt. Printing Office. Washington, D.C. Stonebridge, W. C., and Fletcher. I. C. 1973. Outlook Agr. 7, 155-161. Taylor, R. E., Hays, 0. E., Bay, C. E., and Dixon, R. M. 1964. Soil Sci. Soc. Am. Proc. 28, 123- 125. Unger, P. W. 1969. TEX.Agr. Exp. Sin. MP-933, 1-10. Unger, P. W. 1970. Soil Sci. Soc. Am. Proc. 34, 492-495. Unger, P. W. 1972. Tex. A g r . Exp. Stn. Bull. No. B-1126, 1-20. Unger, P. W. 1975. Soil Sci. Soc. Am. Proc. 39, I 197-1200. Unger, P. W. 1976. Soil Sci. Soc. Am. J . 40, 298-300. Unger, P. W . 1977. Agron. 1. 69, 944-950. Unger, P. W. 1978a. Soil Sci. Soc. Am. J . 42, 486-491. Unger, P. W. 1978b. Agron. J. 70, 858-864. Unger, P. W., and Parker, J. J. 1976. Soil Sci. Soc. A m . J . 40, 938-942. Unger. P. W., and Stewart, B. A. 1976. I n “Multiple Cropping” (M. Stelly, ed.-in-chief), pp. 255-273. Am. SOC.Agron. Spec. Pub. No. 27. Unger, P. W., and Wiese, A. F. 1979. Soil Sci. Soc. Am. J . 43, 582-588. Unger, P. W., Allen, R. R., and Wiese, A. F. 1971. J . Soil Water Conserv. 26, 147-150. Unger, P. W., Wiese, A. F., and Allen, R. R. 1977. J. Soil Water Conserv. 32, 43-48. Unger, P. W., Gerard, C. J., and Wendt, C. W. 1980. I n “Conservation Tillage in Texas” (B. L. Harris and A. E. Coburn, eds.), pp. 18-30. Tex. Agr. Ext. Service B-1290. USDA (United States Department of Agriculture). 1977. “Research Progress and Needs, Conservation Tillage.” ARS-NC-57. U.S. Printing Office, Washington, D.C. Van Bavel, C. H. M., and Schaller, F. W. 1951. Soil Sci. SOC. Am. Proc. 15, 399-404. Van Doren, D. M., Jr., and Allmaras, R. R. 1978. I n “Crop Residue Management Systems” (W. R. Oschwald, ed.), pp. 49-83. Am. SOC.Agron. Spec. Pub. No. 31. Van Doren, D. M., Jr., and Triplett, G. B., Jr. 1969. Ohio Agr. Res. D e v . Cenrer Res. Circ. No. 169, 1-16. Van Wijk, W. R.. Larson, W. E., and Burrows, W. C. 1959. Soil Sci. Am. Proc. 23, 428-434.

58

P. W. UNGER AND T. M. McCALLA

Wicks, G. A,, and Nordquist, P. T. 1976. I n “Conservation Tillage,” pp. 46-55. Great Plains Agr. Council Pub. No. 77. Wicks, G. A., and Smika, D. E. 1973. J. Weed Sci. Soc. Am. 21, 97-102. Wiese, A. F., and Staniforth, D. W. 1973. I n “Conservation Tillage: The Proceedings of a National Conference,” pp. 108-114. Soil Conserv. SOC. Am., Ankeny, Iowa. Wiese, A. F., Unger, P. W., Allen, R. R., and Musick, J. T. 1979. In “What’s New in Water Conservation,” pp. E 1-6. Crop Prod. Util. Symp. Proc., Amarillo, Texas. Williams, J. L., Jr., and Wicks, G. A. 1978. I n “Crop Residue Management Systems’’ (W. R. Oschwald, ed.), pp. 165-172. Am. Soc. Agron. Spec. Pub. No. 31. Willis, W. 0..Larson, W. E., and Kirkham, D. 1957. Agron. J. 49, 323-328. Wilson, H. A., and Browning, G. M. 1946. Soil Sci. Soc. Am. Pror. 10, 51-57. Wischmeier, W. H. 1973. In “Conservation Tillage: The Proceedings of a National Conference,” pp. 133-141. Soil Conserv. SOC.Am., Ankeny, Iowa. Wischmeier, W. H.,and Smith, D. D. 1965. U.S. Dep. Agr. ARS Purdue Agr. Exp. Stn. Agr. Handb. No. 282. U.S. Govt. Printing Office, Washington, D.C. Wischmeier, W. H., and Smith, D. D. 1978. “Predicting Rainfall Erosion Losses.” U.S. Dep. Agr., Agr. Handb. No. 537. U.S. Govt. Printing Office, Washington, D.C. Woodruff, N. P.. and Lyles, L. 1967. I n “Tillage for Greater Crop Production,” pp. 63-67, 70. Am. SOC.Agr. Eng. Pub. No. 168. Woodruff, N. P., and Siddoway, F. H. 1973. I n “Conservation Tillage: The Proceedings of a National Conference,” pp. 156-162. Soil Conserv. SOC.Am., Ankeny, Iowa. Zingg, A. W. 1954. Trans. Am. Geophys. Union 35, 252-258. Zingg, A. W., and Whitfield, C. J . 1957. U.S. Dep. Agr. Tech. Bull. No. 1166, 1-56. Zingg, A. W., Chepil, W. S . , and Woodruff, N. P. 1965. J. Am. SOC. Civil Eng. 91 (Hy 2). 267-287.

ADVANCES IN AGRONOMY. VOL. 33

POTASSIUM IN CROP PRODUCTION Konrad Mengel and Ernest A. Kirkby Institute of Plant Nutrition, Justus Liebig University, Giessen, Federal Republic of Germany and Department of Plant Sciences, The University, Leeds, England

59

. . . . . . . . . . . . . . . . . . . . . . . . . .. . . , . . .. . . . . . . . . A. Soil Potassium Fractions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . Factors and Processes of Potassium Availability . . . . . . . . . . . . . . . . . . . . . . . , . . , . C. Assessment of K Availability in Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

60 60

D. Plant Root Soil Interactions

71 74 74 81 83 85

11. Potassium Availability in the Soil

111. Potassium in Physiology

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

64 70

A. Potassium Transport across Biological Membranes and Cation Competition B. Cell Turgor and Water Economy of Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Energy Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Long-Distance Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Enzyme Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 IV. Potassium Application and Crop Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 A. Crop Response and Potassium Application . . . . . . . . . . . . . . . . . . . . , . . , . . . . , . . . 91 B. Effect of Potassium on Yield Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 C. Secondary Effects of Potassium on Crop Yield.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 103 V. Conclusions 103 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I. INTRODUCTION Potassium was first recognized as an essential element for plant growth following the work of the Englishman Home in 1762 from experiments in which he grew barley in pots of soil and used plant analysis as a means of investigating uptake. Later researchers such as Th. de Saussure and Carl Sprengel recognized that potash was present in plant ash obtained from a large number of different plant species. In reviewing the analytical data of the period, Liebig (1 841) proposed that K was in some way involved in plant metabolism. The experience of farmers around Giessen, the German university town in which Liebig worked, had indicated the beneficial influence of manuring crops with plant ash. Liebig recognized that potash was the essential growth factor in the ash. Furthermore, Liebig was aware +

59 Copyright @ 1980 by Academic Rcss. Inc. All rights of Rpaoducrion in any form reserved. ISBN 0-12600733-9

60

KONRAD MENGEL AND ERNEST A. KIRKBY

that the clay fraction of the soil provided a source of K+ for plant growth. In his book, “Die organische Chemie in ihrer Anwendung auf Agrikultur und Physiologie” he wrote, “There must be a component in clay which has an influence on plant life and which directly participates in plant development. This component is the ever-present potash or sodium. The paramount importance of clay minerals in binding or releasing K+ recognized by Liebig has been confirmed by much subsequent research work. The same is also true of Liebig’s suggestion that K+ was involved in plant metabolism. Potassium is now known to be required by plants in large quantities, and potassium fertilizer application has had a considerable impact on crop production, particularly under conditions where there has been a shift from extensive to intensive agricultural practice (Amon, 1969). In this article, three main aspects of K+ in crop production are reviewed, namely, K availability in the soil, the function of K+ in the plant, and potash fertilizer application. The soil is considered as a source of K+ to plant roots. Pedological and mineralogical problems relating to soil K+ have been reviewed elsewhere by Rich (1968, 1972) and by Schroeder (1976). These aspects are considered here only insofar as they are of direct importance to the availability of K+ in the soil medium and hence to crop growth. The use of K+ in practical crop production is also emphasized in the discussions on the physiological role of K+ in the plant and in fertilizer application. ”

II. POTASSIUM AVAILABILITY IN THE SOIL A. SOILPOTASSIUM FRACTIONS

The potassium status of a soil may be assessed on its content of K+-bearing minerals, since the amount of these minerals present in a soil gives some indication of the potential source of K+ to plants. However, in terms of the ability of the soil to supply K+ to plant roots, the quantity of K+-bearing minerals plays only an indirect role. More important in determining the K+ supply to plants are the soil K+ fractions. These fractions, which have been established experimentally using different extraction techniques, are soil solution K+, K+ adsorbed to clay minerals or humus, and K+ present in minerals. The total quantities of K+ in these three fractions differ considerably between soils. However, in mineral soils in which K+ is present in average amounts, the soil solution K+ makes up about 1 to 3% of the exchangeable K+, which in turn represents only a small fraction-at most a few percent-of the total K+ (Scheffer et al., 1960). Potassium in soil solution tends to equilibrate with K+ in the adsorbed fraction so that these two soil K+ fractions are closely interdependent. The equilibrium between solution and adsorbed K+ is controlled to a large

POTASSIUM IN CROP PRODUCTION

61

extent by the degree of K+ selectivity of the adsorption sites in the exchangeable fraction. Adsorption sites of organic matter and of kaolinitic clay minerals are low in K+ selectivity. Potassium adsorbed on these sites is thus in equilibrium with a relatively high concentration of solution K+ (Ehlers et al., 1968). On the other hand, the 2: 1 clay minerals possess adsorption sites that are much higher in K+ selectivity and that bind K+ very strongly. This is especially true for illitic clay minerals (Ehlers et al., 1967). As shown in Fig. 1 , three types of adsorption sites may be distinguished: planar sites (p-position) with a low K+ selectivity, edge sites (e-position) with a medium K+ selectivity, and inner sites (i-position) with a high K+ selectivity. These highly selective inner sites are of particular interest, since the adsorbed K+ can be considered an integral part of the clay mineral. This is true of the micas, in which the “interlayer K+ ” bridges the adjacent layers by electrostatic bonds. Interlayer K+ is not easily displaced by other cation species and particularly not so by larger cation species such as Ca2+ and Mg2+. For this reason, this particular K+ fraction is termed “nonexchangeable K+.” All three K+ fractions, solution K+, exchangeable K+, and nonexchangeable K+, are interrelated and all play a part in supplying plants with K+. The interrelationshipsbetween the K fractions may be illustrated by considering what happens when a K+ salt such as KCl is added to the soil. At first the salt dissolves and the K+ concentration of the soil solution increases rapidly. Potassium is then removed from the solution by the adsorption sites, the rate at which this occurs depending on the particular equilibrium conditions of the system. This removal of K+ from soil solution is accompanied by an equivalent increase in the soil solution concentration of other cations. The application of K+ to a soil may saturate all three fractions with K+. However, the time required for K+ equilibrium to be reached under field conditions may be as much as several weeks. Saturation of i-positions by K+ is a particularly long-term process since the diffusion of K+ in the interlayer zone is p - position i position

I

I

-

K not exchangeable to IargeCations

*+

I +

‘e-

position

Hydroxy-Al(or Fe) Islands

FIG. 1. Model of an expandable 2: 1 clay mineral with interlayer K+,wedge zones, and p-. e-, and i-positions.

62

KONRAD MENGEL AND ERNEST A. KIRKBY

slow (diffusion coefficient lo-’’ to cm-2 sec-l). In examining this fixation process on a K+-fixing soil (44% clay), Karbachsch (1978) found that even after a very high fertilizer application rate of 1000 kg Wha, the K+ concentration of the soil solution was only on the order of 1 mM K+. As the experiment progressed, this concentration steadily declined until after a period of 90 days it fell to a value of 0.3 mM, which was only about 50% higher than the K concentration of the soil solution prior to K+ fertilizer application. Similar results have been reported by Amberger et al. (1974) who found that the process of K+ adsorption to the i-position of clay minerals extended over a period of about 6 months. From these results it is clear that soils in which the K fixation capacity is high may more or less immobilize fertilizer K+. The agricultural implications of this are considered in the final section of this article. All i-positions, even those occupied by cations other than K+, are involved in the K fixation process. Potassium fixation takes place by means of K+ adsorption to these K+-specific binding sites of the interlayer zone, and in this process cation exchange occurs. As shown in Fig. 1, the replacement of interlayer K+ by larger cation species (Ca2+,MgZ+)expands the lattice and wedge zones are formed. The reverse process occurs when these larger cation species are replaced from interlayer sites by K+ or NH:. The contraction in the mineral is accompanied by a decrease in cation exchange capacity. This is the process by which K+-depleted 2:l clay minerals fix K+. The K+ fixation capacity of soils differs widely and depends much on the type of soil clay minerals present in the soil and on their degree of K depletion. According to investigations of Arifin and Tan (1973) the proportion of wedge zones decreased for the various minerals in the following sequence: micas > illite > vermiculite > montmorillonite. These authors found K+ fixation capacities ranging from 0.3 to 0.6 me K/g clay. The data refer to K+ fixation under dry conditions. The so-called “wet K+ fixation capacity” is a lower value, due to the fact that under wet conditions only micas, illites, and vermiculites fix K+, whereas under dry conditions smectites are also able to fix K+, because of a shrinkage of the mineral. The behavior of clay minerals in relation to K fixation or release on drying, however, is by no means clear-cut (Ahmad and Davies, 1971). Potassium depletion of K+-fixing soil minerals may be of anthropogenic origin or may have occurred during soil development. Potassium fixation is often found on alluvial soils associated with high amounts of organic matter, as is the case for numerous K+-fixing sites in Bavaria in the Federal Republic of Germany. The role of organic matter in K fixation in these soils is not clear, but it is believed that the mineral constituents may have lost substantial amounts of K+ during the period of soil material transportation by water (Niederbudde, 1967), thereby making the soils prone to K+ fixation. Continuous cropping without K+ application may also induce the buildup of a

POTASSIUM IN CROP PRODUCTION

63

K+ fixation potential in the soil. An example of this has been reported by Nielsen (1970). In a long-term field experiment he observed the highest K fixation capacity on a soil which had received no K+ fertilizer for 70 years. Potassium fixation is also dependent on soil acidity, being generally low or absent on more acid soils. Under such conditions more soluble forms of A1 such as AI(OH)?j, Al(OH)2+are available which compete with K+ and are selectively bound to the i-position of the clay minerals (Nemeth and Grimme, 1972). In the long term, low pH conditions also favor the formation of Al-chlorites, a type of mineral which does not fix K+ (Laves, 1978). Under certain circumstances interlayer K+ may contribute to a considerable extent in supplying K+ for plant uptake (Niederbudde et al., 1969; Tabatabai and Hanway, 1969). Mengel and Wiechens (1979) found that under optimum conditions the nonexchangeable K+ fraction of a soil rich in K-bearing minerals (illite, vermiculite) could completely meet the K+ demands of ryegrass. This pot experiment also showed that the proportion of K+ absorbed from the nonexchangeable K+ fraction increased the more the exchangeable K+ fraction was depleted. Below a level of 300 ppm exchangeable K+, most of the K+ absorbed by the ryegrass originated from the nonexchangeable K+ source. Potassium quantities present in this nonexchangeable fraction may be considerable. The availability of this K+, however, decreases as K+ is released. Drews (1978) found that only a very small fraction-at most a few percent-of the interlayer K+ was available to Lolium perenne. In this permanent cropping experiment the rate of K + released from the nonexchangeable soil K + fraction finally became so small that the plants suffered severely from K+ deficiency. These observations of Drews (1978) as well as the experiments of Wiechens ( 1 975) and pot and field experiments of v. Boguslawski and Lach ( 1 97 1) clearly demonstrate that the huge pool of interlayer K can be tapped by plants to only a very limited extent. One may suppose that only interlayer K+ located in the marginal zones of clay minerals is available to plant roots in adequate amounts. The above view is supported by the work of Newman (1969). In studies on the release of K+ by micas he has reported that the K+ concentration in equilibrium with the interlayer K+ decreased as the interlayer K+ is depleted. The process of interlayer K+ release is not yet completely understood. According to v. Reichenbach (1972) it is an exchange process associated with diffusion in which K+ adsorbed to i-positions of the interlayer zone is replaced by other cation species. If the replacing species is a large one (Na+, Mg2+, Ca2+), then K+ exchange results in an expansion of the clay mineral and the formation of wedge zones (see Fig. 1). The resulting widening gap between the two layers of the mineral favors the diffusion of the replaced K+ out of the mineral. A demonstration of this type of behavior has been reported recently by Jackson and During (1979) for New Zealand topsoils of widely different clay mineralogy. Pretreatment of the soils with Ca2+ (as acetate) resulted in an expansion of the +

64

KONRAD MENGEL AND ERNEST A. KIRKBY

clay mineral and an increase in K+ desorption. Potassium desorption is also often associated with the oxidation of F2+ to Fe3+ (Farmer and Wilson, 1970; v. Reichenbach, 1972). Low pH and high moisture conditions are also beneficial to the release of interlayer K'. Net K+ release occurs if the rate of K+ release is higher than the rate of K+ fixation. Since the fixation rate depends directly on the K+ concentration in solution in contact with the clay mineral, a high net rate of release is only likely to occur if the K+ concentration in the solution is extremely low. Diluting the soil solution with distilled water should thus promote the release of interlayer K+. This has been demonstrated in a comparative leaching experiment by Drews (1978). The continuous leaching by water of a soil column containing 300 g soil over a period of 180 days resulted in a loss of 9.8 mg K/100 g soil. However, in the comparative parallel treatment, in which the water was replaced by 15-60 x M KCl no loss of K+ from the nonexchangeable fraction was observed. This experiment shows that net K+ release from clay minerals occurs only if the K+ concentration in the soil solution is extremely low. The question of whether plant roots have a direct influence on the release of interlayer K+ is discussed in the next section. B. FACTORS A N D PROCESSES OF POTASSIUM AVAILABILITY

The contact exchange process as postulated by Jenny and Overstreet (1938) was long held to be the most important means by which plant roots obtain and mobilize K+ adsorbed to clay minerals. This concept was criticized by Lagerwerff (1961), who concluded from his experimental data that the bulk of cations absorbed is taken from the solution and that the exchangeable cation fraction is only indirectly available by means of exchange with cations in solution. Further evidence against the predominant role of contact exchange in K+ uptake came from the findings of Barber et al. (1963). These workers evaluated the amount of K+ accessible to plant roots by interception, or in other words, the K+ in direct contact with the roots as they push their way through the soil. It was concluded that the amount of intercepted K+ was far too small to satisfy the needs of the plant. This conclusion is also consistent with the calculations of Mengel and Kirkby (1978) which show that even for a soil high in exchangeable K+ the amount of K+ in direct contact with plant roots can only satisfy a small fraction of the plant's K requirements. Experiments of Drew and Nye (1969) with Lolium perenne also revealed that only 6% of the total K+ demand was supplied by the soil volume of the root hair cylinder. Ninety-four percent of the K+ taken up therefore originated from beyond the limit of the root hair cylinder. It can thus be concluded that the bulk of K+ required by plants must be transported toward the roots.

POTASSIUM IN CROP PRODUCTION

65

The transport of K+ in the soil medium toward plant roots may take place by mass flow or diffusion. Differentiation between both processes is difficult, and only a rough calculation can be made. According to Barber et al. (1963) only about 10% of the total K+ requirement of crops is transported by mass flow, although the contribution can be somewhat greater when the amount of water transpired by the crop is increased. Generally, however, it is accepted that diffusion is the main process by which K+ is transported to plant roots. Both processes, K+ diffusion and K+ mass flow, have been incorporated into an equation that describes the K+ flux toward plant roots (Barber, 1962).

+ C ~ V+ u

J = Dl(d~l/dr) + Dp(d~2/dr)

(1)

where flux rate (total quantity of ions reaching the root per unit time per unit area of root surface) c, = K+ concentration in the soil solution cp = K+ concentration moving at the soil surfaces (adsorbed K+) c3 = K+ concentration in mass flow water v = velocity of water flowing through the soil toward the roots a = replenishment factor D, = diffusion coefficient of the ions in the soil solution D, = diffusion coefficient for the movement of ions at soil surfaces (exchange diffusion) J

=

This equation shows that quite a number of factors have an influence on the K+ flux rate in the soil medium. If one assumes that the mass flow component ( c 3 v ) is of minor importance and one also neglects the replenishment factor, then the flux rate can be seen to be controlled mainly by factors influencing diffusion. Two kinds of diffusion may be considered, diffusion in the soil solution and diffusion in the zone of adsorbed cations (exchange diffusion). Data for exchange diffusion are very rare in the literature. However, De Lopez-Gonzales and Jenny (1959) reported an exchange diffusion coefficient for Sr2+ of 1.5 x lo-* cm-2 sec-’. This is lower than the diffusion coefficient of S r p + in solution by a factor of lo3. It would therefore seem reasonable to assume that the “exchange diffusion” coefficient for K+ is also by some orders of magnitude lower than the K diffusion coefficient in solution, and that one may neglect the exchange diffusion component of the equation (Nye and Tinker, 1977). If the term for the exchange diffusion in Eq. (1) is dropped, then the equation for the K+ flux may be reduced as follows:

This is Fick’s diffusion law. Nye and Tinker (1977) argue that despite the heterogeneous nature of soil, it is legitimate to treat the soil as a quasi-

66

KONRAD MENGEL AND ERNEST A. KIRKBY

homogeneous body to which this law may be applied and that the diffusivity of such a system is described by the diffusion coefficient. They hold the view that this is valid so long as a representative sample of gas- and liquid-filled pores and adjacent adsorbed phases are included. In a more recent paper Nye (1979) has substituted the diffusion coefficient by a dispersion coefficient (see below). This alteration, however, does not affect the principle of the following deductions. According to Nye and Tinker (1977) the diffusivity may be described by the equation! D = D18fldC1ldC

(2)

where D1 = diffusion coefficient of K+ in free solution 8 = the fraction of the soil volume occupied by solution f , = impedance factor C, = concentration of K+ in soil solution C = concentration of K+ in the whole soil system From this equation it follows that the diffusivity of K+ in the soil media increases with 8, which in turn is closely related to soil moisture. The impedance factor also increases as soil moisture increases. This term represents the tortuous pathway along which K+ has to pass on its way through the soil medium to plant roots. It can readily be visualized that the tortuosity increases as the soil becomes drier. C1 is the K+ concentration in the soil solution, and C is the total K+ directly or indirectly involved in K+ transport. Generally the exchangeable K+ is used when measuring the term C. The ratio dC,ldC is of particular importance since it is the reciprocal of the buffer capacity.

vl)

b = dCIdC, = buffer capacity Substituting b for dCldC, the following equation is obtained: D

= D19fl/b

(3)

From this equation it is clear that the diffusivity of K+ decreases as the K buffer capacity increases. This close relationship between the K+ buffer capacity and the K+ diffusion coefficient has been shown experimentally by Vaidyanathan et al. (1968). The importance of soil moisture for the diffusion of K+ or related cation species has been demonstrated by several authors. Graham-Bryce (1963) found a diffusion coefficient for Rb+ of 1 x lo-' cm+ sec-I at a soil water content of 23%. The coefficient decreased to 5 x lo-* ern+ sec-' when the soil moisture was reduced to 10%.Patrick and Reddy ( 1977) measured a diffusion coefficient of 2.5 x 10-6cm-2sec-l for NH: in paddy soils. In comparison with these values,

POTASSIUM IN CROP PRODUCTION

67

the diffusion coefficient for K+ in pure water is about 1.5 X cm-' sec-'. As it seems likely that the diffusion coefficients for NH:, Rb+ , and K+ should not differ greatly, the coefficients cited above for NH: and Rb+ should also be more or less in the same order of magnitude as those of K+ . It can thus be seen that soil moisture is of crucial importance in K+ availability. From the equations cited above the K+ flux in the soil medium (J) can be described by the following equation: J =

-(*)

(2)

(4)

In this equation exchange diffusion and mass flow are neglected. The rate of K+ diffusion is controlled by the K concentration gradient ( d C , / d r )as well as by diffusivity . If the assumption is made that the rate of K+ absorption by the root is the same as that diffusing to the root surface, then the K+ concentration at the root surface remains constant. This, however, is an exceptional case, and generally the rate of K+ absorption by roots is higher than the rate of K+ transported toward the root surface. For this reason K+ depletion zones develop around the root, and the K+ concentration at the root surface thus declines. The degree of K+ depletion at the root surface can be expressed by the C,/Ci ratio, where Ci represents the initial K+ concentration before uptake begins and C, the K+ concentration at the root surface. Experimental evidence of Rb+ depletion around plant roots has been reported by Barber (1962) and Farr et al. (1969) using autographs. The K+ concentration at the root surface is of crucial importance in relation to K+ uptake, according to the following equation (Drew et al., 1969):

F = 21raaC,

(5)

where F = Flux rate across the root surface (mole cmP sec-') a = radius of the root a = root absorbing power C, = concentration of K+ in solution at the root surface

In this case a single root or single root segment is considered as a cylinder, so that the term 27ra represents the root surface of I-cm root length. The term a is the root absorbing power. A high power means that a high proportion of K+ impinging on the root surface is absorbed and vice versa. The root absorbing power depends much on root metabolism and is thus not a constant term but changes depending on metabolic conditions and plant species involved. From Eq. ( 5 ) it can be derived that the K+ flux across the root surface is related to the K+ concentration at the root surface (C,) in a linear way. This is not

68

KONRAD MENGEL AND ERNEST A . KIRKBY

completely correct, as the K+ uptake rate in relation to C , is rather described by a Michaelis-Menten type of curve, as shown by Barber (1979). However, in cases in which diffusion is the limiting process in transporting K+ to the root surface, the K+ concentration at the root surface is low ( d o p M ) , and in this low concentration range the relationship between K+ uptake and K+ concentration is more or less linear. If the time factor is integrated, K+ uptake can be described by the following equation:

M, = 21ratuC,t

(6)

c,

The term represents the average K+ concentration at the root surface during the uptake period t . The decrease of C , during this time is related to the replenishment of solution K+. The higher the K+ replenishment, the higher is the mean K+ concentration at the root surface. This K+ replenishment is controlled by the K+ buffer capacity (b) of the soil, which can be expressed as the ratio of K+ quantity (exchangeable K+) over K+ intensity (K+ concentration of the soil solution) (Mengel, 1974). Figure 2 from the data of Grimme et al. (197 1) shows the K+ buffer curve of a sandy soil and a loamy soil. From these curves it can be derived that if the same amount of K+ is absorbed by plants (quantity factor) from both soils, the K+ concentration in the soil solution of the sandy soil is depressed to a much higher extent than the K+ in the soil solution of the loamy soil. The steepness of the curve is a direct measure of the K+ buffer capacity. This K+ buffer capacity provides also some information about the K+-supplying power of a soil according

(c,)

me K+ in solution ( intensity )

FIG.2. K+ quantityhntensity relationship of a sandy and a clay soil. The steepness of the curves represents the K+ buffer capacity (after data of Grimme er al., 1971).

69

POTASSIUM IN CROP PRODUCTION

to the following equation (Nye, 1979):

U, = bdC,/dt

(6a)

where U, = K uptake r a t e h i t soil volume. It is evident from Eq. (6a) that the K + concentration at the root surface and the K+ buffer capacity are most important factors controlling K+ uptake of plants. The K+ concentration at the root surface is difficult to measure. It is related to the average K+ soil solution concentration according to an equation established by Baldwin et al. (1973): +

a a av X c, = C,/(l + InD*Bfl 1.65a -

)

(7)

where v x

D*

= flux of water through the root surface = radius of the depletion zone =

dispersion coefficient of the solute in the soil solution (cm' sec-')

This equation takes into account the mass flow (v).The term D* stands for the diffusion coefficient. According to Nye (1979) the dispersion coefficient is more appropriate than the molecular diffusion coefficient. Under normal plant water consumption D* is unlikely to exceed the molecular diffusion coefficient by more than a factor of 2. Major factors controlling C , are the K+ concentration in the bulk soil solution ( C , )and the root absorbing power (a).the former increasing and the latter decreasing the K+ concentration at the root surface. Also the extension of the depletion zone has an influence on C,; the larger the extension, the higher is the K+ concentration at the root surface. The theory of K+ flux toward plant roots as outlined above has been tested with young maize plants by Claassen and Barber (1976), using the following equation for the calculation of plant K+ uptake:

where I, I,,, K, C, E

= = = =

=

K+ uptake rate (influx) influx rate at infinite K+ concentration Michaelis-Menten constant

K+ concentration at the root surface K+ efflux

C , was computed from an equation established by Nye and Maniott (1969). From the plot CJC, versus the radial distance from the root, the K+ concentration

70

KONRAD MENGEL AND ERNEST A. KIRKBY

at the root surface (C,) was derived. Thus it was possible to calculate the K+ uptake according to soil and root parameters and to test this calculation by actual K+ plant uptake. Although four different soils with varying K+ levels were included in this experiment a fairly good correlation (Rz= 0.87) was obtained between the predicted and experimental K+ uptake. Calculated K+ uptake was overestimated by about 50% possibly because competition occurred between roots for soil K+. The good agreement between the predicted Kf uptake and the actual K+ uptake of plants proves that the theory of K+ flux and diffusion in soils is based on sound assumptions. The mathematical model used by Claassen and Barber (1976) takes into account the following factors: effective average diffusion coefficient, initial K+ concentration in the soil solution, and the buffer capacity. Hence these factors are the most important parameters controlling K+ availability in soils. C. ASSESSMENT OF K AVAILABILITY I N SOIL

Potassium concentration in soil solution, K+ buffer capacity, and the soil diffusivity should be considered in estimating K+ availability for practical purposes. Under practical farming conditions the soil diffusivity is difficult to assess in advance since it depends much on soil moisture. Potassium concentration in the soil solution and K+ buffer capacity until now have rarely been used in estimating soil K+ availability. In most cases exchangeable K+ is still regarded as a satisfactory measure of the K+ availability status of soils. This fraction, however; comprises both solution K+ and Kf adsorbed by varying strengths to adsorption sites (p-, e-, and even i-positions). Soils with the same values for exchangeable K+ may thus differ considerably in K+ concentrations in soil solution (Nemeth et al., 1970), because more selectively bound K+ is equilibrated with a relatively low K+ concentration and vice versa. If this specifically bound K+ is taken into account, exchangeable K+ may also be a good indicator of the K+ availability status. Rezk and Amer (1969) thus found a significant correlation between K+ uptake by plants and the “corrected” exchangeable K+ of the soil. This “correction” was obtained by dividing the exchangeable K+ through the Gapon coefficient. By this procedure numerical values are obtained that are closely related to the K+ concentration of the soil solution. Poor correlations between plant response and exchangeable K+ have been obtained especially in investigations where soils of different clay contents and degrees of K+ saturation have been used. Jankovic and Nemeth (1974) even found a negative correlation between the exchangeable soil K+ and sunflower seed yields harvested from five different sites. The same yields, however, were positively correlated with the K+ concentration of the soil solution. A close relationship between K+ concentration of the soil solution and the grain yield of

POTASSIUM IN CROP PRODUCTION

71

wheat, grown under field conditions, has also been reported by Nemeth and Harrach (1974). Similarly in studying the K+ availability of 21 different soils in a pot experiment with oats, v. Braunschweig and Mengel (1971) found a highly significant correlation between the K+ concentration of the soil solution and grain yield. More recently During and Duganzich (1979) have also reported that K+ uptake by white clover was best reflected by the K+ concentration of the soil solution. Exchangeable K+ alone correlated very poorly with uptake except in soils of very low K status. Recent experiments of Wanasuria et al. (1980) have shown that the K+ of paddy soils extracted by electroultrafiltration (EUF) was positively correlated with the grain yield, whereas no significant correlation with the exchangeable K+ was obtained. EUF-extractable K+ does reflect the K+ concentration of the soil solution (Nemeth, 1979). Although the K+ buffer capacity is of paramount importance for K+ availability, little quantitative data are available concerning its influence on K+ supply to crops. Nemeth (1975) in investigating three soils with different K+ buffer capacities found a close negative relationship between the K+ buffer capacity and the decrease in grass yield of four consecutive cuts harvested during the experimental period. Barrow (1966) reported that the correlation between the K+ uptake of clover and the content of exchangeable K+ was improved if in addition to the exchangeable K+, the K+ buffer capacity was also taken into account. Recent experimental results of Busch (1980) obtained in pot experiments with a number of soils differing widely in texture have shown that 50-80% of the variability in K+ uptake could be explained by the K+ buffer capacity and the K+ concentration of the soil solution. Only under extreme K+-deficiency conditions, was K+ uptake much controlled by other, still unknown factors. D. PLANTROOTSOILINTERACTIONS

The quantity of K+ absorbed by crops is also related to root growth, extension, and metabolism. Although root interception contributes only to a minor extent to the total K+ requirement of a crop, root extension and root density in the soil are of importance for the quantity of K+ accessible to plant roots. The extension of the K+ exploitation zone around a plant root represents the soil volume that can be "mined" for K+. Since K+ is mainly transported by diffusion toward plant roots, the bulk of K+ absorbed by plants originates from these zones around roots. It is easy to deduce from this relationship that a dense rooting system can exploit a larger soil volume for K+ than can a poor one. The rooting density ( L , ) can be defined as total root length per unit soil volume. The rooting density has an impact on the extension of the K+ exploitation zone around the root (Nye, 1979) according to the following relationship: x = 1 l(7rL.")~

72

KONRAD MENGEL AND ERNEST A. KIRKBY

where x = radius of exploitation. Thus in dense root systems, the K+ depletion zones around roots are less extended and often an overlapping of the exploitation (depletion) zone occurs. A further factor of importance is the soil volume accessible to plant roots. Pot experiments of Newman and Andrews (1973) have shown, for example, that if only a small soil volume is available, the amount of K+ absorbed by plants is also reduced. When the soil volume was restricted, dense rooting systems were observed and K+ uptake was inadequate. It seems likely that this resulted from root competition for K+ between overlapping depletion zones. Root extension and root mass are both of particular importance if available K+ in the soil is low (Chloupek, 1972). On the other hand, as has been shown by Maertens (1971) using young maize plants, only a small portion of a root system may suffice to ensure ample K+ uptake, provided the K+ availability is high. Thus, in general, the K+ absorption potential of a root segment by far exceeds the rate at which K+ is actually absorbed. In order to assess K+ uptake rates, K+ uptake should be calculated per unit root segment (e.g., cm root length). Such experiments and calculations have been canied out by Mengel and Barber (1974). These workers observed K+ uptake rates per unit length of plant root in the early weeks of plant development. Similar results have also been reported by Adepetu and Akapa (1977), who studied nutrient uptake of five cowpea cultivars. The K+ uptake rate per m root length of 5-day-old plants was four times higher than the K+ uptake rate of 30-day-old plants. These results clearly indicate that especially at an early growth stage, high K + availability is required and that the susceptibility to K deficiency is particularly marked during this period. Barber (1979) reported that maximum K+ uptake rates may differ considerably. Thus for young maize roots maximum K+ uptake rates of 2 pmole and for young soybean roots of 0.4 pmole K + cm-I sec were found (cm refers to root length). This author stresses the fact that the K+ content of the tops rather than the K+ content of the roots has a decisive influence on the maximum K+ uptake rate. Thus in 17-day-old maize plants maximum K+ uptake rates ranged between 1.3 and 4.0 pmole cm-' sec - I according to the K + content of the tops; rates being low in tops with high K+ contents and vice versa. The K+ absorption rate of the root is highly dependent on the root metabolism and particularly on respiration and thus on the carbohydrate content of the root (Mengel, 1967). Generally, younger plants have higher root carbohydrate contents than older plants, and even young root tips of older plants are less capable of absorbing K+ than root tips of younger plants (Vincent et al., 1979). Plant species and even cultivars of the same species may differ in their capability of exploiting soil K+. These differences can be explained in terms of rooting pattern and root metabolism, although the whole complex of K+ uptake by field crops growing in soil is still only poorly understood. Halevy (1977) compared +

POTASSIUM IN CROP PRODUCTION

73

two cotton cultivars differing in their capability of exploiting soil K+. The cultivar with the higher uptake potential for soil K+ was found to maintain vigorous root growth to a later growth stage than the cultivar with the poor K+ exploitation capability. This result suggests that root growth at a later stage in plant development in the higher K+ uptake potential cultivar was the cause of the difference in K+ uptake. A spectacular difference in the K+-exploiting capability exists between grasses and legumes. The legumes are inferior to grasses, and when grown together grasses successfully compete with legumes for soil K+. If abundant K+ is not available, the legumes suffer from K+ deficiency, whereas the grasses still grow vigorously and compete strongly for K+ (Blaser and Brady, 1950). This effect may be explained in part by differences in the rooting patterns of the two plant groups; although this explanation is not completely satisfactory. In this context an experiment of Malquori et al. (1975) is of particular interest, in which wheat and lucerne were grown in nutrient solutions. In one treatment where the K+ source of the nutrient solution was biotite, wheat was able to “extract” K+, whereas lucerne was not. This shows that of the two plant species, interlayer K+ was only accessible to wheat. Similar results have been obtained by Steffens and Mengel (1979), who grew ryegrass and red clover on a soil low in exchangeable K+. It was found that ryegrass was more capable of feeding from the nonexchangeable soil K+ than was clover. One may speculate as to the mechanism by which grasses are more able to utilize this interlayer K+ . In this context the results of Baligar and Barber ( 1978a) are of particular interest. These workers observed that after the addition of Rb+ to a number of different soils the K/Rb ratio of corn plants grown on the soils was more similar to the K/Rb ratio of the exchangeable soil fraction than to the K/Rb ratio of the soil solution. In an analogous experiment with onions the K/Rb ratio in the plants was found to be between the K/Rb ratio of the exchangeable fraction and the K/Rb ratio of the soil solution. Baligar and Barber (1978b) discuss their results in terms of exchange diffusion as proposed by Jenny (1966). Tinker (1978) in commenting on Baligar and Barber’s results suggests that the effect might be related to H+ excretion by roots displacing K+ and Rb+ from adsorbing sites around the root. This question needs to be investigated further. Exchange diffusion from the interlayer K+ of clay minerals to the surface of plant roots does not seem a likely mechanism for K+ release. In neither the experiment of Malquori et al. (1975) mentioned earlier nor the work of Ristori (1973, in which clay mineral was added to a nutrient solution, was evidence provided that close contact between clay mineral and root is essential for exploiting interlayer K+ . It is possible that an extremely low K+ concentration in the soil solution achieved by a high rate of K+ uptake is associated with a net release of interlayer K+ . Research in this direction merits further attention. The results of Drews (1978) are encouraging in this line of approach for he

74

KONRAD MENGEL AND ERNEST A . KIRKBY

found that plants grown under energy stress were less capable of exploiting interlayer K+ than control plants. The control plants also depressed the K+ concentration of the soil solution to a greater extent, and this may have resulted in a higher K+ release by clay minerals. According to Barber (1979) plant roots can deplete the K+ concentration of the nutrient solution to as low a level as 2 pM K+. Clay minerals are also capable of absorbing K+ from plants. In the experiment already mentioned by Malquori et al. (1975) in which biotite was added to a nutrient solution supplying wheat, these authors observed an expansion of biotite to a 14-h; peak indicative of K+ release. The maximum of the peak was obtained at the flowering stage. However, in the period following flowering the 14-h; peak declined, and the authors suggest that K+ released by plant roots after flowering was again fixed by interlayer adsorption sites. This observation is consistent with experiments of Kurdi and Babcock (1970), who found that at low K+ concentrapA4 K+)K was released by the roots and fixed by a tion in the root medium (4 bentonite suspension. The question of whether the excretion of H+ by plant roots can mobilize soil K+ is not yet clear. Newman (1969) found that at a low pH (3.5 to 4.0) the release rate of Kt from biotite was about twice as high as from more neutral pH. If a plant root depresses the pH of the rhizosphere due to H+ excretion, the release of interlayer K+ may therefore be promoted. Net Ht excretion of roots occurs when plants are fed with NHi-N (DeJaegere and Neirinckx, 1978) or in the case of legumes, when they are exclusively dependent on N fixation as an N source (Israel and Jackson, 1978). As yet, however, it is still uncertain whether this process of H+ release can mobilize interlayer K+ in quantities of importance in plant nutrition.

Ill. POTASSIUM IN PHYSIOLOGY A . POTASSIUM TRANSPORT ACROSS BIOLOGICAL MEMBRANES AND

CATION

COMPETITION

Physiology may be considered as a sequence of biochemical and biophysical reactions in living systems. For some of these reactions there is a direct or indirect association with K+, and it is for this reason that the entry of K+ into living systems merits attention. Of all cation species K+ is known to traverse biological membranes most rapidly. It has been shown by numerous authors that K uptake by plant cells is closely associated with metabolism, and especially with root respiration. Whether K+ is actively absorbed as defined by transport against an electrochemical potential is not completely clear. The application of the Nernst equation to studying electrochemical equilibria of ions in roots in bathing solutions has established this for other nutrient ions.

POTASSIUM IN CROP PRODUCTION

75

All anions are transported and accumulated against an electrochemical gradient (Bowling et al., 1966; Higinbotham et d., 1967), whereas the cations Ca 2+ and Mg2+invariably move into roots down an electrochemical gradient. The use of the Nernst equation for K+ transport studies provides evidence of both active accumulation (Pitman and Saddler, 1967; Bowling and Ansari, 1971) and passive equilibrium (Higinbotham et al., 1967; Pallaghy and Scott, 1969). The results of Etherton ( 1963) also indicate that the K+ concentration in the external solution can determine the form of uptake. At low external K+, the internal content was higher than that predicted by the electrochemical gradient indicating active uptake, whereas at high external concentrations the internal K+ was less than predicted and an outpump was suggested. To some extent the apparent discrepancy in the above results reflects the small differences obtained between observed internal concentrations of K+ and those predicted by the Nernst equations. Bowling (1976) has drawn attention to the possibility that since K+ is so mobile in plant tissues, passive diffusion may mask the activity of a K+ pump. The use of the Nernst equation may therefore be an inappropriate test for active K+ transport in higher plants. In reviewing the literature Higinbotham (1973) suggests that although there is good evidence for active K+ influx by algae a clear-cut case for higher plants has still to be made. Recent experiments of Cheeseman and Hanson (1979) with corn roots have shown that K+ can be taken up against an electrochemical gradient. The authors assume that this active K+ uptake is brought about by an ATPase which is inhibited by higher K+ concentrations and thus works only at concentrations <5 mM K+. From the above comments it is not surprising that the mechanism of K+ uptake is still much a matter of speculation. Certain points, however, have been established. The uptake process is known to depend on metabolic energy. ATP generated by respiration or photosynthesis is believed to be closely linked with the transport of K+ across biological membranes. This view is consistent with the observation that K+ uptake and K+ retention by roots responds very sensitively to 0,supply (Hopkins, 1956; Mengel and Pfluger, 1972). Potassium uptake by green leaves is also light dependent as has been shown by Jeschke (1970) with leaves of Elodea densa, by Nobel (1970) with pea leaves, and by Jacoby et a f . (1973) with bean leaves. Energy may be used directly in active uptake to transport K+ against an electrochemical gradient or used indirectly as, for example, by inducing an electrical potential gradient across the plasmalemma down which K+ may move into the cell. It is also well established that K+ uptake is selective in relation to other cations. It is therefore generally accepted that K+ uptake involves the combination of a hypothetical carrier with a K+ ion. This is believed to take place in such a way that the intermediate ion-carrier complex formed at the outer part of the membrane is transported inward. At the internal side of the membrane the ioncarrier complex breaks down, and K+ is released inside the cell.

76

KONRAD MENGEL AND ERNEST A . KIRKBY

A carrier has yet to be isolated from cell membranes. However, within the last 15 years organic compounds have been discovered that are capable of transport-

ing monovalent cations across membranes, and some bind very selectively with

K+. These substances, which are antibiotics, are collectively called ionophores because of their property of acting as ion-carrying agents (Hinkle and McCarty, 1978). Two main groups have been categorized by Pressman (1968). The first group includes valinomycin, gramicidin, and the macrotetralide actins. These are all neutral and form complexes by acquiring the charge of the complexing cation. The second group includes nigericin, which contains a carboxyl group thus giving the molecule a negative charge, which is neutralized by the cation. The structures of all these compounds are extremely complex (Kilbourn et al., 1967; Dobler et al., 1969). All, however, have lipophylic properties, so they are soluble in the lipid matrix of membranes. By selectively binding with K+ they thus facilitate the transport of K+ across biological membranes. The binding is similar to that in the interlayer of micas. In both, the nonactin-K complex and the valinomycin-K complex, K+ is held at the center of a ball and bound by eight 0 atoms. These 0 atoms of the organic complex replace H,O molecules from the K+ ion hydration shell during the process of binding (Kilbourn et al., 1967). The degree of tightness by which the water molecules are held controls the capability of the organic ligand to form a complex cation. Water is held more strongly by the hydrated Na+ ion than by the hydrated K+ ion. The Na+ ion is therefore less readily able to form an Na complex, and it may be supposed that it is this difference by which living organisms are enabled to distinguish very sensitively between K+ and Na+. The considerably higher rates of uptake of K+ than Na+ by most plant species probably also depend on this difference in complex formation. The selective properties of valinomycin in transporting alkali ions across a synthetic membrane were investigated by Mueller and Rudin (1967). In agreement with the above concepts, valinomycin increased membrane permeability to K+ so that the rate of K+ transport was about 300 times higher than that of Na+ . Nigericin also selectively transports K+ but only in exchange for protons (Hinkle and McCarty , 1978). Such or similar compounds to those discussed above may be responsible for K+ selectivity in uptake by higher plants. If the process of K+ complex formation and K+ release were linked directly to metabolism to provide the energy for uphill transport against an electrochemical gradient, the system in effect would be that of active K+ carrier transport. On the other hand, passive K+ transport may also be brought about by these selectively K+-binding compounds by enabling K+ to move down an electrical gradient from the outer solution into the cell. An ion uptake model in which cation influx takes place down an electrochemical gradient has been proposed by Hodges (1973). As shown in Fig. 3, cation uptake is regulated by an ATPase in the plasma membrane. For each ATP +

POTASSIUM IN CROP PRODUCTION

77

KEar$ ar K'

K'

Carrier trcjnsport (Valinomycin) ATP+ HO ,

ADP+ Pi

carrier NOj

FIG.3. Scheme of membrane-located ATPase-driving Kf carrier transport, facilitated K+ diffusion, and carrier-mediated NO, uptake.

molecule split by the ATPase, one H+ is produced which is extruded from the cytoplasm into the outer medium, and one OH- is generated in the cytoplasm. This process induces an electropotential gradient across the membrane, the cytoplasm being more negative compared with the outer medium. The cytoplasm thus attracts cations from the outer medium. Most cation species are probably absorbed by this mechanism, which per se is nonselective. However, as already mentioned, the presence of such substances as nonactin or valinomycin in the membrane could induce selective uptake by preferential binding with K+ in the downhill transport. This passive selective uptake has been called facilitated diffusion. The findings of Ratner and Jacoby (1 976) indicate that the high rate of K+ uptake by plant cells can be accounted for in terms of an ATPase-driven facilitated diffusion of K+ . There is considerable experimental support for the uptake mechanism of K+ as described above. As required in the Hodges scheme, all living plant cells are negatively charged with respect to the outer medium (Dainty, 1962). There is good evidence too that the plasmalemma contains ATPase (Hodges et al., 1972). Fisher et a / . (1970) observed a highly positive correlation between ATPase activity and ion uptake. A striking similarity between the kinetics of ATPase action and the kinetics of K+ absorption by oat roots has also been reported by Leonard and Hodges (1973). Recently an auxin involvement in K+ uptake has been suggested. Erdei et al. ( 1 979) found that auxins stimulated K+ uptake as

78

KONRAD MENGEL AND ERNEST A. KIRKBY

well as ATPase activity. These authors concluded from their in vivo and in vitro experiments with young rice roots that ATPase is directly involved in the K+ uptake process. Travis and Booz (1979) came to the same conclusion in experiments with meristematic and mature root tissue of soya. Thus K+ uptake is directly or indirectly associated with the electrogenic H+ pump (ATPase). Whether ATPase itself functions as a K+ carrier is still a matter of speculation (Poole, 1978). In addition to the carrier type of K+ transport across biological membranes already discussed, another mechanism of ion diffusion involving membrane pores has also been suggested. Such pores can be formed by macrocyclic antibiotics including gramicidin A (Mueller and Rudin, 1967) and may allow “tunnel transport” of cations. In the case of gramicidin A, the pores are each made up of two helical molecules and are permeable to ions with one positive charge (Stryer, 1975). There is now a considerable body of experimental information which strongly suggests that membranes contain selective K+ carriers as well as pores which enable a less selective K+ uptake by tunnel transport. The dual pattern of K+ uptake which Epstein (1966) has described as mechanism I and mechanism I1 may be explained in these terms. Mechanism I probably describes the selective carrier transport, and mechanism I1 the less selective tunnel transport. Cation efflux may also occur through membrane pores, although the process does not appear to be selective. Mengel and Pfluger (1 972) observed similar rates of release of K+ and Na+ from corn roots, whereas in the same tissues the K+ uptake rate was about 10 times higher than that of Na+. In the Hodges’ model anion uptake takes place by exchange for the OHgenerated in the cytoplasm as a result of the ATPase-driven H+ extrusion (see Fig. 3). An additional source of cytoplasmic OH- also arises after the uptake and assimilation of NO; and to a lesser extent also of S q - assimilation. NO; SO:-

+ 8H+ + 8e+ 8H+ + Re-

NH, + 2H20 + OHSH2 2 b O 20H-

+

+

This continuous OH- production resulting from NO-, assimilation may directly provide a means for sustained NO, uptake without accompanying cation uptake by NO;/OH- exchange between the root and the nutrient medium. On the other hand, the OH- produced may stimulate the accumulation of an organic acid anion in the plant via the PEP carboxylase mechanism (Hiatt, 1966; Davies, 1973) with accompanying cation uptake (Kirkby, 1968; Kirkby and Knight, 1977) to balance the organic anion charge. It is also possible that organic acid anions formed in this way may be oxidatively decarboxylated to again give rise to an internal source of OH- to drive anion uptake independent of cation uptake as already described. This mechanism has been considered in relation to longdistance K+ transport and is discussed later. The influence of the accompanying anion such as NO; or C1- on the uptake of

POTASSIUM IN CROP PRODUCTION

79

K+ by young barley plants was studied by Blevins et al. (1974). The results obtained in this experiment are consistent with the concepts discussed above. The uptake and accumulation of K was greater in the NO ;-treated plants particularly toward the end of the experiment (4-36 hours). The authors suggest that one of the main reasons for this finding was a more rapid sustained uptake of NO; which provided a mobile counter-anion for K+ transport. In addition, the synthesis and accumulation of organic acid anions by NO; reduction increased the capacity for K+ accumulation by providing a nondiffusible organic anion source. Similar enhancing effects of NO, on the uptake of K+ and other cations by tomato plants have been observed by Kirkby and Knight (1977). The influence of NO, assimilation in providing a major source of OH- to induce further NO, uptake without accompanying cation uptake, by NO,/OH- exchange, has been reported recently by Kirkby and Armstrong (1980) in the castor oil plant (Ricinus cornmunis). This process also appears to predominate in grasses and cereals where the uptake of anions is about twice that of cations (Kirkby, 1974). The above findings are in agreement with the observation of Johansen and Loneragan ( 1975) who suggest anion-dependent and anion-independent components in the K+ uptake process. The proposed model for cation uptake as considered above is also of relevance to the well-known phenomenon of cation antagonism, which is more precisely described by the term cation competition. Numerous experiments on growing plants in complete nutrient solutions have shown that increasing the concentration of K+ in the outer solution depresses the uptake of other cation species. The reverse effect of other cation species on K+ uptake has also been observed. Since there is little evidence that the uptake of Ca2+, M$+, and Na+ is carrier mediated, it is unlikely that the depression of uptake of these cation species results from a competition for a common carrier site. A more likely explanation for cation competition is that all cation species are attracted by the negative electropotential of the cell, which is continuously being regenerated by H+extrusion and by NO, uptake and assimilation. Cation species that can traverse the plasma membrane more easily, for example, by facilitated diffusion, have a greater chance of saturating the continuously generated negative potential of the cell than cation species for which the plasma membrane represents a banier to uptake. A cation species that enters the cell rapidly thus depresses the uptake of other cation species. Such a mechanism should result in nonspecific cation competition (Mengel, 1973). Competition is nonspecific, because all cation species are involved and the extent of uptake depression depends mainly on the concentration of the cation species in the outer solution and the permeability of the membrane for the individual cations. Most plant species absorb K+ at a rather high rate as compared with the uptake rate of other cation species. From the model described above, therefore, K+ +

80

KONRAD MENGEL AND ERNEST A. KIRKBY

would be expected to be a strong competitor in uptake with the other cation species. This has been demonstrated in numerous experiments (Smith and Robinson, 1971; Mengel and Nemeth, 1971). Nonspecific cation competition becomes especially evident when plants are grown with a suboptimal supply of K+. Such an experiment was conducted by Forster and Mengel (1969) in which the K supply to young barley plants (tillering stage) growing in solution culture was interrupted for a period of one week. This interruption resulted in a dramatic increase in the uptake of the other cation species. The main result of this experiment is given in Table I. Although the cation sum was hardly changed by the K+ interruption, the content of various cation species was much altered. This example demonstrates that in the absence of K+, the uptake of other cation species is considerably enhanced. Similar results have been obtained by Terry and Ulrich (1973a) with leaves of sugar beets and by Dijkshoorn ef al. (1974) in experiments with rice plants. Most plant species absorb K+ at a higher rate than Na+; but there are some, such as Beta species and spinach, that also take up considerable amounts of Na+ (Marschner, 1971). In these species Na+ may partially substitute in the biochemical or biophysical functions of K+. The Na+ uptake mechanisms of these natrophilic species are not yet understood. One may suggest, however, that where Na+ is taken up in such large amounts, this may be achieved by facilitated diffusion through membrane channels formed by gramicidin A or similar compounds. In the carrier-type uptake, the carrier, for example, valinomycin, enniatin A, or related compounds, may act as a shuttle between one side of the membrane and the other. Potassium uptake and K+ release of the carrier should thus be controlled by the electrochemical potential gradient across the membrane. Net K+ uptake should be dependent on the K+ concentration in the cytoplasm, because the net K+ release rate of the carrier at the inner side of the membrane Table I Effect of Interrupted K Supply on the Cation Content (meq/100 g dm) of Roots and Shoots of Young Spring Barley Plants" Roots

Shoots

Control

Intemption

Control

Intemption

K Ca Mg Na

157 9 36 3

28 12 74 78

170 24 54

Traces

52 66 121 12

Sum

205

I92

248

25 1

Forster and Mengel (1969).

POTASSIUM IN CROP PRODUCTION

81

should be depressed if the K+ concentration in the cytoplasm is high and vice versa. Such a relationship between the K+ concentration in the plant and the net rate of K+ uptake has been reported by several authors (Hoffmann, 1966; Johansen et al., 1970; Barber, 1979). The observation that the K+ efflux rate of plant roots is higher if K+ is present in the outer solution (Mengel and Pfliiger, 1972) also supports the concept of a specific K+ carrier working as a shuttle in which K+ loading and unloading is dependent on the electrochemical potential gradient across the membrane. According to Zimmermann (1978), K + uptake is also regulated by the osmotic pressure of the cell, higher turgor pressure decreasing the uptake rate and vice versa. Since a high cell K+ content is generally associated with a high turgor, K+ uptake may also be subjected to osmoregulation. From kinetic studies on uptake it appears that competition does occur for the carrier cation binding sites for chemically closely related cation species such as Ca2+ and Sr2+,Rbf and K + . Organic cations may also react with the carrier binding site. Lepe and Avila (1975) thus reported that alkylguanidines considerably depressed K+ uptake by excised barley roots. The depression in uptake was greater, the longer the alkyl chain. The authors therefore suggest that the positive site of the guanidine complex reacts with the carrier binding site, whereas the lipophilic chain is bound to the membrane. Whether the bulk of NH :-N absorbed by plants is transported through biological membranes by a carrier-type mechanism as outlined above, is in question. For such a mechanism to be operative one might expect a marked competition for binding sites with K +. The uptake of NH: should therefore be depressed by K + . Mengel et al. (1976), however, were not able to detect any such depressing effect of K when studying NH:-N uptake by young rice plants. +

B . CELLTURGOR A N D WATERECONOMY OF PLANTS

A high rate of K+ uptake by root cells depresses the osmotic potential in the cells, and this induces water uptake. The uptake of water by roots and the ability of the plant to exploit soil water therefore depend on the K+ nutritional status of the plant. Water transport into the xylem vessels is also mainly an osmotic process in which K+ in its function as an osmoticum is very important. Electron probe analysis of Lauchli et al. (1971) has shown that parenchyma cells may accumulate K+ to a high extent and that K+ is secreted into the xylem. During this process K+ has to traverse the membrane that separates the symplasm of the living stelar cells from the free space of the conducting vessels. Whether K+ leaks out passively from the symplasm into the xylem vessel or whether it is actively transported across this membrane is not yet clear. It is feasible that this K+ transport mainly occurs by facilitated diffusion dependent on ATPase activity

82

KONRAD MENGEL AND ERNEST A. KIRKBY

and active anion transport. Lauchli (1972) holds the view that the K+ transport is active. This is supported by results that have shown that the xylem parenchyma cells contain numerous plasmodesmata and pits forming a plasmatic pathway across the stele up to the vessels (Lauchli et al., 1974). Whatever the mechanism of K+ release into xylem vessels, K+ is the important osmoticum which drives the water flux from the surrounding cells into the xylem vessels (Baker and Weatherley, 1969). Only in plant species that absorb Na+ in considerable amounts can K+ be substituted in this osmotic function by Na+. The root pressure, which can be of importance for the upward movement of organic and inorganic solutes, is much controlled by the K+ nutritional status of plants. If K+ is low or absent in the root medium, both the quantity of water moved upward by root pressure and the concentration of a number of solutes such as nitrate and amino acids in the xylem sap are considerably depressed. Such observations have been made by Mengel and Simic (1973) in experiments with decapitated sunflower plants. Potassium plays a spectacular role in stomatal opening and closure (Fischer and Hsiao, 1968). Convincing evidence for this essential K+ effect has been provided by electron probe analysis studies of Humble and Raschke (1971) which showed that the increase in turgor in the guard cells associated with stomatal opening resulted from an increase in K+ concentration in the cells. Under light conditions, photophosphorylation seems to provide the ATP required for pumping K+ into the guard cells (Humble and Hsiao, 1969; Turner, 1972). The K+ accumulation is associated with an accumulation of malate which appears to be the major anion charge balancing the accumulated K+. Thus under dark conditions Allaway (1973) found very little malate in the guard cells of Vicia faba, whereas on exposure to light there was a rapid increase in the malate concentration. Similar results have been found by Raschke and Schnabl(l978) who showed that when KCl was applied a proportion of the guard cell K+ was balanced by Cland the rest by malate. Abscisic acid prevents stomatal opening and simultaneously inhibits K+ uptake by guard cells (Horton and Moran, 1972). Fusicoccin, on the other hand, which is a phytotoxin, promotes the uptake of K+ by guard cells and thus induces stomatal opening (Turner, 1972). Since fusicoccin stimulates ATPase selectively (Giaquinta, 1979), it is likely that the K+ uptake is closely related to ATPase activity. The role of K+ in stomatal opening is very specific, and other univalent cations are generally unable to replace K+ in this function, except in a few plant species, for example, Kalanchoe marmorata where Na+ drives the stomatal opening mechanism. The subject of ion transport in stomatal guard cells from the viewpoint of the chemiosmotic theory has recently been considered by Zeigler et al. (1978). Potassium is also a major osmotically active component in other plant cells contributing to cell turgor and enhancing the capacity of plant cells to retain

POTASSIUM IN CROP PRODUCTION

83

water. In this function K+ seems to be of particular importance in young tissues. Thus Arneke (1980) found that the turgor of older leaves of Phaseolus vulgaris was hardly influenced by K+ nutrition, whereas in the younger leaves, the turgor was dependent on the K+ supply to the plants. In the low K+ treatment the turgor of young leaves showed an average value of 5.05 bar which differed significantly from the average turgor (7.17 bar) obtained in the young leaves of the high K+ treatment. This experiment also showed that a first sign of inadequate K+ nutrition was the higher dry weight content of the plant tissues. Turgor appears to be the most sensitive parameter indicating K+ nutritional status. Other K+-related processes such as COz assimilation, phosphorylation, and protein synthesis are less sensitive in registering inadequate K+ supply. Arneke ( 1980) found that the turgor in young leaves had a direct effect on the size of cells and on the growth rate of the entire plant. This work indicates that a high turgor in meristematic cells is a prerequisite for cell expansion, a process which precedes cell division. According to Arneke’s results, K+ seems to be an indispensable osmoticum in young leaves of P haseolus vulgaris. This finding is well supported by experimental findings of Green and Muir (1978) who observed that the expansion of cucumber cotyledons was dependent on the K+ supply. It is feasible that in the natrophilic plant species this osmotic effect of K+ may also be brought about by Na+. Marschner and Possingham ( 1975) thus found that K+ as well as Na+ promoted cell expansion and the production of chloroplasts in the leaves of sugar beet and spinach. The overall effect of K+ on the water economy of plants results from the process cited above and probably also from processes that are not yet known or well understood. This beneficial effect of K+ is of particular importance in practical crop production, since K+ reduces water losses by transpiration (Brag, 1972), so that more organic matter can be produced per unit water consumed by a crop well supplied with K+ (Blanchet et al., 1962; Linser and Herwig, 1968). C. ENERGYMETABOLISM

The basic process of energy metabolism-the conversion of radiation energy into chemical energy-is much controlled by the K+ status of the plant. The beneficial influence of K+ on phosphorylation has thus been reported by various researchers using different plant species (Hartt, 1972; Watanabe and Yoshida, 1970; Pfliiger and Mengel, 1972). Photoreduction (production of NADPH) is also promoted by K+ (Pfliiger and Mengel, 1972). Indirect evidence for this effect has also been provided by experimental findings of Weller and Hofner (1974) and Overnell (1975). Both publications report that K+ increases the photosynthetic release of Oz. The specific function of K + in the energy conversion process is not yet

84

KONRAD MENGEL AND ERNEST A. KIRKBY

completely understood. However, it is recognized that K+ is involved in metabolic reactions including those of energy (ATP), synthesis, and energy transfer. Metabolic energy is generated by the chloroplasts in the process of photophosphorylation. Light-driven electron transport releases protons from the stroma into the inner space of the thylakoid compartment of the chloroplasts (Trebst, 1974). Cations and especially K+ are moved out into the stroma of the chloroplast in exchange for this inward movement of protons. According to Lauchli and Pfluger (1978), K+ in a concentration range of about 100 mM seems to be necessary for high efficiency in energy transfer. Of the other cations Mgz+ is particularly important (Barber, 1977). Portis and Heldt (1 976) have proposed that the light-driven uptake of Mg2+ from the inner space of the thylakoids to the stroma is enough to “switch on” RuBP carboxylase by increasing its affinity for C02. Potassium also promotes CO, fixation but this occurs according to Peoples and Koch (1979) not by direct activation of RuBP carboxylase but rather by favoring the synthesis of this enzyme. The fluxes of protons, electrons, and cations across the thylakoid membrane as described above constitute the major features of the Mitchell (1966) chemiosmosis hypothesis of ATP synthesis. The same is true for charge transfer across the inner mitochondrial membrane. However, in the case of oxidative phosphorylation which occurs during respiration in the mitochondria, the direction of charge transfer is reversed in the synthesis of ATP. Again a proton concentration gradient and an electrical potential difference across the membrane drives ATP synthesis, but in this case the protons are moved from the inside to the outside of the mitochondria. Potassium and calcium thus move inward in exchange for protons. It is generally accepted that K+ plays an analogous role in mitochondrial as in chloroplast energy turnover. However, convincing evidence for the role of K+ in mitochondrial energy transfer is lacking. In the older literature there is some evidence of a beneficial effect of K+ on oxidative phosphorylation (Latzko and Claus, 1958; Latzko, 1961). Jackson and Volk (1968) have also presented data which show that under conditions of suboptimal K + supply a greater proportion of ATP is produced by respiration to compensate for low production under conditions of poor photophosphorylation. This finding is consistent with recent results of Peoples and Koch (1979), who observed an inverse relationship between the rate of dark respiration and the K content of alfalfa leaves. The promoting effect of K+ on photosynthetic ATP synthesis and NADPH production has a general impact on various energy-requiring processes in plant metabolism. The effect of K+ on CO, assimilation has been observed by numerous authors (Ilyashouk and Okanenko, 1970; Estes et al., 1973; Terry and Ulrich, 1973b). Other beneficial influences include those on protein synthesis (Jeanniot et al . , 1970; Hsiao et al., 1970; Koch and Mengel, 1972), on N2 fixation by Rhizobium bacteria (Mengel et al., 1974; Feigenbaum and Mengel, 1979), on amino acid synthesis (SeGer, 1978), on synthesis of vitamin C (Werner, 1957), and on long-distance transport. All these effects can be explained

POTASSIUM IN CROP PRODUCTION

85

in terms of K+ enhancing the production of ATP or NADPH, or both of these compounds. It is well known that K+ activates numerous enzymes, a subject that will be considered in more detail below. The beneficial effect of K+ on protein synthesis may thus originate from K+ enzyme activation, and frequently such effects have been interpreted in these terms. Studies with entire plants, however, often do not allow the distinction of the cause of this beneficial effect between direct enzyme activation and an increase in energy supply associated with higher levels of K nutrition. In this context the findings of Seqer (1978) are of particular interest. In an experiment with spring wheat growing at two levels of K nutrition, she studied the amino acid turnover of the grains during the grain filling period. In the higher K treatment the content of soluble amino acids in the grains during the first weeks of grain filling was about twice as high as that of the low K treatment. In addition the incorporation of amino acids into the grain proteins and especially the turnover of glutamate occurred at a much higher rate in the high K plants. In this treatment too the K+ content was higher in both the roots and the culms. However, no major difference between treatments was observed in the K+ contents of the grain. For this reason it seems unlikely that the higher protein synthesis in the high K treatment was a consequence of K+-stimulated protein synthesis. It seems more feasible that the high level of K nutrition enhanced amino acid translocation from the vegetative plant parts to the grains and that this higher import of amino acids into the grains promoted protein synthesis. This example demonstrates that the beneficial effect of K+ on long-distance transport of assimilates may be of greater importance for protein synthesis than the effect of K+ on enzyme activation. Insufficient K+ in the plant may often result in an increase in protein content of vegetative plant material. This has been observed by Hsiao er al. (1970) in corn seedlings and by Mengel and Koch (1971) in young sunflower plants. The effect cannot be ascribed to improved protein synthesis under suboptimal K+ conditions, for under such conditions the rate of protein synthesis is restricted, as was shown in experiments of Koch and Mengel (1972, 1974) using labeled N. The increased protein content is a secondary effect of suboptimal K+ supply. Hsiao et al. (1970) demonstrated that the growth process was more closely related to the K supply than was protein synthesis. This explanation is consistent with the observations that cell turgor and thus cell expansion are the most K+-sensitive processes. If the growth rate is depressed more than protein synthesis, an accumulation of protein occurs. D. LONG-DISTANCE TRANSPORT

Potassium is known to be very mobile in an upward and downward direction in the entire plant. A high rate of translocation occurs in the xylem because of the

86

KONRAD MENGEL AND ERNEST A. KIRKBY

rapid rate at which K+ is selectively secreted into the root xylem vessels. Of all (.is also by far the most abundant in the phloem sap, where the cation species, I it may reach concentrations of 100 mM and more (Hall and Baker, 1972). This finding indicates that K+ is selectively absorbed by the sieve tubes, and it also explains why K+ can so easily be translocated from the upper plant parts to the basal plant organs and roots. Ben Zioni et al. (197 1) have suggested that K+ circulation in the entire plant is of significance for the upward translocation of nitrate. From their investigations with Nicoriana rustica they concluded that K+ acts as the main counter-ion for the upward translocation of nitrate. These workers propose that on reduction of nitrate in the upper plant parts an equivalent of malate is formed, and some of the K originally accompanying the NO; is then transferred together with malate via the phloem to the roots. Here the malate is oxidized and decarboxylated, and the HCO 3 produced is released into the nutrient medium in exchange for uptake of an NO 3 ion. The K remaining in the root together with this NO ;is then transported upward and the cycle repeated. For the Ben Zioni et al. (1971) K-recirculation scheme to operate, two criteria must be met. In the first place NO; reduction must take place largely in the tops. Second, the OH- resulting from this NO; reduction must be transferred to the nutrient medium in exchange for a net excess anion over cation uptake. From this it follows that in plant species such as tobacco and tomato in which the uptake of anions is not greatly in excess of that of cations, when supplied with NO;-N, that the scheme is probably of little importance. In the case of tomato, for example, about 90% of the OH- charge arising from NO; reduction in the tops is retained in situ as organic acid anions in association with inorganic cations (Kirkby and Knight, 1977). On the other hand, a number of plant species including grasses and cereals do take up a considerable excess of anions over cations (Dijkshoorn, 1962; Kirkby, 1974). Evidence supporting this hypothesis has thus been provided by Blevins er al. (1978), who compared barley seedlings well supplied and poorly supplied with K+. The shoots of the low K+ seedlings had much lower NO; concentrations and lower NO;-reductase activities than the roots, suggesting that K+ plays a major role in NO; translocation. Further support for this scheme has been demonstrated by Kirkby and Armstrong (1980) in the castor oil plant (Ricinus communis). In this species the uptake of anions is again approximately twice that of cations when the plants are fed with NO;-N. Evidence of K recirculation was concluded from an experiment in which a low and high level of NO; nutrition were compared, keeping the cation concentrations in the nutrient solution constant by compensating for the difference in NO 3 concentration by SO:-. In both NO; regimes, the reduction of NO; took place largely in the tops. In agreement with the Ben Zioni et al. (197 1) model the dominant ions in the xylem sap were NO,, K+, and Ca2+,whereas +

+

POTASSIUM IN CROP PRODUCTION

87

those in the phloem were K and organic anions. At the higher level of NO 3 nutrition, N uptake was increased threefold but there was no influence on K uptake. It was therefore concluded that the increase in upward NO, transport was facilitated by K recirculation. Whether K+ is transported in the phloem with malate is still an open question. Mengel and Haeder (1977) reported much lower concentrations of malate (approximately 1 mM) than K+ in the phloem sap of Ricinus communis, a finding consistent with the data of Smith (1978). Since the pH of phloem sap is high (pH 8-9), it would be reasonable to assume that amino acids may provide negative charge to balance the K+ charge. The question of which amino compounds act as counter-ions in K+ transport in xylem and phloem still needs to be investigated. The use of a legume plant would be particularly useful in this respect because it would allow the comparison to be made between NO,, NH:, and molecular N2 fed plants. The significance of an antiport (OH- exchange) as compared to a symport (associated cation uptake) mechanism for NO, uptake has been suggested by Cram (1 976) to relate to the osmotic pressure generation in the plant. Cram argues that if all the N were taken up as KNO, and one K + and one organic acid produced per N reduced to organic form, as in tobacco or tomato, the osmotic pressure of the cells of the Grumineae might be twice the values observed. The switch to NO, uptake without accompanying cation uptake appears to prevent the osmotic pressure from rising and may hence regulate internal osmotic pressure. It is of interest that those plant species which do take up nitrate by an antiport mechanism also take up a relatively high proportion of their cations as K+. For plants taking up ions by the symport mechanism, the cation uptake is higher but so too are the proportions of Ca2+and Mg2+,two cation species that may to a large extent form insoluble salts and thus not contribute to internal cellular osmotic pressure. This difference in behavior lends further support to the Cram ( 1976) internal osmotic regulation hypothesis. The beneficial effect of K+ on the long-distance transport of photosynthates has been shown by several authors for various plant species: by Amir and Reinhold (197 1) for beans, by Hartt (1970) for sugarcane, by Ashley and Goodson (1972) for cotton, by Ilyashouk and Okanenko (1970) for sugar beet, by Haeder et ul. (1973) for potatoes, and by Mengel and Viro (1974) for tomatoes. Barankiewicz ( 1978) reported that under the conditions of suboptimum K+ nutrition, the turnover of C from malate and aspartate to sugar phosphates in corn leaves was affected. From these results it is concluded that K+ promotes the transport of malate and aspartate from the mesophyll tissue toward the bundle sheath cells of C-4 plants. The main organic constituents of the phloem sap consist of sugars, mainly sucrose, and amino acids. For this reason the beneficial effect of K+ on phloem transport also has a direct impact on the long-distance transport of N. This can be +

88

KONRAD MENGEL AND ERNEST A. KIRKBY

of importance for the utilization of N in crop production. Koch and Mengel (1977) thus showed that N stored in vegetative plant parts of wheat could be used for the grain production of wheat to a greater extent when the plants were well supplied with K+ than when they were not. Magnesium transport in the phloem also appears to be promoted by K+. The Mg contents of flax seeds (Linser and Herwig, 1968), of potato tubers (Addiscott, 1974a), and of tomato fruits (Viro, 1973) were thus increased when the level of K nutrition was raised. This finding is of particular interest since generally the Mg content of plant organs is adversely influenced by K+ nutrition as a result of cation competition. The data of Viro (1973) show that although cation uptake competition occurred in tomato plants, retranslocation of Mg2+ from roots, stems, and leaves to the fruits was promoted by K+. The higher K+ treatment thus resulted in a greater MgZ+ concentration in the tomato fruits. The beneficial effect of K+ on phloem transport has also been found where CO, assimilation was not significantly influenced by K+ . This shows that the K+ status of plants is registered more sensitively by phloem transport than by CO, assimilation. The mechanism by which K+ influences phloem transport is not completely understood. Addiscott (1 974b) has discussed several concepts that could explain the positive influence of K+. He concluded that K+ probably favors the process of phloem loading. Mengel and Haeder (1977) in studying the composition of phloem sap in Ricinus communis found that the level of K nutrition did not greatly influence the concentration of phloem sap solutes. In particular the concentrations of the major solutes, sucrose and amino acids, were not significantly altered by different K treatments. The most important finding of this experiment was that K+ raised the flux rate of phloem sap considerably. The authors suggest that K+ stimulates ATP synthesis by its beneficial effect on photophosphorylation. This in turn, they argue, enhances the phloem loading process so that the higher rate of phloem loading also results in a higher rate of phloem flux. This explanation agrees well with the phloem loading concept of Giaquinta (1977) who has suggested that the process of sugar transport across biological membranes is driven by an ATPase. Recent research data of Travis and Booz (1979) provide evidence that the plasma membrane-bound ATPase is activated by K+ . The beneficial effect of K+ on phloem loading could thus also be related to K+ activation of the plasma membrane-bound ATPase. Malek and Baker (1977) have proposed a direct relationship between K+ and the sugar uptake mechanism of the phloem. In this scheme H+ efflux is associated with the uptake of K+, an exchange process similar to that across the thylakoid membranes. The uptake of K+ gives rise to an equivalent release of protons into the apoplast. This raises the H+ concentration at the outer side of the plasma membrane, which according to Giaquinta ( 1977) promotes phloem loading with sugars. This effect of increasing the H+ concentration in the apoplast in

POTASSIUM IN CROP PRODUCTION

89

promoting phloem loading with sugars has been shown experimentally in Ricinus cotyledons by Hutchings (1978). Since phloem loading with amino acids is also enhanced by K+ (Koch and Mengel, 1977), it will be of interest to see whether the transport of amino acids through the plasma membrane of sieve tubes is also brought about by a similar mechanism. Very recent research data of Doman and Geiger ( 1 979) show that the release of photosynthates from the mesophyll cells into the apoplast is dependent on the K+ status of the leaf tissue. In their experiments they were not able to find a direct beneficial effect of K+ on phloem loading. These authors claim that it is the release of photosynthates from the mesophyll cells into the apoplast rather than the phloem loading process itself that is promoted by K+. As most photosynthates have to pass through the apoplast before being collected in the sieve tube companion cell complex, enhanced export of photosynthates out of the mesophyll cells also results in a higher rate of phloem loading. In recent years it has become increasingly clear that hormonal effects may control the movement of nutrients in plants. Such an effect may well account for the findings of Pitman (1972) from an experiment with barley plants in which the growth rates of the plants were varied considerably by altering the photoperiod. It was observed that the net uptake of K by the roots and the transport in the shoots closely reflected changes in shoot growth. Pitman (1972) suggests that the uptake of K+ was regulated by a “feedback” mechanism between roots and shoots, in which the translocation of growth substances may be involved. Direct evidence of growth regulators on the movement of nutrients in intact plants is difficult to obtain because of the problem of interpretation, since the effects of hormones on growth and metabolism may obscure effects on transport. A detailed investigation on the fluxes of K+ in the root of an intact corn plant has been carried out by Richter and Marschner (1973). Influx rates were found to be similar along the length of the primary root. Owing to the higher demand of the growing tip, however, an additional K+ flux occurred toward this site from the basal parts of the primary roots. Newly developed lateral buds were also found to act as sinks for K+. The control of K+ transport within the root by growth substances has been demonstrated by the results of Luttge et al. (1968) for gibberellic acid. Cram and Pitman (1972) have also shown that abscisic acid influences long-distance K+ transport. The results of Mansfield and Jones (1971) indicate that in guard cells ABA increases the permeability of the cell plasmalemma thereby increasing K+ uptake. Collins and Kemgan (1974) also observed that very dilute solutions of ABA (10-8-10-9 M) added to a bathing solution containing detached maize roots increased both the flux of exudate and of K+ into the xylem. Kinetin had the reverse effect. The mechanism by which such hormonal effects operate in the long-distance transport of K+ has yet to be established. +

90

KONRAD MENGEL AND ERNEST A. KIRKBY

E. ENZYMEACTIVATION

More than 60 different enzymes are known that require univalent cations for activation. In most cases K+ is the most efficient cation species in this activation process. This question of enzyme activation has been thoroughly treated by Evans and Sorger (1966), Wilson and Evans (1968), and Evans and Wildes (1971). For this reason the biochemical aspect of enzyme activation by K + is considered here only briefly. A number of K+ specific enzymes are equally activated by NH: and Rb+ under in vitro conditions. This finding supports the concept that the ion radius or the energy of dehydration or both are of direct importance for the activation mechanism. In vivo, however, NH: and Rb+ cannot substitute for K+ in activating enzymes, since these ions are toxic at the concentrations required (El-Sheikh and Ulrich, 1970; Morard, 1973). Potassium activates different groups of enzymes. For crop production the synthates such as starch synthase, phosphorylase, and ADP glucose pyrophosphorylase (Marschner and Doring, 1977; Hawker et a f . , 1979), as well as the enzymes involved in protein synthesis are of particular interest. The general finding that inadequate K+ nutrition results in the accumulation of low molecular weight sugars and amino acids (Okamoto, 1967; Ratner and Yeliseova, 1968; Nowakowski, 1971) in plant tissues may be explained in terms of depressed enzyme activity. In this context, however, it should be remembered that plants poorly supplied with K+ can suffer from a lack of ATP, which in turn can also result in impaired synthesis of polymers such as protein, starch, cellulose, and even nucleic acids. In vitro experiments have shown that maximum K+ activation is obtained within a concentration range of between about 40-80 mM K. Besford and Maw (1976), for example, report optimal activity of pyruvate kinase in vitro at about 45 mM K in fresh tomato leaf tissue. Their findings show that this value is in excess of that required for optimal succinyl CoA synthetases in the same tissue. In many plant tissues it appears that K+ may be present in relatively high concentrations in relation to enzyme activation requirements. Experiments of Pierce and Higinbotham (1970), for example, indicate that the K+ concentration in the cytoplasm of Avena coleoptile cells is in the range of 140-215 mM. Even in K+-deficient tissues the K+ concentration may appear quite high. In an experiment with young K+-deficient bean leaves, Arneke (1980) reported values of 50-70 mM K+ in the press sap, and it seems highly probable that the K+ concentration in the cytoplasm of these leaves was higher. From this evidence it would appear that even under conditions of a suboptimal K+ supply, the K+ concentration for optimum enzyme activation may still be adequate. This implies that for registering inadequate K nutrition, enzyme activation by K+ is a relatively insensitive parameter as compared with K+-stimulated ATP synthesis or cell expansion. This hypothesis needs further investigation, but

POTASSIUM IN CROP PRODUCTION

91

it is well supported by available research data. Its practical consequence is that impairment of enzyme activation by a lack of K+ is unlikely to limit the crop production process. A very good practical example illustrating this point has already been considered in discussing the work of Seser (1978) (see Section 111,

C).

IV. POTASSIUM APPLICATION AND CROP GROWTH A. CROPRESPONSEA N D POTASSIUM APPLICATION

In agricultural practice it is important to know for any given location whether K fertilizer application can significantly increase crop yield. Numerous field experiments have therefore been carried out to investigate crop responses on particular soils. However, it is difficult to extrapolate these results in order to predict fertilizer recommendations because of the many and varying factors influencing the growth of a crop at a particular site. Clearly, though, such generalizations must be made since it is quite impracticable to carry out trials on all the numerous combinations of crops and soils. The main factors influencing crop response to K fertilization are available soil K + , soil moisture, other growth factors, and the particular crop species under consideration. The common practice of expressing K availability in terms of exchangeable K+ is one immediate problem that must be recognized. As already outlined in Section I1 “exchangeable K + ” is not a good indicator of K availability. The equilibrium K+ concentration in the soil solution and the K+ buffer capacity are more reliable parameters for assessing the rate of K+ supply to plant roots in the soil medium. On sandy soils small applications of K+ increase the K+ concentration in the soil solution appreciably and may thus result in substantial yield increases (Mengel and Aksoy, 1971). On more highly textured soils, however, K+ fertilizer applications may scarcely influence K+ concentration in the soil solution, so yield responses are often not obtained. In such soils it must be established whether a lack of response is due to an already high enough level of available K+ in the soil or to too low a K fertilizer application. Generally soils rich in 2:1 clay minerals contain fairly high amounts of exchangeable K+ associated with relatively low K+ concentration in the soil solution (Nemeth et ul., 1970).Solution K+ may thus become a limiting factor in supplying soil K+ to plants (Jankovic and Nemeth, 1974). On such soils therefore the application of K fertilizers must be high enough to increase the K+ concentration of the soil solution if yield increases are to be obtained (Mengel, 1974). v. Braunschweig (1979a) has established from numerous field experiments carried out in West Germany that approximately optimum K+ concentrations in the soil solution are

92

KONRAD MENGEL AND ERNEST A. KIRKBY

obtained when the relationship between lactate-soluble K+ and soil clay is as shown in the following equation: Lactate soluble K+ (ppm)

= %

clay x x

where x is a value between 12 and 20. The value for x increases as the % clay content falls. “Lactate soluble K+ differs from “exchangeable K+,” but the exchangeable K+ value can be used in an analogous equation to that shown above. The equation is valid only in soils in which the clay fraction is made up predominantly by 2: 1 clay minerals. If soils do not differ too greatly in clay content, the value for exchangeable Kt may provide a reliable indicator of the K+ availability and therefore allow comparisons to be made regarding K status. Thus in more than 300 field experiments with wheat, carried out in France, it has been shown that K fertilizer applications resulted in high grain yield increases on soils with values for exchangeable K+ <80 ppm. Medium yield increases were obtained on soils with exchangeable K+ levels between 80 and 160 ppm, and only small yield increases (0.2 ton graidha) were found when the exchangeable Kt was > 160 ppm (LOUC, 1979). Soils capable of fixing large amounts of K+ frequently require particularly high K+ fertilizer dressings to show yield responses. Typical results of K fertilizer applications to a K+-fixing site are presented in Table 11. The data show that the K+ response obtained on a given site can differ considerably and depends on the crop species and on the weather conditions of the particular year. Generally, dry soil conditions during spring provide better K+ responses than moist conditions. The data of Table I1 also indicate that in some cases even rates of 300 kg K,O/ha did not result in spectacular yield increases, and as much as 900 kg K,O/ha were needed to obtain maximum grain yields. The results shown in Table I1 were obtained by Burkart (1975), who camed out field experiments on nine K+-fixing sites located in Southern Bavaria in the ”

Table I1 Effect of K Fertilizer Application on the Grain Yield (ton/ha) on Two Different Sites“ +

Domwag

Weng

K rate (kg K,O/ha)

Spring wheat, 1972

Maize, 1973

Maize, 1972

Spring wheat, 1973

0 300 600 900

3.27 3.96 6.16 4.48

2.48 3.88 5.04 5.48

5.34 5.63 8.66 9.37

4.83 4.62 5.07 5.21

“ Burkart (1 975).

POTASSIUM IN CROP PRODUCTION

93

Federal Republic of Germany. Highest yield responses were found with corn, whereas winter wheat and oats responded only by small to medium yield increases. The extent of the yield response obtained on these sites was not related to the K+ fixation capacity, which ranged from 349 to 783 ppm K as measured as “wet fixation. Dramatic yield increases by high applications of fertilizer K+ have also been reported by Doll and Lucas (1973) on K+-fixing soils in Michigan. Since applied K+ does not immediately equilibrate with soil K+ , but remains for weeks and even months in a readily accessible state, it is opportune to apply potash fertilizer to K+-fixing soils just before sowing. Potassium top dressing is also recommended. During the process of weathering of K-bearing minerals, K+ is released. Whether the rate of release is adequate to meet crop demands depends much on the kind of soil and the intensity of cropping. Young, unweathered soils are mainly rich in K-bearing minerals and may release high quantities of K+. This is borne out by field experiments by Singh and Brar (1977) on young alluvial soils in the Punjab (India). Even though exchangeable K+ values in the soils were low, yields of corn and wheat were high, and responses to fertilizer K+ were not obtained. In the long term, however, it is probable that intensive cropping even on such young soils would result in a major depletion of available soil K+; so fertilizer K+ would be required for optimum grain yields. This is especially the case when the cropping system shifts to a more intensive form and higher amounts of K+ are exported from the field (Kemmler, 1972). Thus Prasad (1977) has reported that numerous field trials carried out in India were highly responsive to N, P, and K fertilizer applications. Potassium responses were particularly high under the conditions of rainfed crops. Stephens (1969) also found with various crops grown in East Africa that during the first two-year cycle of cropping, the response to K+ fertilizer supply was poor or even negative, whereas in the second two-year cycle marked yield increases were obtained. Long-term field experiments with coastal Bermuda grass conducted by Woodhouse (1968) also showed no major K responses in the first five years. After this period, however, spectacular yield increases were obtained in the K-treated plots, and over the eleven years of the experiment K fertilizer application raised yield considerably. Similar results have been reported by Heathcote (1972) in Nigeria and by Anderson (1973) in East Africa. Anderson (1973) found striking K + responses most commonly on very sandy or moderately to strongly acid soils, but also on soils highly saturated with calcium. Crops that especially responded to K fertilizer applications were tea, tobacco, bananas, potatoes, sweet potatoes, coconuts, and grassland cut for silage or hay. DeDatta and Gomez (1975) reported that intensive paddy cropping on vertisols led to a decrease in yield level if K was not applied. The effect of N fertilizer application was poor on these K-deficient soils and could be substantially increased by K+ dressings. These positive effects of K+ were obtained on two sites ”

+

94

KONRAD MENGEL AND ERNEST A. KIRKBY

with a value for exchangeable K+ <200 ppm, whereas on the third site with an exchangeable K+ value of about 350 ppm K, fertilizer application did not result in yield increases. The differences in response between the three sites were attributed to imgation water that contained 5 ppm in the third nonresponsive site, but only 1 ppm in the two responsive sites. The authors noted that K+ responses were higher for rice grown in the dry season than for rice grown in the rainy season. That differences in response should occur between seasons is not completely understood. Generally higher rice grain yields are obtained in the dry season because of better light conditions (Tanaka, 1973). It is feasible that the higher response to K+ under such circumstances relates to the increase in potential yield attributable to the higher light intensity. Recent field experiments on 23 sites in Java in most cases showed considerable rice yield increase as a consequence of K+ fertilizer application. Best results were obtained when K+ was applied during the tillering stage or given in split applications. The most striking result of these field experiments, depicted in Fig. 4, was that yields in the rainy and dry seasons converged with increasing K+ application rates. Each point in Fig. 4 represents the yield average of the 23 sites (Ismunadji et al., 1977). From this pattern of yield trend, it would appear that higher K+ application rates can compensate for the yield-limiting effect of the rainy season. This particular K+ effect during the rainy season is not always observed as has been shown by very recent experimental results of Ismunadji and Partohardjono (1979). In field experiments carried out on a broad range of different soil types in Indonesia, K+ application resulted in remarkable grain yield increases. In con-

Grain yield

( kglha

30 K

60

x lo-')

90 120 150 180

Treatment ( k g K&/ ha)

FIG.4. Effect of K fertilizer application on rice yield during the rainy season and the dry season (Ismunadji er al., 1977).

POTASSIUM IN CROP PRODUCTION

95

trast to the curves shown in Fig. 4, however, the K+ response curves diverged rather than converged. Response to K+ fertilizer application is especially likely to occur when high-yielding rice varieties are grown and when N- and P-containing fertilizer is applied. According to Tanaka (1973) crops grown on sandy latosols and in calcareous habitats are particularly susceptible to K+ deficiency. On these soils, responses to K+ application are likely to occur. Whether a crop responds to potassium is also dependent on other growth factors, and particularly on nitrogen and water supply. Highest K+ responses are likely if these two factors are not growth limiting. This relationship has been clearly demonstrated by Gartner (1969) in an experiment with tropical grasses. In the low N treatment (1 12 kg N/ha), K application had hardly any effect on dry matter yield, whereas in the highest rate treatment (448 kg N/ha) a more than 40% yield increase was attributable to K+. Similar relationships between N and K supply have been reported for grain yields of barley by Macleod (1969), for grain yield of oats by Mengel and Helal ( 1 968), and for grain yield of wheat by Helal and Mengel (1968). In a recent paper, George et al. (1979) reported that K+ application only resulted in a significant yield increase of smooth grass (Bromus inermis) if N was abundantly supplied. The results of this investigation also indicate that the highest K response was obtained in a year with almost no deficit in precipitation during the growth period. The effect of soil moisture on K+ response is often difficult to interpret. When water supply directly limits growth, high rates of K+ application are without effect. On the other hand, when low soil moisture conditions are limiting K+ availability, the application of K+ may result in a yield response, particularly if it encourages the root system to take up the water in deeper soil layers. As already discussed in Section 11, K+ diffusion and the release of nonexchangeable K+ depend greatly on soil moisture. Optimum soil moisture conditions favor K + diffusion in soil medium and are thus beneficial to the supply of K+ to plant roots (Mengel and v. Braunschweig, 1972). Under such conditions indigenous soil K+ may suffice for optimum plant growth, and so fertilizer K+ dressings may be without response. However, under drier conditions on the same soil with the same crop, K+ fertilizer responses may be obtained. This is the main reason why K+ fertilizer responses may differ considerably from one year to the next for the same crop growing on the same soil. The importance of soil moisture conditions in regulating K+ fertilizer response of potatoes and wheat has been demonstrated in long-term field trials (19351949) by van der Paauw (1958). Yields were found to be highly dependent on the total rainfall during the period from May to July. Potassium responses were higher the lower the rainfall during this period. The relationship between rainfall and the effect of a K+ application on grassland is also documented in long-term field experiments of Kuntze and Bartels (1975). These workers found that in dry years highest yields were obtained for a soil with a K+ level of 21 mg lactate-

96

KONRAD MENGEL AND ERNEST A. KIRKBY

soluble K+/100 ml soil. In rainy years the corresponding value was only 13 mg lactate-soluble K+/100ml soil. High K+ responses are especially likely to occur if the top soil layer is rather dry during the early development of a crop. At this stage of growth the K+ demand per unit root length is particularly high (Mengel and Barber, 1974) and thus K+ diffusion toward plant roots may well become a growth-limitingprocess. For most crops K+ nutrition in the early growth stage is decisive in determining the yield level harvested later, so K+ availability conditions during this period of plant growth merit special attention. The release of interlayer K+ much depends on soil moisture. High soil moisture conditions favor a net release, and this may play an important part in supplying plants with K+. If the soil medium is relatively dry, however, K+ release is much restricted. Under such dry conditions, fertilizer K+ may give rise to significant responses that would very infrequently be obtained under more moist conditions. This relationship between K+ release of interlayer K+, soil moisture conditions, and K+ response has been considered by Mengel and Wiechens (1979) in pot experiments with ryegrass. Whether a K+ response is obtained or not also depends on the crop species and even sometimes on different cultivars of the same species. Why such interspecies differences exist is not yet completely understood. Probably several factors are involved. One important factor especially during the vegetative period is the growth rate. If the growth rate is high, so too is K+ demand. For a given soil with given environmental conditions a response to fertilizer K+ is more likely to occur with a fast- than a slow-growing crop. This is probably the reason that corn generally responds better than small grain cereals to K+ fertilization (Burkart, 1975). Another factor about which not much is known is rooting pattern and root metabolism in relation to K nutrition. In this respect, differences between grasses and dicotyledons are of particular interest. Blaser and Brady (1950) found that K+ application preferentially favored the growth of leguminous species in a grass-legume sward and that potassium depletion of the soil resulted in a considerable decrease of the proportion of legumes in the sward. This observation that K+ fertilization especially favors the growth of legumes in grassland was noted over 40 years ago by Konig (1935) and has since been confirmed by various authors (Kemmler et al., 1977; Schmitt and Brauer, 1979). Steffens and Mengel (1979) reported that when ryegrass and clover were grown together, the K+ content of the ryegrass was higher and that of the red clover lower than the corresponding values when the two species were grown separately. These authors suggest that the better response of red clover to ryegrass to K application reflects the successful competition of ryegrass over clover for soil K+. Field experiments also support the results discussed above. Thus van der Paauw (1958) reported that potatoes responded much better to K+ fertilizer dressings than wheat did.

POTASSIUM IN CROP PRODUCTION

97

Similarly Schon et al. (1976) in a 20-year field experiment carried out on a loess soil also confirmed these interspecies differences in response to K fertilizer. Cereal responses were rather low, but with broad beans and potatoes spectacular yield increases were obtained. The mechanism of these differences in response between species is not yet understood. The large variation in response of crops to K fertilization has been used by Greenwood et al. (1974) to formulate fertilizer recommendations for vegetable crops. These workers argue that the large numbers of vegetable crops grown on widely different soils make it impossible to carry out trials to cover the possible combinations of crops and soil. They have therefore proposed a shortcut developed from the concept that a crop which is more responsive than another to fertilizer on one particular site should be more responsive on other sites. Using this approach Greenwood et al. (1974) have made measurements of response curves of yield against level of fertilizer for many crops at one site, and the response curves of one of the crops on a range of sites. From this information the response curves for all crops on all sites are predicted and then tested against the results of independent experiments. This approach may also prove useful for agricultural crops. From the above discussion it is clear that it is not possible to provide general K fertilizer recommendations that are applicable to different soils, climates, and crops. The basis for recommendations should be made using K soil analysis and determining critical levels by K fertilizer field experiments for representative soils and crops. Of the analytical methods used at the present time to formulate K+ fertilizer requirements much stress is placed on exchangeable K+. This is not always satisfactory. Future soil analysis should include two main factors that control K availability, namely the K buffer capacity of the soil and the K concentration of the soil solution, both measured under standard conditions. The optimum soil K level can differ for various crops; for example, it will be higher for potatoes or sugar beets than for cereals. For this reason the crop species in a rotation with the highest K requirement needs particular attention in K fertilizer practice. If the level of available soil K+ is much below the optimum, higher amounts of K fertilizer must be applied than those removed by the crop. Under K+-fixing soil conditions these K application rates may be extremely high, as has been outlined in Section IV, A. If the soil K+ level is higher than optimum, lower K fertilizer rates should be applied. In some cases even the omission of K fertilization may be opportune. Soils with an optimum K+ level should receive amounts of K+ that maintain this K+ status. These quantities can be calculated from the removal of K+ by the harvested produce, provided that losses due to K+ leaching are negligible. Potassium fertilizer policy also depends much on soil texture and the types of clay minerals present in the soil. Medium to highly textured soils have a medium to high K buffer capacity, and are not prone to K leaching. For such soils K+ can

98

KONRAD MENGEL AND ERNEST A . KIRKBY

be applied in autumn, winter, or before the rainy season without danger of K losses by leaching. Two or even three crops of a rotation can be supplied by only one K treatment. The amount of K+ given in the application, however, should meet the total K requirement of the crops in question. This technique of K fertilizer application has been tested by Ansorge (1967) in numerous field experiments on medium-textured soils in Germany. No yield depressions were observed when K+ was applied on a two-year cycle as compared with K treatments applied each year. On poorly buffered soils, especially on sandy soils under humid conditions, a K fertilizer policy must take into account the possibility of loss through leaching. These soils should be treated with fertilizer K+ just before the crop is sown or planted. In this context it is also worth pointing out that the K+ in crop residues (straw, leaves, and roots) can also be leached by winter rainfall or under monsoon conditions. B. EFFECTOF POTASSIUM ON YIELDCOMFQNENTS

It is well known that K+ is mainly taken up during the vegetative period of plant growth. Pitman (1972) has demonstrated that in barley plants the rate of uptake is directly related to the growth rate. As already indicated, there is evidence that the K+ uptake is dependent on the hormonal status of plants (Cram and Pitman, 1972). Abscisic acid is known to inhibit K+ uptake, whereas indole acetic acid promotes the uptake of K+ (Erdei et al., 1979). Inadequate K+ supply retards the vegetative development of the plant, which may not only affect the production of vegetative plant material but also the development of reproductive organs and the filling of storage tissues with photosynthates. Mengel and Forster (1968) in studying the effect of interrupting the K+ supply on the development of spring barley found that grain yields were more depressed the earlier the K+ interruption took place. Thus a 16-day interruption between the tillering and the stem elongation stages resulted in a grain yield depression of about 40%. This yield depression was brought about mainly by a reduction in the number of ears per plant and a lower single grain weight. The interruption in K+ supply during the stage of ear emergence also depressed grain yield. In this case, however, depression resulted largely from a decrease in single grain weight. In a further treatment of this solution culture experiment in which plants were grown without K+ from pollination until maturity, no significant grain yield reduction was obtained. Analogous results have been reported by Forster (1973b) for spring wheat and oats and by Chapman and Keay (1971) for wheat. These results obtained in solution culture experiments are consistent with observations in the field. Thus Ralph (1976) found in field experiments on clay soils in England that the grain yield of winter wheat was increased as a consequence of K+ application by improving the single grain weight. In some cases

99

POTASSIUM IN CROP PRODUCTION

this treatment also increased the number of grains per ear. Ralph (1976) observed that K+ especially promoted the development of the proximal, central, and distal spikelets. Baier and Smetankova (1974) who carried out a large number of field experiments also found that K+ mainly increased the grain yield of cereals by improving the single grain weight. Single grain weight depends much on the K+ status of the plant at the flowering stage. Late K+ application has little effect on the grain development (Mengel and Forster, 1971), and the rate of K+ uptake by cereals after flowering is probably very low. However, the K+ status of leaves and culms at the grain filling period has a substantial impact on photosynthesis and on the translocation of photosynthates from these organs toward the ears (Koch and Mengel, 1977). Ralph (1976) and Forster (1976) made the observation that under optimum K nutritional conditions the senescence of the flag leaf is delayed. This results in a prolonged “leaf area duration” which according to Evans et al. (1975) is important for grain development. Forster’s (1976) main results, shown in Table 111, reveal that K+ had a beneficial effect on increasing leaf area, chlorophyll content, and succulence of the flag leaf. This significance of succulence is not yet understood. It has been suggested that it may have a beneficial effect on phloem loading and probably also on the mobilization of photosynthates deposited in the leaves prior to the grain filling period. This view is supported by experiments of Seqer (1978) who observed that shortly after pollination a substantial amount of nitrogen was still stored in the culms and leaves of wheat in the form of protein. These proteins were mobilized at the stage of highest grain growth and used for the synthesis of grain proteins. Plants with a high K+ status were found to be more efficient in mobilizing the stored leaf proteins and in translocating the resulting amino acids toward the grains. From Seqer’s results it is also clear that the beneficial effect of K+ on grain filling was not related to the K+ content of the grains but resulted excluTable I11

Effect of K + Supply on Grain Yield, Grain Yield Components, Chlorophyll Content, Flag Leaf Area, and Succulence of the Flag Leaves”

Grain yield, g/plant Single grain weight, mg/grain Number of graindear Number of eardplant Flag leaf area, cm2/leaf Chlorophyll content, mg/cm2 leaf area Succulence, mg HZO/lOOcm’ leaf area

1.86 25.1 24.6 3.05 30.4 1.4 349

2.85 32.1 28.1 3.17 41.6 8.3 380

“The figures represent an average of five cultivars of spring wheat (Forster, 1976).

3.48 34.1 31.8 3.22 50.2 34.2 442

100

KONRAD MENGEL AND ERNEST A. KIRKBY

sively from the influence of K+ on the translocation of assimilates from the vegetative plant parts toward the ears. Grain growth is not only dependent on the provision of a source of assimilates to the ears but also on the ability of the ears to provide a “sink” for these assimilates. Shading, for example, has a depressive effect on grain development, which cannot be alleviated by increasing the level of K nutrition (Mengel and Haeder, 1976). On the other hand, the depressing effect of shading on vegetative growth can be reduced to some extent by a high K+ supply (Haeder and Mengel, 1975). This demonstrates that K nutrition can beneficially influence the “source” but not the sink. If the sink metabolism is the limiting factor in grain development, then increasing K supply is without effect. This has been shown by Mengel and Haeder ( 1976) with spring wheat and by Beringer and Koch (1977) in experiments with the high-lysine barley variety “Riso 1508.” It is very likely that the production of seeds, tubers, and roots is influenced by K+ in the same way as has been shown for cereals. The beneficial effect of K+ is to promote phloem loading and phloem transport and thus provide the “physiological sink” with assimilate. Thus K nutrition enhances tuber growth of potatoes (Haeder et al., 1973). Obigbesan (1977) has drawn attention to the observation that K+ application to cassava grown at two different locations in Nigeria particularly increased the storage roothop ratio. An analogous observation for potatoes has been reported by Haeder et al. (1973). These authors also found that the number of potato tubers per plant were highest in the low K+ treatment, but the tubers remained very small. The effect of K+ particularly was to increase tuber size. From experiments of Haeder (1975) it appears that the negative effect of chloride on tuber filling results from impaired translocation of photosynthates from the leaves toward the tubers. This is the main reason that the application of potassium sulfate produces better potato starch yields than treatment with potassium chloride. It is generally accepted that K+ increases the starch content of tubers and the sugar content of roots of sugar beet. This positive effect, however, is not always observed and depends much on the crop yield and the degree of K” deficiency in the soil. In perennial crops a substantial amount of K+ is retranslocated from leaves into stems and twigs before leaf fall begins. This K+ is mobilized again in the following spring when new leaves are formed. For these crops the early development also depends considerably on K+ nutrition. Experiments of Fremond and Ouvrier (1971) on sandy soils in the Ivory Coast have shown that young coconut palm trees which received a K+ application of about 1 kg Wtree grew more vigorously than the control plants without K fertilizer treatment. The K+treated plants fruited earlier and produced much higher yields than the untreated palms. A late application of KS to these untreated palms did result in a yield increase, but the yields obtained were not as high as those obtained when K+ was applied at an early stage. This example demonstrates the importance of K+ in

POTASSIUM IN CROP PRODUCTION

101

establishing high-yield crop stands of plantation crops. According to investigations of Ollagnier and Ochs (1971) and v. Uexkull(l972) oil and coconut palms have a special chloride requirement, and KCI applications may increase yield not only as a consequence of K+ but also because of the beneficial effect of chloride. C . SECONDARY EFFECTS OF POTASSIUM ON CROPYIELD

A number of more indirect effects of K+ occur which are pertinent in discussing crop yields. These secondary effects cannot be considered here in detail, but some major points are presented. Much research work has been done which shows that under conditions of an inadequate K+ nutrition, the susceptibility of crops to plant disease is increased. This topic has been thoroughly treated by Goss (1968), and the 12th Colloquium of the International Potash Institute, Berne, was particularly devoted to this subject. Ismunadji (1976) in discussing rice diseases and physiological disorders draws attention to the observation that K+ improves the resistance of rice to fungal diseases. Analogous observations have been reported for other crops (Kriiger, 1976). Why optimum K+ status especially improves fungal disease resistance is not yet completely understood. Trolldenier and Zehler (1976) have considered the relationship between plant nutrition and rice diseases. These authors suggest that under the conditions of insufficient K+ supply the formation of the cuticle and epidermal cell walls is affected so that fungal hyphae may more easily penetrate these barriers than through the wellestablished cuticles and epidermal cell walls of plants adequately supplied with K+ . The higher contents of sugars and soluble amino acids, frequently found in K+-deficient plant tissues, are also considered to provide a good nutrient medium for fungi and bacteria. The findings of Baule (1969) also indicate that forest trees well supplied with K+ are more resistant to sucking insect infestation than are trees that are deficient in K+. In paddy rice growing, Fe toxicity associated with high Fez+concentrations in the root zone is often associated with K+ deficiency. Trolldenier (1977) has reported that under conditions of suboptimal K+ nutrition the redox potential in the rice rhizosphere decreases and the relative proportion of Fez+ to Fe3+ increases. The resulting Fe toxicity can be alleviated or even completely eradicated by K+ application. It appears that in plants well supplied with K+ that 0, transport is enhanced from the plant tops via the stems to the roots, thus enabling the oxidation of Fez+ to Fe3+ in the rhizosphere. The partial precipitation of Fe3+ at the root surface gives rise to a red color which is indicative of healthy rice roots. The disease “greenback” is a physiological disorder of tomato fruits which is highly dependent on K nutrition as well as on tomato cultivar (Forster, 1973a). The results of Forster and Venter (1975) have shown that the proportion of

102

KONRAD MENGEL AND ERNEST A. KIRKBY

tomato fruits with greenback can be considerably reduced at higher levels of K+ nutrition. On the other hand, if too high a level of K nutrition is supplied, this can bring about the occurrence of “blossom end rot” in tomatoes and “bitter pit” in apples. These Ca-related disorders can be induced by the effects of K-impairing calcium uptake and translocation. Vertregt (1968) reported a close relationship between “black spot” in potato tubers and their K content. Black spot susceptibility disappeared if the K content exceeded 600 me Wkg dry matter. According to Macklon and DeKock (1967) the K content of tubers is positvely correlated with the citric acid content of tubers. Citric acid is known to inhibit the formation of an Fe chlorogenic complex, which induces black spot formation. This relationship may therefore account for K+-black spot association. V.Braunschweig (1979b) also suggests that high turgor in potato tubers prevents the disease. Since K+ is a major osmoticum in potato tubers, the beneficial effect of K+ may also be explained in terms of increased turgor. There are several indications in the literature that a high K content in plant tissues increases frost resistance. This relationship has been demonstrated recently by Eifert and Eifert (1976) in grapes. These workers found an inverse relationship between the K content of vine leaves and frost damage. A beneficial effect in this respect of K fertilizer application to forest trees has also been found by Koskela (1970). The mechanism of this direct effect of K+ on frost resistance is not yet completely understood. It is supposed that K+ not only has an influence as an osmoticum but also improves resistance by affecting other biochemical reactions. Salt tolerance can also be enhanced by K+. In solution culture experiments with barley, Helal et al. (1975) reported that negative effect of NaCl salinzation (60 mM NaCI) on growth could be completely suppressed by the addition of KCI (5 or 10 mM) to the nutrient medium. In further experiments Helal and Mengel (1979) observed that Na salinity impaired N turnover and that this effect could be alleviated by the presence of K+ in the medium. Unpublished data of Held indicate that salt tolerance is related to the energy status of plants and that plants grown under saline conditions consume more energy than plants that are not suffering from salt stress. Helal suggests that as a result of this improvement of K+ on the energy status of the plants, the resistance to salinity is increased. It is well known that Na+ can substitute for K+ to some degree in plant nutrition. In this respect considerable differences occur between crop species. This question has been thoroughly treated by Marschner ( I 97 1) and for this reason will not be considered here in detail. In some species, however, Na+ has an additional influence on growth and crop production. This has been shown for sugar beets by Draycott et al. (1970), El-Sheikh and Ulrich (1970), and Jude1 and Kuhn (1975). The K efficiency of crops has been a subject of recent interest. An “efficiency ratio” has been measured relating the quantity of plant material produced to the

POTASSIUM IN CROP PRODUCTION

103

amount of K+ taken up. Gerloff (1976) reported significant intercultivar differences for this efficiency ratio in Phaseolus vulgaris and tomato. High efficiency was frequently related to a high uptake of Na+, presumably because Na+ was to some extent replacing the nonspecific physiological role of K+ in plant metabolism.

V. CONCLUSIONS Potassium is characterized by unique behavior in soils and living systems and contrasts greatly with related cation species such as Na+ or Ca2+. It is suggested that this unique property of K+ is dependent on the relatively low energy required for the dehydration of K+ and on the high affinity of the dehydration K+ to oxygen atoms. Interlayer K+ in micas or mica-related minerals is embedded at the center of a structure formed by six peripherally arranged oxygen atoms. Analogous structures are found in living systems such as enniatins, nonactin, and valinomycin. It is feasible that these K+ complexes are essential for specific K+ functions in living systems. It appears that typical K+ reactions in the soil as well as those in the plant basically originate from the same property of K+, namely its tendency to substitute the hydration water by other oxygen-containing ligands. This property thus controls K+ fixation and release by clay minerals as well as the K+ buffer capacity of soils, on one hand, and the selective K+ transport through biological membranes and all processes related to this transport, on the other. Both the characteristic behavior of K+ in soils as well as in plants are of agronomic relevance. The understanding of the K relationships and interactions in soils is pertinent of estimating optimum fertilizer application rates. The importance of K fertilizer application will increase the more plant cultivation and production shifts from an extensive to an intensive form. The question of the economics of K fertilization thus requires increasing attention especially in developing countries. The understanding of the physiological functions of K+ in crops is relevant to the production of plant products of high quality. In developing new cultivars or even new crops, knowledge of the physiological role of K+ in plants will also be useful. The fact that K+ is especially involved in the conversion of solar energy into chemical energy could be of some importance in developing crops that are solely grown for energy production. REFERENCES Addiscott, T. M. 1974a. J . Sci. Food Agric. 25, 1173-1 183. Addiscott, T. M. 1974b. I n “Potassium Research and Agricultural Production,” pp. 175-190. Proc. 10th Congr. Int. Potash Inst., Berne.

104

KONRAD MENGEL AND ERNEST A. KIRKBY

Adepetu, J. A., and Akapa, L. K. 1977. Agron. J . 69, 940-943. Ahmad, N., and Davis, C. E. 1971. SoilSci. 112, 100-106. Allaway, W. G. 1973. PIanra 110, 63-70. Amberger, A., Gutser, R., and Teicher, K. 1974. Planr Soil 40, 269-284. Amir, S., and Reinhold, L. 1971. Physiol. Plant. 24, 226-231. Anderson, G. D. 1973. In “Potassium in Tropical Crops and Soils,” pp. 413-437. Proc. 10th Colloq. Int. Potash Inst., Berne. Ansorge, H. 1967. Drsch. Akad. Landwirtschafiswiss. (Berlin) Rep. No. 76. Arifin, H. F., and Tan, K. H. 1973. Soil Sci. 116, 31-35. Arneke, W. 1980. Ph.D. Thesis, FB 19, Justus Liebig University, Giessen. Arnon, I. 1969. In “Transition from Extensive to Intensive Agriculture with Fertilizers,” pp. 13-25. Proc. 7th Colloq. Int. Potash Inst., Beme. Ashley, D. A . , and Goodson, R. D. 1972. Crop Sci. 12, 686-690. Baier, J . , and Smetankova, M. 1974. In “Potassium Research and Agricultural Production,’’ pp. 161-170. Proc. 10th Congr. Int. Potash Inst., Beme. Baker, D. A,, and Weatherley, P. E. 1969. J. Exp. Bor. 20, 485-596. Baldwin, J. P., Nye, P. H., and Tinker, P. B. 1973. PIanr Soil 38, 621-635. Baligar. V. C., and Barber, S. A. 1978a. Soil Sci. Soc. Am. J . 42, 575-579. Baligar, V. C., and Barber, S. A. 1978b. Soil Sci. SOC.Am. J . 42, 618-622. Barankiewicz, T. J. 1978. Z . Pflunzenphysiol. 89, 11-20. Barber, J . 1977. In “Fertilizer Use and Production of Carbohydrates and Lipids,” pp. 83-93. Proc. 13th Colloq. Int. Potash Inst., Berne. Barber, S. A. 1962. SoilSci. 93, 39-49. Barber, S . A. 1979. In “The Soil-Root Interface” (J. L. Harley and R. Scott Russell, eds.), pp. 5-20. Academic Press, New York. Barber, S . A., Walker, J. M., and Vasey, E. H. 1963. Agric. Food Chem. 11, 204-207. Barrow, N. J. 1966. Aust. J . Agric. Res. 17, 849-861. Baule, H. 1969. Landw. Forsch. 23(I), 92-104. Ben Zioni, A,, Vaadia, Y., and Lips, S . H. 1971. Physiol. Planr. 24, 288-290. Beringer, H . , and Koch, K. 1977. Landw. Forsch., Sonderh. 34(II), 36-44. Besford, R. T., and Maw, G. A. 1976. Ann. Bor. 40, 461-471. Blanchet, R., Studer, R., and Chaumont, C. 1962. Ann. Agron. 13, 93-1 10. Blaser, R. E., and Brady, N. C. 1950. Agron. J . 42, 128-135. Blevins, D. G., Hiatt, A. J., and Lowe, R. H. 1974. PIanr Physiol. 54, 82-87. Blevins. D. G., Barnett, N. M., and Frost, W. B. 1978. Plant Physiol. 62, 784-788. Boguslawski, E. v., and Lach, G. 1971. Z. Acker Pflanzenbau 134, 135-164. Bowling, D. J. F. 1976. “Uptake of Ions by Plant Roots.” Chapman & Hall, London. Bowling, D. J. F., and Ansari, A. Q. 1971. PIanra 98, 323-329. Bowling, D. J. F., Macklon, A. E. S., and Spanswick, R. M. 1966. J . Exp. Bor. 17, 410-416. Brad, J. 1971. Biochemistry 14, 127-134. Brag, H. 1972. Physiol. Planr. 26, 250-257. Braunschweig, L. C., v. 1979a. Landw. Forsch. Sonderh. 35, 219-231. Braunschweig, L. C., v. 1979b. Karroffelbau, Hefi Jan. Braunschweig, L. C., v., and Mengel, K. 1971. Landw. Forsch. Sonderh. 26(I), 65-72. Burkart, R. 1975. Ph.D. Thesis, FB Agriculture and Horticulture, Technische Universitat, Miinchen. Busch, R. 1980. Ph.D. Thesis, FB 19, Justus Liebig Universitat, Giessen. Chapman, M. A., and Keay, J. 1971. Ausr. J . Exp. Agric. Anim. Hush 11, 223-228. Cheeseman, J. M., and Hanson, J. B. 1979. PIanr Physiol. 64, 842-845. Chloupek, 0. 1972. 2. Acker Pflanzenbau 136, 164-169. Claassen, N . , and Barber, S . A. 1976. Agron. J . 68, 961-964.

POTASSIUM IN CROP PRODUCTION

105

Collins, J . C., and Kenigan, A. P. 1974. New Phytol. 73, 309-314. Cram, W. J . 1976. In “Transport and Transfer Processes in Plants” (I. F. Wardlaw and J. B. Passioura, eds.), pp. 113-124. Academic Press, New York. Cram, W. J., and Pitman, M. G. 1972. Aust. J. Biol. Sci. 6, 1125-1 132. Dainty, J . 1962. Annu. Rev. Plant Physiol. 13, 379-402. Davies, D. D. 1973. In “Biosynthesis and its Control in Plants” (E. V. Millborrow, ed.), pp. 1-20. Academic Press, New York. Davis, R. F., and Higinbotham, H. 1976. Plant Physiol. 57, 129-136. De Datta, S. K., and Gomez, K. A. 1975. Soil Sci. 120, 361-366. DeJaegere, R., and Neirinckx, L. 1978. Z. Pflan7enphysiol. 89, 129-140. De Lopez, J . , and Jenny, H. 1959. J. C o h i d Sci. 14, 533-542. Dijkshoorn, W. 1962. Nature (London) 194, 165-167. Dijkshoom, W., Sujitno, J. S. A., and Ismunadji, M. 1974. Plant Soil 40, 525-534. Dobler, M., Dunitz, J . D., and Krajewski, J. 1969. J. Mol. Biol. 42, 603-606. Doll, E. G., and Lucas, R. E. 1973. In “Soil Testing and Plant Analysis” (L.M. Walsh and J. D. Beaton, eds.), pp. 133-151. Soil Sci. SOC.America, Madison, Wisconsin. Doman, D. C., and Geiger, D. R. 1979. Plant Physiol. 64, 528-553. Draycott, A. P., Marsh, J . A. P., and Tinker, P. B. H. 1970. J. Agric. Sci. 74, 567-573. Drew, M. C.. and Nye, P. H. 1969. Plant Soil 31, 407-424. Drew, M. C . , Nye, P. H., and Vaidyanathan, L. V. 1969. Plant Soil 30, 252-270. Drews, J . U. 1978. Ph.D. Thesis, FB 19, Justus Liebig Universitat. Giessen. During, C., and Duganzich, D. M. 1979. Plant Soil 51, 167-176. Ehlers, W., Gebhardt, H., and Meyer, B. 1967. Z. Pflanzenernahr. Bodenk. 117, 29-52. Ehlers, W., Gebhardt, H., and Meyer, B. 1968. Z. Pflanzenernahr. Bodenk. 119, 173-186. Eifert, A., and Eifert, J . 1976. Potash Rev. ( B e m e ) Subj. 29, 8. Suite 10. El-Scheikh, A. M.. and Ulrich, A. 1970. Plant Physiol. 46, 645-649. Epstein, E. 1966. Nature (London) 212, 1324-1327. Erdei, L., Toth, I . , and Zsoldos, F. 1979. Physiol. Plant. 45, 448-452. Estes, G. 0.. Koch, D. W., and Bruetsch, T. F. 1973. Agron. J. 65, 972-975. Etherton, B. 1963. Plant Physiol. 38, 581-585. Evans, H. J . , and Sorger, G. J. 1966. Annu. Rev. Plant Physiol. 17, 47-77. Evans, H. J . , and Wildes, R. A. 1971. In “Potassium in Biochemistry and Physiology,”pp. 13-39. Proc. 8th Colloq. Int. Potash Inst., Berne. Evans, L. T . , Wardlaw, I. F., and Fischer, R. A. 1975. In “Crop Physiology” (L.T. Evans, ed.), pp. 101-149. Cambridge Univ. Press, London and New York. Farmer, V. C., and Wilson, M. J. 1970. Nature (London) 226, 841-842. Fan, E., Vaidyanathan, L. V., and Nye, P. H. 1969. Soil Sci. 107, 385-391. Feigenbaum, S., and Mengel, K. 1979. Physiol. Plant. 45, 245-249. Fischer, R. A., and Hsiao, T. C. 1968. Plant Physiol. 43, 1953-1958. Fisher, J . D., Hansen, D., and Hodges, T. K. 1970. Plant Physiol. 46, 812-814. Forster. H. 1973a. Tech. Commun. Int. Sor. Hortic. Sci. 29, 319-325. Forster, H. 1973b. Landw. Forsch. 26, 221-227. Forster, H. 1976. Z. Acker Pflanzenbau 143, 169-178. Forster, H., and Mengel, K. 1969. Z. Acker Pflanzenbau 130, 203-213. Forster, H., and Venter, F. 1975. Gartenbauwiss. 40, 75-78. Fremond, Y., and Ouvrier, M. 1971. Oleagineux 10, 609-616. Gartner, J . A. 1969. Queensl. J. Agric. Anim. Sci. 26, 365-372. George, J . R . , Pinheiro, M. E., and Bailey, T. B. 1979. Agron. J. 71, 586-591. Gerloff, G. C. 1976. In “Plant Adaption to Mineral Stress in Problem Soils” (M. J. Wright and S. A. Ferrari, eds.), pp. 161-173. Beltsville, Maryland.

106

KONRAD MENGEL AND ERNEST A. KIRKBY

Giaquinta, R. T. 1977. Narure (London) 267, 369-370. Giaquinta. R. T. 1979. Plant Physiol. 63, 744-748. Goss, R. L. 1968. In “The Role of Potassium in Agriculture” (V.J. Kilmer, S. E. Younts, and N. G. Brady, eds.). pp. 221-241. Am. Soc. Agron., Madison, Wisconsin. Graham, E. R.. and Fox, R. L. 1971. SoilSci. 111, 318-322. Graham-Bryce, I. J. 1963. J . Agric. Sci. 60,239-244. Green, J. F., and Muir, R. M. 1978. Physiol. Plant. 43, 213-218. Greenwood, D.J.. Cleaver, T. J.. and Turner, M. K. 1974. Proc. Ferr. Soc. (London) 145, 4-30.

Grimme, H., Nemeth, K.. and Braunschweig, L. C., v. 1971. Landw. Forsch. Sonderh. 26(I), 165-1 76.

Haeder, H. E. 1975. Landw. Forsch. Sonderh. 32(I), 122-131. Haeder, H. E., and Mengel, K. 1975. Z. Pflanzenerniihr. Bodenk. 6, 573-582. Haeder, H. E., Mengel, K., and Forster, H. 1973. J. Sci. Food Agric. 24, 1479-1487. Halevy, J. 1977. Porash Rev. (Berne) Subj. 27, 79. Suite 5 , 1-13. Hall, S . M.. and Baker, D. A. 1972. Planra 106, 131-140. Hartt, C. E. 1970. Planr Physiol. 46, 419-422. Hartt, C. E. 1972. Plant Physiol. 49, 569-571. Hawker, J. S., Marschner, H., and Krauss, A. 1979. Physiol. Planr. 46, 25-30. Heathcote, R. C. 1972. Porash Rev. (Berne) Subj. 16, 57. Suite. Helal, H. M., and Mengel, K. 1968. Z. Pjanzenerniihr. Bodenk. 120, 89-98. Helal, H. M., and Mengel, K . 1979. Planr Soil 51, 457-462. Helal, H. M., Koch, K., and Mengel, K. 1975. Physiol. Plant. 35, 310-313. Hiatt, A. J. 1966. Z. Pjanzenphysiol. 56, 233-245. Higinbotham, N. 1973. Bor. Rev. 39, 15-69. Higinbotham, N., Etherton, B., and Foster, R. J. 1967. PIanr Physiol. 42, 37-46. Hinkle, P. C., and McCarty, R. E. 1978. Sci, Am. 238, 104-123. Hodges, T. K . 1973. Adv. Agron. 25, 163-207. Hodges, T. K., Leonard, R. T., Bracker, C. E., and Keenan, T . W. 1972. Proc. Narl. Acad. Sci. U . S . A . 69, 3307-331 1 . Hoffmann, W. E. 1966. Z. Pjlanzenernahr., Dung. Bodenk. 113, 112-20. Home, E. 1762. Quoted by E. J. Hewitt and T. A. Smith 1975. In “Plant Mineral Nutrition,” p. 15. English Univ. Press, London. Hopkins, H. T. 1956. Plant Physiol. 31, 155-161. Horton, R. F., and Moran, L. 1972. Z. Pflanzenphysiol. 66, 193-196. Hsiao, T. C., Hageman, R. H., and Tyner, E. H. 1970. Crop Sci. 10, 78-82. Humble, G. D., and Hsiao, T. C. 1969. Plant Physiol. 44, 230-234. Humble, G . D., and Raschke, K. 1971. Plant Physiol. 48, 447-453. Hutchings, M. 1978. Planra 138, 237-241. Ilyashouk, E. M., and Okanenko, A. S. 1970. Fiziol. Rasr. 17, 445-451. Ismunadji, M .1976. In “Fertilizer Use and Plant Health,”pp. 47-60. Proc. 12th Colloq. Int. Potash Inst., Berne. Ismunadji, M., and Partohardjono, S. 1979. Porash Rev. (Berne) Subj. 9, 43rd Suite. Ismunadji, M., Partohardjono. S., Makarim, A. K., and Supardi, G. 1977. Workshop Paper Porassium (Tugu)D e c . Israel, D. W., and Jackson, W. A. 1978. In “Mineral Nutrition of Legumes in Tropical and Subtropical Soils”(C. S . Andrew and E. J. Kamparath, eds.), pp. 113-129. CSIRO, Australia. Jackson, B. L. J., and During, C. 1979. P l m r Soil 51, 197-204. Jackson, W. A , , and Volk, R. J. 1968. I n “The Role of Potassium in Agriculture” (V.J. Kilmer, S. E. Younts, and N. C. Brady, eds.), pp. 109-145. Am. Soc. Agron., Madison, Wisconsin. Jacoby, B., Abas, S . , and Steinitz, B. 1973. Physiol. Planr. 28, 209-214.

POTASSIUM IN CROP PRODUCTION

107

Jankovic, M., and Nemeth, K. 1974. In “Potassium Research and Agricultural Production,’’ pp. 75-83. Proc. 10th Congr. Int. Potash Inst., Berne. Jeanniot, A., Dupaigne, G.,and Coic, Y . 1970. Agrochimica 15, 61-73. Jenny, H. 1966. Plant Soil 25, 265-289. Jenny, H., and Overstreet, R. 1938. Proc. Natl. Acad. Sci. V.S.A. 24, 384-392. Jeschke, W. D. 1970. Planru 91, 111-128. Johansen, C., and Loneragan, J. F. 1975. Aust. J . Plant Physiol. 2, 75-83. Johansen, C., Edwards, D. G.,and Loneragan, J. F. 1970. Plant Physiol. 45, 601-603. Judel, G . K., and Kiihn, H. 1975. Zucker 28, 68-71. Karbachsch, M. 1978. 2. Pflanzenerniihr. Bodenk. 141, 513-522. Kemmler, G. 1972. Proc. Fert. World Congr., 7th, Vienna pp. 545-563. Kemmler, G.,Oohara, H.,and Drake, M. 1977. Potash Rev. Subj. 7, 23rd Suite. Kilbourn, B. T . , Dunitz, J. D., Pioda, L. A. R.. and Simon, W. 1967. J . Mol. Biol. 30, 559-563. Kirkby, E. A. 1968. SoilSci. 105, 133-141. Kirkby, E. A. 1974. In “Plant Analysis and Fertilizer Problems” (J. Wehrmann, ed.), Vol. 2, pp. 557-568. Ger. SOC.Plant Nutrition, Hannover. Kirkby, E. A., and Armstrong, M. J. 1980. Plant Physiol. 65, 286-290. Kirkby, E. A., and Knight, A. H. 1977. Plant Physiol. 60, 349-353. Koch, K., and Mengel, K. 1972. J . Sci. Food Agric. 23, 1107-1112. Koch, K., and Mengel, K. 1974. In “Plant Analysis and Fertilizer Problems” (J. Wehrmann, ed.). Vol. I , pp. 209-218. Ger. SOC.Plant Nutrition, Hannover. Koch, K., and Mengel, K. 1977. Agron. J . 6 9 , 477-480. Konig, F. 1935. “Der Einfluss der Kalisalzdiingung auf Wert und Wirkung des Wirtschaftsfutters.” Parey, Berlin. Koskela, V. 1970. Folia Forest. 7 8 , Metsiintutkimuslaitos, Institutum Forestale Fenniae, Helsinki. Kriiger, W. 1976. In “Fertilizer Use and Plant Health,”pp. 145-156. Proc. 12th Colloq. Int. Potash Inst., Berne. Kuntze, H.,and Bartels, R. 1975. Landw. Forsch. Sonderh. 31(1), 208-219. Kurdi, F. H.,and Babcock, K. L. 1970. Plant Soil 33, 413-424. huchli, A. 1972. Annu. Rev. Plant Physiol. 23, 197-218. Lguchli, A.. and Pfliiger, R. 1978. In “Potassium Research-Review and Trends,” pp. 1 1 1-163. Roc.11th Congr. Int. Potash Inst., Berne. Lauchli, A., Spurr, A. R., and Epstein, E. 1971. Plant Physiol. 48, 118-124. Uuchli, A., Kramer. D., Pitman, M. G., and Liittge, U. 1974. Planta 119, 85-99. Lagerwerff, J. V. 1961. Plant Soil 13, 253-263. Latzko, E. 1961. Biochim. Biophys. Acta 4 7 , 350-357. Latzko, E., and Claus, D. 1958. Landw. Forsch. Sonderh. 11, 101-109. Laves, D. 1978. Arch. Acker Pflanzenbau Bodenk. 22(8), 521-528. Leonard, R. T., and Hodges, T. K. 1973. Plant Physiol. 52, 6-12. Lepe, B. G., and Avila, E. J. 1975. Plant Physiol. 56, 460-463. Liebig, J. V. 1841. “Die organische Chemie in ihrer Anwendung auf Agrikultur und Physiologie.” Verl. Viehweg, Braunschweig. Linser, H., and Herwig, K . 1968. Kali Briefe, 9 , Fachgeb. 2 , (2). h u e , A. 1979. Potash Rev. (Berne) No. 4 , Subj. 16 79th Suite. Liittge, U., Bauer, K . . and Kohler, D. 1968. Biochim. Biophys. Acro 150, 452-459. Macklon, A. E. S., and DeKock, P. C. 1967. Physiol. Plant. 20, 421-429. Macleod, L. B. 1969. Agron. J. 61, 26-29. Maextens, M. C. 1971. C.R. Acad. Sci. (Paris)273, 682-684. Malek, F., and Baker, D. A. 1977. Planta 135, 297-299. Malquori, A., Ristori, G.,and Vidmch, V. 1975. Agrochimica 19, 522-529. Mansfield, T. A., and Jones, R. J. 1971. Planra 101, 147-158.

108

KONRAD MENGEL AND ERNEST A. KIRKBY

Marschner, H. 1971. In “Potassium in Biochemistry and Physiology,” pp. 50-63. Proc. 8th Colloq. Int. Potash Inst., Berne. Marschner. H., and Diiring, H. W. 1977. In “Fertilizer Use and Production of Carbohydrates and Lipids,” pp. 101-113. Roc. 13th Colloq. Int. Potash Inst., Berne. Marschner, H., and Possingham, J. V. 1975. Z. Pflanzenphysiol. 75, 6-16. Mengel, D.B., and Barber, S . A. 1974. Agron. J. 66, 399-402. Mengel, K. 1967. Ergebnisse der landwirtschaftlichen Forschungen an der Justus Liebig Universitat, Giessen. Heft IX, pp. 73-82. Mengel, K. 1973. Bull. Rech. Agron. Gemblowr pp. 168-179. Mengel, K. 1974. Landw. Forsch. Sonderh. 31(I) 45-58. Mengel, K., and Aksoy, T. 1971. Z. Pflanzenerniihr. Bodenk. 128, 28-41. Mengel, K., and Braunschweig, L. C., v. 1972. Soil Sci. 134, 142-148. Mengel, K., and Forster, H. 1968. Z. Acker Pflanzenbau 127, 317-326. Mengel, K., and Forster, H. 1971. Plant Soil 35, 65-75. Mengel, K., and Haeder, H. E. 1976. Proc. Int. Colloq. ConrrolPIant Nutr., 4th, Gent pp. 463-475. Mengel, K., and Haeder, H. E. 1977. Plant Physiol. 59, 282-284. Mengel, K., and Helal, H. M. 1968. Z. Pflanzenernahr. Bodenk. 120, 12-20. Mengel, K., and Kirkby, E. A. 1978. “Principles of Plant Nutrition.” Int. Potash Inst., Berne. Mengel, K., and Koch, K. 1971. Z. Pflanzenernahr. Bodenk. 130, 224-233. Mengel, K., and Nemeth, K. 1971. Landw. Forsch. 24, 152-158. Mengel, K., and Pfliiger, R. 1972. Plant Physiol. 49, 16-19. Mengel, K., and Simic, R. 1973. Physiol. Plant. 28, 232-236. Mengel, K.,and Viro, M. 1974. Physiol. Plant. 30, 295-300. Mengel, K., and Wiechens, B. 1979. Z. Pflanzenernahr. Bodenk. 142, 836-847. Mengel, K., Haghparast, M., and Koch, K. 1974. Plant Physiol. 54, 535-538. Mengel, K., Viro, M., and Hehl, G. 1976. Plant Soil 44, 547-558. Mitchell, P. 1966. B i d . Rev. 41, 445-502. Morard, P. 1973. Ph.D. Thesis, University of Toulouse. Mueller, P., and Rudin, D. 0. 1967. Biochem. Biophys. Res. Commun. 26, 398-405. Nemeth, K. 1975. Plant Soil 42, 97-107. Nemeth, K. 1979. Adv. Agron. 31, 155-187. Nemeth, K., and Grimme, H. 1972. Soil Sci. 114, 349-354. Nemeth, K.,and Harrach. T. 1974. Landw. Forsch. Sonderh. 30(I), 131-137. Nemeth, K., Mengel, K., and Grimme, H. 1970. Soil Sci. 109, 179-185. Newman, A. C. D. 1969. J. Soil Sci. 20, 357-373. Newman, E. I . , and Andrews, E. 1973. Plant Soil 38, 49-69. Niederbudde, E. A. 1967. Z. Pflanzenernahr. Dung. Bodenk. 115, 28-43. Niederbudde, E. A., Todorcic, B., and Welte, E. 1969. Z. Pflanzenernlir. Bodenk. 123, 85-100. Nielsen, J. D. 1970. Tidsskr. Planteavl. 74, 24-43. Nobel, P. S. 1970. Plant Physiol. 46, 491-493. Nowakowski, T. 2. 1971. In “Potassium in Biochemistry and Physiology,” pp. 45-49. Proc. 8th Colloq. Int. Potash Inst., Berne. Nye, P. H. 1979. In “The Soil-Root Interface” (J. L. Harley and R. Scott Russell. eds.), pp. 39-49. Academic Press, New York. Nye, P. H., and Maniott, F. H. G. 1969. Plant Soil 30, 459-472. Nye, P. H., and Tinker, P. B. 1977. / n “Solute Movement in the Soil-Root System.” Blackwell. Oxford. Obigbesan, G. 0. 1973. In “Potassium in Tropical Crops and Soils,” pp. 311-322. Proc. 10th Colloq. Int. Potash Inst., Berne. Obigbesan, G. 0. 1977. J . Agric. Sci. Camb. 89, 23-27.

POTASSIUM IN CROP PRODUCTION

109

Okamoto, S. 1967. Soil Sci. Plant Nutr. 13, 143-150. Ollagnier, M., and Ochs, R. 1971. Oleagineux (6). 367-372. Overnell. J. 1975. Physiol. Plant. 35, 217-224. Paauw, F., van der 1958. Plant Soil 3, 254-268. Pallaghy, C. K., and Scott, B. I. H. 1969. Aust. J. Biol. Sci. 22, 585-600. Patrick, W. H., Jr., and Reddy, K . R. 1977. In “Proceedings of the International Seminar on Soil Environment and Fertility Management in Intensive Agriculture,” pp. 275-281. Soc. Sci. Soil Manure, Japan, Nippon Dojohiro Gakkai, Tokyo. Peoples, T. R., and Koch, D. W. 1979. Plant Physiol. 63, 878-881. Pfluger, R., and Mengel, K. 1972. Plant Soil 36,417-425. Pierce, W. S . , and Higinbotham, N. 1970. Plant Physiol. 46. Pitman, M. G. 1972. Aust. J . Biol. Sci. 25, 905-919. Pitman, M. G., and Saddler, H. D. W. 1967. Proc. Narl. Acad. Sci. U.S.A. 57, 4 4 4 9 . Poole, R. J. 1978. Annu. Rev. Plant Physiul. 29, 437-460. Portis, A. R., and Heldt, H. W. 1976. Biochim. Biophys. Acta 449, 434-446. Prasad, R. 1977. 2nd FAO/SIDA Seminar on field food crops in Africa and the Near East, Lahore, Pakistan. Pressman, B. C. 1968. Proc. Am. Soc. Exp. Biol. 27, 1283-1288. Ralph, R. L. 1976. Ph.D. Thesis, Faculty of Biology A. Cambridge University. Raschke, K., and Schnabl, H. 1978. Plant Physiol. 62, 84-87. Ratner, A., and Jacoby. B. 1976. J. Exp. E d . 27, 884-852. Ratner. E. I . , and Yeliseova, 0. I. 1968. Fiziol. Rust. 15, 488-497. Reichenbach, H., v. 1972. In “Potassium in Soil,” pp. 33-42. Proc. 9th Colloq. Int. Potash Inst., Berne. Rezk, A. I., and Amer, F. 1969. Soil Sci. Am. Proc. 33, 876-880. Rich, C. I. 1968. In “The Role of Potassium in Agriculture” (V.J. Kilmer. S. E., Younts, and N. C. Brady, eds.), pp. 79-108. Am. SOC.Agron., Madison, Wisconsin. Rich, C. 1. 1972. In “Potassium in Soil,” pp. 15-31. Proc. 9th Colloq. Int. Potash Inst., Berne. Richter, C. H., and Marschner, H. 1973. Z. Pflnnzenphysiol. 70, 211-221. Ristori, G. G. 1975. Potash Rev. (Berne) Subj. 3 , 53. Suite No. 9. Scheffer, F., Welte, E., and Reichenbach, H., v. 1960. Z. PJanzenerniihr. Dung. Bodenk. 88, 115-128. Schmitt, L., and Brduer, A. 1979. “75 Jahre Darmstadter Wiesendungungsversuche. ” Sauerlanders, Frankfurt am Main. Schon, M., Niederbudde, E. A., and Mahkorn, A. 1976. Z. Acker PJanrenbau 143, 27-37. Schroeder, D. 1976. Kali Briefe. Fachgeb. 1,(3). Seser, M. 1978. Kali Briefe (Biintehofl 14(6), 393-402. Singh, B., and Brar, S. P. S. 1977. Porash Rev. Subj. 9 , 35th Suite, No. 6. Smith, F. A., and Robinson, J. B. 1971. Ausr. J . B i d . Sci. 24, 861-871. Smith, J . A. C. 1978. Ph.D. Thesis, University of Glasgow. Steffens, D., and Mengel, K. 1979. Landw. Forsch. (in press). Stephens, 0. 1969. Exp. Agric. 5, 263-269. Stryer, L. 1975. “Biochemistry.” Freeman, San Francisco. Tabatabai, M. A,, and Hanway, J. J. 1969. Soil Sci. Soc. A m . Proc. 33, 105-109. Tanaka, A. 1973. I n “Potassium in Tropical Crops and Soils,” pp. 147-167. Proc. 10th Colloq. Int. Potash Inst., Berne. Terry, N., and Ulrich, A. 1973a. Plant Physiol. 51, 783-786. Terry, N., and Ulrich, A. 1973b. Planr Physiol. 51, 1099-1101. Tinker, P. B. 1978. In “Potassium in Soils and Crops” (G. S. Sekhon, ed.). Potash Res. Inst. of India, New Dehli.

110

KONRAD MENGEL AND ERNEST A. KlRKBY

Travis, R. L.. and Booz, M. L. 1979. Plant Physiol. 63, 573-577. Trebst. A. 1974. Annu. Rev. Plant Physiol. 25, 423-458. Trolldenier, G. 1977. Plant Soil 47, 193-202. Trolldenier, G., and Zehler, E. 1976. In “Fertilizer Use and Plant Health,” pp. 85-93. Roc. 12th Colloq. Int. Potash Inst., Berne. Turner, N. C. 1972. Nature (London) 235, 341-342. Uexk~ill,H. R., v. 1972. Oleagineux ( I ) , 13-19. Vaidyanathan, L. V., Drew, M. C., and Nye, P. H. 1968. J. Soil Sci. 19, 94-107. Vertregt, N. 1968. Eur. Porato J . 11, 34-44. Vincent, J., Leggett, 1. E., and Egli, D. B. 1979. In “The Soil-Root Interface” (J. L. Harley and R. Scott Russell, eds.), p. 440.Academic Press, New York. Viro, M. 1973. Ph.D. Thesis, FB 19, Justus Liebig University, Giessen. Wanasuria, S., De Datta, S . K., and Mengel, K. 1980. Plant Soil (in press). Watanabe, H., and Yoshida, S. 1970. Soil Sci. Plant Nurr. 16, 163-166. Weller, H., and Hofner, W. 1974. Kali Briefe 12, Fachgeb. 2 . (4). Werner, W. 1957. Ph.D. Thesis, Agricultural Faculty, Justus Liebig University, Giessen. Wiechens, B. 1975. Ph.D. Thesis, FB 19, Justus Liebig University, Giessen. Wilson, H. W., and Evans, H. J. 1968. In “The Role of Potassium in Agriculture”(V. J. Kilmer, S . E. Younts, and N. C. Brady, eds.), pp. 189-202. Am. Soc. Agron., Madison, Wisconsin. Woodhouse, W. W., Jr. 1968. Agron. J . 60, 508-512. Zeigler, E., Bloom, A. J., and Hepler, P. K. 1978. What’s New Plant Physiol. 9, 29-32. Zimmermann, U. 1978. Annu. Rev. Plant Physiol. 29, 121-148.

ADVANCES IN AGRONOMY, VOL 33

UTILIZATION OF WILD SPECIES FOR CROP IMPROVEMENT' H. T. Stalker Department of Crop Science, North Carolina State University, Raleigh, North Carolina

I. Introduction

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

112 ,113 A. Collection and Reservation 113 B. Species Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . 114 C. Barriers to Hybridization.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 D. Genotype Buffering and Chromosome Homologies 117 111. The Gap between Hybridization and Utilization . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . 118 IV. Approaches for Utilizing Wild Germplasm Resources . . . . . . . . . . . . . 119 A. Direct Hybridization .......................................... 120 122 B. Bridge Crosses .......... 123 C. Chromosome M D. Physiological M ............................................ 125 E. Other Methods . . . . . . . . . . . . . . . . . . . . . . . . . 125 126 127 ............................... B. Nicotiana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 C. Saccharum . . . . . . . . . . . . . . . . . . . . . 130 D. Solanum . . . . . . . . . . . . . . . . . . . . . . . 131 E. Triticum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 F. Zea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 VI. Specific Uses of Wild Species for Crop Improvement 135 A. Disease and Insect Resistances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 B. Yield. . . . . . . . . . , . 136 C. Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 11. Biosystematics

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

D. Earliness and Adaptation

E. Modes of Reproduction . , . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . 138

F. Miscellaneous Uses ................................ G. New Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Summary and Conclusions ................................ References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

138 139 140 140 14 1

'Paper no. 6188 of the Journal series of the North Carolina Agricultural Research Service, Raleigh, North Carolina 27650. 111 Copyright 0 1980 by Acndcmic Press. Inc.

All rights of reproduction in my f a m nscrvcd. ISBN 0-lZMXn33-9

112

H. T. STALKER

1. INTRODUCTION

Plant improvement has progressed significantly during the past 10,OOO- 15,000 years. Early cultivators not only exploited their environment but brought under domestication most modem crop plants. The harvesting-sowing-harvesting cycle with associated selection pressures transformed low-yielding shattering plants into present-day productive cultivars. During the centuries of cultivation, a wild-weed- domesticate species complex has evolved for most groups of cultivated species. Plant breeding techniques have resulted in increased yields and solved many problems associated with quality, diseases, insects, and harvest. The plant breeder has, historically, utilized the variability in land races for selection and improvement of crops. However, as modem varieties are planted on much of the cultivated acreage and as population centers expand, many land races are no longer grown and the associated wild species are becoming extinct. In addition, the variability and germplasm resources available for many cultivated varieties are extremely limited (Harlan, 1972). As additional genetic resources are required to fill voids in breeding populations, unique and imaginative procedures are required to exploit fully the potential of our crop plants. Utilization of wild species is one method designed to introduce additional germplasm into cultivated varieties. The first recorded interspecific hybrid was made in 1717 between carnation and sweet William by Thomas Fairchild (Allard, 1960). Thousands of interspecific crosses have been attempted since that time. Most attempts have probably been made by investigators who were simply curious about progenies of species hybrids. However, incorporating desired genes into existing cultivated varieties has become increasingly important. Unique hybrid genotypes often produce unexpected plant types that have economic value. Although the literature is filled with reports of interspecific hybrids, the number of their descendants actually utilized by the farmer are rather limited (see Harlan, 1976a; Hawkes, 1977; Sanchez-Monge and Garcia-Olmedo, 1977). Obtaining hybrids between cultivated and wild species often requires a great effort. First-generation hybrids are often partially sterile, and many programs are abandoned after a few cycles of selection because of continued sterility, low yields, or poor quality characteristics of hybrid derivatives. Chromosomal, genetic, cytoplasmic, or mechanical isolation barriers can present severe handicaps for utilization. Exploitation of species related to crop plants requires the integration of many disciplines. Expertise in the fields of botany, taxonomy, cytology, genetics, ecology, plant breeding, and biochemistry greatly increases the probability of eventual success. An understanding of gene centers, centers of diversity, and species relationships also enhances the progress of germplasm utilization. The effort required to transfer even a single gene from a wild to a cultivated species is often very great, and quantitative traits are even more troublesome to transfer.

WILD SPECIES USE FOR CROP IMPROVEMENT

113

The integrity of the crop’s phenotype must also be maintained to assure highyielding varieties within a production regime. Although great effort may be required, wild species have contributed germplasm to several crops with great economic rewards. The following review is intended to present an overview of the successful transfers to useful genes from wild species to related field crop plants. This review of selected examples will illustrate different points related to utilization of wild species for crop plants and will not attempt to review the entire area of interspecific hybridization. Although species hybrids have been used more extensively with ornamentals than with field crops, ornamentals will not be included and horticultural and vegetable crops will be discussed very briefly. Emphasis will be placed on problems associated with utilizing interspecific hybrids for improving crop species, methods used to overcome these obstacles, an assessment of the value of making the effort to produce something valuable, and the potential of utilizing other wild species germplasm. Attempts to utilize wild species germplasm to improve a crop species depend on species relationships, modes of reproduction, the extent to which the crop can be changed genetically without reducing the economic value, the number of genes controlling the trait in question, methods to overcome undesirable linkage groups, ease and power of screening procedures, and the amount of effort that can be devoted to the problem. Before presenting a review of utilizing wild species for crop improvement, biosystematic considerations (such as taxonomy and species relationships), barriers to hybridization, and approaches used for germplasm exploitation will be discussed.

11. BIOSYSTEMATICS Much has been written on the subjects of speciation, centers of origin, and the need to preserve genetic resources. Discussion of these topics in this article will be rather brief with references to more in-depth reviews noted in the text. The first concern of the breeder who wishes to utilize wild species is the collection of germplasm and the relationships of the taxon with which he or she is working. This is followed closely by chromosome numbers, crossability , fertility, and chromosome homology of species. This section will present discussions conceming collection of germplasm, species concepts, barriers to hybridization, and genotype buffering and chromosome homologies. A . COLLECTION A N D PRESERVATION OF GERMPLASM

While most plant breeders select desired genotypes within the cultivated species, making interspecific hybrids depends on species collections for pro-

114

H.T. STALKER

gress. As human population centers close to gene centers expand and demand more food, vast areas of land are planted with a small number of high-yielding varieties. Not only are land races being replaced by new varieties, but many wild species are becoming extinct (Harlan, 1975a). The loss of some species may be of little importance for crop improvement, but potentials are unknown until collections are screened and tested. Even then they remain unknown for traits not of current interest or for diseases not yet developed. For example, only one accession of Oryza nivara Sharma and Shastry was found resistant to grassy stunt virus after thousands of cultivated and hundreds of wild accessions were screened (Anonymous, 1974). Germplasm resources of every type, including cultivated land races, weeds, and wild species, need to be collected, evaluated, and maintained before being lost forever. The pioneering works of de Candolle ( I 959) and Vavilov (1949/1950) concerning plant geography presented landmark works on origin and diversity of cultivated plants. Although the gene centers of Vavilov do not strictly coincide with crop origins of Harlan (1975b), the primary and secondary gene centers of Vavilov are often the best sources of plant pest resistances (Leppik, 1970). Systematic collection of the wild species related to crop plants is still needed for most crop species. B . SPECIES CONCEFTS

Before a group can be fully exploited, the available plant materials and their relationships must be understood. Taxonomy is of central importance in this goal. Classification systems are based on judgments of individuals and should not define rigid boundaries for taxonomic categories. The classical binomial system of naming organisms traces back to the foundations laid by Linnaeus in 1753 and was bound on morphological similarities and differences. Numerical, cytological, biochemical, and hybridization tools have been used to delineate groups of organisms. Conceptual problems arise when two or more morphological species are fully cross-compatible and produce fertile hybrids or when sibling species are not cross-compatible. Polyploids, asexual reproduction, and genes affecting crossing relationships (including genes that control pairing and cytoplasmic traits) further complicate the picture of determining species relationships. Artificial selection within crop species makes classification of them even more difficult, and classical taxonomic systems often are inadequate (Harlan and deWet, 1971). Because of the inherent problems associated with classical taxonomic classification, biological species concepts have been developed. Biological species consist of cross-compatible taxa (for review, see Scudder, 1974). Naming plants as biological species has the advantage of knowing compatible parents for hybrid programs and thus aids in selection of compatible species for a cross-

WILD SPECIES USE FOR CROP IMPROVEMENT

115

ing program. Although names associated with biological species are often impractical for filing plants in herbaria, members of the same biological species will usually produce viable offspring. To describe more adequately the germplasm available for hybridization with cultivated plants, while not getting bogged down in nomenclature, Harlan and deWet (1963) devised a gene pool concept which groups cross-compatible taxa. They preserve the classical taxonomic species, but create three gene pool levels (Harlan and deWet, 1971). The primary gene pool includes all forms that hybridize freely, produce viable hybrids, and have chromosomes that pair and exchange genes freely. Most hybrids between subspecies or wild-weed-cultigen complexes belong to this group. The secondary gene pool includes species that can be used as germplasm resources; but hybrids are difficult to make because of polyploid level differences, chromosome alterations, or genetic barriers to hybridization. Some degree of sterility is associated with the first-generation hybrids in the secondary gene pool. Members of the tertiary gene pool are difficult to utilize and plants often belong to different genera. Crosses are usually made only at high polyploid levels in complex hybrids, and sterility is always associated with offspring. Fertility can sometimes be restored, but usually the percentage of recovered viable zygotes is extremely small. C. BARRIERS TO HYBRIDIZATION

In reviews by Stebbins (1958), Levin (1971), and Bates and Deyoe (1973), many of the problems of interspecific hybridization were outlined. Although these reviews did not specifically concern utilization of wild species, many of the principles are applicable to hybrids between wild and crop species. In light of the numerous examples presented in these papers, only a summary of barriers to hybridization between species will be presented in this section. Reproductive barriers to hybridization can be divided into two broad groups-premating and postmating. Premating barriers due to geographic isolation are usually unimportant for utilizing wild species. However, apomixis or pollen-pistil incompatibilities may restrict the use of some species. While asexual reproduction can often be circumvented by making reciprocal crosses, pollen-pistil incompatibilities usually present greater problems. Screening for compatible parents, applying immunosuppressants, mechanical manipulations of styles, and bud pollination offer avenues to circumvent these problems. Postmating barriers to hybridization usually present the greatest handicap for obtaining interspecific hybrids. Included in this group are barriers due to ploidy differences, chromosome alterations, chromosome loss or elimination, cytoplasmic incompatibilities, seed dormancy, andor hybrid breakdown (i.e., lethals or loss of vigor in later generations). Cytological barriers to hybridization are many times easily identified.

116

H.T. STALKER

Chromosomal variation between two species caused by inversions, translocations, duplications, deletions, aneuploidy, or polyploidy may be observed under the microscope. Manipulation of cytological barriers to hybridization offers one of the greatest challenges to utilizing wild species for improvement of crop species. Approaches to altering the genomic makeup of species and hybrids are discussed in a later section of this review. Chromosome elimination or loss in interspecific hybrids can be classified as a secondary barrier to gene transfer, but an important one for consideration in some species. Lange (1971) and Orton and Tai (1977) reported the elimination of Hordeum bulbosum L. chromosomes in first-generation hybrids of H . vulgare L. x H . bulbosum without apparent gene exchange. Selective chromosome elimination has also been reported in Nicotiana (Gupta and Gupta, 1973). Hybridizing species with different euploid or aneuploid chromosome series can result in a similar effect. The end-product of a crossing program to incorporate wild species germplasm should be cytologically stable plants at the chromosome number of the cultigen. After species at different ploidy levels are hybridized and gene exchange occurs, then repeated backcrossing to the crop species will usually restore stability in hybrid progenies and eliminate most or all wild species chromosomes. For other species crosses, such as Zea mays L. (2n = 20) x Tripsacum dactyloides (L.) L. (2n = 36 or 72) or Nicotiana tabacum L. ( 2 n = 48) x N . plumbaginifolia Vivians (2n = 20), the chromosomes are nonhomologous and crossing-over is rare (Stalker et a l . , 1977; Apple, 1962). When 46-chromosome (10 Z. mays 36 T . dactyloides) maize-Tripsacum hybrids are backcrossed to maize as the pollen parent, 46-chromosome plants are again obtained because the Tripsacum chromosomes preferentially pair and the 36 T . maize univalents are eliminated. Fifty-six chromosome (20 Z. mays dactyloides) plants may be produced at a low frequency as a result of unreduced female gametes, and Tripsacum chromosomes may then be selectively eliminated during repeated pollinations with maize to give a 2n = 20, maize-like plant (deWet et a l . , 1970). Cytoplasmic incompatibilities can preclude successful reciprocal crosses (Harvey et al., 1972), lead to sterility (Maan, 1977), or cause degeneration of F, plants. The literature contains many reports of interspecific hybrids being successful only when one of the parents is used as the female. An explanation for this unidirectional hybridization is cytoplasmic incompatibility (Harvey et al., 1972). Other possibilities such as sporophytic or gametophytic incompatibilities should also be investigated. Other barriers to hybridization include nonflowering of hybrids, seed dormancy, and/or hybrid breakdown. When hybrids are made between members of section Rhizomatosae and section Aruchis of peanuts, most first-generation plants are vigorous, but fail to produce flowers (W. C. Gregory, personal communication). Techniques to induce flowering in these hybrids have not been found. A

+

+

WILD SPECIES USE FOR CROP IMPROVEMENT

117

similar situation occurs in a few interspecific hybrids in the genus Glycine, but grafting hybrid stems onto cultivated varieties induces flowering (Newel1 and Hymowitz, 1979). Although seed dormancy is not encountered as frequently as other barriers to hybridization, it can restrict rapid introgression of germplasm from wild to cultivated species. Almost all wild species and many cultivated plants have some seed dormancy. Like most other traits, genes controlling this physiological process can usually be transferred between species. In the genus Aruchis a few hybrid combinations remain dormant for 5 to 10 years regardless of the techniques applied to initiate germination (W. C. Gregory, personal communication). Hybrid breakdown frequently occurs in the second or later generations after hybridization of some species. Plant lethality may be due to differences in genetic content between species or it may occur because some F, plants can tolerate chromosomal duplications or deficiencies which then cause lethals in offspring (Stebbins, 1958). Rarely do all plants die in segregating generations, but large numbers of progenies may be required to break linkages of desiredand lethal genes. D. GENOTYPE BUFFERING AND CHROMOSOME HOMOLOGIES

Collecting and classifying species are only the first steps for utilization of wild species germplasm. The initial wild species by cultivated species hybrids must be produced, and then the transfer of desired genes to commercially acceptable types is necessary. The difficulties encountered during these procedures depend on the cultigen’s acceptance of foreign germplasm, the chromosome homology between the species, and many other factors. Polyploids are generally more likely to accept genes from other species than are diploids because polyploids have greater genetic buffering and because one dose of an additive gene is tolerated more easily than two genes. Outcrossing crop species are similarly more likely to be successful parents in a crossing program with wild species because they are heterozygous for many genes (Harlan, 1975a). Manipulation of ploidy levels and reproductive systems has played a key role in introducing foreign germplasm into commercially acceptable types. High chromosome homologies among cultivated and wild species enhance the probability of successful germplasm utilization. When wild species have genomes nonhomologous with the cultivated species, it may be necessary to create addition or substitution lines, induce translocations, or manipulate genetic systems that control chromosome pairing in order to introgress wild germplasm into cultivated varieties. Rare crossovers between nonhomologous chromosomes of interspecific or intergeneric hybrids may also occur. Methods available for determining chromosome homologies include the ob-

118

H.T. STALKER

servation of chiasma frequency in diploid hybrids, multivalent formation in alloploids (Gerstel and Mann, 1965), nuclear gene segregations, and statistical procedures (Sficas and Gerstel, 1962). Simple genetic systems appear to control meiotic processes in some species such as Trificurn (Riley and Chapman, 1958) and possibly in Nicotiana (Smith, 1977). Bivalent formation is usually a good indication of chromosome homologies in diploids, whereas multivalent formation in polyploids is less definitive (deWet and Harlan, 1972). See Gerstel and Mann (1965) or Riley and Law (1965) for additional information concerning chromosome homologies. Finally, although chromosomes may pair, the important criterion is crossing-over and gene exchange. Bodmer and Parsons (1962) and Lindsley et al. (1968) presented more comprehensive reviews of genetic control of crossing-over.

111. THE GAP BETWEEN HYBRIDIZATION AND UTILIZATION A key observation when reviewing the literature on utilizing wild species germplasm for crop improvement is the predominance of sterility in F, hybrids. Many barriers to hybridization exist, and combining the genomes of two species in one nucleus is often a considerable accomplishment in its own right. Even when partially fertile interspecific hybrids are produced, linkages with undesirable genes many times limit the usefulness of these hybrids. A variety of methods for restoring fertility in hybrids between wild and cultivated species is available. The most common fertility restoration method is colchicine treatment of sterile F, plants, but naturally occurring unreduced gametes and parthenogenesis have also played a role in fertility restoration for a few groups. Many examples could be cited where the initial hybrids were made, fertility was restored, and then the hybrids were abandoned after a few generations. In addition to sterility barriers, there are several reasons for this-for example, the effort required to propagate species hybrids is often great, and associated problems of unfavorable linkages related to poor quality or yield often restrict the usefulness of wild species germplasm. Considering the expenses of resources, the list of commercial varieties with wild species germplasm in their pedigree is meager for most crops. However, as additional specific needs arise and new techniques are developed, increased utilization of wild species is anticipated. Delineation between wild species hybrids with commercially acceptable varieties and hybrids with land races is often difficult when reviewing the subject of utilization of wild species for crop improvement. The introduction of genes from land races to a commercially acceptable crop variety often requires much devel-

WILD SPECIES USE FOR CROP IMPROVEMENT

119

opmental effort because of the poor agronomic types found in the land races. Whenever possible, hybrids between wild and cultivated species should be made with acceptable commercial varities of the cultivated species where the crop will be cultivated. This is, however, not feasible for many crops, because many wild species hybridize more readily with cultivated lines that have not been highly inbred and selected. The clearest plant breeding objective is to find a single dominant gene in a wild species, hybridize plants, obtain fertile offspring, and backcross to the cultivated species. Many times hybridization programs have been initiated merely for observing unique progeny or to study biosystematic relationships. Many hybrids have been produced under the collective idea of “adding genetic variability,” when in reality this usually has relatively little value in itself. Making hybrids and then reporting their origin and unique characteristics are of interest to the biosystematist, but of little commercial value. It is quite another thing to make the hybrids and then carry the crossing and selection program to the degree where the material can be utilized by the farmer. Both successful and unsuccessful attempts at utilizing wild species for improvement of crop plants have been experienced. Creating superior types does not always guarantee commercial production of the hybrid derivatives. For example, even as early as 1920, hybrids between alfalfa species resulted in a 40% increase in yield (Waldron, 1920). Hybridization techniques between alfalfa species are difficult, however, and the higher yielding plants have never been used by the farmer. Breeders have pressures to produce better agronomic varieties and do not have the luxury of hybridizing only to observe unique progenies. Programs directed toward utilizing the vast resources available in the related wild species germplasm can be worthwhile. The successes have been relatively few to date compared to the extensive efforts put forth. However, many critical problems have been solved when the desired germplasm could only be found in wild species. New and novel ideas in the future for utilization of wild species germplasm will aid in the improvement of crop plant species.

IV. APPROACHES FOR UTILIZING WILD GERMPLASM RESOURCES The wild relatives of crop plants differ greatly in their potential usefulness as a germplasm resource. Many collections are often distantly related to the crop, and many severe problems must be overcome before introgression of germplasm is successful. In order to overcome sterility barriers, unique and imaginative methods have been employed to manipulate chromosomes. Many of the suc-

120

H. T. STALKER

cesses have come only after extensive cytogenetic, genetic, and reproductive investigations. Consideration of approaches for utilizing wild germplasm resources will be made in this section. A. DIRECT HYBRIDIZATION

1 . Same Chromosome Number

The easiest route for wild species germplasm incorporation into the cultigen’s genome is by direct hybridization of two species at the same ploidy level. When both species have homologous chromosomes, success is often accomplished with little difficulty. Although two compatible taxa may be considered members of the same species, the relationships of many crops and related species are not clearly defined. Many weed-crop complexes exist, and the advantages of transfemng germplasm from closely related species (sometimes referred to as subspecies) rather than distantly related taxa enhance the chances of success. 2 . DifSerent Chromosome Numbers a . Differences in the Euploid Chromosome Complement. Great difficulty is encountered when interspecific hybrids are attempted between two taxa with different ploidy levels. Many crop plants are polyploids, such as cotton, potato, peanut, wheat, tobacco, oat, sugarcane, and most forages. Many of the wild relatives of these plants have lower chromosome levels. Even when the chromosomes of diploid and tetraploid species are homologous, some degree of sterility is encountered in triploid hybrids because of irregular chromosome segregation during meiosis. Four basic methods have been employed to overcome sterility arising from ploidy differences: (1) directly hybridizing taxa at the different ploidy levels; (2) raising the ploidy level of the wild species (or species hybrids) to the same ploidy level as the cultigen before hybridizing with it; (3) doubling the chromosome number of the species at the higher ploidy level (usually the crop species) before hybridizing with the wild species; and (4) reducing the ploidy level of the cultigen to that of the wild species, making the hybrid, and then resynthesizing the chromosome number to equal the cultigen. Each method affords advantages and disadvantages, and each pathway will therefore be discussed individually. 1. Direct hybridization. Direct hybridization at different ploidy levels is the most common method of hybridizing wild diploids with their polyploid crop counterparts. Numerous examples are found in species of tobacco, potato, peanut, cotton, wheat, and others (Knot and Dvorak, 1976; Hawkes, 1977;

WILD SPECIES USE FOR CROP IMPROVEMENT

121

Sanchez-Monge and Garcia-Olmedo, 1977). A tetraploid X diploid species hybrid usually results in a partially sterile triploid F , plant. Fertility is most commonly restored by colchicine-treating vegetative tissues, thus producing hexaploid offspring. The triploids may also naturally produce unreduced gametes thus creating hexaploids if both sets of unreduced gametes combine. Sometimes pollinating the F 1 with 2x or 4x pollen will restore full fertility. Backcrossing the hexaploid with either the crop or wild species is a logical method to reestablish chromosomally stable progenies at the same ploidy level as the crop. The most successful pathway depends on the amount of chromosome pairing between the species and on whether lines with wild species chromosomes substituted for those of the crop species can be tolerated. Plants can then be backcrossed to commercially acceptable varieties. In theory, this is a simple procedure, but in practice, overcoming the sterility barriers and breaking adverse linkage groups can make introgression of the desirable traits difficult (Gerstel and Mann, 1966). 2. Doubling chromosome number of the species at lower ploidy level. Synthesizing autotetraploidsor amphidiploids from the wild species before hybridizing with cultivated species offers a second pathway for introgressing wild species germplasm to the cultigen. When unreduced gametes are frequently produced, such as in the genus Solanurn, direct diploid x tetraploid pollinations will often produce the desired tetraploid progenies (Quinn and Peloquin, 1973). In most other species colchicine is used to raise the chromosome number of the diploid taxa or species hybrids to the ploidy level of the crop. Producing autotetraploids or amphidiploids can have the immediate advantage of avoiding sterile triploid hybrids. Furthermore, when genes are desired from two or more diploid species, selection can often be conducted at the lower ploidy level. However, autotetraploids or amphidiploids derived from diploid species are not always vigorous or fertile, and selection for unique hybrid combinations may be necessary. Vavilova (1975) and Kryuchkova (1972) doubled the chromosome numbers of several diploid potato species before successfully hybridizing with Solanum tuberosum L. in order to transfer frost resistance to the cultigen. Amphidiploids of tobacco species have been successfully crossed with N. tabacum to transfer disease and nematode resistances (Stavely et al., 1973). Selection at the diploid level has been suggested specifically for Solanurn (Hougas and Peloquin, 1960), Gossypium (Meyer, 1974), Arachis (Stalker and Wynne, 1979), and for polyploid species in general by Chase (1964). 3 . Doubling chromosome number of species at higher ploidy level. Doubling the chromosome number of a polyploid cultivated species before crossing with a diploid species presents another method for utilizing germplasm from species at different ploidy levels. Although first-generation hybrids are semisterile, the method has the advantage of avoiding colchicine-treating vegetative plant parts of sterile F, hybrids. Interspecific hybrids thus produced can either be selfpollinated or backcrossed to the cultivated species. Wernsman and Matzinger

122

H.T. STALKER

(1966) proposed crossing 8x N . tabucum

X 2x N . otophora (Griseb.) Comes and backcrossing the pentaploid hybrid with 4x N. tabacum. Synthetics produced by this method had nearly a 13% yield increase above the N . tabacum parent (Wernsman et al., 1976). 4. Reducing chromosome number of species at the higher ploidy level. Reducing the ploidy level of the cultivated species to that of the wild species is usually more difficult than the reverse process. With the use of anther culture the procedure may have more success in the future. Selecting homozygotes at the diploid level usually is easier than at the polyploid level because simpler genetic ratios exist, dosage effects are absent or lessened, and homozygosity can more easily be attained for desired genes. Haploid plant production by reduced parthenogenesis affords a unique and successful method of creating polyhaploids in the genus Solanum. Haploidy frequency in potatoes can be increased to 35-80% by using elite combinations of seed parents (Peloquin et al., 1966). Potato species can be hybridized at the diploid level, selections made, and then polyploids reformed by tetraploid X diploid matings. Another example of successfully reducing the chromosome number of a species is found in the rangegrass Eragrostis curvulu (Schrad.) Nees (Voight and Bashaw, 1972). The tetraploid species is apomictic, but a sexual diploid plant originating from the tetraploid species was discovered. Since variability is limited in the asexually reproducing species, the sexual plant was used to introduce new gene combinationsto the apomictic forage grass (Stalker and Wright, 1975).

b . Different Basic Genomes. Hybridizing species with different basic genomes creates added difficulties for utilizing wild species germplasm. Several examples in the genera Nicotiana, Saccharum, Zea, Sorghum, and Triticum are found in the literature. Extremely rare recombinational events [i.e., Zea (2n = 20) x Tripsacum (2n = 36 or 72) (deWet and Harlan, 1974) or Saccharum ( 2 n = 112) x Sorghum (2n = 20) (Gupta et al., 1978)] may facilitate gene exchange. Special techniques such as induced translocations [i.e., Triticum x Aegilops or Triticum X Agropyron (see Knott, 1971)] have been successful. Chromosome addition or substitution lines may also be used [i.e., Triticum x Agropyron (Kimber, 1967; Khush, 1973)] when the chromosomes of two species are only partially homologous. B . BRIDGECROSSES

Alternate pathways for germplasm introgression have been suggested when direct hybridization at the same or different ploidy levels is difficult or impossible. Bridge crosses have been used when gene transfer by simpler methods have

WILD SPECIES USE FOR CROP IMPROVEMENT

123

failed. The method has been shown to work only under the special condition where species A hybridizes with species B but not with species C, and species B and C form viable hybrids, and when desired traits are dominant or partially dominant. Bridge crosses have most often been utilized in species of tobacco and wheat. Nicotiana repanda Wild. et Lehmann will hybridize with N . sylvestris Speg. and Comes, but not with N. tabacum. Since N . sylvestris and N . tabacum will hybridize, the potential exists for N. sylvestris to serve as a bridge. Burk (1967) attempted to use N. sylvestris as an intermediate between N. repanda and N . tabacum to transfer mosaic resistance to the cultigen, but difficulties were encountered because resistance was a recessive trait. Aegilops ventricosa has eyespot (Cercosporella herpotrichoides) resistance and incorporation of this character into Triticum aestivum L. is desired. However, the hybrid between T. aestivum and A . ventricosa is sterile. When the hybrid (T. turgidum x A . ventricosa) x T . aestivum was produced and backcrossed with T. aestivum, eyespot resistance was transferred to the cultigen (Kimber, 1967). Bridge crosses have also been utilized in Cucurbita (Rhodes, 1959) and Solanum (Hermsen and Ramanna, 1973; Vavilova, 1975; Dionne, 1963). The use of bridging species is complicated, and the method makes selection for desirable genes difficult. Unless a good screening procedure is available, the desired genes will probably be lost. The probability of successfully transferring quantitatively inherited traits has been found to be manyfold less than for monogenic dominant characters. However, when needs are great and other routes for germplasm incorporation into the cultigen are lacking, utilizing bridge crosses deserves attention. C . CHROMOSOME MANI PU LATIONs

I . Chromosome Additions and Substitutions Chromosome addition lines have been usually found to be of little commercial value. Extra univalents or bivalents from either a cultivated or wild species generally have adverse effects on fertility. Polyploids usually tolerate extra chromosomes more successfully than diploids, but the delicately balanced genome complexes which create the agronomically productive genotype are often disrupted by even a single extra chromosome. Several addition lines have been produced in wheat with improved qualities for disease resistance, earliness, winter hardiness, and protein content; however, none of these addition lines have had acceptance as commercial varieties (Khush, 1973). A notable exception is in the genus Saccharum where the addition of entire or partial genomes from wild species to S. ofticinarum L. (2n = 80) is a common breeding procedure. Modem sugarcane varieties are complex allopolyploids with chromosome numbers ranging from 2n = 100 to 125. A heterozygous chromosome addition line has been

124

H. T.STALKER

used to transfer resistance to Meloidogyne javanica (rootknot nematode) from Nicotiana longijlora Cavanilles to N . tabacum (Schweppenhauser, 1968). Galinat (1976) has proposed introducing a Tripsacum chromosome into Z . mays to produce sweet corn. Chromosome substitutions many times have a lesser effect on the plant’s phenotype than chromosome addition lines (Bumham, 1962).The value of interspecific chromosome substitutions depends on the nature of the desired character, adverse linkages of the substituted chromosomes, and the divergence of the wild species chromosome from the cultivated one. Utilization of substitution lines has mostly been restricted to species where extensive cytogenetic knowledge has been accumulated. Substitution lines have been used to transfer rust resistances from Agropyron species to Triticum aestivum (Kimber, 1967; Johnson, 1966). Chaplin and Mann (1961) also proposed substituting nonhomologous chromosomes of tobacco species (e.g., N . paniculata L., N . plumbaginifolia, or N . rustica L.) for the chromosomes of the cultivated species of N. tabacum. However, reduced quality will probably restrict the use of these hybrids.

2 . Chromosomal Translocations Extreme measures may be required to incorporate desired genes into the cultivated species when species have nonhomologous chromosomes and when chromosome substitutions are undesirable. Utilization of chromosomal translocations provides a mechanism by which this transfer can occur. Sears (1956) reported the f i s t successful interspecific gene transfer utilizing a chromosomal translocation in wheat. He induced an exchange from Aegilops umbellulata Zurk to T . aestivum with the aid of radiation, and the transferred segment carried genes for resistance to Puccinia triticina Eriks (leaf rust). Homozygous lines which were distinguished from the normal wheat only by their resistance to rust and by slightly later maturity were then selected. The use of chromosomal translocations to transfer genes from species of Agropyron or Secale to Triticum has been reviewed by Knott (197 1 ) . Gill and Kimber (1977) have used Giemsa stain to identify interspecific translocated chromosome segments. Transfers between species of tobacco have also been successful by using translocations to establish disease resistance (Gerstel and Burk, 1960). Chromosome translocations between species may also occur without the aid of radiation, although at a much lower frequency. For example, a spontaneous translocation between maize chromosome 2 and a Tripsacum chromosome was reported by Maguire (1957), but no economic use has yet been made of the exchange. 3 . Pairing Control

The manipulation of genes to facilitate nonhomologous chromosome pairing in interspecific hybrids has great potential. Riley and Chapman (1958) reported that

WILD SPECIES USE FOR CROP IMPROVEMENT

125

chromosome pairing was under genetic control in Triticum, and the subject has been reviewed by Riley and Law (1965), Sears (1969, 1976), and Riley (1974). Triticum represents the only cultivated genus where genes have successfully been manipulated to permit homoeologous chromosome pairing and gene transfers between species. D. PHYSIOLOGICAL MANIPULATIONS

I . Embryo Culture Embryo culture can be advantageously used to obtain mature plants when the endosperm degenerates or when seeds do not reach physiological maturity. The technique is becoming widely utilized as a tool for obtaining viable interspecific or intergeneric hybrids. A condensed list of genera in which successful applications of embryo culture used to obtain interspecific hybrids includes: Phaseolus (Braak and Koovistra, 1975), Trifolium (Keim, 1953), Gossypium (Anonymous, 1977), Cucurbita (Wall, 1954), Lycopersicon (Smith, 1944), Hordeum (Kruse, 1974), and Triticum (Ahokas, 1970). Murashige (1977) has presented an indepth discussion concerning the prospects of utilizing embryo culture for plant breeding.

2 . Grafting Grafting one plant onto another does not in itself produce an interspecific hybrid. However, the technique has been applied to enhance the probability for successful hybridization. Grafting has promoted survival of interspecific Beta hybrids (Coe, 1954; Johnson, 1956). Evans and Denward (1955) reported an apparent correlation of grafts between species and the success ratio of interspecific hybridization in the genus Trifolium, and Starzychi (1959) successfully crossed T . repens L. and T . pratense L. only after the two species had been grafted onto each other. Grafting technique has also been used as a tool to predict compatible parents in this genus, although Evans (1962) reported that grafting did not increase the frequency of hybrids in Trifolium. For nonflowering interspecific Glycine hybrids, Newel1 and Hymowitz ( 1979) grafted interspecific crosses onto G. max (L.) Merrill in order to overcome the inhibition to flowering. E. OTHERMETHODS

A variety of growth regulators and hormones have been applied to reproductive structures to initiate seed production after pollinating with diverse germplasm. A few examples will illustrate the general uses of growth regulators.

I26

H.T. STALKER

Honma and Heeckt (1960) applied N-m-tolyphthalamic acid to the flower pedicel at the time of pollination to obtain Brassica peheninsis (Lour.) Rups. x B. oleracea L. hybrids. Repeated applications of naphthalene acetamide and potassium gibberellate to the pedicel of pollinated flowers improved pod set in Phaseolus vulgaris L. x P . acutifolius Gray hybrids (Al-Yasiri et al., 1964). Pittagelli and Stavely (1975) used indole acetic acid in lanolin to help overcome sterility barriers between Nicotiana repanda and N . tabacum. Reciprocal crosses between Vigna radiatu and V . umbellatu were successful when the immunosuppressant €-amino caproic acid (EACA) was applied to flower buds before pollination (Baker et a l . , 1975). Finally, Kruse (1974) used 2,4-D (dimethylamine) applications prior to pollination and followed with gibberellic acid treatments to induce intergeneric hybridization between Hordeum and species of Avena, Phleum, Dactylis, Alopercurus, Triticum, Lolium, and Festuca. Additional methods that have been applied to utilize wild species germplasm include manipulation of sporophytic or gametophytic incompatibilities (deWet et al., 1973; Stalker and Wright, 1975); use of pollen mixtures (Kryuchkova, 1972; C. 0. Grassl, personal communication);application of killed pollen of the female species mixed with viable pollen from the desired male parent; and mechanically shortening the style to allow the elongating pollen tube to reach the ovule (deWet et al., 1976) and making reciprocal crosses (Maan, 1977). Tissue culture of Nicotiana suaveolens Lehm x N . tabacum cotyledons to produce viable plants resistant to brown spot has also been used (Lloyd, 1975). As technology advances in the field of cell biology of crop species, additional hybridization techniques have been developed. Somatic hybridization is a technique whereby genomes from different species and genera have been combined without pollination. Fusions in the genus Nicotiana [for example, N . glauca Grah X N . langsdor$ii Weim (Carlson et a l . , 1972)] have been successfully made using somatic cells. However, the method thus far has not been practical for crop improvement programs, and recovery of viable plants from protoplast fusion has only been successful between species in which standard pollinations also produce hybrids. Somatic hybridization does, however, offer a possible method for obtaining hybrids, but more conventional procedures will continue to be the most productive routes for gene transfer in the near future (Gamborg et al., 1977). For additional information concerning protoplast fusion, see Kao et al. (1974), Cocking (1977), or Gamborg et al. (1977).

V. EXAMPLES OF SPECIES USED IN WILD SPECIES HYBRIDIZATION PROGRAMS Man uses 3000 or more plant species for food and cultivates some 150 of these to the extent that they have entered into commerce. The crops that produce at

WILD SPECIES USE FOR CROP IMPROVEMENT

127

least 20 million metric tons include: wheat, rice, corn, potato, barley, sweet potato, cassava, grape, soybean, oat, sorghum, sugarcane, millet, banana, tomato, sugar beet, rye, orange, coconut, cotton, apple, yam, peanut, and watermelon (Harlan, 1976b). Although interspecific hybridization has been attempted in all these crops, the transfer of wild species genes into commercial varieties has had varying success. The list of crops in which great effort has been expended to utilize wild species germplasm includes cotton, tobacco, sugarcane, potato, wheat, and corn. Brief reviews of these six crops will be presented in this section. The number of journal pages filled with interspecific hybridization reports of these six crops is very extensive, and reference to other reviews will be made whenever possible. The following discussions will concentrate on techniques and accomplishments that have had practical applications in plant breeding. A. Gossypium

The genus Gossypium has four cultivated species-G. hirsutum L. (2n = 52), G. barbadense L. (2n = 52), G. herbaceum L. (2n = 26), and G. arboreum L. (2n = 26), plus 30 short-fibered or lintless wild diploid species (Phillips, 1976). The genus has seven genomic groups-six groups of diploids (A, B, C, D, E, and F genomes) and one tetraploid group (AD genome). The wild species are scattered in subtropical areas around the world. Gossypium hirsutum (upland cotton) is the most important cultivated cotton species and furnishes most of the world’s cotton fiber. All tetraploid species have one genome of 13 chromosomes similar to cultivated Old World diploids and another set of 13 chromosomes similar to New World diploid species (Meyer, 1974). The incorporation of germplasm from wild species with either the A or the D genome of the cultivated tetraploids would thus appear to be relatively easy. However, many cotton breeders do not recommend interspecific hybridization programs for improvement of cultivated cotton. Successful incorporation of wild species germplasm is difficult because of the finely tuned genetic system controlling lint quality, strength and texture (Harland, 1970), hybrid breakdown among interspecific hybrids (Stephens, 1950; Lewis, 1957), high negative correlations between fiber strength and high yield (Al-Jibouri et al., 1958; Cooper, 1969), and reduced chiasma frequencies in many of the interspecific hybrids (Stephens, 1949; Rhyne, 1958). One method of utilizing wild cotton species is to hybridize taxa at the diploid level, treat F, plants with colchicine, and then cross the amphidiploid with the cultivated tetraploid species. Some wild diploid species will hybridize with G. hirsutum only when complex interspecific hybrids are used as parents in a crossing program (Meyer, 1974). For example, after chromosome doubling, Beasley (1942) crossed the amphidiploid (G. thurberi Tod x G. arboreum) x G. hir-

128

H.T. STALKER

sutum and backcrossed to G. hirsutum. Lines were developed with superior lint strength but otherwise poor quality. Gossypium thurberi has more recently been the source of increased lint strength in variety releases (Anonymous, 1968; Sappenfield, 1970). Other breeding procedures have included producing hybrids between the wild diploid and cultivated tetraploid species, increasing the chromosome number to the hexaploid chromosome level, and then backcrossing to a cultivated variety. Utilizing wild species for breeding disease and insect resistances into cultivated cotton has met with limited success (Phillips, 1976). Gossypium armourianum Kearney has served as a source of the D2 smoothless gene to G. hirsutum (Meyer, 1957). Smooth cottons are resistant to boll weevils (Lukefahr et al., 1971). Blank et al. (1972) used G . raimondii as a source of rust resistance, and Knight (1948, 1963) used G. arboreum as a source of blackarm resistance for G . barbadense improvement. Introducing cytoplasmic male sterility to cultivars from wild species is another potential use of wild Gossypium germplasm (Meyer, 1973b; Meyer and Meyer, 1965; Murthi and Weaver, 1974). The nucleus of G. hirsutum has been incorporated into a number of the cytoplasms of wild species, and male fertility restorer genes have been identified in several cases. Meyer (1973a) registered 16 G. hirsutum cytoplasmic male sterile germplasm lines that had wild species in their pedigree.

B. Nicotiana The genus Nicotiana includes two cultivated species, N . rustica and N . tabacum, plus approximately 60 wild species (Goodspeed, 1954). Both diploids and tetraploid wild species exist in nature, with basic genomes of n = 8, 9, 10, and 12. The tetraploid N . tabacum (2n = 48) is the most widely cultivated tobacco, and probable progenitors of this species are N . sylvestris and N . tomentosiformis Goodspeed (Gerstel, 1960; Gray et al., 1974). Species of tobacco are particularly well suited for interspecific crosses because more than half of the wild species will hybridize with N. tabacum, pollinations result in large numbers of seeds, and plant propagation is relatively easy. As early as 1912, East and Hayes (1912) listed a great number of interspecific hybrids with N . fabacum, and more than 300 interspecific hybrids have now been reported (Smith, 1968). Interspecific hybrids within subgenera usually show greater fertility than hybrids between subgenera. Methods utilized for germplasm introgression have included direct hybridization, hybridization with induced polyploids, bridge crosses, chromosome addition lines, chromosome substitution lines, chromosomal translocations, isolation of rare crossover events between normally nonhomologous chromosomes, hybridization of N . tubacum monosomics with wild species to induce fertility when

WILD SPECIES USE FOR CROP IMPROVEMENT

129

cultivated varieties with whole genomes would not hybridize with the wild species, and growth regulators (Mann et al., 1963; Gerstel, 1946; Burk, 1972; Pittagelli and Stavely, 1975). Mann et al. (1963) reported that all successful gene transfers to N . tabacum from related species have been in the form of dominant genes. Only the five wild species in the subgenus tabacum ( n = 12) have been shown to exhibit good chromosome pairing with N . tabacum, whereas species of other subgenera have few chromosome homologies to N. tabacum (Moav, 1958; Chaplin and Mann, 1961). Ploidy differences create a major obstacle for incorporating germplasm from these species to the cultigen. When closely related species were crossed with the cultigen, increased leaf production was observed by Edmondson (1969), Matzinger and Wernsman ( 1967), Wernsman and Matzinger (1966), and Wernsman ef al. ( 1976). Wernsman and Matzinger (1966) proposed using octaploid N . tabacum as a parent with the wild diploid species to avoid producing sterile triploids. The pentaploid (2n = 60) thus created could then be backcrossed to N . tabacum and the hybrid derivatives put into a recurrent selection program for specific combining ability. An alternative for rapid utilization is to colchicine-treat diploid interspecific hybrids, that is, produce 4x ( N . sylvestris X N . otophora), then hybridize the amphidiploid with N . tabacum. The interspecific hybrids thus produced could then potentially be used for commercial production (Wernsman and Matzinger, 1966). Although interspecific hybrids between the species closely related to the cultigen and N . tabacum result in faster growth rates, Mann and Weybrew (1958) and Oupadissakoon and Wernsman (1977) concluded that the hybrids were commercially unacceptable because the alkaloid and sugar contents were significantly less in hybrids than in cultivated lines. The greatest accomplishment obtained from utilization of the wild Nicotiana species has been high levels of resistance transferred to cultivated tobacco against many of the most serious diseases. The first successful transfer was for resistance to mosaic virus from N . glutinosa L. (2n = 24) to N . tabacum (Holmes, 1938). Gerstel ( 1943) reported .that the genes conditioning mosaic disease resistance were located on a N . glutinosa chromosome substituted for one of N . tabacum, and Gerstel (1946) concluded that the substitution occurred during the early generations of hybridization. Gerstel and Burk (1960) later found that the gene was transferred to an N . tabucum chromosome. Nicotiana glutinosa has continued to be a source of resistance to mosaic virus for several commercial highyielding varieties (Sand and Taylor, 1961). Clayton (1947) developed wildfire [Pseudomonus tabaci (Wolf and Foster) Stapp] and blackfire [ P . angulata (Fromme and Murray) Stapp] resistant breeding lines using 4x [ N . longijlora (2n = 20) x N . tabacum (2n = 48)] hybrid derivatives. Most F, seedlings died, but after 18 months three seedlings formed callus tissue at ground level and produced shoots. Only one plant was able to be crossed as a female parent with N . tabacum, and a single resistant line was iso-

130

H. T.STALKER

lated. The first wildfire-resistant variety was released in 1955 (Heggestad et al., 1960). Octaploid N . rabacum X N . plumbaginifolia (2n = 20) was used as a source of black shank (Phyrophrhora parasiticu var. nicotiana) resistance in flue-cured tobacco (Apple, 1962). Black shank-resistant varieties had lower yields but acceptable quality characters. Nicoriuna debneyi Domin. (2n = 48) has been used as a source of blue mold resistance in a cigar wrapper tobacco (Clayton, 1967, 1968; Clayton et al., 1967). The transfers were only successful when breeding stocks or South American collections were used as backcross parents instead of commercial U.S. varieties. Nicotiana debneyi has also been a source of resistance for black root rot disease [Thielaviopsis basicola (Beck. and Br.) Fern.], wildfire and Fusarium wilt, plus tolerance to black shank in Burley 49 (Clayton, 1969). Nicotiana repanda (2n = 48) is resistant to 13 diseases and 2 insect pests of tobacco (Pittagelli and Stavely, 1975). Bridge crosses were utilized to transfer resistance for mosaic virus from N . repanda to N . tabacum, with N . sylvestris serving as the intermediate (Burk, 1967). Transfers of this type are not easily made as evidenced by the difficulty encountered when attempting to transfer resistance to rootknot nematodes from N . repanda to cultivated tobacco by the same method. Clayton (1950) obtained male-sterile lines in N . debneyi x N . tabacum hybrids; and male-sterile cigar, flue-cured, and burley tobacco lines were developed. Legg et al. (1974) found that male-sterile burley cultivars derived from N . megalosiphon Heurck and Mueller compared favorably with male-fertile plants and could be used commercially without loss of vigor or quality. They proposed using the male-sterile lines only as temporary solutions to problems because of the effort required for maintaining male-sterile lines. Hosfield and Wemsman (1974) concluded that yields were depressed in male-sterile lines, but no changes were noted in chemical constituents of harvested leaves. Progress has been made in transferring male fertility restorer genes, and Gerstel et al. (1978) reported that a satellited chromosome of N . repandu induced fertility in N. tubacum. Chaplin and Ford ( 1 965) released a male-sterile line for breeding purposes which was derived from N . undulafa Ruiz and Pavon (2n = 24). For a further discussion of utilizing male sterility genes of wild tobacco species, the reader is directed to the review by Berbec (1974) who reported that 11 wild species X N . tabucum hybrids resulted in male-sterile lines. C. Saccharum

The genus Saccharum belongs to the tribe Andropogoneae, family Poaceae. Members of the genus include two cultivated species, S . oflcinarum L. (noblecanes) and S . edule Hassh, plus four wild species (Stevenson, 1965). Saccharurn oflcinarum ( 2 n = 80) is considered an ancient allopolyploid and was almost used as the world’s sole commercial sugarcane before the 1900s. Out-

WILD SPECIES USE FOR CROP IMPROVEMENT

131

breaks of disease such as sereh during the mid- and late 1800s caused radical changes in sugarcane breeding (Abbott, 1953). Utilization of wild sugarcane species has been an integral part of sugarcane improvement since the late 1800s, especially for incorporating disease resistances into cultivars. Wakker made the first reported interspecific cross between S. oflcinarum and “Kassoer,” a wild taxon which was identified in 1916 by Jesweet to be S. spontaneum L. (Bremer, 1961a). “Kassoer” was most important for introducing sereh disease resistance into noblecanes. Other early work with wild species introductions X S. ofwinarum hybrids resulted in canes with high sugar content. These interspecific selections were also found to be resistant to sereh and mosaic diseases (Brandes and Sartoris, 1936). All modem sugarcane varieties are derived from interspecific hybrids and contain three to five species in their pedigrees (Price, 1963). Chromosome number increases are common when species of Saccharum are hybridized, and modem varieties have very high aneuploid chromosome numbers ranging from 2n = 100 to 125 (Price, 1963). Since vegetative propagation is practiced for commercial production, fertility is not required in hybrid plants. The increase in chromosome number is important for adding desired characters from wild species because there is little gene exchange between S. oflcinarum and other species of the genus (Price, 1963). Saccharum spontaneum x S . ofticinarum crosses have led to plants with as many as 160 chromosomes (Bremer, 1961b). Sugarcane has probably been hybridized with species of more genera than has any other taxon. Grassl (1963, 1977) lists eight genera in which intergeneric hybrids have been made with Saccharum, including: Eccoilopus, Erianthus, Miscanthidium, Miscanthus, Narenga, Ripidium, Sclerostachya, and Sorghum. The cultivated sugarcane clones most receptive to foreign pollen are generally not pure species, but very complex interspecific hybrids. Many of the intergeneric hybrids are easily made with the aid of male sterility genes (Grassl, 1963). Although intergeneric hybrids have contributed little to the development of modem commercial varieties, they have great potential for germplasm improvement. Grassl hybridized S. ofticinarum (2n = 112) with S. bicolor L. Moench (2n = 20) and produced a hybrid with about 132 chromosomes. After backcrossing with S. bicolor, 40-chromosome hybrids were recovered with at least some gene exchange reported (Gupta et a l . , 1976). The progenies have great yield potential because the pedicellate spikelets in Sorghum are sterile, but in some recovered Saccharurn-Sorghum progenies both the sessile and pedicellate spikelets are fertile (deWet et al., 1976). A twofold yield for grain increase is potentially present in these intergeneric hybrid derivatives.

D. Solanurn The genus Solanum contains more than 2000 species. The tuberous species of the genus grow in the central Andes of South America and Mexico (Hawkes,

132

H. T.STALKER

1972). Taxa are separated geographically but genetic barriers are seldom complete among the 100 or more tuberous potato species in the genus (Hawkes, 1963). Seven taxa are cultivated: three diploids (2n = 24), two triploids, one tetraploid, and one pentaploid (Hawkes, 1958), with the most economically important species being the tetraploid S . tuberosum (2n = 48). The polyploid series in the genus from 2n = 24 to 72 makes introgression from wild to cultivated species difficult. Dodds (1966) and Ross (1966) reviewed work with wild potato species used for improvement of S . tuberosum. Dodds concluded that the wild species will probably only be utilized as sources of disease resistance, but suggested using these germplasm resources for genes conditioning yield and other agronomic characters. Transferring disease resistance from the wild species to S . tuberosum has had a great favorable economic impact on society. Few disease epidemics caused as much hardship as late blight (Phytophtora infestans) in Europe during the 1840s. More than one million Irish died and another one and a half million people emigrated to the Americas because of the loss of the potato crop (Salaman, 1949). Breeders first used the Mexican hexaploid species S . demissum Lindl. as a source of disease resistance to combat late blight. Like most other miracle solutions, genes from S . demissum turned out to offer only temporary resistance against the disease. At least nine races of P . infestuns have been reported and S . demissum contains genes conditioning resistance to only six of these (Salaman, 1949; Van der Plank, 1968). Most varieties have two or more wild species in their pedigrees (Hougas and Ross, 1956). Solanum demissum carries genes conditioning resistance to late blight, leaf roll, and virus X (Ross, 1966), and has been the most common germplasm resource for potato improvement. Other potato species have been used in variety development for resistances to viruses A, X , and Y, Leprenorarsa decernbineuru, and nematodes (Ross, 1966). Although wild potato species have been utilized primarily as a source of genes conditioning disease resistance, other traits found in wild species also have potential uses. Dionne (1961) and Grun and Aubertin (1965) indicated that genes conditioning cytoplasmic male sterility have been extracted from several species. Wild species have also been used in breeding programs to increase starch content in potato tubers (Ospchuck, 1970). Although yields were reduced in these hybrid derivatives, Motskaitis and Vinitskus (1975) used several wild potato species to increase both yield and starch content of S . tuberosum. Solanum acaule has contributed genes for frost resistance to several commercial varieties (Ross, 1966). Several methods have been employed to incorporate wild and cultivated species germplasm into S . tuberosum. Direct hybridization and backcrossing with S . tuberosum have most often been used, but bridge crosses (Dionne, 1963), pollen mixtures and chemical treatments (Kryuchkova, 1972), hybridization of induced autotetraploids with the cultigen (Livermore and Johnstone,

WILD SPECIES USE FOR CROP IMPROVEMENT

133

1940), and haploid breeding techniques (Hougas et al., 1958) have all been employed for crop improvement. The genus is rather unique in its high frequency of haploidy. Chase (1963) and Hanneman and Ruhde (1978) reviewed haploidy breeding in potatoes in which haploids are produced from tetraploid potatoes, selection takes place at the diploid level, and polyploids are then resynthesized.

E. Triticum The genus Triticum includes a polyploid series of species and hybrids ranging from diploids (2n = 14) to hexaploids (2n = 6x = 42). The cultivated hexaploid species T. aestivum contains the genomes of three different species and is a classical example of utilizing wild species for crop improvement through interspecific hybridization and genome building. Although the genus contains 22 species, hexaploid bread wheat ( T . aestivum) and tetraploid durum wheat (T. durum Desf.) constitute most of the cultivated wheat acreage (Zohary, 1970). Possibly more has been written concerning utilizing wild species of wheat than any other crop. Sears (1969, 1975), Riley (1965), and Riley and Kimber (1966) have presented reviews of the extensive work in wheat cytogenetics and interspecific hybridization. Only a brief discussion will be presented here. Resistance to several diseases has been incorporated into cultivated wheats from closely related species of Triticum, Aegilops, and Agropyron. McFadden (1930) was the first to transfer disease resistance to a commercial wheat variety from a related species. According to Sears (1972b), three methods have been developed for transferring genetic material to wheat from wild relatives. These include the ( 1) induction of homoeologous pairing and crossing-over between wild and cultivated species; (2) use of ionizing radiation to translocate alien chromosome segments; and (3) exploitation of the tendency of univalent chromosomes in wheat to misdivide, resulting in teleocentrics of two univalents which reunite and produce a chromosome with an arm from each of the two univalents. The induction of homoeologous pairing is by far the easiest method of transferring germplasm (Sears, 1972b; Riley et al., 1968), but induced translocations have also produced favorable results (Sears, 1956; Knott, 1961; Sharma and Knott, 1966; Johnson, 1966). Substitution lines have also been successfully utilized for transferring germplasm into the cultivated wheat genomes. Several chromosomes of Agropyron species carrying genes for disease resistance have been substituted for those of the wheat genomes (Knott, 1964; Sears, 1967; Schultz-Schaefferand McNeal, 1977; Knott et al., 1977). In addition to serving as a source of genes conditioning disease resistance, wild species have been used as a source for improved winter hardiness (Tsenov et al., 1974; Yakovlev, 1972), lodging resistance conditioned by short stature (Vasilenko, 1973), and cytoplasmic male sterility (Maan, 1973; Franckowiak et al., 1976).

I34

H.T.STALKER

F. Zea The genus Zea contains four members: Z. mays ssp. mays (cultivated maize), Z. mays ssp. rnexicana (Schrad.) Iltes (annual teosinte), Z. perennis (Hitchc.) Reeves et Mangelsdorf (tetraploid perennial teosinte), and Z. diploperennis Iltis, Doebley , and Guzman (diploid perennial teosinte). Maize and annual teosinte hybridize naturally where they grow together in Mexico or Central America (Wilkes, 1977), and teosinte-maize introgression occurs in a few areas. Emerson and Beadle (1930) reported that the chromosomes of maize and teosinte pair normally, and many maize-teosinte hybrids have been produced (Mangelsdorf, 1974). The successful improvement of cultivated maize varieties with the addition of teosinte germplasm has been limited. Mangelsdorf ( I 958) indicated that teosinte has a mutagenic effect on maize, and this instability may have contributed to the limited usefulness of teosinte to improve yields of corn. Interest in species of Tripsacurn as a germplasm resource for improvement of maize began when Mangelsdorf and Reeves (1931) made the first maizeTripsacurn intergeneric hybrids. Tripsacurn has a base chromosome number of 9 as opposed to 10 for maize. Crosses can be made in either direction, but hybrids are often more easily produced when Tripsacurn serves as the female parent. When the reciprocal cross is attempted, the maize styles (ear silks) must be shortened to allow the Tripsacurn pollen to reach the ovules (deWet et al., 1973). Studies of the morphology and genetics of Tripsacurn x maize hybrids have been extensive (see Mangelsdorf, 1974; deWet and Harlan, 1974; Harlan and deWet, 1977, for reviews). Diploid maize X 2x Tripsacurn produces viable F, hybrids, but in the second or third backcross generation, using maize as a pollen parent, 20-chromosome maize plants have been recovered without apparent intergenomic gene exchange. However, Maguire ( 1957) observed a translocation from the long arm of a Tripsacurn chromosome to the short arm of chromosome 2 of maize. She later reported evidence for crossing-over between the short arm of maize chromosome 2 and a homoeologous segment of a Tripsacurn chromosome (Maguire, 1963). Tripsacurn chromosomes have loci in common with at least 8 of the 10 maize chromosomes (Galinat, 1974; Rao and Galinat, 1976), and Simone and Hooker (1976) reported the transfer of genetic resistance to northern corn leaf blight from Tripsacurn to maize. Cytological pathways from the initial hybrid to recovered 20-chromosome maize are more complex when tetraploid versus diploid Tripsacurn is used (Newel1 and deWet, 1973; Harlan and deWet, 1977). For example, the Fl hybrid 10 Z. mays) or 82 chromomay have 46 chromosomes (36 T . ductyloides somes (72 T . dactyloides 10 Z . mays). When 2n = 46, elimination of maize chromosomes has been observed to take place during anaphase I of meiosis, but the Tripsacurn chromosomes segregate normally (Harlan ef a l . , 1970). How-

+

+

WILD SPECIES USE FOR CROP IMPROVEMENT

135

ever, after anaphase I1 no cytokinesis occurs, so a doubled Tripsacurn haploid chromosome number results. Backcrossing with maize again results in 46chromosome plants, and this process had been shown to be a sexual rather than an apomictic one (deWet et a l . , 1970). Although at a rare frequency, maize chromosomes have been shown to associate with those of Tripsacurn during meiosis and then migrate to the poles along with the Tripsacurn chromosomes. The chromosome number of the hybrids can then be advanced from 46 to a higher ploidy level (deWet and Harlan, 1974). The Tripsacurn chromosomes are eliminated when maize is used as a recurrent parent, and 20-chromosome recovered resulted in some, but not all, hybrid derivatives after six or more generations. Stalker et al. (1977, 1978) concluded that the probability of gene transfers between maize and Tripsacurn chromosomes was high during the early generations of backcrossing. Hybrid derivatives of tetraploid Tripsacurn x diploid maize are now being utilized in commercial breeding programs.

VI. SPECIFIC USES OF WILD SPECIES FOR CROP IMPROVEMENT A. DISEASE A N D INSECT RESISTANCES

By far the most common reason for attempts of utilizing wild species is to transfer disease resistance. Most successes have been with single genes where adequate selection pressures could be applied. Clayton (1954) made the following conclusions about using wild species for crop improvement: (1) the most valuable genes conditioning disease resistances are found in species distantly related to the cultigen with which interspecific hybrids are difficult to make; (2) gene conditioning resistances that would justify interspecific transfers often behave as monogenic dominant genes for immunity or very high resistance; (3) transfers often require 10 or more years of continuous work; and (4) if possible, a species that produces an F, or allopolyploid which exhibits full resistance to the pathogen should be used. Much has been written concerning horizontal (many genes) and vertical (single genes) disease resistances. Although horizontal resistance is desirable, transferring gene complexes only compounds the problem of producing commercially acceptable, disease-free plant types. Watson (1970) reviewed general considerations of disease transfer from wild to cultivated species, and Knott and Dvorak (1976) presented an extensive bibliography of specific examples. Utilization of the rootstocks of wild species has also eliminated many diseases and insect pests of plants that are now commonly grafted, such as grape, citrus, and rubber (Wellman, 1972). Programs aimed at breeding for insect resistance have often met with less

136

H. T. STALKER

success than programs aimed at disease resistance. This has been in part due to lack of adequate screening procedures for pest resistances in germplasm collections, the damage done by insects being less severe than diseases in many regions, application of insecticides by farmers to control insect pests economically, and general placement of priorities in other areas. However, genetic resistance to many insect pests has been found in wild species. An example of successful gene transfer is red stele and aphid resistances from Frugaria chiloensis Duch to the cultivated strawberry (Jones, 1976). Hammons (1970) also reported resistance to leaf-chewing insects in Arachis hypoguea derived from A . monticolu Krap. et Rigoni.

B. YIELD Increased yield can be separated into two parts-that due to increased production of vegetative plant parts and that due to increased seed yields. When leaf or stem plant parts are of commercial value, such as in forages or tobacco, increasing vegetative plant structures has great economic importance. Hybrids between wild and cultivated species of wheat, maize, sorghum, pearl millet, rice, tobacco, and several forage plants have resulted in increased production of leaves and stems. However, increasing yields while maintaining the required quality associated with commercial varieties in cultivation can be difficult. For example, Wernsman et al. (1976) found that whereas tobacco leaf production could be increased by using progenitor species, the advantage gained was negated by poor leaf quality. Cotton fibers may be lengthened as the result of interspecific hybridization (Schwendiman and Lefort, 1974). However, there appear to be severe barriers to increasing both yield and cotton fiber length at the same time. Tam and Tai (1977) also observed increased tuber yields after interspecific hybridization between wild and cultivated potato species. The probability of increasing seed yields from utilization of wild species would seem more unlikely than observed gains in vegetative plant parts because of sterility associated with many interspecific hybrids. Although breeding programs for yield have historically been secondary to breeding for diseases or specific morphological traits when attempting to utilize wild species, the number of successes are increasing. Several examples of yield increases have been reported for species of Vigna (Bruter, 1971), Zea (Reeves and Bockholt, 1964), Ribes (Glebova, 1976), Vanilla (Anonymous, 1972), Arachis (Hammons, 1970; Stalker et al., 1979), and Avena (Frey, 1976). Frey (1976) reported that A . sativu x A . sterilis resulted in a 25-30% yield increase over the recurrent parent. This was twice the yield increase obtained from 1905 to 1960 with germplasm of midwestern United States A . sativa varieties (Frey, 1976). Lawrence and Frey (1975) concluded that the second to the fourth backcross generations appeared to

WILD SPECIES USE FOR CROP IMPROVEMENT

137

be the best ones for selecting high-yielding transgressive segregates from interspecific oat hybrids. Possibly the best example of utilizing wild species germplasm to obtain yield increases for commercial production is in the genus Saccharurn. All modem varieties are progenies of S. oficinarum (2n = 80) x wild species of the genus. With the resulting resistance to sugarcane pests, the high polyploid (2n = 100 to 125) derived varieties greatly outyielded the old noblecanes (Bremer, 1961b). C.

QUALITY

Improvement of quality is of great importance for many crops. Wild species germplasm has had limited use in this area because of the genetic complexity of most quality characters. However, the potential exists for making alterations in chemical composition and morphological traits by using wild species germplasm. Wild grasses usually have a higher protein percentage'than related cultivars (Harlan, 1967), and protein quality or quantity has been altered in several species of the Poaceae. However, Harlan et a f . (1973) also noted that protein percentage and seed size are negatively correlated. The advantage sought from wild germplasm is often lost by the time large-seeded types are selected from hybrid derivatives. Seeds with an increased protein percentage have been selected from crosses with wild species of rye (Yakovlev, 1972), rice (Anonymous, 1974), and oats (Frey, 1975; Lyrene and Shands, 1975). In addition, oil quality has been improved by utilization of wild oil palm species (Musso et a f . , 1977), and A . sativu x A . sterifis hybrid derivatives produced transgressive segregates for high oil percentage in seeds (Frey et a l . , 1975). In tomatoes, the soluble-solids content of commercial varieties has been increased substantially by hybridizing cultivated varieties with a wild green-fruited species (Rick, 1974). Palakarcheva and Edreva (1972) and Palakarcheva and Bailov (1973) reported improved leaf quality of tobacco by using N. debneyi. Several reports of increased fiber quality of cotton have been made (Anonymous, 1968; Sappenfield, 1970). Starch content has also been improved in potato tubers by utilizing wild species germplasm (Osipchuck, 1970; Motshaitis and Vinitskas, 1975). Although few forage species are completely domesticated, the quality of plant products has been improved by utilization of wild species in this group. Improvement includes both the manipulation of wild species for economic use and the utilization of species hybrids for crop improvement. Hybrids between widely separated species may be better adapted in forages than in other groups because 50% or more of the grass species are polyploids, and some sterility can be accepted (Dewey, 1977). For example, the Cynodon interspecific hybrid derivative coastal Bermuda grass is widely used as a forage grass in the southeastern region of the United States. In range and pasture grasses, a number of cases have

138

H. T. STALKER

been reported where derivatives of interspecific hybrids were more palatable than parental species. For example, tall fescue is grown on 10 to 14 million hectares in North America. Quality of tall fescue-giant fescue hybrid derivatives is superior to the parental species and has promise as a new forage (Buchner et al., 1976). Other improved grasses include maize used as silage with teosinte in its pedigree (Sidorov and Shulakov, 1962), cultivated races of sorghum used for fodder (Yakushevskii and Ivanyikovich, 1974), sorghum-Sudan grass hybrids (Anonymous, 1975), and pearl millet-elephant grass hybrids (Chheda et al., 1973; Singh et a l . , 1972). D. EARLINESS AND ADAFTATION

As land for agriculture becomes more limited and crops are moved into less suitable environments, variability for new adaptations must be exploited in order to maintain high yields. Although breeding programs utilizing only the cultivated species have done much for moving crops into new environments, the variability found in wild species offers a valuable resource for genes conferring wider adaptations. Because of severe climatic conditions in the Soviet Union, breeders there have probably utilized wild species to improve the winter hardiness of their crops more than breeders elsewhere (Harlan, 1976a). Cold tolerance has been successfully transferred from wild species of peppermint (Bugaenko et al., 197% tomato (Robinson and Koualewski, 1974), grape (Filippenko and Lebedev, 1971), strawberry (Jones, 1976), wheat (Kuvarin, 1973), rye (Yakovlev, 1972), onion (Meer, 1975), and potato (Ross, 1966) to their cultivated relatives. A crop’s range can also be extended by breeding for shorter growing seasons. Wild Brassica (Ellerstrom, 1977) and Glycine (Harlan, 1976a) species have been used to introgress genes for earliness into cultivated varieties. Breeding for drought and heat tolerance has been accomplished by utilizing wild species of peas (Drozd, 1965) and wheat (Rao, 1974). Other environmental characters for which wild species have been utilized include breeding for increased salt tolerance in tomato (Frederickson and Epstein, 1975; Rush and Epstein, 1978), tolerance to calcareous soils in grape (Malikova and Zenkova, 1977), and lack of photosensitivity in Pennisetum (Chheda et al., 1973). E. MODESOF REPRODUCTION

Sterility is the most common alteration in reproduction which results from interspecific hybridization. Cytoplasmic male-sterility is one economically important derivative of this result, especially for hybrid seed production. An abbreviated list of crops in which cytoplasmic male-sterility has been discovered in crosses between wild and cultivated species include: wheat (Maan, 1973;

WILD SPECIES USE FOR CROP IMPROVEMENT

139

Franckowiak et al., 1976), cotton (Meyer, 1973a,b; Murthi and Weaver, 1974), barley (Schooler, 1967; Ahokas, 1975), tobacco (Burk, 1960; Chaplin and Ford, 1965), ryegrass (Wit, 1974), potato (Dionne, 1961; Grun and Aubertin, 1965), and sunflower (Gimenez and Fick, 1975). Duvick (1959) and Harvey et al. ( 1972) reviewed the uses and potentials of cytoplasmic male-sterility for several crops. In addition to cytoplasmic male-sterility, other modes of reproduction have been altered in crops by introducing germplasm from wild species. Petrov et al. ( 1977) reported apomictic plants of maize-Tripsacurn hybrid derivatives. Likewise, apomixis has been transferred from wild Beta species to cultivated species of sugar beet (Filotowicz and Dalke, n.d.). Parthenogenesis in wild and cultivated potato (Hanneman and Ruhde, 1978) and sorghum (Doggett, 1964) has been manipulated to overcome ploidy sterility barriers among diploids and tetraploid species. Taliaferro and Bashaw (1966) have outlined a scheme for utilizing apomictic and sexual forage grass species for crop improvement. The cleistogamy and self-fertility traits of wild Secale species may also be transferred to cultivated rye (Kuckuck, 1973).

F. MISCELLANEOUS USES Wild relatives of crop plants have been utilized for a number of different traits that do not conveniently fit into other categories. Wild species have been used to transfer genes conditioning hard-seededness to cultivated bramble fruits (Hull, 1968), to introgress dark green color and excellent leaf texture into lettuce (Ryder and Whitaker, 1976), and to transfer a gene for carotenoid synthesis to cultivated tomato (Rick, 1967). Plant stature and other morphological traits can also be improved by utilizing wild species germplasm. For example, Hurtado et al. (1970) reported interspecific hybrids of oil palms resulted in shorter plants. Easier collection of fruits during the 20-year growth period of the crop was then made possible. Selections among Triticurn x Agropyron hybrid derivatives have also resulted in semidwarf wheats (Vasilenko, 1973). From species hybrids, Daskalov et al. (1971) obtained bright red, thin-fleshed red peppers with the fruits drooping in bunches. One by-product of species hybridization has sometimes been naturally occurring mutations which might one day serve as sources of breeding materials. Instability can result when cultivated tobacco and N. plurnbaginifolia are hybridized. Somatic variegations were observed in tobacco by Moav and Cameron (1960). Mangelsdorf (1958) observed 42 mutations in maize x teosinte hybrids after four backcross generations. Rick ( 1967) likewise reported mutagenic effects in interspecific tomato hybrids. Whether these mutations can be utilized is debatable, but apparently the genomes of the crops have been altered.

140

H.T. STALKER G. NEW CROPS

Possibly the only new crops developed in modem times are the garden strawberry and triticale. Cultivated strawberries were derived from the American octaploids Fragaria chiloensis and F . virgeniana Ehrh. by way of an accidental cross in a botanical garden (Jones, 1976). Remarkably, the hybrid had high yield potential while combining many favorable quality traits from each of the parental species. Modem varieties can yield up to 40 tons per acre (Wilhelm, 1974). Triticale resulted from Triticum x Secale hybrid derivatives (see Tsen, 1974, or Larter, 1976, for review) and is now in limited commercial production. Although this new crop still has several drawbacks, not the least of which is sterility, more commercially acceptable types can be anticipated in the future.

VII. SUMMARY AND CONCLUSIONS Germplasm resources of wild species provide no miracle cures for combating plant pests or improving quality, yield, or specialized traits for crop improvement. Valuable germplasm resources are available in wild species, however, and their exploitation will continue to improve the quantity and quality of cultivated species. The greatest benefits from wild relatives of crops have been in the families Poaceae (i.e., wheat and sugarcane) and in the Solanaceae (i.e., tobacco and potato). This has been in part due to the many polyploid species in these groups of ancient hybrid origin and the extensive research effort devoted to aspects of crop improvement involving wild species. Successfully utilizing wild species in other groups is rapidly increasing. There are many reports of successful species hybrids with many crop species. Although difficulty is encountered when attempting to infer from these reports whether or not the “useful germplasm” will actually make it to commercial production, progress is being made. As new techniques are developed and greater efforts are devoted to improving crop species with the aid of wild species, greater numbers of successful transfers will be utilized by the farmer. Germplasm resources in most wild species up to now have been difficult to exploit. A comprehensive knowledge about the taxonomy, biosystematics, and genetic makeup of specific groups not only presents the possibilities and limitations, but also greatly enhances the probability of success in a program to utilize the wild species germplasm. Germplasm resources in wild species have been most utilized when desirable genes were unavailable in the crop species. Although wild species have commonly been used as a source for disease resistance in some crops, a great void of knowledge exists for most groups. Many plant collections need to be screened for disease and insect resistances, genes for quality and yield, and other desirable characteristics.

WILD SPECIES USE FOR CROP IMPROVEMENT

141

Germplasm transfer from wild to cultivated species is often difficult. Furthermore, elimination of unfavorable linkages between desired and unfavorable genes may be quite difficult to break. The most severe bottleneck for germplasm utilization is not usually production of the F, hybrid, but sterility or hybrid breakdown may cause formidable obstacles to introgression from wild to cultivated species. Clear objectives for introgressing specific genes and the development of techniques to overcome sterility barriers aid in the eventual transfer and use of germplasm to improve commercial varieties. Utilization of wild species germplasm is many times best accomplished when the researcher making wild and cultivated species hybrids is part of a conventional breeding team where he is responsive to the needs of specific crops. The use of wild species germplasm holds a dynamic place in crop improvement and will become increasingly important as new variability is required to meet the needs of conventional breeding programs. ACKNOWLEDGMENTS I am grateful to Drs. P. J. Buescher, J. M. J. deWet, M. M. Goodman, J. R. Harlan, L. L. Phillips, and E. A. Wernsman for reading the original manuscript and for their useful suggestions.

REFERENCES Abbott, E. V. 1953. I n “Plant Diseases,” pp. 526-539. USDA Yearbook. Ahokas, H . 1970. Ann. Bor. Fenn. 7, 182-192. Ahokas, H . 1975. Ann. Bor. Fenn. 11, 25-28. Al-Jibouri, H . , Miller, P. A,, and Robinson, H. F. 1958. Agron. J . 50, 633-636. Allard, R. W. 1960. “Principles of Plant Breeding.” Wiley, New York. Al-Yasiri, Salih Azia, and Coyne, D. P. 1964. Crop Sci. 4, 433-435. Anonymous. 1968. “Notice to Growers and Seedsmen Relative to the Release of a Commercial Variety of Upland Cotton, TH 149.” N.C. Agric. Exp. Sin. and USDA Memo. Anonymous. 1972. Annu. Rep. IRAT. 1970 27, 183. Anonymous. 1974. In!. Rice Res. Insr. Annu. Rep., 1973, Los Banos, Philippines. Anonymous. 1975. Ausr. Div. Trop. Agron., CSIRO, Annu. Rep., 1974-1975, Brisbane. Anonymous. 1977. Arra Genet. Senica 4, 3 I - 4 1. Apple, 1. L. 1962. Phyroparhology 52, 1 . Baker, L. R., Chen, N. C., and Park, H. G. 1975. Horric. Sci. 10, 313. Bates, L. S . , and Deyoe, C. W. 1973. Econ. Bor. 27, 401-412. Beasley, J . 0. 1942. Generics 27, 25-54. Berbec, J . 1974. Z. Pj7anzenzuechr. 73, 204-216. Blank, L. M., Fisher, W. D., and Stith, L. S. 1972. “Notice of release of three noncommercial stocks of cotton, Arizona 22, 383 and 495, resistant to Southwestern cotton rust.” Ariz. Agric. Exp. Sin. and USDA Memo. Bodmer, W. F., and Parsons, D. A. 1962. Adv. Genet. 11, 2-100. Braak, J. P., and Kooistra, E. 1975. Euphytica 24, 669-679. Brandes, E. W., and Sartoris, G. B. 1936. In “Yearbook of Agriculture,” pp. 561-623. USDA Yearbook.

142

H. T. STALKER

Bremer, G. 1961a. Euphytica 10, 59-78. Bremer, G. 1961b. Euphyrica 10, 229-243. Brown, W. L. 1961. Iowa Acad. Sci. 68, 90-94. Bruter, D. P. 1971. Metody Issled. Zernobob. kul ’rurami 1, 288-296. Buchner, R. C., Webster. G. T., Burrus. P.B., 11, and Bush, L. P. 1976. Crop Sci. 16, 81 1-816. Bugaenko, L. A., Reznikova, S. A., and Popovich, A. L. 1975. Tr. Prikl. Bot. Genet. Sel. 54, 267-274. Burk, L. G. 1960. J. Hered. 51, 27-31. Burk, L. G. 1967. J. Hered. 58, 215-218. Burk, L. G. 1972. Tob. Sci. 16, 43-45. Burnham, C. R. 1962. “Discussions in Cytogenetics” (3rd ed.). Published by the author, St. Paul, Minnesota. Candolle, Alphonse de. 1959. “Origin of Cultivated Plants” (1886 trans.). Hafner, New York. Carlson, P. S . , Smith, H. H., and Dearing, R. D. 1972. Proc. Narl. Acad. Sci. U.S.A. 69, 2292-2294. Carson, H. L. 1976. Nature (London)259, 395-396. Chaplin, J. F., and Ford, 2. T. 1965. Crop Sci. 5, 436-438. Chaplin, J . F., and Mann, T. J. 1961. N.C. Exp. Sfn. Tech. Bull. 145. Chase, S . S. 1963. Can. J . Genet. Cytol. 5 , 359-363. Chase, S. S. 1964. Crop Sci. 4, 334-337. Chheda, H. R., Aken’Ova, M. E., and Crowder, L. V. 1973. Crop Sci. 13, 122-123. Clayton, E. E. 1947. J . Hered. 38, 35-40, Clayton, E. E. 1950. J . Hered. 41, 171-175. Clayton, E. E. 1954. J. Hered. 45, 273-277. Clayton, E. E. 1967. Tob. Sci. 11, 107-110. Clayton, E. E. 1968. Tob. Sci. 12, 112-124. Clayton, E. E. 1969. Tob. Sci. 13, 30-37. Clayton, E. E., Heggstad, H. E., Grosso, J. J., and Burk, L. G. 1967. Tob. Sci. 11, 91-99. Cocking, E. C. 1977. In “Interspecific Hybridization in Plant Breeding” (E. Sanchez-Monge and F. Garcia-Olmedo, eds.), pp. 229-235. Proc. 8th EUCARPIA Congress, Madrid. Coe, G. E. 1954. Proc. Am. Soc. Sugar Beet Technol. 8, 157-160. Cooper, D. T. 1969. Ph.D. Thesis, Dept. of Crop Sci., N.C. State Univ.,Raleigh. Daskalov, Kh., Rusenova, I., and Milkova, L. 1971. Gradinaratvo 13, 5-7. deWet, J . M. J., and Harlan, J. R. 1972. Taron 21, 67-70. deWet. J . M. J., and Harlan, J. R. 1974. Genetics 78, 493-502. deWet, J. M. J., Lambed, R. J., Harlan, J. R., and Naik, S. M. 1970. Caryologia 23, 183-187. deWet, 1. M. J., Harlan, J. R., Engle, L. M.,and Grant, C. A. 1973. Crop Sci. 13, 254-256. deWet, J. M. J., Gupta, S. C., Harlan, J. R., and Grassl, C. 0. 1976. Crop Sci. 16, 568-572. Dewey, D. R. 1977. Genet. Led. 5, 7-18. Dionne, L. A. 1961. Am. Poraro J . 38, 117-120. Dionne, L. A. 1963. Euphytica 12, 263-269. Dodds, K . S. 1966. Endeavour 25, 83-88. Doggett, H. 1964. Heredity 19, 543-558. Donnelly, E. D. 1971. Crop Sci. 11, 721-724. Drozd, A. M. 1965. Bull. Appl. Bor. Genet. PIanr Breed. 37, 130-145. Duvick, D. N. 1959. Econ. Bor. 13, 167-195. East, E. M . , and Hayes, H. K. 1912. USDA Bur. Plant lnd. Bull. 243, 1-58. Edmondson, E. R. 1969. M.S. Thesis, Dept. of Crop Sci., N.C. State Univ.. Raleigh. Ellerstrom, S. 1977. Sver. Ursadesforen. Tidskr. 87, 363-367. Emerson, R . A., and Beadle, G. W. 1930. Am. N a . 64, 190-192.

WILD SPECIES USE FOR CROP IMPROVEMENT

143

Evans, A. M. 1962. Euphytica 11, 164-176. Evans, A. M., and Denward, T. 1955. Nature (London) 175, 687-689. Filippenko, I. M., and Lebedev, A. V. 1971. Tr. Tsentr. Genet. Lab. I.V. Michurina 12, 46-56. Filutowicz, A., and Dalke, L. n.d. Biul. Inst. Hodowli Aklemal. Ros. (5-6), 15-23. Franckowiak. J . D., Maan. S. S., and Williams, N. D. 1976. Crop Sci. 16, 725-728. Fredrickson, D. L., and Epstein, E. 1975. Plant Phys. 56(Suppl. 2), 4. Frey, K. J. 1975. Crop Sri. 15, 277-279. Frey, K. J. 1976. Egypt. J. Genet. Cytol. 5 , 460-482. Frey, K. J., Hammond, E. G.,and Lawrence, P. K. 1975. Crop Sci. 15, 94-95. Galinat, W. C. 1971. Annu. Rev. Genet. 5 , 447-478. Galinat, W. C. 1974. Evolution 27, 644-655. Galinat, W. C. 1976. Maize Genet. Coop. Newslett. 50, 72-73. Gamborg, 0. L., Constabel, F., Kao, K. N., Fowke, L. C., Ohyama, K., Pelcher, L., and Kartha, K. K. 1977. In “Molecular Genetic Modification of Eucaryotes” (I. Rubenstein, R. L. Phillips, C. E. Green and R. J . Pesnick, eds.), pp. 1-12. Academic Press, New York. Gerstel, D. U. 1943. Genetics 28, 533-536. Gerstel, D. U. 1946. Genetics 31, 421-427. Gerstel, D. U. 1960. Genetics 45, 1724-1734. Gerstel, D. U., and Burk, L. G. 1960. Tob. Sci. 4, 147-150. Gerstel, D. U., and Mann, T. J. 1965. In “Chromosome Manipulations and Plant Genetics” (R. Riley and K. R. Lewis, eds.), pp. 113-120. Heredity (Suppl.) 20. Gerstel, D. U., Burns, J. A., and Burk, L. G. 1978. Genetics 89, 157-169. Gill, B. S., and Kimber, G. 1977. Crop Sci. 17, 264-266. Gimenez, J. D., and Fick, G.N. 1975. Crop Sci. 15, 724-726. Glebova, E. 1. 1976. Nauch. Tr. Leningr. S-Kh. Inta 299, 25-29 (Ru). Goodspeed, T. M. 1954. “The Genus Nicotiana. ” Chronica Botanica, Waltham, Massachusetts. Grassl, C. 0. 1963. Proc. Congr. Int. SOC.Sugar Cane Technol.. 11th pp. 447-456. Grassl, C . 0. 1977. Sugarcane Breeders Newslett. 39, 8-33. Gray, J. C., Kung, S. D., Wildman, S. G.,and Sheen, S.J . 1974. Nature (London)252,226-227. Grun, P . , and Aubertin, M. 1965. Genetics 51, 399-409. Gupta, S. C., and Gupta, P. 1973. Genetics 73, 605-612. Gupta, S . C., Harlan, J . R., deWet, J. M. J., and Grassl, C. 0. 1976. Caryologia 29, 351-359. Gupta, S . C., Harlan, J . R.. and deWet. J . M. J. 1978, Crop Sci. 18, 879-883. Hammons, R. 0. 1970. Crop Sci. 10, 459. Hanneman. R. E., and Ruhde, R. W. 1978. Am. Potato J. 55, 259-263. Harlan, J. R. 1967. Archaeology 20, 197-201. Harlan, J. R. 1972. J . Environ. Qwl. 1, 212-215. Harlan, J. R. 1975a. “Crops and Man.” Amer. Soc. Agron., Madison, Wisconsin. Harlan, J. R. 1975b. J. Hered. 66, 182-191. Harlan, J. R. 1976a. Crop Sci. 16, 329-333. Harlan, J . R. 1976b. Sci. Am. 235, 88-97. Harlan, J. R., and deWet, J. M. J. 1963. Evolution 17, 497-501. Harlan, J. R., and deWet, J. M. J. 1971. Taxon 20, 509-517. Harlan, J. R., and deWet, J. M. J. 1977. Proc. Natl. Acad. Sci. U.S.A. 74, 3494-3497. Harlan, 1. R . , deWet. 1 . M. J., Naik, S. M., and Lambert, R. I . 1970. Science 167, 1247-1248. Harlan, J. R., deWet, J. M. J.. and Price, E. G. 1973. Evolution 27, 31 1-325. Harland, S. C. 1970. In “Genetic Resources in Plants-Their Exploration and Conservation” (0. H. Frankel and E. Bennett, eds.), pp. 335-340. Davis, Philadelphia, Pennsylvania. Harvey, D. H., Levings, C. S., 111, and Wemsman, E. A. 1972. Adv. Agron. 24, 1-27. Hawkes, J. G. 1958. Euphytica 7, 257-270.

144

H. T. STALKER

Hawkes, J. G. 1963. “A Revision of the Tuber-Bearing Solanums” (2nd ed.). Scottish Plant Breeding Station Record, Pentlandfield, Roslin, Miotlothian. Hawkes, J. G. 1972. Symp. B i d . Hung. 12, 183-188. Hawkes, J. G. 1977. Euphyrica 26, 615-621. Heggestad, H. E., Clayton, E. E., Neas, M. O., and Skoog, H. A. 1960. Tenn. Agric. Exp. Sin. Bull. 321. Hennsen, J. G. Th., and Ramanna, M. S. 1973. Euphytica 22, 457-466. Holmes, F. 0. 1938. Phyropurhology 28, 553-561. Honma, S.,and Heecht, 0. 1960. Euphytica 9, 243-246. Hosfield, G. L., and Wemsman, E. A. 1974. Crop Sci. 14, 575-577. Hougas, R. W., and Peloquin, S. J. 1960. Eur. Poraro J . 3, 325-330. Hougas, R. W., and Ross, R. W. 1956. Am. Potato J . 33, 328-339. Hougas, R. W., Peloquin, S. J., and Ross, R. W. 1958. J. Hered. 49, 103-106. Hull, J. W. 1968. Proc. Am. Soc. Horric. Sci. 93, 280-288. Hurtado, M., Reinaldo, J., and Nunez, G. R. 1970. Acra Agron. 20, 9-23. Johnson, R. 1966. Can. J. Genet. Cyrol. 8, 279-292. Johnson, R. T. 1956. J. Am. SOC. Sugar Beer Technol. 9, 25-31. Jones, J. K. 1976. In “Evolution of Crop Plants” (N. W. Simmonds, ed.), pp. 237-242. Longman, London. Kao. K. N., Constabel, F., Michayluk, M. R., and Gamborg, 0. L. 1974. Planra 120, 215-217. Keim, W. F. 1953. Agron. J . 45, 601-606. Khush, G. S. 1973. “Cytogenetics of Aneuploids.” Academic Press, New York. Kimber, G. 1967. J. Agric. Sci. Camb. 68, 373-376. Knight, R. L. 1948. J. Genet. 48, 359-369. Knight, R. L. 1963. J. Genet. 58, 328-346. Knott, D. R. 1961. Can. J. Plant Sci. 41, 109-123. Knott, D. R. 1964. Can J . Genet. Cyrol. 6, 500-507. Knott, D. R. 1971. In “Mutation Breeding for Disease Resistance,” pp. 67-77. IAEA, Vienna. Knott, D. R., and Dvorak, J. 1976. Annu. Rev. Phyroparhol. 14, 21 1-235. Knott, D. R., Dvorak, J., and Nanda, J. S. 1977. Can. J . Genet. Cyrol. 19, 75-79. Kruse, A. 1974. Herediras 78, 319. Kryuchkova, A. F. 1972. Tr. Sredneaz. Opyr. S t . VNII Rastenievod. 2, 234-241. Kuckuck, H. 1973. Wheat lnf. Serv. 36, 21-26. Kuvarin, V. V. 1973. Sel. Semenovod. 6, 69-71. Lange, W. 1971. Euphyrica 20, 181-194. Larter, E. N. 1976. In “Evolution of Crop Plants” (W. D. Simmonds. ed.), pp. 117-120. Longman, London. Lawrence, P. K., and Frey, K. J. 1975. Euphyrica 24, 77-85. Legg, P. D., Collins, G. B., and Litton, C. C. 1974. Tob. Sci. 18, 160-162. Leppik, E. E. 1970. Annu. Rev. Phyroparhol. 8, 323-344. Levin, D. A. 1971. Taxon 20, 91-113. Lewis, C. F. 1957. Agron. J . 49, 455-460. Lindsley, D. L., Sandler, L.. Nicoletti, B., and Tnppa, G. 1968. In “Replication and Recombination of Genetic Material” (W. J. Peacock and R. D. Brock, eds.), pp. 253-269. Australian Acad. Sci., Canberra. Livermore, J. R., and Johnstone, F. E., Jr. 1940. Am. Poraro J . 17, 170-173. Lloyd, R. 1975. Tob. Sci. 19, 4-6. Lukefahr, M. J.. Martin, D. F., and Meyer, J. R. 1965. J. Econ. Enromol. 58, 516-518. Lukefahr, M. J., Houghtaling, J. E., and Graham, H. M. 1971. J . Econ. Enromol. 64, 486-488. Lyrene, P. M., and Shands, H. L. 1975. Crop Sci. 15, 398-400.

WILD SPECIES USE FOR CROP IMPROVEMENT

145

Maan, S. S. 1973. Euphytica 22, 287-300. Maan, S. S. 1977. In “Interspecific Hybridization in Plant Breeding” (E. Sanchez-Monge and F. Garcia-Olmedo, eds.), pp. 201-214. Proc. 8th EUCARPIA Congress, Madrid. McFadden, E. S. 1930. J . Am. Soc. Agron. 22, 1020-1034. Maguire, M. P. 1957. Genetics 42, 473-486. Magwire, M. P. 1963. Can. J . Genet. Cytol. 5 , 414-420. Malikova, V. M., and Zenkova, R. A. 1977. Vinogradarstvo 1976, pp. 118-122. Mangelsdorf, P. C. 1958. Cold Spring Harbor Symp. Quant. B i d . 23, 409-421. Mangelsdorf, P. C. 1974. “Corn, Its Origin, Evolution and Improvement.” The Belknap Press of Harvard Univ., Cambridge, Massachusetts. Mangelsdorf, P. C., and Reeves, R. G. 1931. J . Hered. 22, 329-343. Mann, T. J.. and Weybrew, J . A. 1958. Tob. Sci. 2, 120-125. Mann, T. J., Gerstel, D. U.,and Apple, J. L. 1963. Proc. In:. Tob. Sci. Congr., 3rd, Salisbury pp. 20 1-207. Matzinger, D. F., and Wernsman, E. A. 1967. Der Zuechrer 37, 188-191. Meer, Q. P., van der. 1975. Zaadbelangen 29, 178. Meyer, J . R. 1957. J. Hered. 48, 249-250. Meyer, V. G . 1973a. Crop Sci. 13, 439-444. Meyer, V . G. 1973b. Crop Sci. 13, 778. Meyer, V. G. 1974. Econ. Bot. 28, 56-60. Meyer, V . G..and Meyer, J. R. 1965. Crop Sci. 5,444-448. Moav, R. 1958. Am. Nat. 92, 267-278. Moav, R . , and Cameron, D. R. 1960. Am. J . Bor. 47, 87-93. Motskaitis, Yu. Yu., and Vinitskas, Z. V. 1975. Schlechtd. Dokl. Vsesorjuznoi Ordena Lenina Akad. Sel‘skokhozyaistvennykh Nauk. Imeni V . I . Lenina 7, 17-19. Murashige, T. 1977. Hortic. Sci. 12, 127-130. Murthi, A,, and Weaver, J. B., Jr. 1974. Crop Sci. 14, 658-663. Musso, S., Bouron, C., Archier, P., and Yohou, H. 1977. Oleagineux 32, 489-491. Newell, C. A,, and deWet, J. M. J . 1973. Can. J . Genet. Cytol. 15, 763-778. Newell, C. A,, and Hymowitz. T. 1979. Crop Sci. 19, 121-123. Orton, T. J., and Tai, W. 1977. Can. J . Bot. 55, 3023-3033. Osipchuck, A. A. 1970. Kartoplyarstvo. Resp. Mizhvid. Temat. Nauk. Zb. 1, 32-37. Oupadissakoon, S., and Wernsman, E. A. 1977. Crop Sci. 17, 843-847. Palakarcheva, M., and Bailov. D. 1973. Nauchna sesiya Inst. Genet. Selekktsiya Rasteniyara Sofia 15 i 16 mart 1971, g BAN 1973, pp. 59-70. Palakarcheva, M., and Edreva, A. 1972. Genet. Sel. 5, 379-389. Peloquin, S. J., Hougas, R. W., and Gabert, A. C. 1966. In “Chromosome Manipulations and Plant Genetics” (R. Riley and K. R. Lewis, eds.), pp. 21-28. Heredity Suppl., Vol. 20. Oliver and Boyd, Edinburgh and London. Petrov, D. F., Belousava, N. I . , Ivanova, L. S., Laikova, L. I., Lukina, L. A., Sorkina, T. P., Sukhareva, N. B., Fokina, E. S., and Yudin, B. F. 1977. Fundam. Issled. Biol. Nauki. Novasibirsk pp. 149- 154. Phillips, L. L. 1976. In “Evolution of Crop Plants” (N. W. Simmonds, ed.), pp. 196-200. Longman, London. Pittagelli, G. W., and Stavely, J. R. 1975. J . Hered. 66, 281-284. Price, S. 1963. Econ. Bor. 17, 97-106. Quinn, A. A,, and Peloquin. S. J. 1973. Am. Poraro J . 50, 415-420. Rao, B. G.S., and Galinat, W. C. 1976. J. Hered. 67, 235-240. Rao, M. V. 1974. In “Diversity and Change on the Indian Subcontinent” (Sir Joseph Hutchinson, ed.), pp. 33-45. Cambridge Univ. Press, London and New York.

146

H. T. STALKER

Reeves, R. G., and Bockholt, A. J. 1964. CropSci. 4, 7-10. Rhodes, A. M. 1959. Proc. Am. Sor. Hortic. Sci. 74, 546-551. Rhyne, C. L. 1958. Genetics 43, 822-834. Rick, C. M. 1967. Proc. Inr. Horric. Congr., XVII 3, 217-229. Rick, C. M. 1974. Hilgardia 42, 493-510. Riley, R. 1965. In “Essays on Crop Plant Evolution” (Sir Joseph Hutchinson, ed.), pp. 103-122. Cambridge Univ. Press, London and New York. Riley, R. 1974. Generics 78, 193-203. Riley, R., and Chapman, V. 1958. Narure (London) 182, 713-715. Riley, R., and Kimber, G. 1966. Ann. Rep. Planr Breed. Insr. Cambridge 1964-65, pp. 6-36. Riley, R., and Law, C. N. 1965. Adv. Genet. 13, 57-114. Riley, R., Chapman, V., and Johnson, R. 1968. Nafure (London) 217, 383-384. Robinson, W., and Koualewski, E. 1974. Proc. Inr. Horric. Congr., XIX, I . Sect. VII, p. 716. Ross, H. 1966. Am. Poraro J . 43, 63-80. Rush, D. W., and Epstein, E. 1978. Planr Phys. 61(Suppl.), 94. Ryder, E. J., and Whitaker, T. W. 1976. In “Evolution of Crop Plants’’ (N. W. Simmonds, ed.), pp. 39-41. Longman, London. Salaman, R. N. 1949. “The History and Social Influence of the Potato.” Cambridge Univ. Press, London and New York. Sanchez-Monge, E., and Garcia-Olmedo, F. (eds.). 1977. “Interspecific Hybridization in Plant Breeding.” Roc. 8th Congress of EUCARPIA, Madrid. Sand, S. A., and Taylor, G. S. 1961. Connecricur Agri. Exp. Sra. Bull. 636, 4 p. Sappenfield, W. D. 1970. “Notice of Release of a Variety of Upland Cotton, Delcot 277.” Mo. Agric. Exp. Stn. and USDA Memo. Schooler, A. B. 1967. J. Hered. 58, 207-21 1 . Schulz-Schaeffer, J., and McNeal, F. H. 1977. Crop Sci. 17, 891-896. Schwendiman. J., and Lefort, P. L. 1974. Coron Fibres Trop. 29, 231-236. Schweppenhauser. M. A. 1968. Coresra Inf. Bull. 1, 9-20. Scudder, G. G . E. 1974. Can. J. 2001.52, 1121-1134. Sears, E. R. 1956. Brookhaven Symp. Biol. 9, 1-21. Sears, E. R. 1967. Cienc. Cult. 19, 175-178. Sears, E. R. 1969. Annu. Rev. Genet. 3, 451-468. Sears, E. R. 1972a. Can. J. Gener. Cyrol. 14, 736. Sears, E. R. 1972h. Sradler Genet. Symp. 4, 23-38. Sears, E. R. 1975. In “Handbook of Genetics” (R. C. King, ed.), Vol. 2, pp. 59-91. Plenum, New York. Sears, E. R. 1976. Annu. Rev. Genet. 10, 31-51. Sficas, A. G., and Gerstel, D. U. 1962. Generics 47, 1171-1185. Sharma, D., and Knott, D. R. 1966. Can. J. Genet. Cyrol. 8, 137-143. Sidorov, F. F.. and Schulakov, I. K. 1962. Bull. Appl. Eor. Genet. Planr Breed. 34, 76-85. Simone, G.W., and Hooker, A. L. 1976. Proc. Am. Phyrhoparhol. Soc. 3, 207. Singh, R. D.. Premchand, and Rahaman, A. 1972. Indian J . Agric. Sci. 42, 218-222. Smith, B. 1977. M.S. Thesis, Dept. of Crop Sci., N.C. State Univ., Raleigh. Smith, H. H. 1968. Adv. Genet. 14, 1-54. Smith, P. G . 1944. Proc. Am. Soc. Horric. Sci. 44, 413-416. Stalker, H. T., and Wright, L. N. 1975. Ariz. Acad. Sci. 10, 106-110. Stalker, H. T., and Wynne, J. C. 1979. PeanutSci. 6, 110-114. Stalker, H. T., Harlan, J. R., and deWet, J. M. J . 1977. Am. J. Eor. 64, 1162-1 169. Stalker, H. T.. Harlan, J. R., and deWet, J. M. J. 1978. Caryologia 31, 271-282. Stalker, H. T., Wynne, J. C., and Company, M. 1979. Euphyrica 28, 675-684. Stanychi, S. 1959. Hodowla Rod. Aklim. Nasienn. 3, 277-319.

WILD SPECIES USE FOR CROP IMPROVEMENT

147

Stavely, 1. R., Pittarelli, G. W., and Burk, L. G. 1973. J. Hered. 64, 265-271. Stebbins, G. L. 1958. Adv. Gener. 9, 147-215. Stephens, S. G. 1949. Generics 34, 627-637. Stephens, S. G. 1950. Eof. Rev. 16, 115-149. Stevenson, G. C. 1965. “Genetics and Breeding of Sugarcane.” Longmans, Green, New York. Taliaferro, C. M., and Bashaw, E. C. 1966. Crop Sci. 6, 473-476. Tarn, T. R., and Tai, G. C. C. 1977. Crop Sci. 17, 517-521. Tsen, C. C. (ed.). 1974. “Triticale: First Man-Made Cereal.” The Am. Assoc. of Cereal Chemists, St. Paul, Minnesota. Tsenov. A., Stankov, I . , and Stankova, P. 1974. C. R. Acud. Agric. Georgi Dimirrov 7, 51-54. Van der Plank, J. E. 1968. “Disease Resistance in Plants.” Academic Press, New York. Vasilenko, I. I. 1973. Sel. Semenovod. 6, 71-73. Vavilov, N . 1. 1949/1950. In “The Origin, Variation, Immunity and Breeding of Cultivated Plants” (K. Starr Chester, trans.). Chronica Botanica, Waltham, Massachusetts. Vavilova, M. 1975. Karrofel Ovoschi 10, 7-9. Voight. P. W., and Bashaw, E. C. 1972. Crop Sci. 12, 843-847. Waldron, L. R. 1920. J . Am. SOC. Agron. 12, 133-143. Wall, 1. R . 1954. Proc. Am. Soc. Hortic. Sci. 63, 427-430. Watson, 1. A. 1970. In “Genetic Resources in Plants-Their Exploration and Conservation” (0.H. Frankel and E. Bennett, eds.), pp. 441-457. Davis, Philadelphia, Pennsylvania. Wellman, F. L. 1972. “Tropical American Plant Disease.” Scarecrow Press, Metuchen, New Jersey. Wernsman, E. A,, and Matzinger, D. F. 1966. Crop Sci. 6, 298-300. Wernsman. E. A., Matzinger, D. F., and Mann, T. J. 1976. Crop Sci. 16, 800-803. Wilhelm, S. 1974. Am. Sci. 62, 264-271. Wilkes, H. G. 1977. Econ. Bnr. 31, 254-293. Wit, F. 1974. Euphytica 23, 31-38. Yakovlev, G. V. 1972. Tr. Prikl. Eor. Gener. Sel. 50, 229-238. Yakushevskii, E. S., and Ivanyukovich, L. K. 1974. Tr. Prikl. Eor. Genet. Sel. 52, 253-264. Zohary, D. 1970. In “Genetic Resources in Plants-Their Exploration and Conservation” (0.H. Frankel and E. Bennett, eds.), pp. 239-247. Davis, Philadelphia, Pennsylvania.

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY, VOL. 33

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS, A REVIEW R. J. Buresh,' M. E. Casselman, and W. H. Patrick, Jr. Laboratory for Wetland Soils and Sediments, Center for Wetland Resources, Louisiana State University, Baton Rouge, Louisiana

I. Introduction 11. Nitrogen Fix

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

A. Rice Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Coastal Zones.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Freshwater Ecosystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Nitrogen Fixation in the Aerobic Layer of Flooded Soil ................ IV. Nitrogen Fixation in the Anaerobic Layer of Flooded Soil. ...................... V. Nitrogen Fixation in the Root Zone of Nonnodulated Plants ..................... A. Rice ........................... C. Overview

150

the Water Column and on the Soil Surface

.........

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

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

159 160 161 163 165

. . . . . . . . . . . . 169

VI .

A. Plants Other Than Rice B. Rice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

173

v11. Environmental Factors Influencing Nitrogen Fixation in Flooded Soil . . . . . . . . . . . . . 174 A.

B. C. D. ............... E. F. G. H. Salinity . . . . . . . . . . . . VIII. Comparison of Acetylene Reduction and I5N Methodology ...................... IX . Contribution of Fixed Nitrogen to the Nitrogen Requirements of Plants . . . . . . . . . . . . X. Perspectives ............................................................ References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

180 183 185 187

'Present address: International Fertilizer Development Center, P. 0. Box 2040, Muscle Shoals, Alabama 35660. *.

I 49 Copyright 0 1980 by Academic Ress. Inc. All rights of reproduction in any form rrscncd. ISBN 0-12-000733-9

150

R. 1. BURESH ET AL.

1. INTRODUCTION Flooded soils and natural wetlands are two of the Earth’s most productive land types. Both types of ecosystems play an important role in food production. Lowland rice (Oryza sariva), the major food source for approximately one-half of the world’s population, is grown for the most part on flooded soil. Wetlands such as coastal salt marshes and freshwater swamps and marshes associated with lakes and streams form the base of the food chain that produces fish and other seafood which constitute another major source of man’s food. Nitrogen is usually the limiting nutrient in both flooded soils and wetlands. The major role of fertilizer nitrogen in increasing the production of lowland rice is well documented with hundreds of experiments throughout the world showing response to nitrogen. Studies in natural wetlands show that nitrogen is usually the major nutrient limiting plant growth (Valiela and Teal, 1974; Broome et al., 1975; Patrick and DeLaune, 1976). This is unlike the situation in lakes where phosphorus may sometimes be the limiting nutrient, since the nitrogen requirement in such systems can apparently be largely satisfied by nitrogen fixation in the water column. Rice grown in the United States is heavily fertilized with nitrogen, but much of the rice in Asia receives very little fertilizer nitrogen. Rice yields are lower in nonfertilized than in fertilized fields, yet through the years consistent yields have been obtained in successive rice crops without the benefit of nitrogen fertilizer and with no apparent decrease in the nitrogen content of the soil. The maintenance of nitrogen fertility in these soils has been attributed to nitrogen fixation (App et al., 1978; Koyama and App, 1979). In fact, lowland rice appears to rely more heavily than other food crops on nonsymbiotic nitrogen fixation as a source of nitrogen. In subsistence agricultural systems, where grain is grown every year with no added fertilizer, yields of lowland rice are usually higher than yields of upland grain crops. Part of this yield difference can be attributed to a better water supply in flooded soils, but part of the difference is likely due to either greater nitrogen fixation or greater nitrogen conservation in the flooded soil. Organic nitrogen tends to accumulate more in flooded soils than in well-drained soils. The accumulation of nitrogen in flooded soils in most cases is not solely the result of increased nitrogen fixation. It can also be due to slower decomposition of organic nitrogen compounds under poor aeration conditions which prevent rapid turnover of the soil nitrogen. This slow mineralization rate allows a larger fraction of the fixed nitrogen to accumulate as soil organic nitrogen. The restricted penetration of oxygen into flooded soils results in the formation of an aerobic surface layer of soil and an underlying anaerobic zone. Dissolved oxygen in the overlying water reaches the soil surface and barely penetrates the soil before it is rapidly consumed. The depth to which oxygen penetrates the soil

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

151

is determined by the net effect of the oxygen supply to the soil and the oxygen consumption rate. A high consumption rate results in a thin oxidized soil surface layer, whereas a low consumption rate results in a thicker oxidized layer. A large quantity of readily decomposable organic matter in the soil leads to a high oxygen demand and consequently a thin oxidized layer. If oxygen consumption in the overlying water is very high, oxygen might not reach the soil; the soil would then be completely anaerobic. In addition to the soil surface an aerobic-anaerobic interfacial area can also exist in the plant root zone. Wetland plants can transport atmospheric oxygen through internal porous tissue to their roots (Van Raalte, 1941; Teal and Kanwisher, 1966). Some of the oxygen reaching the root can diffuse outward into the surrounding soil thereby creating a thin aerobic soil layer surrounded by a much more extensive anaerobic zone. Flooded soil systems tend to be more favorable sites for nonsymbiotic nitrogen fixation than well-drained soils. Many of the characteristic features of flooded soils indicate the suitability of these systems for nitrogen fixation. Their poor

FIG. 1. Schematic representation of the components of an idealized flooded soil system.

152

R. 3 . BURESH ET AL.

aeration, for example, suggests a favorable environment for fixation because nitrogenase, the enzyme system that catalyzes nitrogen fixation, is oxygen sensitive. The near neutral pH and low redox potential found in most flooded soils (see, for example, Ponnampemma, 1972) are apparently favorable for heterotrophic nitrogen fixation. Nitrogen-fixing blue-green algae can readily proliferate in the photic zone of flooded soil systems. This review examines nitrogen fixation in the component parts of a flooded soil system: the water column and soil surface, the aerobic soil layer, the anaerobic soil layer, the plant root zone (rhizosphere), and the leaf and stem surface of plants. These component parts of a flooded soil system are illustrated in Fig. 1. Additional sections will examine the environmental factors influencing fixation and discuss the assumptions and limitations associated with the indirect acetylene reduction assay as compared with the direct I5N method for fixation in flooded soils. The contribution of biologically fixed nitrogen to the nitrogen requirements of plants in flooded soil systems will also be examined.

II. NITROGEN FIXATION IN THE WATER COLUMN AND ON THE SOIL SURFACE The portion of a flooded soil system exposed to sunlight can be a suitable site for autotrophic nitrogen fixation. The main autotrophic nitrogen fixers in flooded soil systems are blue-green algae, cyanobacteria. They grow both in the water column and on the surface of the soil layer in rice fields (Bunt, 1961; Watanabe and Yamamoto, 1971; Watanabe et al., 1977b, 1978a,b), marshes (Jones, 1974; Carpenter et al., 1978), and intertidal zones (Stewart and Pugh, 1963; Burris, 1976). They are also present in lakes (Rusness and Burris, 1970; Fogg, 1971) and the ocean (Dugdale et al., 1964; Fogg, 1978). The distribution of autotrophic nitrogen fixers is usually erratic and dependent upon environmental conditions. The following sections examine nitrogen fixation in the water column and on the soil surface of rice fields, coastal zones, and freshwater ecosystems. A. RICEFIELDS

Blue-green algae have been recognized for many years as important agents of nitrogen fixation in rice fields (De, 1936, 1939; Singh, 1961). They are widely distributed throughout the world (Watanabe, 1959; Fogg et al., 1973) but tend to be more abundant in tropical and subtropical regions than in temperate regions (Watanabe and Yamamoto, 1971). Nitrogen fixation by blue-green algae in rice fields fluctuates greatly because of heterogeneity in algal growth and rapid ap-

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

153

pearance and disappearance of algal blooms. The importance of nitrogen fixation by blue-green algae is suggested by the findings that nitrogen-fixing activity is greatly reduced when the floodwater and surface soil is removed from a field and replaced with distilled water (Watanabe et al., 1977b, 1978a,b) and that nitrogenase activity is not present on paddy soil covered with a black cloth (Alimagno and Yoshida, 1977). Researchers at the International Rice Research Institute (IRRI) in the Philippines (Watanabe et al., 1977b, 1978a) found that nitrogen-fixing activity in a planted rice field, measured by an in situ assay, was related to the biomass and nitrogenase activity of blue-green algae in the floodwater. Two peaks in algal fixation were observed during both the dry and wet season; a small one occurred at the early stage of rice growth and a larger one was present near or after harvest (Watanabe et al., 1978b). Nitrogen-fixing activity of blue-green algae was greater in planted than in unplanted plots (Watanabe et al., 1978b) and greater in unfertilized plots than in those fertilized with NPK (Watanabe et al., 1977b, 1978a). It was also greater in the dry season than in the wet season (Yoshida and Ancajas, 1973a; Watanabe et al., 1977b, 1978a). Alimagno and Yoshida (1977) found maximum nitrogenase activity near midday for laboratory-grown algae and near 1700 hours for algae in a field plot. Reynaud and Roger (1978) observed four types of diurnal curves for nitrogenfixing activity by algae in rice fields in Senegal. They concluded that light intensity was an important factor regulating these diurnal variations. Some nitrogen-fixing activity by blue-green algae was observed in fields at night (Alimagno and Yoshida, 1977). Watanabe et al. (1978b) estimated the contribution of nitrogen fixers in the water column and on the soil surface by removal of the floodwater and surface soil and subsequent replacement with algal-free water. They concluded that nitrogen-fixing activity of blue-green algae and perhaps bacteria associated with the algal biomass was greater than that of microorganisms associated with the rice rhizosphere during both the wet and dry season. It should be noted that removal of the floodwater and surface soil eliminated fixation by algae attached to weeds in the water column as well as that by free-living blue-green algae. Nitrogen fixation by blue-green algae associated with aquatic plants will be discussed in Section VI. Wada et al. (1978), on the other hand, reported very little nitrogen-fixing activity associated with photosynthetic microorganisms in a Japanese rice field and concluded that the floodwater and soil surface were not important sites for nitrogen fixation in this rice field. The floodwater was a relatively more important location for nitrogen fixation in unfertilized than in fertilized plots. In a subsequent study, this research group (Panichsakpatana et al., 1978) found that nitrogen fixation by blue-green algae in the rice field was substantially less than that in flooded soil samples incubated within a growth chamber under simulated

154

R. J . BURESH ET AL.

field conditions. Reynaud and Roger (1978) determined a relationship between algal biomass and its nitrogen-fixing rate and from that relationship estimated an algal fixation rate of a few kg N/ha per cultivation cycle in Senegal rice fields. Nitrogen fixation by blue-green algae in paddy fields was recently reviewed by Roger and Reynaud (1979), Stewart et a f . (1979), and Watanabe and Cholitkul (1979) in a symposium at IRRI. Selected estimates for nitrogen fixation in the overlying water and on the soil surface of rice fields are listed in the top portion of Table I. Photosynthetic bacteria capable of fixing nitrogen occur in rice fields (Kobayashi er al., 1967). They normally require light and anaerobiosis for growth, but these conditions seldom coexist in flooded soil systems. However, the photosynthetic bacteria Rhodopseudomonas capsulatus can coexist with heterotrophic bacteria in cultures and fix nitrogen under aerobic light conditions (Kobayashi and Haque, 1971). Certain nitrogen-fixing photosynthetic bacteria can develop an association with Azorobacter (Okuda er a f . , 1961; Kobayashi et a f . , 1967), but no symbiotic relationship is found between the photosynthetic bacteria R. capsulatus and Clostridium in mixed cultures (Okuda er a f . , 1961). Habte and Alexander (1980) observed nitrogenase activity attributable to photosynthetic bacteria in a flooded soil that was amended with ethanol and treated with a herbicide to inhibit algal fixation. Their findings suggest that nitrogen fixation by photosynthetic bacteria and blue-green algae did not occur simultaneously in the rice-soil system they examined. The distribution of photosynthetic bacteria in flooded soil systems is generally erratic, and their actual contribution to nitrogen fixation in rice fields is uncertain (Watanabe, 1978). B . COASTAL ZONES

Regions in coastal wetlands that are uncolonized by higher plants tend to be favorable sites for growth of blue-green algae (Stewart and h g h , 1963; Jones, 1974). High rates of algal fixation have been reported on algal mats and mudflats in numerous salt marshes (Jones, 1974; Whitney et a f . , 1975; Carpenter et al., 1978; Casselman, 1979). However, these marsh regions unvegetated by higher plants and exhibiting high nitrogen-fixing activity occupy only a small fraction of the total area of most marshes (Whitney et al., 1975; Casselman, 1979). Carpenter et al. (1978) found that algal mats and pannes in a Massachusetts salt marsh exhibited the highest photosynthetic nitrogen-fixing activity (see Table 11). Nitrogenase activity, attributable to blue-green algae, was also present on the soil surface in regions of the marsh vegetated by Spartina afterniflora, a common salt marsh macrophyte. The algal mats and pannes comprised only a small portion of the total area in this marsh; consequently, the authors concluded that the overall contribution of fixed nitrogen by blue-green algae was greatest in

Table I Estimated Rates of Nitrogen Fixation in Rice Systems Location Philippines

Philippines

Philippines

Ivory coast Senegal

Thailand

System Paddy water, planted field, wet season Paddy water, unplanted field, wet season Paddy water, planted field, dry season Paddy water, unplanted field, dry season Paddy water and surface soil, unfertilized planted field Paddy water and surface soil, fertilized planted field Paddy water and surface soil, wet season Paddy water and surface soil, dry season Algae, waterlogged soil Algae, overlying water and soil &ace Paddy water and soil from throughout the country

Rate

Method

Reference

CzHz

Yoshida and Ancajas, 1973a

CZHZ.in siru

Alimagno and Yoshida, 1977

204 mmoles CzH,/m2/163 days (19 kg N/ha/163 days)" 307 mmoles CZH,/m2/168days (29 kg N/ha/168 days)" 4-8 pg N/g soil/day

C2Hz, in situ

Watanabe et a'..1978b

CZH2.anaerobic, light assay

Rinaudo et

0-60 nmoles CzH,/cmZ/hr

CzHz

Reynaud and Roger, 1978

CZHZ

Matsuguchi et al., 1975

3 kg N/ha/l19 days 1 1 kg N/ha/l19 days

14 kg N/ha/l19 days

14 kg N/ha/l19 days 18-33 kg Nlhalcrop

2-6 kg N/ha/crop

(a few kg N/ha/crop, 10-30 kg N/ha/crop is exceptional) 0.5-54kg N/ha/yr (avg. = 6.9 kg N/ha/yr)

41..

1971

(continued)

Table I (continued)

C;

cn

Location

System

Louisiana, U.S. Ivory coast

Soil, substantial algal growth Rice, laboratory grown

Philippines

Rice, rhizosphere

Philippines

Rice, rhizosphere and stem, wet season

Philippines

Philippines

Rice, rhizosphere and stem, dry season Rice, associative fixation, variety IR26 Rice, associative fixation, variety IR36 Planted field, flooded soil, wet season Unplanted field, flooded soil, wet season Planted field, upland soil, wet season Unplanted field, upland soil, wet season

Rate 57 pg N/g soiVyr Up to 6OOO nmoles C,H,/ g dry mMhr 0.05 kg Nlhalday (18 kg N/ha/yr) 90 mmoles C,HJm2/163 days (8 kg N/ha/163 days)a 50 mmoles C,H,/m2/168 days (5 kg N/ha/168 days)" 5.9 kg N/ha/l07 days

Method

Reference

2-year incubation under light C,H,

Reddy and Patrick, 1979

C2H2. in situ, algal activity removed C,H,, in situ. algal activity removed

Watanabe et al.. 19771,

C,H2, in situ. algal activity removed

Watanabe et al.. 1979

C,H,, laboratory assay of mixed soil samples

Yoshida and Ancajas, 1973a

I5N2,

Dommergues et al.. 1973

Watanabe et ul.. 1978b

4.8 kg N/ha/95 days

57 kg N/ha/l19 days

22 kg N/ha/ 119 days

7 kg N/ha/l19 days 3 kg N/ha/ll9 days

Ivory coast Ivory Coast

Philippines

Louisiana, U.S. Philippines

Planted field, flooded soil, dry Season Unplanted field, flooded soil, dry season Planted field, upland soil, dry season Unplanted field, upland soil. dry season Planted soil Planted soil

Planted soil, unfertilized, wet season Planted soil, fertilized, wet season Planted soil, unfertilized, dry Season Planted soil, fertilized, dry season Planted field, panicle initiation stage Planted field

63 kg N/ha/ll9 days 28 kg N/ha/ll9 days

5 kg N/ha/ll9 days

3 kg N/ha/ll9 days 72 kg N/ha/yr 246.6 pnoles CZHJ700 cmVday (0.32 kg N/ha/day) 14.0 kg N/ha/216 days

C,H,, in siru C,H,, in siru integrated daily average C,H,, in situ ( 4 1 factor)

Balandreau er al., 1976 Balandreau er al., 1974

0.2-0.4 kg N/ha/day

CPHp,in situ

Reddy and Patrick, 1979

15-50 kg N/ha/crop

N balance

Koyama and App, 1979

W a t a ~ b eer al., 1978a

10.8 kg N/ha/216 days

1 1 . 1 kg N/ha/119days

3.7 kg N/ha/l19 days

Values were calculated from the reported rate of acetylene reduction with a conversion factor of 3 moles of acetylene reduced for each mole of dinitrogen gas fvted.

Table I1 Estimated Rates of Nitrogen Fixation in the Photic Zone of Various Coastal Habitats Location New York, U.S.

Massachusetts, U.S.

System Blue-green algal mat, 2-month summer average Intertidal mudflat, 2-month summer average Unvegetated (panne) area, 2-month summer average Stagnant pools of water, summer Algal mat Unvegetated pannes Soil surface under short Spartina

Rate" 2870 pg N/m2/hr

Reference Whitney et al., 1975

136 p g N/mz/hr 63.5 p g N/mz/hr 8.6-4820 p g Nlliterlhr 22.8 kg Nlhatyr 23.8 kg Nhalyr 8.7 kg N/ha/yr

Carpenter et al., 1978

alterngora

Louisiana, U .S . England

Georgia, U.S. Nova Scotia, Canada

Creek banks and bottoms Entire salt marsh, yearly average Mudflat Soil surface under S. afterniflora Mudflat Algal mat Unvegetated creek banks Stagnant pools of water Soil surface under short S. alternijlora Soil surface under short S. alterniflora

3.5 kg Nhalyr 6.4 kg Nhatyr 15.6 kg N/ha/yr 171 kg Nihalyr 4.3 kg Nlhalyr 201 kg Nha/yr 316 kg N/ha/yr 462 kg Nihalyr 4-9 kg Nihalyr 22 kg Nlhatgrowing season

All the reported rates were determined with the acetylene reduction method.

Casselman, 1979 Jones. 1974

Hanson, 197% Patriquin and McClung, 1978

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

159

marsh areas vegetated with Spartina and on creek banks and bottoms, because these regions comprised the vast majority of the total land area. The authors also indicated that shading by Spartina reduced nitrogen fixation on the marsh surface. Carpenter et a f . (1978) found that algal fixation in a Massachusetts salt marsh was considerably higher in the summer than in the winter. It also exhibited a daily peak between 1000 and 1200 hours followed by a sharp drop near midday. Jones (1974) estimated nitrogen fixation rates ranging from 4.3 to 462 kg N/ ha/year for zones of an English salt marsh dominated by blue-green algae. An algal fixation rate of 23 kg N/ha/yr for algal mats in a Massachusetts salt marsh can be calculated from the results of Carpenter et af. (1978). The corresponding rate for the entire salt marsh was only 6.4 kg N/ha/yr. This rate is only about one-tenth that reported for bacterial nitrogen fixation, rhizosphere plus nonrhizosphere, in this marsh (Teal et a f . , 1979). Casselman (1979) found large seasonal variations in nitrogen-fixing activity by blue-green algae in mudflats in a Louisiana salt marsh. She concluded that algal nitrogenase activity was strongly influenced by water levels in the marsh. Algal activity was suppressed during periods of the year when the mudflat was inundated by water. Maximum fixation by blue-green algae in coastal wetlands occurs when the algae are moist but not submerged by water (Jones, 1974; Casselman, 1979). Reported rates of nitrogen fixation on the soil surface of various coastal habitats are shown in Table 11. The results of Patriquin and McClung (1978) suggest that nitrogenase activity on the soil surface of a salt marsh along the coast of Nova Scotia, Canada, was largely associated with nonheterocystous blue-green algae and perhaps photosynthetic bacteria. Jones ( 1974) also indicated that nonheterocystous blue-green algae were possibly active fixers in an English salt marsh. Estimates of nitrogen fixation in the water column of estuaries and coastal wetlands are generally small or negligible. Brooks et a f . (1971) reported low levels of nitrogen-fixing activity in the water column of a Florida estuary, and Marsho et al. (1975) observed no fixation in the surface waters of a Chesapeake Bay estuary over a 15-month period. Negligible nitrogenase activity was found in the water column of Florida mangrove habitats (Zuberer and Silver, 1978), the tidal waters of a Florida salt marsh (Green and Edmisten, 1974) and a Long Island salt marsh (Whitney et al., 1975), and the water column of a stream in a Louisiana salt marsh (Casselman, 1979). C. FRESHWATER ECOSYSTEMS

Numerous studies have attributed nitrogen-fixing activity in freshwater lakes to blooms of blue-green algae in the photic zone (Home and Fogg, 1970; Rusness

I60

R. J. BURESH ET AL.

and Bums, 1970; Granhall and Lundgren, I97 1 ; Torrey and Lee, 1976; Home, 1979). Low rates of fixation attributed to heterotrophic nitrogen-fixing bacteria have been found in anoxic water from the aphotic zone (Brezonik and Harper, 1969; Keirn and Brezonik, 1971). Saralov and Dzyuban (1978) reported that blue-green algae were mainly responsible for nitrogen-fixing activity in the trophic layer of a eutrophic Russian lake, but fixation in the anoxic hypolimnion which contained hydrogen sulfide was due to anaerobic bacteria. Bacterial nitrogen fixation in the water column is generally of minor significance when compared to fixation by blue-green algae (Torrey and Lee, 1976). Nitrogen-fixing activity in the water column of lakes exhibits substantial variations with location and time (Stewart et al., 1971). The abundance and seasonal variation of blue-green algae in lakes is influenced by phosphorus availability (Stewart et al., 1971; Liao, 1977). Greater algal fixation is generally found in eutrophic and mesotrophic waters rather than in oligotrophic waters (Stewart et al., 1971; Tison et al., 1977). Algal nitrogenase activity in lakes displays striking diurnal patterns that appear to be related to light intensity (Peterson et al., 1977; Home, 1979). The reported contributions of fixation in the water column to the total nitrogen income of freshwater ecosystems vary greatly. Many researchers (Home and Fogg, 1970; Stewart et al., 1971; Tison et al., 1977) have concluded that nitrogen fixation in the water column represents a minor contribution to the total nitrogen input in lakes. However, a recent review by Mague (1977) revealed that in some lakes 40-78% of the total nitrogen income can originate from fixation by bluegreen algae. Blue-green algae colonizing the surface of lake bottoms can be agents of nitrogen fixation. Moeller and Roskoski ( 1978) measured the nitrogen-fixing activity in samples of the surface soil layer obtained from the littoral zone of an oligotrophic lake. They observed considerable variability among samples but found nitrogenase activity associated with benthic blue-green algae. Their results were not sufficient to determine the contribution of nitrogen fixation by benthic blue-green algae to the total nitrogen income of the lake. Liao (1977) speculated that attached algae in the littoral zone of a Canadian lake fixed nitrogen. The distribution and significance of benthic nitrogen-fixing blue-green algae in freshwater ecosystems appears to merit further investigations.

Ill. NITROGEN FIXATION IN THE AEROBIC LAYER OF FLOODED SOIL Magdoff and Bouldin ( 1970) reported substantially greater nitrogen fixation in the surface layer than in the lower layers of flooded, cellulose-enriched soil-sand

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

161

media. These authors hypothesized that the products of anaerobic cellulose decomposition move from the anaerobic to the aerobic zone where they serve as substrates for aerobic heterotrophic nitrogen fixers such as Azotobacter. Aerobic bacteria are more energy efficient than anaerobic bacteria, so they are able to fix more nitrogen with a given quantity of substrate than the latter. According to these authors a large aerobic-anaerobic interfacial area and alternate flooding and drying favor intense nitrogen fixation. Rice et al. (1967) found high rates of fixation in a thin layer of saturated wheat straw-amended soil with a large aerobic-anaerobic interfacial area. In contrast to Magdoff and Bouldin (1970), these researchers (Rice et al., 1967; Rice and Paul, 1972) concluded that the products of plant residue decomposition in the aerobic soil layer were utilized as energy sources by anaerobic nitrogen fixers in the lower layer. A strict anaerobe, Clostridiurn, was presumed to be the active nitrogen fixer in their waterlogged soil system. Unlike Magdoff and Bouldin (1970), Rice et al. (1967) and Rice and Paul (1972) were unable to isolate Azotobacter in their system, and they employed wheat straw rather than cellulose as an energy source. Regardless of the explanation, the findings of both research groups and Waughman (1976) suggested that the presence of both an aerobic and an anaerobic soil zone can stimulate nitrogenase activity. A mutualistic relationship between Azotobacter and Clostridium is well known in cultures (Jensen and Holm, 1975). From an estimation of microbial biomass, turnover, and nitrogen content, Becking (1978) calculated a fixation rate of 45 kg N/ha/yr for free living Beijerinckia, an aerobe common in tropical soils. This calculation may be an overestimation of actual rates in nature because all the nitrogen in the bacterial cells was assumed to originate from molecular nitrogen. No attempt was made to measure an in situ fixation rate. Evans and Barber (1977) in a review article attributed a fixation rate of only 0.3 kg N/ha/yr to Azotobacter, generally considered the main free-living aerobic nitrogen-fixing bacteria. The reported occurrence of aerobic nitrogen-fixing bacteria in nature was summarized by Knowles (1977).

IV. NITROGEN FIXATION IN THE ANAEROBIC LAYER OF FLOODED SOIL Nitrogen-fixing activity associated with free-living heterotrophic bacteria has been reported in freshwater (Howard et al., 1970; Keirn and Brezonik, 1971; MacGregor et al., 1973), estuarine (Brooks et al., 1971; Herbert, 1975; Marsho et al., 1975), marine (Hartwig and Stanley, 1978), and salt marsh sediments (Teal et al., 1979). and rice soil (MacRae and Castro, 1967). Most of the

162

R. J . BURESH ET AL.

researchers cited did not mention the oxidation-reduction status of the sediments and soils under investigation, and many of them measured nitrogenase activity on composite samples; consequently it is difficult to determine the nitrogen-fixing activity within the anaerobic layer relative to the aerobic layer. Brooks et al. (1971) was unable to detect nitrogenase activity in the flocculent, unconsolidated 1-2 cm surface layer of an estuarine sediment. Nitrogen-fixing activity was present below this layer; and the measured rates were higher at the 2-5 cm depth than at the 5-20 cm depth. Fixation was also observed to decrease with depths down the profile of lake sediments (Keirn and Brezonik, 1971) and unvegetated tidal creeks in a Massachusetts salt marsh (Teal et al., 1979). Evidence indicates that bacterial fixation is probably greater in the anaerobic than in the aerobic layer of flooded soils. Rice and Paul (1972) attributed the nitrogen fixation in a waterlogged soil-wheat straw system to microorganisms in the anaerobic rather than the aerobic layer. Sediment cores from a Scottish estuary exhibited higher fixing activity when incubated under anaerobic than under aerobic conditions (Herbert, 1975). Hartwig and Stanley (1978) found significant nitrogenase activity in coastal sediments that were reduced and high in organic matter. However, Atlantic deep-sea sediments that were oxidized and low in organic matter exhibited no detectable nitrogen-fixing activity. The soil factors influencing heterotrophic nitrogen fixation in soil are discussed in Section VII. In a recent symposium at IRRI, Matsuguchi (1979) stated that current interest in nitrogen fixation in rice fields has focused largely on blue-green algae in the photic zone and heterotrophic bacteria associated with rice roots, but little attention has been paid to heterotrophic fixation in the portion of the soil layer without roots. He believed that the significance of heterotrophic bacterial fixation in soil zones free of rice roots has been underrated, and he presented data indicating that the soil layer free from roots was the most important site for nitrogen fixation in a Japanese rice field. Additional research in Japan by Wada et al. (1978) confirms the conclusions of Matsuguchi (1978). Wada et al. (1978) found greater nitrogen-fixing activity in the anaerobic soil layer without roots than in either the rice rhizosphere or the upper 0-2 cm layer of soil and the floodwater. Algal growth in the field studied by Wada et al. (1978), unlike those examined by many others (Watanabe, 1977b, 1978a,b), was not sufficient to contribute significant fixed nitrogen. More recently, Wada et al. (1979) suggested that organic debris may be an important microsite for heterotropic nitrogen fixation in paddy soils. Nitrogen fixation in flooded soils and sediments has been predominantly attributed to the anaerobe Clostridium (Rice et al., 1967; Brooks e? al., 1971; Rice and Paul, 1972), but other anaerobes, such as Desulfovibrio (Herbert, 197% and facultative anaerobes, such as Klebsiella and Enterobacter (Werner et al., 1974), may also be important. The occurrence of anaerobic and facultative nitrogen fixers in nature was recently reviewed by Knowles (1977). In summary, nitrogen fixation by free-living bacteria in unamended soil is

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

163

generally presumed to be low compared to fixation by blue-green algae. Evans and Barber ( 1977) estimated that free-living Clostridium pasteurianum fix from 0.1 to 0.5 kg N/ha/yr, whereas free-living blue-green algae fix an estimated 25 kg N/ha/yr. Even though the fixation rate by heterotrophic bacteria in flooded soil systems may be low, freshwater, estuarine, and marine habitats on the Earth constitute such a large area that the total amount of nitrogen fixed over a period of time by heterotrophic bacteria may be significant relative to that fixed by blue-green algae.

V. NITROGEN FIXATION IN THE ROOT ZONE OF NONNODULATED PLANTS Nitrogen-fixing heterotrophic bacteria in nonsymbiotic association with the roots of various nonnodulated plants are important agents of nitrogen fixation. Dobereiner et al. ( 1972) reported that the association of a bacteria, Azotobacter paspali, with the roots of a tropical grass, Paspalum notatum, fixed an estimated 90 kg N/ha/yr. De-Polli et al. (1977) confirmed both nitrogen fixation by bacteria associated with the roots of tropical grasses and plant assimilation of the fixed nitrogen with “N-labeled nitrogen gas. Nonsymbiotic nitrogen fixation by bacteria has also been found in the rhizosphere of rice (Rinaudo er a l . , 1971; Yoshida and Ancajas, 1971, 1973a,b; Dommergues eral., 1973) and freshwater (Bristow, 1974; Silver and Jump, 1975), marine (Patriquin and Knowles, 1972), and salt marsh angiosperms (Jones, 1974; Patriquin, 1978a; Patriquin and McClung, 1978). The soil-root interface can be divided into three regions: ( I ) the outer rhizosphere comprising the soil immediately surrounding the root (2) the rhizoplane or the actual root surface, and (3) the inner rhizosphere (also called the endorhizosphere or histosphere) comprising the cortical tissue of the root. Each region contains heterotrophic bacteria, some of which are able to fix nitrogen (Dommergues and Rinaudo, 1979). The rice rhizosphere is apparently inhabited mainly by microaerophilic nitrogen fixers rather than facultative or strict anaerobes (Trolldenier, 1977; Watanabe et al., 1979). The high proportion of microaerophiles is presumably a result of an oxidized zone surrounding the roots. Research on the salt marsh macrophyte S . alterniflora revealed microaerophilic and anaerobic nitrogen-fixing bacteria in the rhizosphere, but the microaerophiles were associated mainly with the endorhizosphere rather than with the rhizoplane (Patriquin, 1978a; Patriquin and McClung, 1978). Larger populations of nitrogen-fixing bacteria are present in rhizosphere than in nonrhizosphere soil in rice (Balandreau et al., 1975), the sea grass Thalassia testudinum (Patriquin and Knowles, 1972), and S . alternzjlora (Patriquin and McClung, 1978). Nitrogenase activity in salt marshes has been shown to approximate the root distribution of S . alternijlora (Hanson, 1977b; Teal et al., 1979;

164

R. J . BURESH ET AL.

Casselman, 1979). Researchers have suggested that root exudates of rice (Yoshida and Ancajas, 1973), mangroves (Zuberer and Silver, 1978), and S. alterniflora (Hanson, 1977b) supply nitrogen-fixing rhizosphere bacteria with an energy source. Balandreau et al. (1975) speculated that the response of nitrogen fixation in the rice rhizosphere to illumination results from root exudation promoted by increased plant photosynthesis. Root exudates are known to stimulate microbial growth and activity in the rhizosphere of plants (Rovira, 1965). A. RICE

Yoshida and Ancajas (1973a,b) observed more nitrogen-fixing activity in soil planted with rice than in unplanted soil (see Table I for fixation rates reported by these authors). Lee et al. (1977a) found greater nitrogenase activity in the soil of planted areas of a rice field than in the soil of unplanted areas between plant rows. Use of 15N-labeled nitrogen gas in a sealed chamber has confirmed the presence of nitrogen fixation in the root zone and incorporation of the fixed nitrogen into rice plants (Eskew et a l . , 1979; Ito et a l . , 1980). Watanabe and Barraquio (1979) concluded that most bacteria present in the rice rhizosphere are nitrogen fixers, but many of them are overlooked because they require low levels of mineral or organic nitrogen for growth. Bacteria that have been isolated from rice roots include Azotobacter (Balandreau et al., 1975), Beijerinckia (Diem et a l . , 1978), Enterobacter (Watanabe et a l . , 1977b), and Spirillum (Kumari et a l . , 1976; Nayak and Rao, 1977) among others (Balandreau et a l . , 1975; Watanabe et a l . , 1977b). Nitrogenase activity in the rhizosphere varies with location on the root. More activity is associated with the older basal portion than with the younger part near the tip (Watanabe and Lee, 1977); Diem et al., 1978; Panichsakpatana et a l . , 1979). Some estimated rates of nitrogen fixation by microorganisms associated with rice are given in Table I. Early measurements of associative nitrogen fixation in the rice rhizosphere were based on assays for nitrogenase activity of excised roots (Yoshida and Ancajas, 1971, 1973a), but more recent research has attempted to measure in situ nitrogen-fixing activity with enclosures that cover the plants (Lee et a l . , 1977a; Lee and Yoshida, 1977; Watanabe et a l . , 1977b). The effectiveness of in situ assays with the acetylene reduction method is limited by the transfer of acetylene and ethylene between the atmosphere and saturated soil. Lee and Watanabe (1977) indicated that acetylene in the atmosphere is transported through the rice plant to the root zone and ethylene is transported upward from the root zone, but this process can be limited by closure of the plant stomata (Rinaudo et al., 1977). The in situ field methods probably measure fixation only in the root zone and not the surrounding anaerobic soil because acetylene and ethylene diffusion through saturated soil is slow (Watanabe et a l . , 1978a). Pres-

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

165

ent measurements by in situ techniques suggest that nitrogen fixation by heterotrophic bacteria in the root zone may be less important than previously indicated. Watanabe et al. (1977b) estimated that this process contributed 0.05 kg N/ha/day in a rice field at IRRI. Watanabe et al. (1978b) considered this process less significant than algal fixation, although this may not always be the case because algal populations in rice paddies can fluctuate greatly with time and location. The recent observation that nigrogen-fixing activity associated with rice plants grown in water culture is higher than that measured by the in situ assay suggests that field measurements of associative fixation in the rice rhizosphere might be low (IRRI, 1977). Nitrogen-fixing activity in the rhizosphere varies between different varieties of rice (Lee and Yoshida, 1977; Rinaudo et al., 1977) and also exhibits both seasonal and diurnal fluctuations. Lee et al. (1977b) observed an increase in nitrogen-fixing activity from 4 weeks after transplanting until the heading stage and then a rapid decrease. Panichsakpatana et al. (1979) found an increase in nitrogenase activity in soil adhering to roots with an increase in age of the plant until the flowering stage. Balandreau et al. (1974) reported a peak in nitrogenase activity near midday. However, they employed an in situ technique without elimination of algal activity, and Watanabe and Cholitkul (1979) suspected that some of the measured nitrogenase activity could be attributable to algae. Trolldenier (1977) found maximum activity in the afternoon at 1500 hours on root-soil samples. The dependence of the nitrogen-fixing activity on light intensity suggests a photosynthetic effect on nitrogen fixation in the rhizosphere (Dommergues et al., 1973). Nitrogen fixation in the root zone is possible only if nitrogen gas is present. Rice plants have long been known to contain a transporting system whereby atmospheric oxygen is supplied to the roots (Van Raalte, 1941). Yoshida and Broadbent (1975) confirmed by use of I5N that atmospheric nitrogen is also transported by rice to its roots where it diffuses into the rhizosphere for fixation by bacteria. The rate of nitrogen gas transport by the rice plant was greatest at the heading and flowering stages (Yoshida and Broadbent, 1975), the growth stages which correspond to greatest nitrogen-fixing activity in the rhizosphere (Yoshida and Ancajas, 1973a). The ability of rice to transport atmospheric nitrogen gas to the root zone is also suggested by the observation of Yoshida et al. (1975) that planted rice soils contained more nitrogen gas than unplanted soils. B. OTHER PLANTS

A wide range of estimates have been reported for nitrogen fixation associated with the rhizosphere of Thalassia testudinurn, the predominant sea grass in the tropical and subtropical regions of the Atlantic Ocean (see Table 111). Patriquin

Table III Estimated Rates of Nitrogen Fixation by Microorganisms in the Root Zone of Nonnodulated Plants Plant

Location

Thahssia tesrudinwn

Barbados, West Indies

Florida, U.S. Florida, U.S.

-8

Spanina alterniflora

New York, U.S.

Nova Scotia, Canada

Georgia, U.S.

Method of estimation Plant tissue and rhizosphere sediment, CzHz Roots and rhizomes, CzHz, January Sediment, washed plant tissue, C,H, Rhizosphere sediment, tall growth form, C,H,, summer Rhizosphere sediment, short growth form, CZH2.summer In situ plant associated activity, CZHZ. summer, 5 sampling times Washed roots and rhizomes, C,H, Rhizosphere soil, CZH2.summer Rhizosphere sediment, 0-22 cm depth, short growth form, C,H,

Rate"

Reference

100-500 kg N/ha/yr

Patriquin and Knowles, 1972

0-0.08 ng/&r

McRoy et al., 1973

12-34 kg N/ha/yr

Capone and Taylor, 1978

22.8-282 pg N/mz/hr (average = 116 p g N/mz/hr)

Whitney et al., 1975

31.1-99 pg N/m% (average = 65.1 p g N/m2/hr) 0-146 pmoles C,HJm2/hr

Patriquin and Denike, 1978

(0-1.36 mg NlmVhr)

76-234 pmoles C,HJmPihr (0.71-2.18 mg N/mzihr) 173-707 pmoles C,HJm*/hr (1.61-6.60 mg N/m2/hr) 21.6-51.4 g N/mz/yr (216-514 kg N/ha/yr)

Hanson, 1977L1

Massachusetts, U.S.

Louisiana, U.S.

Nova Scotia. Canada

-

Glycerin borealis

Maryland, U.S. Ontario, Canada

TYPh SP.

Ontario, Canada

Typha larifolia

Massachusetts, U.S.

Juncur balticur

Massachusetts, U.S.

Carex scoparia

Oregon, U.S. Massachusetts, U.S.

Lysimachia terrestris

Massachusetts, U.S.

Lythrum salicaris

Massachusetts, U.S.

Scirpus polyphyllus

Massachusetts. U.S.

21 m

a

Rhizosphere sediment, short low marsh, C,H, Rhizosphere sediment, tall low marsh, CZH2 Rhizosphere sediment, streamside marsh, CzHz Washed roots and rhizomes, rhizosphere sediment, C,H, ( 4 1 factor), growing season Washed roots, C,H2 Roots and rhizomes, C,H,, anaerobic Roots and rhizomes, C2H2.anaerobic Cores, CZHZ,aerobic, August Cores, C2Hz. aerobic, August Cores, CZHZranaerobic Cores, C,H,, aerobic, August Cores, C,H,, aerobic, July Cores, C2H2,aerobic, JdY Cores, C2H2. aerobic, July

121 kg N/ha/yr

Teal et al., 1979

84 kg N/ha/yr 109 kg Nihalyr

93 kg N/ha/growing Season

6.0 t- 2.7 nmoles C,H,/g/hr 37 pmoles C,HJg/day

Casselman, 1979

Pabiquin and McClung, 1978

Van Berkum and Sloger, 1979 Bristow, 1974

(60kg N/ha/yr) 3.6 pmoles C,HJg/day (34 pg N/g/day) 181 2 33 g N/ha/day 65

5

9 g Niha/day

Bristow, 1974 Kana and Tjepkema, 1978 Kana and Tjepkema, 1978

800 g N/ha/day 177 ? 58 g Nhalday

Tjepkema and Evans, 1976 Kana and Tjepkema, 1978

20 t- 5 g N/ha/day

Kana and Tjepkema, 1978

27 ? 4 g Nihalday

Kana and Tjepkema, 1978

100 t- 16 g N/ha/day

Kana and Tjepkema, 1978

Unless otherwise indicated a factor of 3 was used to convert moles of acetylene reduced to moles of dinitrogen gas fixed.

168

R. J . BURESH ET AL.

and Knowles (1972) estimated a fixation rate of 100-500 kg N/ha/yr for Thalasfrom assays with excised rhizomes and roots. McRoy et al. (1973) found much lower levels of nitrogen fixation associated with Thalassia and did not support the belief of Patriquin (1 972) that inorganic nitrogen in the root zone of this plant was derived entirely from nitrogen fixation. McRoy et al. (1973) employed shorter incubation periods and lower incubation temperatures than Patriquin and Knowles (1972). Capone and Taylor (1978) recently presented a revised estimate of 12-34 kg N/ha/yr for nitrogen fixation in the rhizosphere of Thalassia. They found the largest nitrogenase activity in the rhizosphere sediment near midday. Jones (1974) reported that Spartina anglica and Puccinellia maritima in an English salt marsh stimulated nitrogen fixation in surrounding soil. Studies with 15N2revealed that fixed nitrogen was incorporated into plant tissue. Casselman ( 1979) observed significantly higher nitrogenase activity in soil vegetated with S. alternij7ora than in unvegetated soil in a Louisiana salt marsh. Nitrogenase activity in samples of soil with associated roots from this marsh were highest during the plant flowering stage in late summer and fall. Nitrogenase activity was also greater in the samples from the upper 10 cm soil layer than in those from the 10-20 cm layer. The results of Casselman (1979) indicate that the root zone of Spartina was the most important site for nitrogen fixation in this marsh. Valiela and Teal (1979) also indicated that bacteria associated with the plant rhizosphere were the major agents of nitrogen fixation in a Massachusetts salt marsh. Patriquin and McClung (1978) found that Spartina roots more than 2 years old exhibit little or no nitrogenase activity. This finding indicates that nitrogen fixation in the root zone is dependent upon plant metabolism. Nitrogenase activity in vegetated soil is significantly greater in streamside than in inland salt marshes in Louisiana (Casselman, 1979). The streamside marshes support more vigorous and productive stands of Spartina than the inland marshes (DeLaune er al., 1979). This observation that the most productive stands of plants had the largest nitrogen-fixing activity in the root zone further suggests a dependence of fixation in vegetated soil upon plant metabolism. Patriquin and Keddy ( 1978) measured nitrogenase activity associated with excised roots of 33 angiosperms and soil-plant cores from a salt marsh in Nova Scotia, Canada. The observed nitrogen-fixing activities were variable but generally indicate the presence of low to moderate levels of nitrogenase activity associated with roots throughout the marsh. The authors concluded that nitrogenase activity associated with roots of these temperate-latitude salt marsh angiosperms was apparently not influenced by whether the plant possessed a C, or C, pathway of photosynthetic metabolism. Kana and Tjepkema (1978) determined nitrogenase activity in soil cores containing nonnodulated herbaceous plants from well-drained and wetland habitats in Massachusetts and found that the measured nitrogen-fixing activities, presumably associated with the plant root zone, were considerably higher for six wetland plant species than for six species

sia

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

I69

from well-drained fields. Nitrogen-fixing activity has also been observed in the root zone of mangroves (Zuberer and Silver, 1978) and a number of other aquatic plants (Bristow, 1974; Silver and Jump, 1975; Tjepkema and Evans, 1976). Reported rates of nitrogen fixation in the root zone of various plants are listed in Table 111. C. OVERVIEW

Comparison and interpretation of reported values for nitrogen fixation in associative systems is difficult because different experimental methods and assay conditions have been employed (Evans and Barber, 1977; Knowles, 1977). Many estimates of nitrogen fixation associated with rice (Yoshida and Ancajas, 1971, 1973a), marine plants (Patriquin and Knowles, 1972), and marsh plants (Patriquin, 1978a; Patriquin and Keddy, 1978) have been made on excised plant roots, and only recently have efforts been made to measure in situ nitrogenfixing activity (Watanabe et a l . , 1977b; Patriquin and Denike, 1978). Activities obtained on excised roots do not necessarily represent actual rates of nitrogen fixation in the field. Evans and Barber (1 977) concluded that many reported fixation values, especially those for tropical grasses, were suspect because the researchers preincubated plant roots before exposure to acetylene. This technique has been shown to increase the population of nitrogen-fixing bacteria associated with the roots and thereby overestimate nitrogen fixation (Barber et a l . , 1976; Koch, 1977). Gaskins and Carter (1975) in a review article evaluating methods for measurement of nitrogenase activity recommended cautious use of any procedure that determines nitrogen-fixing activity in detached plant parts. They concluded that such assays were of only qualitative interest. The strong inhibitory effect of chloramphenicol, an inhibitor of protein synthesis, on nitrogenfixing activity of excised rice roots suggests that nitrogen-fixing microorganisms proliferate during the assay period (Watanabe and Cabrera, 1979). Patriquin ( 1978b) suggested that maximum nitrogenase activity in excised roots of S . alterniflora incubated for 1-2 days under a low level of oxygen could be considered an estimate of “potential nitrogenase activity” rather than in situ nitrogenase activity. Nitrogenase activity associated with washed plant roots incubated under low levels of oxygen was greater than that measured with in situ assay systems (Patriquin and Denike, 1978). The very high fixation rates of 100-500 kg Nlhalyr for Thalassia determined by Patriquin and Knowles (1972) from incubations of excised roots for 1 or more days, therefore, are very likely estimates of potential rather than actual in situ nitrogen fixation. In situ techniques are nonetheless not without faults. Poor gaseous exchange between the atmosphere and root zone within the enclosures employed for in situ measurements can result in an underestimation of nitrogen fixation. Flooded soil systems are favorable environments for nitrogen fixation by

170

R. J . BURESH ET AL.

heterotrophic bacteria associated with the plant root zone, and indeed higher rates of associativefutation have been observed in the rhizosphereof wetland than dryland plants (Yoshidaand Ancajas, 1973a;Kana and Tjepkema, 1978).The high moisture content and associated low oxygen tension as well as the supply of root exudates are features of flooded soil systems that favor associative nitrogen fixation in the rhizosphere. Transport of atmospheric oxygen by wetland plants to their roots can create aerobic microzones in the rhizoplane and the soil surrounding the roots (Van Raalte, 1941; Teal and Kanwisher, 1966). Therefore, the concept of an aerobic-anaerobic interface as a zone for stimulated nitrogenase activity, which was discussed in Section 111, also applies to the plant rhizosphere. The large area of the aerobic-anaerobic interface, the supply of carbon and energy substrates from roots that are available to microorganisms (Rovira, 1965), and the possibility for associations between nitrogen fixers in the rhizosphere with other nitrogen fixers or nonfixing microorganisms (Dommergues and Rinaudo, 1979) suggest that the root zone is a privileged habitat for nitrogen fixation. Although nitrogen fixation by bacteria in the rhizosphere might be less than that of blue-green algae in some rice fields, it is nonetheless quantitatively significant (Knowles, 1977).

VI. NITROGEN FIXATION ON THE LEAF AND STEM SURFACE OF PLANTS Ruinen ( 1970, 1974) observed nitrogen-fixing microorganisms in the leaf surface habitat (the phyllosphere) of grasses and concluded that leaf and stem surfaces of selected plants are a site for nitrogen fixation. Subsequent studies with wetland plants have reported nitrogenase activity in the phyllosphere of sea grasses (Goering and Parker, 1972; Capone and Taylor, 1977), freshwater macrophytes (Finke and Seeley, 1978), and salt marsh macrophytes (Green and Edmisten, 1974; Casselman, 1979). A. PLANTSOTHERTHANRICE

The leaf and stem surfaces of both living and decaying S. alternijlora, a common salt marsh macrophyte, are sites for nitrogen fixation. Nitrogenase activity is greater in the phyllosphere of dead plants than in that of living plants (Green and Edmisten, 1974; Patriquin, 1978a; Casselman, 1979) (see Table IV). Green and Edmisten (1974) reported an extremely high nitrogenase activity associated with old S. alternijloru stems in a Florida marsh during the month of July. They estimated that potential nitrogen fixation associated with the aboveground portion of S. alternijlora was equivalent to about 1550 kg Nlha during

Table IV Estimated Rates of Nitrogen Fixation by Microorganisms on the Leaf and Stem Surface of Plants Plant Thalarsia testudium

Spartina alterniflora

Location

Method of estimation

Florida, U.S. Texas, U.S.

Leaves, C,H,, January Leaves, C,H,, August

Florida, U.S.

Leaves, C,H,

BahalXX3.S

Leaves, C,H,

Georgia, U.S.

Leaves, C,H,

Nova Scotia, Canada

Leaves, C2H2

Old stems, C,H,

Rhizophora mangle

Florida, U.S.

Myriophyllum spicatum

New York, U.S.

Living leaves and stems, CZH2.yearly average Dead leaves and stems, C,H,, yearly average Leaves decayed for 2-3 weeks, C,H,, light, anaerobic Leaves, C,H,

Maryland, U.S.

Aboveground plant, C,H,

Maryland, U.S.

Aboveground plant, C,H,

Louisiana, U.S.

Potamogeton perfoliatus

Rate" 0-0.4 ng N/g/hr 6-32 mg N/m2/hr (90-470 mg N/rnz/day)b 0-20 kg N/ha/yr (0-5 mg N/m2/day) 1.15 and 3.20 mg N/m2/day 0.2-0.5 g N/m2/yr (2-5 kg N/ha/yr) 0.52 and 5.85 nmoles CzHJg/hr (4.9 and 54.6 ng N/g/hr) 24.2 nmoles C,HJg/hr (226 ng N/&) 0.03 g NlmVyr (0.3 kg N/hdyr) 0.66 g N/m2/yr (6.6 kg N/ha/yr) 11 P8 N/&

McRoy et al., 1973 Goering and Parker, 1972

0.15-0.60 nmoles/mg/hr (1.4-5.6 PLg N/g/hr) 29 ng N/g/hr (August) 16 ng N/& (September) 30 ng N/g/hr (August) 3 ng N/g/hr (September)

Finke and Seeley, 1978

In all cases a factor of 3 was used to convert moles of acetylene reduced to moles of dinitrogen gas fured. bThe calculation was based on a 12-hour light period and a 12-hour dark period. a

Reference

Capone and Taylor, 1977 Capone et al.. 1979 Hanson, 1977b Patriquin, 1978a

Casselman, 1979

Gotto and Taylor, 1976

Lipschultz et al., 1979 Lipschultz et al., 1979

172

R. J . BURESH ET AL.

July. The validity of this value is questionable because it is several hundred times greater than comparable values reported by Patriquin (1978a) in Nova Scotia and Casselman (1979) in Louisiana, and it is much larger than reported fixation rates for all components of marsh ecosystems (Carpenter et al., 1978; Casselman, 1979). Nitrogen fixation associated with the phyllosphere of S. alternij7ora fluctuates seasonally. Rates reported in a Florida marsh peaked in July and August and were near negligible in winter (Green and Edmisten, 1974). Casselman (1979) observed higher epiphytic nitrogen fixation on both live and dead plants during the spring and fall than at midsummer. Nitrogen-fixing blue-green algae (Hanson, 1977b) and photosynthetic bacteria (Green and Edmisten, 1974; Hanson, 1977b) have been observed on the leaves and stems of S. alterniflora. Nitrogenase activity in the phyllosphere of sea grasses in Redfish Bay, Texas (Goering and Parker, 1972) and near the Florida coast (Capone and Taylor, 1977) has been attributed to epiphytic blue-green algae, particularly Calothrin. The presence of blue-green algae on the leaves of the sea grass T. testudium was apparently necessary for significant nitrogen-fixing activity in the phyllosphere (McRoy et al., 1973; Capone and Taylor, 1977). Nitrogen fixation associated with the leaves of Thalassia was greater during the day than during the night (Capone and Taylor, 1977). Capone et al. (1979) speculated that nitrogen fixed in the phyllosphere of Thalassia contributed primarily to leaf epiphyte rather than macrophyte production. Gotto and Taylor ( 1976) reported nitrogenase activity associated with decaying leaves of red mangrove (Rhizophora mangle). The observed activity was mainly light dependent and anaerobic, hence the authors concluded that photosynthetic bacteria and blue-green algae were the predominant nitrogen fixers, whereas facultative and anaerobic bacteria accounted for a smaller portion of the fixation. Epiphytic nitrogen fixation on the decaying leaves and stems of wetland plants may increase the nitrogen content of the plant material. This is of major significance because decaying plant tissues are degraded to organic detritus which forms the base of the food web in many aquatic ecosystems (Odum, 1971). Finke and Seeley ( 1 978) showed that microorganisms capable of fixing nitrogen are associated with the leaves of Myriophyllum spicatum, a submerged freshwater macrophyte. Blue-green algae were apparently the major nitrogen fixers, but nitrogen-fixing photosynthetic bacteria were isolated from the phyllosphere of Myriophyllurn. Nitrogen-fixing activity of the epiphytic microorganisms was maximal between 1200 and 1800 hours, but some fixing activity was present throughout the night. Fixation was greater in the summer than in the winter months. Purchase (1977) reported that nitrogen fixation by Azotobacter on the leaves of Eichhornia crassipes (water hyacinth) was negligible. Conditions of extreme nitrogen deficiency were required for nitrogen fixation by Azotobacter associated with Eichhornia, and even then the nitrogen-fixing activ-

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

173

ity was small. Reported rates of nitrogen fixation by microorganisms on the leaf and stem surface of various plants are given in Table IV. B. RICE

Lee el al. (1977a) concluded that nitrogen fixation in the phyllosphere of rice probably was negligible because detachment of the aerial parts of the rice plant did not decrease the nitrogen-fixing activity. Watanabe et al. (1978b) observed nitrogenase activity, presumably associated with blue-green algae, on the lower portion of rice stems in a flooded field. Fixation in the phyllosphere, excluding the stem part up to 10 cm from the soil surface, was negligible. In more recent research Watanabe el ul. (1 979) showed that the lower portion of the rice stem was inhabited by nitrogen-fixing heterotrophic bacteria. Use of 15N2confirmed that the lower portions of rice stems, including the outer leaf sheath, inner leaf sheath, and basal node, are sites for nitrogen fixation (Ito et a l . , 1980). Nitrogen fixation associated with the rice stem can complicate attempts to measure fixation associated with the plant rhizosphere (Watanabe et a l . , 1978b, 1979). Ratoons sprouting from rice stubble exhibited higher rates of nitrogenase activity than either uncut plants or stubble in which ratooning was prevented (Watanabe et a l . , 1977b). Watanabe and Cholitkul (1979) attributed the high nitrogen-fixing activity associated with the ratoon to blue-green algae attached to the ratooning hill. C . Azolla-Anabaena ASSOCIATION

The blue-green alga Anabaena azollue in symbiotic association with the freshwater fern Azolla fixes nitrogen. Azolla is widely distributed on the surface of paddy fields and water bodies within tropical and temperate regions (Moore, 1969). The algae inhabit the leaf cavity of Azolla and obtain a degree of physical protection and probably metabolites from the fern (Moore, 1969; Peters and Mayne, 1974). Peters and Mayne (1974) confirmed that the alga was the agent of nitrogenase activity. The algal-free fern cannot fix nitrogen and requires a combined nitrogen source for growth (Ashton and Walmsley, 1976). Azoffa containing Anabaena is able to assimilate atmospheric nitrogen. In fact, the fern assimilates nitrogen more efficiently from the algal symbiont than from the medium on which it is grown (Ashton and Walmsley, 1976). Growth of Azolla can generally be promoted with phosphorus fertilization (Watanabe et al., 1977a; Singh, 1979). Azolla is able to cover quickly the surface of the floodwater in rice fields without interference to normal cultivation practices. Actively growing Azolla can

174

R. J . BURESH ET AL.

double its mass within a few days. Watanabe et al. (1977a) found doubling times of about 5- 10 days in a rice field; the doubling times were even shorter for plants grown in culture solutions in the greenhouse. Growth of the fern in rice fields and water bodies is usually not continuous throughout the year. For example, a depressed growth rate attributed to high temperature and high light intensity was observed from April through June at IRRI (IRRI, 1977). The life cycle of Azolla and the influence of environmental factors upon its growth has been reviewed by others (Moore, 1969; Ashton and Walmsley, 1976; Becking, 1979; Peters et al., 1979). The Azolla-Anabaena association is of agronomic interest as a green manure and possible partial substitute for nitrogen fertilizer in rice fields. Becking (1972) reported a maximum nitrogen input of 335-670 kg N/ha/yr for the symbiotic association, but Becking (1979) later mentioned that a more reasonable estimate for annual nitrogen fixation by the association under natural conditions in a rice field was 103-162 kg N/ha. Research at IRRI in the Philippines (IRRI, 1975) revealed a nitrogen-fixing activity of about 1 kg N/ha/day for the AzollaAnabaena association when the fern nearly covered the surface of a paddy field. Later research (IRRI, 1977) found a production of 330 kg N/ha by Azolla in 220 days. Kellar (1979) estimated an annual nitrogen input of 164 kg N/ha by the Azolla-Anabaena association in a New Zealand lake. This rate of nitrogen fixation exceeded that attributed to free-living Anabaena. A single measurement of diurnal nitrogenase activity by Kellar (1979) indicated that 24% of the total daily fixation associated with Azolla in the lake occurred at night.

VII. ENVIRONMENTAL FACTORS INFLUENCING NITROGEN FIXATION IN FLOODED SOIL A. ENERGY SOURCE

The availability of carbon compounds as an energy source appears to be the primary factor limiting growth of photosynthetic nitrogen-fixing bacteria because these microorganisms must obtain their energy from carbon compounds synthesized by other organisms (Stewart, 1969; Hanson, 1977a). Many carbon substrates, including glucose (Okafor and MacRae, 1973; Jones, 1974; Zuberer and Silver, 1978), sucrose (Keirn and Brezonik, 1971; Marsho et al., 1975), and rice straw (Rao, 1976; Matsuguchi, 1977; Reddy and Patrick, 1979), have been shown to stimulate nitrogen fixation in soils and sediments. The addition of rice straw to a paddy soil can increase the population of Azotobacter (Araragi and Tangcham, 1979). Matsuguchi et al. (1975) in a survey of rice soils in Thailand observed that soils rich in organic matter generally supported higher populations of heterotrophic nitrogen-fixing bacteria. Large applications of organic matter

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

175

can depress nitrogen-fixing activity; Rao (1976) and Yoneyama et al. (1977) observed retarded fixation in waterlogged soils amended with 2% rice straw as compared to lower levels of amendment. They speculated that compounds toxic to microorganisms could have formed during the decomposition of large quantities of applied rice straw. Matsuguchi (1977) suggested that the addition of nitrogen fertilizer to rice fields receiving rice straw can further enhance nitrogen fixation. The added nitrogen apparently stimulates rice straw decomposition thus resulting in an accumulation of available energy source for nitrogen-fixing microorganisms. The excretion of organic compounds from plant roots helps make the rhizosphere a favorable environment for heterotrophic nitrogen fixation. Nitrogenfixing bacteria are able to use these exudates as an energy source (Rovira, 1965). Periods of high C/N plant residue degradation or high C/N root exudate production favor nitrogen fixation by heterotrophic bacteria (Knowles, 1977). Photosynthetic nitrogen-fixing bacteria and blue-green algae normally do not rely on an exogenous carbon supply as long as photosynthesis occurs (Stewart, 1969). However, blue-green algae are capable of heterotrophic nitrogen fixation (Fay, 1976).

B. NITROGEN It is well documented that inorganic nitrogen inhibits nitrogen fixation. However, results in the literature are very diverse; they indicate that inhibition of nitrogen fixation by inorganic nitrogen does not always occur and that the inhibitory effect is dependent upon the amount of inorganic nitrogen and perhaps other factors in the soil such as the nitrogen-fixing microflora. Patriquin and Keddy (1978) found that nitrogenase activity associated with the root zone of angiosperms in a salt marsh was inversely correlated with the ammonium concentration in the groundwater. Teal et al. (1979) indicated that high levels of ammonium in marsh soil receiving fertilizer can inhibit bacterial nitrogen fixation in the soil, but they speculated that the levels of interstitial ammonium normally present in unfertilized marsh soil were probably not sufficient to inhibit nitrogen fixation. Wada et al. (1978) and Panichsakpatana et al. (1978) were unable to find a correlation between ammonium levels and nitrogen-fixing activity in Japanese rice soil. However, Casselman (1979) found an inverse relationship between concentration of water-soluble plus exchangeable ammonium and nitrogenase activity in vegetated salt marsh soils. She concluded that the soil zones in this vegetated marsh with the lowest levels of native ammonium were the most favorable for nitrogen-fixing activity. Application of either ammonium sulfate or potassium nitrate at rates of 50 and 200 p g N/g during the early phases of rice plant development were found to

176

R. J . BURESH ET AL.

inhibit nitrogen fixation significantly (MacRae, 1975). Yoshida et al. (1973) observed nearly complete inhibition of nitrogen fixation with 160 pg/g of applied inorganic nitrogen in a rice soil, and Van Raalte et al. (1974) found inhibition of fixation by addition of sewage sludge and urea to a salt marsh soil. Balandreau et al. (1975) reported that application of up to 40 p g ammonium sulfate-N/g stimulated nitrogenase activity in the rice rhizosphere, but higher additions resulted in decreased activity. Midseason application of nitrogen fertilizer to rice has been reported to result in an initial decrease in nitrogen-fixing activity in the root rhizosphere followed by a period of enhanced activity (Gilmour et al., 1978). The repression of nitrogenase activity after fertilizer addition might be only temporary because rice plants can rapidly assimilate ammonium in the soil (Trolldenier, 1977). Jurgensen ( 1973) speculated that added fertilizer might stimulate nitrogen fixation indirectly through enhancement of higher plant production which subsequently can result in a greater supply of root exudates and residues for heterotrophic nitrogen fixers. Hanson (1977b) reported that monthly additions of ammonium nitrate to a salt marsh soil stimulated nitrogen-fixing activity presumably by enhancing plant production and root exudation. Nayak and Rao (1977) showed that Spirillum isolated from the roots of rice plants receiving ammonium sulfate had greater nitrogen-fixing efficiency than those isolated from roots of plants that received no added nitrogen. These authors found that the application of high levels of inorganic nitrogen to rice suppressed the nitrogen-fixing potential of Spirillum. Roots of plants exposed to 20 and 40 kg N/ha generally harbored Spirillum possessing a higher nitrogen-fixing efficiency than those from plants receiving higher levels of nitrogen. In summary, ammonium in sufficiently high levels definitely does inhibit nitrogen fixation. The stimulatory effect of low levels of ammonium on nitrogen-fixing activity in flooded soil systems containing higher plants can be somewhat misleading. The actual concentration of ammonium in the plant rhizosphere in these cases is not known, but it might very likely be considerably less than the concentration in the soil outside the rhizosphere because available ammonium is rapidly assimilated by plant roots. The stimulatory effect of ammonium is probably indirect, whereby the ammonium results in enhanced plant production and greater subsequent exudation of carbon compounds that can serve as an energy source for nitrogen fixers in the rhizosphere. Stewart (1969) speculated that the levels of combined nitrogen in most natural ecosystems are typically too low to inhibit nitrogen fixation by blue-green algae. Liao (1977) found that nitrogen-fixing algal species were dominant in lake water receiving no nitrate enrichment, but that algal species not able to fix nitrogen became predominant after the addition of nitrate. Yoshida et al. (1973) observed that although addition of nitrogen fertilizer to rice soil increased algal growth, more blue-green algae were generally present in unfertilized soil treatments. Recent research by Roger et al. (1980) showed that the surface broadcast appli-

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

177

cation of urea fertilizer to a rice field inhibits nitrogen fixation and favors growth of green algae. However, deep placement of urea supergranules in the soil does not alter the natural algal nitrogen-fixing activity on the soil surface. C . MINERAL NUTRIENTS

A review of the literature by Jurgensen (1973) indicated that the effect of added mineral nutrients upon nonphotosynthetic nitrogen-fixing bacteria has been variable. Response of these microorganisms to fertilizer treatments is probably indirect and associated with stimulated higher plant growth which results in increased plant residues and root exudates for microbial assimilation. Growth of heterotrophic bacteria is probably limited more by a lack of energy source than by a lack of mineral nutrients. Photosynthetic nitrogen fixers, on the other hand, are not normally limited by the availability of carbon compounds since they obtain most of their energy by photosynthesis. The growth of these microorganisms is influenced by soil nutrient status more than the growth of heterotrophic bacteria is (Jurgensen, 1973). Applied phosphorus has been shown to stimulate nitrogen fixation by blue-green algae in rice soils low in phosphorus (De and Mandal, 1956), but researchers at IRRI found greater phototrophic nitrogenase activity in unfertilized plots than in those receiving NPK fertilizer (Watanabe, 1977b, 1978a). Phosphorus levels in aquatic ecosystems have been related to algal growth and nitrogen fixation (Vanderhoef et al., 1974; Liao, 1977). The results of Stewart et al. ( 1 97 1) suggest that blooms of nitrogen-fixing blue-green algae develop in aquatic systems when phosphorus is available and combined nitrogen is low. D. LIGHT

Light-dependent activity in flooded soil systems is attributed mostly to bluegreen algae and in some cases to photosynthetic bacteria. Photosynthetic nitrogen fixers are confined to the photic zone of flooded soils which includes the column of water overlying the soil, the surface of the soil layer, and the leaf and stem surfaces of plants. Carpenter et al. ( 1978) concluded that the availability of light is the main factor affecting nitrogen fixation on the surface of a Massachusetts marsh. These authors found that shading by plants significantly reduced nitrogen fixation by blue-green algae. Diurnal variations in nitrogen-fixing activity of blue-green algae are largely influenced by light intensity (Reynaud and Roger, 1978). Certain strains of blue-green algae can fix small amounts of nitrogen in the dark when an energy source is present (Watanabe and Yamamoto, 1967; Fay,

178

R. J . BURESH ET AL.

1976). Heterotrophic nitrogen fixation under dark conditions by blue-green algae has been generally considered to be insignificant in nature (Alexander, 1977); but Fay (1976) recently suggested that dark fixation of nitrogen by blue-green algae may be greater than previously suspected, and Finke and Seeley (1978) speculated that this reaction may be significant among epiphytic blue-green algae capable of assimilating carbonaceous plant exudates. Nonsymbiotic nitrogen fixation associated with the root zone of plants exhibits diurnal variations, suggesting that it is influenced by light intensity (Dommergues et al., 1973; Trolldenier, 1977). Balandreau et al. (1975) speculated that plant photosynthesis promotes root exudation which in turn acts as a source of carbon and energy for heterotrophic nitrogen fixers in the root zone. E. OXYGEN

Nitrogenase is vulnerable to inhibition by oxygen. Even though the ability to fix nitrogen is distributed among aerobic, facultative, and strict anaerobic bacteria, most studies indicate that nigrogen fixation in soil is greater under anaerobic than under aerobic conditions (Brouzes et al., 1969; Rao, 1976). Aerobic nitrogen-fixers such as Azotobacter require oxygen for metabolism; however, Azotobacter has been shown to fix nitrogen most efficiently at low oxygen tensions (Stewart, 1969). Subatmospheric oxygen tensions appear to favor proliferation of aerobic nitrogen-fixing bacteria in the rhizosphere. Trolldenier (1977) found that an intermediate oxygen level of 3% was more favorable for nitrogenase activity in the rice rhizosphere than either 21 % oxygen or oxygen-free conditions. The aerobic-anaerobic interface in soil has been shown to be an ideal location for nitrogen fixation (Magdoff and Bouldin, 1970; Rice and Paul, 1972). The water content of flooded soils influences the oxygen status of the soil by limiting the movement of oxygen into and through the soil. Heterocystous blue-green algae, the major group of nitrogen-fixing blue-green algae, fix nitrogen under aerobic conditions. However, low oxygen tensions can enhance nitrogen fixation by these microorganisms. Reduced oxygen tension also encourages the development of nitrogenase activity in certain nonheterocystous blue-green algae (Mague, 1977). Nitrogen-fixing photosynthetic bacteria require anaerobic conditions for growth. F. REWX POTENTIAL

Nitrogen fixation within the soil layer of flooded soil systems is greater under reduced than under oxidized conditions. Wada et al. (1978) and Panichsakpatana et al. (1 978) concluded that the development of reduced conditions in the soil is

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

179

the second most important factor, after the availability of an energy source, controlling nitrogenase activity in a rice soil. Yoneyama ef al. (1977) found the greatest nitrogen-fixing activity in a flooded soil within the redox range from -200 to -260 mV. DeLaune et al. (1976) found more nitrogen-fixingactivity in a salt marsh soil at - 250 mV near pH 7 than under more oxidizing conditions at lower pH values. Incubation of salt marsh sediment under controlled redox potential and pH conditions by Casselman ( 1979) indicated that nitrogenase activity was greatest under anaerobic conditions. Trolldenier (1977) found that redox potential influenced nitrogenase activity on the roots of rice grown in nutrient solution. Nitrogen-fixing activity increased to a maximum value with decreasing redox potential and then declined rapidly with further decreases in redox potential. G . pH

Jurgensen and Davey (1970) in a review article concluded that absence of a nitrogen-fixing microflora is not likely the factor limiting fixation in soils. According to these authors many nitrogen-fixing microorganisms, especially Clostridiurn species, are acid-tolerant and widely distributed in nature. Species of Beijerinckia, aerobic nitrogen-fixing bacteria, characteristically occur under acid conditions. Azotobacter and blue-green algae are normally restricted to nearly neutral or slightly alkaline environments; they are generally of little importance in soils below pH 6.0 (Jurgensen and Davey, 1970; Wilson and Alexander, 1979). It is often difficult to distinguish the effects of pH on blue-green algae that are directly due to hydrogen ion concentration from those that are due to other chemical factors such as the solubility of trace nutrients (Fogg et al., 1973). Yoneyama et al. (1977) found that peak nitrogen-fixing activity in a flooded soil amended with either rice straw or cellulose corresponded to pH 7.1-7.5 in the overlying water. Controlled pH and redox potential incubations of salt marsh sediment by Casselman (1979) revealed that nitrogenase activity was greater at pH 6.5 than at either pH 5.0 or 8.0.

H. SALINITY Nitrogenase activity has been observed in soils and sediments from freshwater, estuarine (Herbert, 1975), and marine (Hartwig and Stanley, 1978) flooded soil systems. Nitrogen fixation is an important process for nitrogen input in numerous salt marshes (Valiela and Teal, 1979; Casselman, 1979). These environments are normally influenced by both seawater and freshwater inputs, and the salinity of the water in the soil is usually intermediate between that of freshwater and

180

R. J. BURESH ET AL.

seawater. Patriquin and Keddy (1978) concluded that groundwater salinity in a Nova Scotia marsh is not a factor influencing the nitrogenase activity associated with the roots of the plants. Herbert (1975) examined the effect of salinity upon nitrogen-fixing capacity of bacteria isolated from estuarine sediment. The Clostridium isolates were not affected by salt concentrations up to 0.5 M . The Desulfowibrio species studied had an obligate salt requirement for nitrogen fixation; greatest fixation occurred at salt concentrations between 0.2 and 0.4 M. None of the Azotobacter, Klebsiella, and Enterobacter strains fixed nitrogen in the presence of salt. Werner et al. (1974), on the other hand, isolated Klebsiella and Enterobacter species capable of fixing nitrogen in a saline medium. Jones (1974) found that cultures of blue-green algae, Nostoc species, fix nitrogen optimally at low salinities.

VIII. COMPARISON OF ACETYLENE REDUCTION AND 15N METHODOLOGY Nitrogen fixation is currently measured by the acetylene reduction method and direct fixation of 15N-enriched NO.The uptake of 15N2is the most satisfactory method for measuring nitrogen fixation; but this method is much less sensitive than acetylene reduction, and it requires expensive 15N2and analysis with either a mass spectrometer or an emission spectrometer. Acetylene reduction is the more commonly employed method since it is simple, inexpensive, rapid, and very sensitive. Acetylene is reduced by nitrogenase to ethylene which is measured by gas chromatography. Various problems and precautions are associated with the acetylene reduction method. Valid use of the method requires that acetylene does not alter other metabolic activities within the system under investigation and that ethylene is stable during the assay period. However, acetylene has been shown to influence certain microbial activities. It inhibits methane formation (Oremland and Taylor, 1975), methane oxidation (de Bont and Mulder, 1976), and growth of the nitrogen-fixing bacteria, Clostridium pasteurianum (Brouzes and Knowles, 1971). Even though some methane oxidizers are known to fix nitrogen, the acetylene reduction method cannot be employed to measure nitrogenase activity in these bacteria because acetylene inhibits methane oxidation. Consequently, the acetylene reduction assay should not be used in habitats containing methaneutilizing bacteria. The low rates of methane oxidation observed in a rice field by de Bont et al. (1978) indicate that methane-dependent nitrogen fixation is presumably of minor consequence in rice fields. Although ethylene reportedly can be taken up by rice roots (Yoshida and Ancajas, 1971; Yoshida and Suzuki, 1975) and soil in the presence of oxygen

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

181

(Abeles et al., 1971), the findings of de Bont (1976) and Witty (1979) indicate that degradation of ethylene produced from acetylene during the acetylene reduction assay is not significant and does not invalidate the assay. Methane-utilizing bacteria can cooxidize ethylene, but oxidation of methane and cooxidation of ethylene are inhibited in the presence of acetylene; thus removal of ethylene by methane-oxidizing bacteria in the acetylene reduction assay is not likely (de Bont and Mulder, 1976). Proper assay of nitrogen fixation with the acetylene reduction method requires that the acetylene reaches the nitrogen-fixing sites and that the ethylene formed from the acetylene is quantitatively recovered. Even though acetylene is soluble in water, its transfer from the gas phase to the aqueous phase may be slow during assay of water (Flett er al., 1976) or water-saturated soil (Lee and Watanabe, 1977). Flett et al. (1976) recommended the agitation of water samples immediately after the addition of acetylene, and Matsuguchi et al. (1978) proposed an initial evacuation procedure to promote diffusion of acetylene into waterlogged soil. Ethylene is soluble in water, and a portion of the ethylene formed during the acetylene reduction assay may remain in the aqueous phase. The error associated with failure to consider soluble ethylene becomes greater as the aqueous phase comprises a larger fraction of the total volume of the assay chamber. In laboratory assays of waterlogged soil the samples should be agitated prior to gas analysis in order to release trapped ethylene to the gas phase (Lee and Watanabe, 1977; Matsuguchi et al., 1978, 1979). In field assays of waterlogged soil systems the quantity of ethylene remaining in the water-saturated phase can be reduced by stirring the soil prior to gas sampling (Lee and Watanabe, 1977). A time lag in the evolution of ethylene has been observed during the incubation of wetland plant-soil cores (Tjepkema and Evans, 1976; Kana and Tjepkema, 1978) and rice-soil systems (Lee and Yoshida, 1977). A corresponding lag period was not observed in plant-soil cores from well-drained fields (Kana and Tjepkema, 1978). Rinaudo et al. (1977) indicated that the lag in a rice-soil system could be suppressed by addition of ethylene along with acetylene at the start of the assay. Nonetheless, the presence of a lag in ethylene formation and the dependence of ethylene formation upon the composition of the gas atmosphere during incubation (Lee and Yoshida, 1977; Kana and Tjepkema, 1978) complicate the estimation of nitrogen fixation with the acetylene reduction method. Diurnal changes of nitrogenase activity in associative fixation indicate that acetylene-reducing activity obtained at a given time of the day, rather than over a complete 24-hour period, cannot be extrapolated to daily rates (Balandreau et al., 1974). The acetylene reduction assay requires short incubation periods since long-term incubations with acetylene can lead to anomalous results (Hardy et al., 1973; David and Fay, 1977).

182

R. J . BURESH ET AL.

The validity of the acetylene reduction assay in soil systems with low rates of nitrogen fixation has been seriously questioned by Witty (1979). He used I4Clabeled acetylene to show that in highly active nitrogen-fixing systems [rate estimated at 360 g N/ha/day (131 kg N/ha/yr) based on a theoretical conversion factor of 31 the ethylene formed during incubation was from the added acetylene. But in a less active pasture grass-soil core, more than one-half the ethylene formed during the incubation was not derived from the added acetylene. Total production of ethylene indicated that the pasture grass-soil core fixed an estimated 62 g N/ha/day, but recovery of labeled carbon revealed that only 43% of the ethylene was derived from the added acetylene. Ethylene produced in soil is rapidly decomposed in the absence of acetylene, but in the presence of acetylene the ethylene produced in the soil is not decomposed because acetylene inhibits ethylene oxidation (de Bont and Mulder, 1976). Consequently, the measurement of total ethylene production during the assay overestimates nitrogenase activity. The degree of overestimation depends upon endogenous ethylene production and may vary between different soils and plant-soil systems. Control incubations without added acetylene do not measure endogenous ethylene production. Witty (1979) believed that errors in the acetylene reduction method could be significant for soil-containing systems fixing less than an estimated 100 g N/ ha/day (36 kg N/ha/yr) based on a theoretical conversion factor of 3. This casts doubts on the accuracy of many fixation rates obtained with the acetylene reduction method for flooded soil systems because many reported values are less than 36 kg N/ha/yr. The acetylene reduction assay constitutes an indirect rather than a direct measurement of nitrogen fixation. Reduction of 3 moles of acetylene to ethylene is theoretically equivalent to reduction of 1 mole of dinitrogen gas. However, reported conversion factors for acetylene reduction to nitrogen fixation vary from this factor of 3. For soils they generally range from 3.0 to 6.9, but values up to 25 have been reported in anaerobic soils (Hardy er al., 1973). The acetylene reduction method must be calibrated with the direct isotopic method (Bergerson, 1970; Steyn and Delwiche, 1970; Bums, 1974). Bums ( 1974) emphasized that the measurement of I5N2reduction and acetylene reduction should be under identical conditions and incubation periods in order to obtain an appropriate conversion factor. However, this is very difficult because long incubations are required with I5N in order to obtain measurable atom % I5N enrichments in the soil or plant tissue, but long incubations with acetylene can lead to erroneous estimates of fixation. Rice and Paul (1971) recommended a normal atmosphere of 80%N2 when calibrating the two methods in waterlogged soil systems. Recent conversion factors experimentally determined for blue-green algae (Peterson and Burris, 1976; Potts er al., 1978) and surface marine sediment (Potts et al., 1978) were less variable than those in the earlier literature reviewed

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

183

by Hardy et al. (1973), but they were consistently higher than 3. Conversion factors obtained by Liao (1977) for nitrogen-fixing blue-green algae in a lake averaged higher than 3, but unlike the values reported by Peterson and Bums (1976) and Potts et al. (1978), they were highly variable ranging from 1.3 to 170. Liao (1977), however, did not calibrate the 15N method and acetylene reduction method under identical incubation conditions. The depression of hydrogen production in the presence of acetylene increases the “theoretical” conversion factor to values above 3 (Burns and Hardy, 1975; Hill, 1978). Watanabe et al. (1978a) assumed a “theoretical” factor of 4 based on the following two equations for the nitrogenase reaction: N,

+ 8H+ + 8e-+2NH3 + H,

C2H2+ 2H+ + 2e-+C,H

I

In a later paper Watanabe et al. (1978b) expressed their results as ethylene formation and made no attempt to convert to a nitrogen equivalent. In the absence of a calibration, the acetylene reduction method is only an approximation of nitrogen fixation. Results should be reported as acetylene reduction without conversion to an equivalent quantity of nitrogen fixation (Bums, 1974). Few researchers have calibrated the acetylene reduction method with l5NZ reduction; most continue to use the conversion factor of 3 (Hauck and Bremner, 1976). Recent observations of conversion factors greater than 3 (Peterson and Burris, 1976; Potts et al., 1978), accumulation of ethylene from endogenous sources in the presence of acetylene (Witty, 1979), and greater acetylene reduction for excised roots than for in situ systems (Patriquin and Denike, 1978) suggest that reported values for nitrogen fixation in flooded soil systems based on the acetylene reduction assay may be prone to overestimation.

IX. CONTRIBUTION OF FIXED NITROGEN TO THE NITROGEN REQUIREMENTS OF PLANTS Indirect evidence for the contribution of nitrogen fixation to nitrogen fertility in rice soils has come from nitrogen balance studies. Long-term studies at IRRI in the Philippines revealed that rice plots were able to sustain similar yields for as many as 24 successive crops without nitrogen fertilization or an apparent decrease in total nitrogen content in the soil (App et al., 1978). Nitrogen inputs estimated at as much as 50 kg N/ha were required to replace the nitrogen removed by the rice crop and thereby maintain the nitrogen content of the soil. Biological nitrogen fixation was presumably responsible for the addition of most of this nitrogen and the maintenance of soil fertility (App et al., 1978; Koyama and App, 1979).

184

R. J. BURESH ET AL.

Actual measurements of nitrogen fixation rates in rice fields and other flooded soil systems vary greatly. In fact, in some instances different research groups attribute the major source of fixed nitrogen to different processes. For example, Watanabe et al. (1978b) found algal fixation to be the major nitrogen-fixing process in rice fields at IRRI, but Japanese workers (Wada et al., 1978) concluded that algal fixation in their fields was minimal and bacterial fixation in nonrhizosphere soil was most important. A large portion of the variability for estimates of nitrogen fixation is attributable to differences in location, time, and method of measurement. Regardless of the variations and limitations, the data of most researchers reveal that even if all the nitrogen estimated to be fixed in a flooded soil system were available to higher plants this would probably be insufficient to meet all the nitrogen requirements of the plants. The rates of nitrogen fixation measured by Reddy and Patrick (1979) in rice soil were not adequate to support high yields of rice. Valiela and Teal (1979) estimated that the total nitrogen input by fixation in a salt marsh represented about 37% of the plant nitrogen requirements. Nitrogen fixation in most cases is considered to represent only a minor contribution to the nitrogen input of lakes (see Section 11,C). The availability of fixed nitrogen to higher plants is poorly understood. Measurements of nitrogen-fixing activity and plant production in communities of the sea grass Thalassia by Capone et a / . (1979) suggest that the nitrogen fixed by epiphytes on the grass contributes mostly to the production of the epiphytes rather than to that of the plant. However, nitrogen fixation in the plant rhizosphere presumably is a source of nitrogen for the plant. Estimates of nitrogen fixation in the rhizosphere varied from 12-48% of the nitrogen requirement for leaf production in this study. Confirmation of actual incorporation of fixed nitrogen into plant tissue requires use of l5N-labeled nitrogen gas as a tracer. Incorporation of atmospheric nitrogen into rice (Eskew et al., 1979; Ito et al., 1980) and several plant species in an English salt marsh (Jones, 1974) has been confirmed with 15N, but for the most part few I5N tracer studies have been conducted. The nitrogen fixed by blue-green algae for the most part is not immediately available to plants; rather it becomes available only gradually after the algae die and undergo decomposition, whereby the fixed nitrogen in the algal biomass is mineralized to forms that can be assimilated by plants. Stewart et al. (1979) in a symposium at IRRI mentioned that 15N tracer studies in the United Kingdom suggest that much longer than a few months is required for most of the nitrogen in blue-green algae to be released in forms readily available to plants. The Azolla-Anabaena association is a major potential source of fixed nitrogen to rice. The fixed nitrogen in the Azolla-Anabaena complex, as with that in free-living blue-green algae, for the most part is released slowly during decom-

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

185

position following death. Watanabe et al. (1977a) in pot experiments found that after 6 weeks 62 and 75% of the total nitrogen in fresh and dried Azolla, respectively, incorporated into soil was released as ammonium. Incorporation of dried Azolla in soil was also observed to enhance rice growth, but the availability of nitrogen in the Azolla was 40% lower than that of added ammonium sulfate. Singh (1 979) found similar release rates of 56 and 80% of the Azolla nitrogen after 3 and 6 weeks, respectively. The recommended agronomic practice is to incorporate Azolla into the soil because Azolla is more beneficial to the rice crop when it is incorporated than when it is allowed to remain on the soil surface (Singh, 1979). The use of Azolla in rice cultivation in China (Liu Chung Chu, 1979), India (Singh, 1979), Vietnam (Dao The Tuan and Tran Quang Thuyet, 1979), and North America (Rains and Talley, 1979) was reviewed at a recent IRRI symposium. Shanmugam et al. (1978) encouraged research into the development of genetic variants of Azolla-blue-green algae associations that can readily release fixed nitrogen as ammonium while actively growing. The contribution of nitrogen fixation to the fertility of rice fields can be enhanced with selected management practices. The reported stimulation of nitrogen fixation by rice straw (Rao, 1976; Reddy and Patrick, 1979) suggests that the incorporation of high C/N plant residue might increase the amount of nitrogen fixed. Watanabe (1978) estimated that incorporation of 1 ton of rice straw per hectare can increase nitrogen fixation by 2-5 kg N/ha. Matsuguchi (1979) showed that the application of compost and rice straw stimulates nitrogen fixation. The increase in fixation was greater in nonrhizosphere soil than in the rice root zone. Data of Matsuguchi (1977) indicated that addition of nitrogen fertilizer increases the stimulatory effect of rice straw upon nitrogen fixation. Although the added nitrogen might initially inhibit fixation, it encourages decomposition of the rice straw which then supplies carbon substrate suitable as an energy source for heterotrophic fixation. After the added nitrogen disappears in the soil, a favorable environment for nitrogen fixation is present. Inoculation of rice fields with blue-green algae is another technique to increase the input of fixed nitrogen (Venkataraman and Goyal, 1968). Its importance has been examined by Venkataraman (1975, 1979) and Yamaguchi (1979). The utilization of Azolla, as previously mentioned, is an important means for enhancement of fixed nitrogen input in rice cultivation.

X. PERSPECTIVES Flooded soil systems tend to be suitable environments for both autotrophic and heterotrophic nitrogen fixation. Nitrogen fixation by blue-green algae in the

186

R. J . BURESH ET AL.

photic zone is known to provide significant amounts of fixed nitrogen and can be the major process by which nitrogen is supplied to flooded rice soils and unvegetated areas in wetlands. However. Mague (1977) in a review article mentioned that much of the information on algal fixation in aquatic ecosystems has been obtained from work with cultures of blue-green algae, and he suggested more emphasis on the measurements of in situ nitrogen fixation rates and on experiments utilizing 15N. In the soil column the types of nitrogen-fixing microorganisms present in the aerobic surface layer differ from those in the anaerobic subsurface layer. In the oxygenated surface layer, conditions are favorable for aerobic nitrogen fixers such as Azotobacter, whereas obligate anaerobes such as Clostridium are restricted to the underlying anaerobic zone. Anaerobic nitrogen fixation is usually considered only a minor contribution to the nitrogen economy of flooded soils (Evans and Barber, 1977), but recent Japanese research (Matsuguchi, 1979; Wada et al., 1978) indicated that anaerobic soil zones without rice roots can be the most significant region for fixation in rice fields without significant algal growth. Heterotrophic nitrogen-fixing bacteria associated with the rhizosphere of rice and other wetland plants are important agents of nitrogen fixation in flooded soil systems, but present evidence (Watanabe et a l . , 1977b, 1978b) suggests that this associative fixation is probably less important than previously suggested in the literature. The phyllosphere and stems of wetland plants are zones for autotrophic and heterotrophic nitrogen fixation that have only recently been investigated. The Azolla-Anabeana symbiotic association is a promising source of nitrogen in rice cultivation. The comparison and interpretation of reported nitrogen fixation rates in the literature is difficult because of differences in methodology and possible errors associated with the assay techniques. Most reported fixation rates have been obtained by the indirect acetylene reduction method. The factor for conversion of acetylene reduction to actual nitrogen fixation can vary, yet it is seldom determined experimentally. Most researchers assume a theoretical factor of 3, although some recent reports (Patriquin and McClung, 1978; Watanabe et a l . , 1978a) have used a theoretical factor of 4. Even though rice and many wetland plants are known to stimulate nitrogen fixation, the nitrogenase activity of nitrogen fixers associated with the plant rhizosphere, phyllosphere, and stems in flooded soils is not well understood. The quantity of nitrogen fixed per hectare per year in flooded soil systems has not been satisfactorily determined, and the availability of fixed nitrogen to plants remains largely unanswered. A recent review by Watanabe (1978) stressed the need for new experimental techniques to measure nitrogen fixation in the field. An understanding of the relative importance of the various processes of nitrogen fixation in both agronomic and nonagronomic flooded soil systems can aid in development of management practices that will enhance productivity.

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

187

ACKNOWLEDGMENTS Partial support during the preparation of this review was provided by the Louisiana Sea Grant Program, a part of the National Sea Grant program maintained by the National Oceanic and Atmospheric Administration, U.S. Department of Commerce. The authors extend their appreciation to R. D. DeLaune for his helpful suggestions during the preparation of this manuscript.

REFERENCES Abeles, F. B., Craker. L. E., Forrence, L. E., and Leather, G. R. 1971. Science 173, 914-916. Alexander, M. 1977. “Introduction to Soil Microbiology.” Wiley, New York. Alimagno, B. V., and Yoshida, T. 1977. Plant Soil 47, 239-244. App, A,, Watanabe, I., Alexander, M., Ventura, W., and DeDatta, S. K. 1978. Agron. Abstr. p. 149.

Araragi, M., and Tangcham. B. 1979. Soil Sci. Plant Nufr. 25, 283-295. Ashton, P. J., and Walmsley, R. D. 1976. Endeavour 35, 39-43. Balandreau, J. P., Millier, C. R., and Dommergues, Y. R. 1974. Appl. Microbiol. 27, 662-665. Balandreau, J . , Rinaudo, G., Fares-Hamad, I., and Dommergues, Y. 1975. In “Nitrogen Fixation by Free-Living Micro-Organisms”(W. D. P. Stewart, ed.), pp. 57-70. Cambridge Univ. Press, London and New York. Balandreau, J. P., Rinaudo, G., Oumarov, M. M., and Dommergues, Y. R. 1976. In “Proceedings of the First International Symposium on Nitrogen Fixation” (W.E. Newton and C. J. Nyman, eds.), Vol. 2, pp. 61 1-628. Washington State Univ. Press, Pullman. Barber, L. E., Tjepkema, J. D., Russell, S. A., and Evans, H.J. 1976. Appl. Environ. Microbiol. 32, 108-113. Becking, J. H. 1972. Fortschr. Bot. 34, 459-467. Becking, J . H. 1978. Ecol. Bull. (Stockholm)26, 116-129. Becking, J. H. 1979. In “Nitrogen and Rice,” pp. 345-373. IRRI, Los Ba~ios,Philippines. Bergersen, F. J. 1970. Aust. J . Biol. Sci. 23, 1015-1025. Brezonik, P. L., and Harper, C. L. 1969. Science 164, 1277-1279. Bristow, J. M. 1974. Can. J . Bof. 52, 217-221. Brooks, R. H., Brezonik, P. L.. Putnam, H. D., and Keirn, M.A. 1971. Limnol. Oceanogr. 16, 701-7 10. Broome, S . W., Woodhouse, W. W., Jr., and Seneca, E. D. 1975. Soil Sci. SOC. Am. Proc. 39, 301-307.

Brouzes, R., and Knowles, R. 1971. Can. J. Microbiol. 17, 1483-1489. Brouzes. R., Lasik, J., and Knowles, R. 1969. Can. J . Microbiol. 15, 899-905. Bunt, J. S. 1961. Nature (London) 192, 479-480. Bums, R. C., and Hardy, R. W. F. 1975. “Nitrogen Fixation in Bacteria and Higher Plants.” Springer-Verlag, Berlin and New York. Bums, R. H. 1974. In “The Biology of Nitrogen Fixation” (A. Quispel, ed.), pp. 9-23. NorthHolland Publ., Amsterdam. Bums, R. H. 1976. Aust. J . Plant Physiol. 3, 41-51. Capone, D. G., andTaylor, B. F. 1977. Mar. Biol. 40, 19-28. Capone, D. G.. and Taylor, B. F. 1978. Annu. Meet. Am. SOC. Limnol. Oceanogr. Victoria, B.C. (Abstr. ) . Capone, D. G . , Penhale, P.A., Oremland, R. S.,and Taylor, B. F. 1979. Limnol. Oceanogr. 24, 117-125.

188

R. J. BURESH ET AL.

Carpenter, E. J . , Van Raalte, C. D., and Valiela, I. 1978. Limnol. Oceanogr. 23, 318-327. Casselman, M. E. 1979. M.S. Thesis, Louisiana State University, Baton Rouge. Dao The Tuan, and Tran Quang Thuyet. 1979. In “Nitrogen and Rice,” pp. 395-405. IRRI, Los Ba~ios,Philippines. David, K. A. V., and Fay, P. 1977. Appl. Environ. Microbiol. 34, 640-646. De, P. K. 1936. Indian J . Agr. Sci. 6, 1237-1245. De, P. K . 1939. Proc. R. Soc. London B 127, 121-139. De, P. K . , and Mandal, L. N. 1956. Soil Sci. 81, 453-458. De Bont, J . A. M. 1976. Can. J. Microbiol. 22, 1060-1062. De Bont, J . A. M., and Mulder, E. G. 1976. Appl. Environ. Microbiol. 31, 640-647. De Bont, J . A. M., Lee, K. K., and Bouldin, D. F. 1978. Ecol. Bull. (Stockholm) 26, 91-96. DeLaune, R. D., Buresh, R. J . , and Patrick, W. H.. Jr. 1979. Est. Coasr. Mar. Sci. 8,477-487. De-Polli, H., Matsui, E., Dobereiner, J., and Salati, E. 1977. Soil B i d . Biochem. 9, 119-123. Diem, G . , Rougier, M., Hamad-Fares, I., Balandreau, J. P., and Dommergues, V. R. 1978. Ecol. Bull. (Stockholm) 26, 305-3 1 I . Dobereiner, J., Day, J . M., and Dart, P. J. 1972. J . Gen. Microbiol. 71, 103-1 16. Dommergues, Y. R., and Rinaudo, G. 1979. In “Nitrogen and Rice,” pp. 241-260. IRRI, Los BaIios, Philippines. Dommergues, Y.,Balandreau, J . , Rinaudo, G . , and Weinhard, P. 1973. Soil Biol. Biochem. 5 , 83-89. Dugdale, R. C., Goering, J. J., and Ryther, J. H. 1964. Limnol. Oceanogr. 9, 507-510. Eskew, D. L.. Eaglesham, A. R. J.. and App, A. A. 1979. Agron. Abstr. p. 87. Evans, H. J . , and Barber, L. E. 1977. Science 197, 332-339. Fay, P. 1976. Appl. Environ. Microbiol. 31, 376-379. Finke, L. R., and Seeley, H. W., Jr. 1978. Appl. Environ. Microbiol. 36, 129-138. Flett, R. J., Hamilton, R. D., and Campbell, N. E. R. 1976. Can. J . Microbiol. 22, 43-51. Fogg, G. E. 1971. Plant Soil Spec. Vol. pp. 393-401. Fogg. G. E. 1978. Ecol. Bull. (Stockholm) 26, 11-19, Fogg, G. E., Stewart, W. D. P., Fay, P.,and Walsby, A. E. 1973. “The Blue-Green Algae.” Academic Press, New York. Gaskins, M. H., and Carter, J. L. 1975. Soil Crop Sci. Soc. Fl. Proc. 35, 10-16. Gilmour, J . T., Gilmour, C. M., and Johnston, T. H. 1978. Soil Biol. Biochem. 10, 261-264. Goering, J . J . , and Parker, P. L. 1972. Limnol. Oceanogr. 17, 320-323. Gotto, J . W., and Taylor, B. F. 1976. Appl. Environ. Microbiol. 31, 781-783. Granhall, U., and Lundgren, A. 1971. Limnol.Oceanogr. 16, 711-719. Green, F., and Edmisten, J. 1974. Ecol. Stud. 8, I 13-126. Habte, M., and Alexander, M. 1980. Appl. Environ. Microbiol. 39, 342-347. Hanson, R. B. 1977a. Appl. Environ. Microbiol. 33, 596-602. Hanson, R. B. 1977b. Appl. Environ. Microbiol. 33, 846-852. Hardy, R. W. F., Bums, R. C., and Holsten, R. D. 1973. Soil Biol. Biochem. 5 , 47-81. Hartwig, E. O . , and Stanley, S. 0. 1978. Deep Sea Res. 25, 41 1-417. Hauck, R. D., and Bremner, J. M. 1976. Adv. Agron. 28, 219-266. Herbert, R. A. 1975. J. Exp. Mar. Biol. Ecol. 18, 215-225. Hill, S. 1978. Ecol. Bull. (Srockholm) 26, 130-136. Home, A. J . 1979. Limnol. Oceanogr. 24, 329-341. Home, A. J . , and Fogg, G. E. 1970. Proc. R. Sue. London B 175, 357-366. Howard, D. L.. Frea, J. I., Ptister, R. M., and Dugan, P. R. 1970. Science 169, 61-62. IRRI. 1975. Annu. Rep. Int. Rice Res. Inst. Los Barios, Philippines. IRRI. 1977. Annu. Rep. Inr. Rice Res. Inst. Los Barios, Philippines. Ito, 0.. Cabrera, D., and Watanabe, I. 1980. Appl. Environ. Microbiol. 39, 554-558.

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

189

Jensen, V., and Holm, E. 1975. I n “Nitrogen Fixation by Free-Living Micro-Organisms” (W. D. P. Stewart, ed.), pp. 101-119. Cambridge Univ. Press, London and New York. Jones, K. 1974. J . Ecol. 62, 553-565. Jurgensen, M. F. 1973. J . Soil Sci. 24, 512-522. Jurgensen, M. F., and Davey, C. B. 1970. Soils Ferr. 33, 435-446. Kana, T. M., and Tjepkema, J . D. 1978. Can. J. Bor. 56, 2636-2640. Keim, M. A., and Brezonik, P. L. 1971. Limnol. Uceanogr. 16, 720-731. Kellar, P. E. 1979. Annu. Meet. Am. SOC. Limnol. Oceanogr.. Corpus Christi, Tex. (Abstr.). Knowles, R. 1977. I n “A Treatise on Dinitrogen Fixation. Section IV: Agronomy and Ecology” (R. W. F. Hardy and A. H. Gibson, eds.), pp. 33-83. Wiley, New York. Kobayashi. M., and Haque, M. Z. 1971. Plant Soil Spec. Vol. pp. 443-456. Kobayashi, M., Takahashi, E., and Kawaguchi, K. 1967. Soil Sci. 104, 113-1 18. Koch, B. L. 1977. Planr Soil 47, 703-706. Koyama. T., and App. A. 1979. In “Nitrogen and Rice,” pp. 95-104. IRRI, L o s Bafios, Philippines. Kumari, M. L., Kavimandan, S. K., and Rao, N. S. S. 1976. Indian J . Exp. Biol. 14, 638-639. Lee, K. K., and Watanabe, I. 1977. Appl. Environ. Microbiol. 34, 654-660. Lee. K. K., and Yoshida, T. 1977. Plant Soil 46, 127-134. Lee, K. K., Alimagno, B.. and Yoshida, T. 1977a. Plant Soil 47, 519-526. Lee, K. K., Castro, T., and Yoshida, T. 1977b. Planr Soil 48, 613-619. Liao, C. F. H. 1977. Hydrobiology 56, 273-279. Lipschultz, F., Cunningham, J. J., and Stevenson, J. C. 1979. Esr. Coasr. Mar. Sci. 9, 813-818. Liu Chung Chu. 1979. In “Nitrogen and Rice,” pp. 375-394. IRRI, Los Bailos, Philippines. MacGregor, A. N., Keeney, D. R., and Chen, K. L. 1973. Environ. Lerr. 4, 21-26. MacRae, I. C. 1975. Soil Biol. Biochem. 7, 337-338. MacRae, I. C., and Castro, T. F. 1967. Soil Sci. 103, 277-280. McRoy, C. P., Goering, J. J., and Chaney, B. 1973. Limnol. Uceanogr. 18, 998-1002. Magdoff, F. R., and Bouldin. D. R. 1970. PIanr Soil 33, 49-61. Mague, T. H . 1977. In “A Treatise on Dinitrogen Fixation. Section IV: Agronomy and Ecology” (R. W. F. Hardy and A. H.Gibson, eds.), pp. 85-140. Wiley, New York. Marsho, T. V., Burchard. R. P., and Fleming, R. 1975. Can. J . Microbiol. 21, 1348-1356. Matsuguchi, T. 1977. Jpn. Agr. Res. Q 11, 204-210. Matsuguchi, T. 1979. In “Nitrogen and Rice,” pp. 207-222. IRRI, Los Bafios, Philippines. Matsuguchi, T., Tangcham, B., and Patiyuth, S. 1975. Jpn. Agr. Res. Q 8, 253-256. Matsuguchi, T . , Shimomura, T., and Lee, S. K. 1978. Ecol. Bull. (Srockholm) 26, 137-147. Matsuguchi. T., Shimomura, T., and Lee, S. K. 1979. Soil Sci. Plant Nurr. 25, 323-336. Moeller, R. E., and Roskoski, J. P. 1978. Hydrobiology 60, 13-16. Moore, A. W. 1969. Bor. Rev. 35, 17-33. Nayak, D. N., and Rao, V. R. 1977. Arch. Microbiol. 115, 359-360. Odum, E. P. 1971. “Fundamentals of Ecology,” 3rd ed. Saunders, Philadelphia, Pennsylvania. Okafor, N., and MacRae, I. C. 1973. Soil Biol. Biochem. 5 , 181-186. Okuda, A., Yamaguchi, M., Kobayashi, M., and Katayama, T. 1961. Soil Sci. Planr Nurr. 7, 146-151.

Oremland, R. S., and Taylor, B. F. 1975. Appl. Microbiol. 30, 707-709. Panichsakpatana, S., Wada, H.,Kimura, M., and Takai, Y. 1978. Soil Sci. PIanr Nurr. 24, 367373. Panichsakpatana, S . , Wada, H.,Kimura, M., and Takai, Y. 1979. Soil Sci. PIanr Nurr. 25, 165171. Patrick, W. H., Jr., and DeLaune, R. D. 1976. Esr. Coasr. Mar. Sci. 4, 59-64. Patriquin, D. G . 1972. Mar. Biol. 15, 35-46.

190

R. J. BURESH ET AL.

Patriquin, D. G. 1978a. Ecol. Bull. (Srockholm) 26, 20-27. Patriquin, D. G. 1978b. A q u a . Bor. 4, 193-210. Patriquin, D. G . , and Denike, D. 1978. Aquut. Bor. 4, 21 1-226. Patriquin, D. G., and Keddy, C. 1978. Aquat. Bot. 4, 227-244. Patriquin, D. G., and Knowles, R. 1972. Mar. Biol. 16, 49-58. Patriquin, D. G., and McClung, C. R. 1978. Mar. Biol. 47, 227-242. Peters, G. A., and Mayne, B. C. 1974. Planr Physiol. 53, 820-824. Peters, G. A,, Mayne, B. C., Ray, T. B., and Toia, R. E., Jr. 1979. In “Nitrogen and Rice,” pp. 325-344. IRRI, Los Ba~ios,Philippines. Peterson, R. B., and Bums, R. H. 1976. Anal. Biochem. 73, 404-410. Peterson, R. B., Friberg, E. E., and Burris, R. H. 1977. Plant Physiol. 59, 74-80. Ponnamperuma, F. N. 1972. Adv. Agron. 24, 29-96. Pons, M., Krumbein, W. E.. and Metzger, J. 1978. In “Environmental Biogeochemistry and Geomicrobiology” (W. E. Krumbein, ed.), pp. 753-769. Ann Arbor Science Publ., Ann Arbor, Michigan. Purchase, B. S . 1977. Planr Soil 46, 283-286. Rains, D. W., and Talley, S. N. 1979. In “Nitrogeh and Rice,” pp. 419-431. IRRI, Los Baiios, Philippines. Rao, V. R. 1976. Soil Biol. Biochem. 8, 445-448. Reddy, K. R., and Patrick, W. H., Jr. 1979. Soil Sci. 128, 80-85. Reynaud, P. A., and Roger, P. A. 1978. Ecol. Bull. (Stockholm) 26, 148-157. Rice, W. A., and Paul, E. A. 1971. Can. J . Microbiol. 17, 1049-1056. Rice, W. A,, and Paul, E. A. 1972. Can. J . Microbiol. 18, 715-723. Rice, W. A., Paul, E. A., and Wetter, L. R. 1967. Can. J . Microbiol. 13, 829-836. Rinaudo. G . , Balandreau, J., and Dommergues, Y. 1971. Plant Soil Spec. Vol. pp. 471-479. Rinaudo, G., Hamad-Fares, I., and Dommergues, Y . R. 1977. In “Biological Nitrogen Fixation in Farming Systems of the Tropics” (A. Ayanaba and P. J. Dart, eds.), pp. 313-322. Wiley, New York. Roger, P. A., and Reynaud, P. A. 1979. In “Nitrogen and Rice,” pp. 287-310. IRRI, Los Batios, Philippines. Roger, P. A,, Kulasooriya, S . A., Tirol, A., and Craswell, E. T. 1980. Planr Soil (in press). Rovira, A. D. 1965. Annu. Rev. Microbiol. 19, 241-266. Ruinen, J. 1970. Planr Soil 33, 661-671. Ruinen, I . 1974. I n “The Biology of Nitrogen Fixation” (A. Quispel, ed.), pp. 121-167. NorthHolland Publ., Amsterdam. Rusness, D., and Bunis, R. H. 1970. Limnol. Oceanogr. 15, 808-813. Saralov, A. I., and Dzyuban, A. N. 1978. Microbiology 47, 110-1 12. Shanmugam, K. T., O’Gara, F., Anderson, K., Morandi, C., and Valentine, R. C. 1978. In “Nitrogen in the Environment. Vol. 2. Soil-Plant-Nitrogen Relationships” (D. R. Nielsen and J. G. MacDonald, eds.), pp. 393-416. Academic Press, New York. Silver, W. S . , and Jump, A. 1975. In “Nitrogen Fixation by Free-Living Micro-Organisms” (W. D. P. Stewart, ed.), pp. 121-125. Cambridge Univ. Press, London and New York. Singh, P. K. 1979. In “Nitrogen and Rice,” pp. 407-418. IRRI, Los Baiios, Philippines. Singh, R. N. 1961. “Role of Blue-Green Algae in Nitrogen Economy of Indian Agriculture.” Indian Council Agric. Res., New Delhi. Stewart, W. D. P. 1969. Proc. R. Soc. London B 172, 367-388. Stewart, W. D. P., and Pugh, G. J. F. 1963. J. Mar. Biol. Assoc. U . K . 43, 309-317. Stewart, W. D. P., Mague,T.. Fitzgerald, G. P., and Burris, R. H. 1971. New Phyrol. 70,497-509. Stewart, W. D. P., Rowell. P., Ladha, J. K., and Sampaio, M. J. A. M. 1979. In “Nitrogen and Rice,” pp. 263-285. IRRI, Los BaIios, Philippines.

NITROGEN FIXATION IN FLOODED SOIL SYSTEMS

191

Steyn, P. L., and Delwiche, C. C. 1970. Environ. Sci. Technol. 4, 1122-1128. Teal, J . M . , and Kanwisher, J. W. 1966. J. Exp. Bor. 17, 355-361. Teal, J . M., Valiela, T., and Berlo, D. 1979. Limnol. Oceanogr. 24, 126-132. Tison, D. L., Palmer, F. E., and Staley, J. T. 1977. Water Res. 11, 843-847. Tjepkema, J . D., and Evans, H. J. 1976. Soil Biol. Biochem. 8, 505-509. Torrey, M . S . , and Lee, G . F. 1976. Limnol. Oceanogr. 21, 365-378. Trolldenier. G. 1977. Plant Soil 47, 203-217. Valiela, I., and Teal, J. M. 1974. In “Ecology of Halophytes” (R. J. Reimold and W. H. Queen, eds.), pp. 547-563. Academic Press, New York. Valiela, I., and Teal, J. M. 1979. Nature (London) 280, 652-656. Van Berkum, P., and Sloger, C. 1979. Plant Physiol. 64, 739-743. Vanderhoef, L. N., Huang, C. Y., Musil, R., and Williams, J. 1974. Limnol. Oceanogr. 19, 119-125. Van Raalte, C. D., Valiela, I., Carpenter, E. J., and Teal, J. M. 1974. Est. Coast. Mar. Sci. 2, 301-305. Van Raalte, M. H. 1941. Ann. Bot. Gardens Buitenzorg. 51, 43-58. Venkataraman, G. S. 1975. In “Nitrogen Fixation by Free-Living Micro-Organisms” (W. D. P. Stewart, ed.), pp. 207-218. Cambridge Univ. Press, London and New York. Venkataraman, G. S . 1979. In “Nitrogen and Rice.” pp. 31 1-321. IRRI, Los Bafios, Philippines. Venkataraman, G. S., and Goyal, S. K. 1968. Soil Sci. Plant Nurr. 14, 249-251. Wada, H., Panichsakpatana, S., Kimura, M., and Takai, Y. 1978. Soil Sci. Plant Nutr. 24, 357365. Wada, H., Panichsakpatana, S., Kimura, M.. and Takai, Y. 1979. Soil Sci. Plant Nurr. 25, 453456. Watanabe, A. 1959. J. Gen. Appl. Microbiol. 5 , 21-29. Watanabe, A , , and Yamamoto, Y. 1967. Narure (London) 214, 738. Watanabe, A., and Yamamoto, Y. 1971. Plant Soil Spec. Vol. pp. 403-413. Watanabe, I. 1978. In “Soils and Rice,” pp. 465-478. IRRI, Los Ba~ios,Philippines. Watanabe, I., and Barraquio, W. L. 1979. Nature (London) 277, 565-566. Watanabe, I., and Cabrera, D. R. 1979. Appl. Environ. Microbiol. 37, 373-378. Watanabe, I., and Cholitkul, W. 1979. In “Nitrogen and Rice,” pp. 223-239. IRRI, Los Batios, Philippines. Watanabe, I., and Lee, K . K. 1977. In ‘‘Biological Nitrogen Fixation in Farming Systems of the Tropics” (A. Ayanaba and P. J. Dart, eds.), pp. 289-305. Wiley, New York. Watanabe, I . , Espinas, C. R., Berja, N. S., and Alimagno, B. V. 1977a. Int. Rice Res. Insr. Res. Pap. Ser. 11, Los Bafios, Philippines. Watanabe, I., Lee, K. K., Alimagno, B. V., Sato, M., del Rosario, D. C., and de G u m a n , M. R. 1977b. Int. Rice Res. Inst. Res. Pap. Ser. 3, Los Bafios, Philippines. Watanabe, I., Lee, K. K . , and Alimagno, B. V. 1978a. Soil Sci. Plant Nurr. 24, 1-13. Watanabe. I., Lee, K. K . , and de G u m a n , M. 1978b. Soil Sci. Plant Nurr. 24, 465-471. Watanabe, I., Barraquio, W. L., de G u m a n , M. R., and Cabrera, D. A. 1979. Appl. Environ. Microbiol. 37, 813-819. Waughman, G. J . 1976. Can. J . Microbiol. 22, 1561-1566. Werner, D., Evans, H. J.. and Seidler, R. J. 1974. Can. J . Microbiol. 20, 59-64. Whitney, D. E., Woodwell, G. M., and Howarth, R. W. 1975. Limnol. Oceanogr. 20, 640-643. Wilson, J . T., and Alexander, M. 1979. Soil Sci. SOC. Am. J . 43, 936-939. Witty, J . F. 1979. Soil Biol. Biochem. 11, 209-210. Yamaguchi, M. 1979. In “Nitrogen and Rice,” pp. 193-204. IRRI, Los Bafios, Philippines. Yoneyama, T., Lee, K. K., and Yoshida, T. 1977. Soil Sci. Plant Nurr. 23, 287-295. Yoshida, T., and Ancajas. R . R. 1971. Soil Sci. Soc. Am. Proc. 35, 156-158.

192

R. J . BURESH ET AL.

Yoshida, T., and Ancajas, R. R. 1973a. Soil Sci. SOC.Am. Proc. 37, 42-46. Yoshida. T., and Ancajas, R. R. 1973b. Soil Biol. Biochem. 5, 153-155. Yoshida, T., and Broadbent, F. E. 1975. Soil Sci. 120, 288-291. Yoshida, T., and Suzuki, M. 1975. Soil Sci. Plant Nurr. 21, 129-135. Yoshida, T., Roncal, R. A , , and Bautista, E. M. 1973. Soil Sci. Plant Nutr. 19, 117-123. Yoshida, T . , Takai, Y . , and del Rosario, D. C. 1975. Plant Soil 42, 653-660. Zuberer, D. A . , and Silver, W . S . 1978. Appl. Environ. Microbiol. 35, 561-575.

ADVANCES IN AGRONOMY. VOL. 33

EXPERIENCE WITH SOIL TAXONOMY OF THE UNITED STATES Marlin G.Cline Department of Agronomy, Cornell University, Ithaca, New York

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

IV. Problems for Users of Soil Taxonomy . . . . . . . . . . . . . .

. . . . . . . . . . . . . 201

V. Taxonomic Problems

E. Vertisols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

VI. Impact of Soil Taxonom A. Soil Classification .

C. Interpretations for Applied Objectives

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

218

I. INTRODUCTION

Readers should recognize that much of the literature cited in this article is based on an incomplete preliminary draft of Soil Taxonomy, known widely as “The Seventh Approximation” (Soil Survey Staff, 1960), and on four supplements issued in 1964, 1966, 1967, and 1968. During the 15 years from release of ’Department of Agronomy Series Paper No. 1333. I93 Copyright @ 1980 by Academic Press. Inc. All rights Of reproduction in any form reserved. ISBN 0-12-ooO733-9

194

MARLIN G. CLINE

the Seventh Approximation until the complete text was published for general distribution (Soil Survey Staff, 1975), the number of great groups recognized in the system increased from 112 to 230, and the number of subgroups, from 28 1 to 1251. The basic concepts and principles, however, changed much less. The National Cooperative Soil Survey of the United States has used the system as its official taxonomic classification since January 1965. The date of an article cited may be used as an approximate indicator of the stage of development of the system on which it is based. The author has been selective to avoid citing conclusions that would have been clearly outdated by subsequent changes in the system. To bring the information available up to date, the author requested leading soil scientists of 46 countries to respond to specific questions about use of the system in the areas they represent, including (1) how much and in what ways it is used; (2) problems encountered in the mechanics of its application; (3) problems of classifying soils by its criteria; (4) its impact on soil surveys; and (5) its usefulness for development of soil interpretations for applied purposes. Fifty-three individuals, representing 3 1 foreign countries, responded. Four others responded for multicountry areas. Sixteen responded for five agencies of the United States. The individuals are identified by the countries or areas they represent in footnote 2. Information from these sources is identified in the text as “personal communication” by the convention (PC) and is cited in the present tense to distinguish it from that derived from published literature.* *The author is most grateful to the individuals named below for invaluable assistance. All responded to the author’s request for information, either by answering specific questions or by referring the inquiry to others. Many added relevant comments or supplied documents that would not otherwise have been seen. For foreign countries: A r g e n t i n a x . D. Scoppa; Australia-K. H. Northcote and R. F. Isbell; Belgium-R. Langohr; Brazil-M. Camargo; Canada-G. M. Coen, L. M. Lavkulich, W. W. Pettapiece, I. Snedden. and C. Tarnocai; Chile-M. Carrasco, J. Munito, W. Luzio, and J. Salgado; Colombia-R. Guerrero; Costa Rica-A. Alvarado; England and Wales-B. W. Avery, P. Bullock, and B. Clayden; Federal Republic of Germany-E. Mueckenhausen; Hungary-I. Szabolcs; India-S. B. Deshpande, R. S. Murthy, and S. S. K. Nanda; Iraq-F. H. Altaie; Ireland-T. Walsh and M. Gardiner; Japan-M. Oyama; Kenya-F. N. Muchena and P. M. A h ; Netherlands-J. Schelling and W.G. Sombroek; New Zealand-B. C. Barrat, E. Griffiths, M. L. Leamy, and W. C. Rijkse; Norway-J. Lag; Pakistan-M. B. Choudhri; Peoples Republic of China-Lien-Chieh Li; Romania-A. Canarache and C. Rauta; Scotland-J. M. Ragg; South Africa<. N. MacVicar and M. C. Laker; Sri Lanka-K. A. De Alwis and R. Tinsley; Sudan-M. A. Ali; Thailand-S. Panichapong; Trinidad-N. Ahmad; USSR-I. P. Gerasimov; Venezuela-J. Comerma; Zimbabwe Rhodesia-J. M. DeVillers and W. D. Purves. For the United States: Soil Conservation Service-K. W. Flach, R. L. Guthrie, J. E. McClelland, M. Stout, J . M. Williams, and J. E. Witty; Forest Service-R. G. Cline, D. Eagleston, M. Kaplan, R. Poff, W. Sexton, A. Sherrell, and G. Warrington; Iowa-T. E. Fenton; North Carolina-% W. Buol; Texas-L. P. Wilding. For multicountry areas: G. D. Smith, general perspective; C. Charreau, French-speaking Africa; F. Moorman, Africa and Southeast Asia; S. W. Buol, Latin America; W. B. Peters, World Bank. Citations of the personal communications from these individuals are identified in the text by

(W.

SOIL TAXONOMY OF THE UNITED STATES

195

II. GENERAL REACTIONS TO THE SYSTEM General reactions to Soil Taxonomy in the literature have ranged from absolute rejection to substantial endorsement, each in the context of the author’s precepts of principles of classification and of order in relationships among soils. Some of the literature cited here illustrates an observation by Mulcahy and Humphries (1967) that “It is the subjective, and therefore, emotionally involved nature of most soil classifications which turns the scientist from the narrow path of detached reason. ” The extremely negative reaction of Russian soil science is not easily rationalized. Gerasimov’s (1962) initial review was a critical but reasoned statement. He noted some merits of the Seventh Approximation and was critical of several of its attributes, including his interpretation that it had abandoned the principles of Dokuchaiev. In 1964, a series of I 1 articles in Pochvovedenie (No. 4, pp. 14-48) found little merit in the system’s logic or in its applications to soil classification of specific areas, to soil genesis, to soil survey, or to soil geography. A translation of the summary paper (Gerasimov, 1964) characterized the system as an empirical scheme “justified by references to the most widespread modern bourgeois-philosophical, subjective-idealistic trends” and of ‘‘limited positive interest. ” Later reviews have been similarly critical (Gerasimov, 1969, 1978). Soil Taxonomy is a major departure from the philosophy of classification followed by the Russian school. D’Hoore (1968) noted that Russian soil classification relies heavily on factors of soil formation as criteria of the higher categories. Soil Taxonomy uses intrinsic soil properties selected with the effects of soil formation as criteria, but Yerokhina and Sokolova (1964) appear to consider this a violation of Dokuchaiev’s principles. Fridland (1964) denied that soil classifications are not truth but are contrivances to suit people’s purposes, as declared by the authors of Soil Taxonomy, apparently accepting genetic theory as truth in detail. These kinds of differences and the fact that translations of English to Russian may not convey the bases of Soil Taxonomy fully could account for parts of the critical reviews, but they do not explain them fully. Muir (1962) was critical of the Seventh Approximation, largely on grounds that it provided no statement of basic principles as he conceived them. Webster (1968a) found the strictly defined limits of the criteria of the Seventh Approximation unjustified in view of errors of their measurement and, like Avery (1968), maintained that a hierarchal system is not appropriate for soils. He found no relevance in the concepts of pedon and soil individual (polypedon) and concluded that the system exhibits circular reasoning. Mitchell (1973) published a rebuttal to both Webster and Avery, questioning the validity of their analyses. Webster (1968b) also discussed faults of the system for use by geographers and advised them to avoid using it. Bunting (1969) promptly published a note in the same

196

MARLIN G. CLINE

journal characterizing Webster’s review as most subjective and unjust. These are examples of diametrically opposing viewpoints related to concepts and biases of individuals. Australian soil scientists generally have been critical of Soil Taxonomy, though judgments have varied among individuals. A committee of the Australian Society of Soil Science reviewed the Seventh Approximation in detail. They reported it a great improvement over earlier schemes but with serious problems for soils of Australia (Butler et al., 1961). Stephens (1963) took exception to a number of assumptions, criteria, and definitions and considered the system more nearly a key than a classification. He too found serious deficiencies for soils of Australia. At a time when factual keys were gaining acceptance in Australia, Mulcahy and Humphries (1967) reported that the Seventh Approximation was an entirely commendable attempt to maximize information content but that the use of criteria reflecting soil genesis was logically indefensible. They characterized the choice and weighting of criteria as subjective and so biased by conditions in the United States that application in Australia was not satisfactory. Currently, Northcote (PC) reports that few Australian soil scientists are inclined to devote time to use of Soil Taxonomy. In contrast, Moore ( 1 978) has concluded that Soil Taxonomy has good potential as a medium for transfer of technology in Australia. MacVicar et a[. (1977) have not been completely satisfied with the results of application of Soil Taxonomy in South Africa. They commented, however, that the system is “refreshing” and has “loosened the shackles of traditionalism and stimulated rethinking on soil classification. ” Laker (PC) writes that the system has had a revolutionary impact on ideas about soil classification and soil survey in South Africa, although it is not used as a national system. Sehgal and Sys (1970) found a need for changes in detail of some criteria to accommodate important properties of soils of the Punjab. They emphasized, however, that criticism based on departure from theories of soil genesis is invalid and that genetic relationships are shown more specifically than ever before. They noted the lack of geographic bias for Entisols, Vertisols, Histosols, and Inceptisols but commended the strengthening of relationships between soil classification and soil geography through the concepts of the pedon and polypedon. Langhor (PC) disagrees with ideas expressed in Soil Taxonomy about genesis of fragipans and argillic horizons, which influence choice and weighting of criteria strongly. He would construct some parts of the system differently. Nevertheless, he considers it the most precise and best system for international communication though not well suited to definition of mapping units in the soil survey of Belgium. Like many others, he deplores the complicated writing of the published system. One of the most objective and comprehensive studies of the system has been reported by Ragg and Clayden (1973). They concluded that use of quantitative

SOIL TAXONOMY OF THE UNITED STATES

197

criteria, the concept of diagnostic horizons, emphasis on criteria not readily altered by man, and the nomenclature are valuable contributions. They saw great value in the system as a medium for international reference but found problems for its application in Britain. They identified the complexity created by large numbers of taxa above soil families, the complexity of definitions, emphasis on the mollic epipedon and base status as criteria, and need to amend definitions of some criteria as deficiencies for classification of British soils. They concluded, however, that the system had brought traditional British groupings rightfully into question and had placed soil classification on a more acceptable scientific level. Leahey (1 963) reported that Canadians were greatly impressed by the Seventh Approximation but for national purposes preferred to continue with the system they were constructing. The two systems have been developed concurrently with considerable interaction and commonality of language and structure. Coen (PC) writes that many, if not most, criteria of the Canadian system (Clayton et al., 1977) have been strongly influenced by Soil Taxonomy; some are identical. It should also be noted that Canadians have contributed to Soil Taxonomy. Rauta (PC) reports that although Romania uses its own classification, Soil Taxonomy is respected as a secondary system. He cites its use of intrinsic soil properties, quantitative diagnostic criteria, and connotative nomenclature as important attributes. He is critical of the use of the same properties as diagnostic criteria at different levels in the system and of the complexity of both definitions and their presentation. Some Romanian soil scientists have contributed to Soil Taxonomy and take special interest in it. Duchaufour (1963) noted deficiencies of the Seventh Approximation for use in France but considered it a mark of real progress in soil classification. Kesseba et al. (1972) found it the most comprehensive framework for assessing Tanzanian soil resources even though it did not fit well everywhere. Conflicting viewpoints involving national pride are revealed by a resolution of the New Zealand Society of Soil Science requesting reconsideration of a decision by the Soil Bureau to use Soil Taxonomy on a trial basis. Miller (1978) provided a concise review of experience in New Zealand, the level of interest in Soil Taxonomy, and reasons for experimenting with it.

111. USE OF SOIL TAXONOMY INTERNATIONALLY Table I summarizes the predominant intensity and frequency of use of Soil Taxonomy by scientists of the major institutions concerned with soils in countries for which information is available. Scientists of most countries that are not listed probably use the system infrequently if at all.

I98

MARLIN G. CLlNE Table I The Intensity and Frequency of Use of Soil Taxonomy by Country' Used continually as a primary system

Used frequently as a secondary system

Argentina Chile Colombia EcuadoP Guyana' India Iraq New Zealand Pakistan Sudan United States Venezuela

Belgium Boliviab Brazil Canada Costa Rica England and Wales Ghanad Irand Ireland Japan Kenya Nigeriad Perub Romania Sierra Leoned Sri Lmka Thailand Tanzaniad Trinidad

Used infrequently or not used Australia China, Peoples Republic Franced Germany, Federal Republic Guatemalac Haitic Hungary MaliP Mauritania' Mexico' Netherlands Niger' Norway Panama" Scotland Senegal' South Africa Upper Volta' USSR Zimbabwe Rhodesia

Based on personal communication with scientists in the countries, except as noted. Information provided by S. Buol, North Carolina State University. 'From an unpublished survey of "Soil Taxonomy in the Tropics" by R. Guerrero, University of Puerto Rico. Based on evidence in the literature. Information provided by C. Charreau, ICRISAT, Dakar,Senegal.

A.

COUNTRIES WHERESOILTAXONOMY Is USEDAS

A

PRIMARY

SYSTEM

Soil Taxonomy has been used as the official taxonomic system of the National Cooperative Soil Survey of the United States since January 1965. The United States Forest Service also uses the system for taxonomic classification of soils, although their mapping units are identified in other terms for in-service use in some parts of the country. The system is taught in soils courses of major universities. Soils are identified in terms of its taxa in most articles concerned with pedology in the major journals and bulletins of the United States. Guerrero reports (in an unpublished survey of Soil Taxonomy in the Tropics) that four Latin American countries in addition to the six listed in Table I used Soil

SOIL TAXONOMY OF THE UNITED STATES

199

Taxonomy exclusively in their national soil survey operations, but he did not identify them. He reports that Guyana is currently converting from the 1938 USDA system to Soil Taxonomy. C. Scoppa (PC) reports that soil series and their phases are used for soil mapping at scales larger than 1: 100,OOO in Argentina; subgroups and families are used at smaller scales. Westin (1963) and Comerma (197 1) have discussed the advantages of the Seventh Approximation for soil surveys in Venezuela; Comerma also identified some problems that were encountered in its application. Murthy (PC) reports that 700 copies of an Indian edition of Soil Taxonomy have been distributed and that short courses and regional workshops are being conducted to facilitate its use. In spite of some problems associated with an inadequate data base, Murthy el al. (1977) have identified major advantages of the system for India. Articles by Deshpande et a l . (1971) and Murthy (1979) provide examples of its use. M. B. Choudhri (PC) reports that Soil Taxonomy is used as a primary taxonomic system for soil surveys of Pakistan in conjunction with FA0 units (FAO-UNESCO, 1974). Ahmad et al. (1977) identified soils of the Pakistan Punjab by taxa of Soil Taxonomy without comment. The initial steps of converting from a national soil classification for Iraq to Soil Taxonomy have been described by Altaie et al. (1969). Cook ( 1975) found Soil Taxonomy generally applicable to soils of the Sudan. He reported problems related mainly to deficiencies of information and training of personnel, but he noted that reasonable approximations of classification by the system can be made even though data necessary for precision are lacking. M. A. Ali (PC) reports that the Seventh Approximation was introduced in the Sudan in the 1960s during development of the Soil Survey Administration under FA0 auspices. The soil survey of the Sudan is currently correlating taxa of Soil Taxonomy with units of the FAO/UNESCO legend, as is done in Pakistan. Some problems have been encountered with criteria of Vertisols (Ali, 1972). Ali (PC) also reports that the legend for the soil map of Arab countries is based on subgroups of Soil Taxonomy. The Soil Bureau of New Zealand adopted Soil Taxonomy in 1977 for an extended trial (Miller, 1978). A final decision on its continued use depends on results of that trial. Some individuals in most of the countries that use Soil Taxonomy as a primary system for soil surveys still prefer other schemes. Individuals contacted by the author in Chile, India, and New Zealand indicate that they use Soil Taxonomy mainly as a secondary system for communication. Guerrero’s unpublished survey, cited above, showed that teaching of Soil Taxonomy in universities of many Latin American countries is weak and that a significant number of individuals still rely mainly on systems that were used previously. Guerrero notes in his report that use of these systems is decreasing as workers gain experience with Soil Taxonomy.

200

MARLIN G. CLINE

B. COUNTRIES WHERESOILTAXONOMY Is USED AS A SECONDARY SYSTEM In the 19 countries listed in the second column of Table I, national soils programs depend on taxonomies designed for their conditions or on more comprehensive schemes, such as the FAO/UNESCO legend, the French system, or the 1938 USDA classification. Soil scientists of these countries, however, commonly use Soil Taxonomy for international communication and for correlation of their soils with those of other countries. Butler et al. (1961), Deckers (1966), Ragg and Clayden (1973), Beinroth (1975), and Moore ( 1978) recommended its use for these purposes because of its precision and comprehensive scope, though they preferred other schemes for national purposes. C. Rauta (PC) reports that Soil Taxonomy is not used in Romania for national programs of soil survey and classification but is used frequently in scientific papers, reports for both national and international meetings and symposia, guidebooks, manuals, and Ph.D. theses. De Coninck et al. (1976), for example, identified the soils in terms of Soil Taxonomy for a study of clay mineralogy in Romania. In some of the countries listed, Soil Taxonomy is used in national programs as a secondary system for people of high technical competence, though a less complex scheme is given preference for general use. Dent and Changprai (1973) outlined Soil Taxonomy as the “primary” system for the country’s most competent soil scientists in the Soil Survey Handbook of Thailand, but they retained the national system as a scheme more consistent with the technical competence of most workers. Dewan and Famouri (1964) outlined the Seventh Approximation in their publication on soils of Iran but identified soils of the country in terms of the 1938 USDA classification. Authors who use Soil Taxonomy as a secondary system use taxa of categorical levels ranging from orders to families, depending on the purpose and scale of the study, the level at which soils can be identified with the data available, and the familiarity of the individual with the system. Orders, suborders, and great groups are used most commonly for studies involving broad perspective. Harris et al. (197 1) correlated soil zones of Costa Rica at the order level for a very general picture of the soil geography of that country. Aubert and Tavernier (1972) used suborders and great groups as reference taxa for their small-scale soil map of the humid tropics. Moorman and Van Breemen ( 1978) used great groups to identify soils of the principal rice-growing areas of the world. Camargo and Falesi (1975) and Zamora (1975) identified large areas of Brazil and Peru, respectively, in terms of great groups for broad perspective. Sanchez and Buol (1974) used Soil Taxonomy for their study of soils of the upper Amazon basin, correlating the taxa with units of both the FA0 and Peru systems. Subgroups are used commonly for geographic studies of intermediate intensity

SOIL TAXONOMY OF THE UNITED STATES

20 1

and scale, as in Dijkerman’s (1969) report on soil resources of Sierra Leone. Many authors use subgroups in preference to taxa of lower categories to identify soils used for studies of morphology and genesis. Examples include those by Lepsch er al. (1977) in Brazil, Walmsley and Lavkulich (1975) and Hendershot and Lavkulich (1978) in Canada, Ojanuga et al. (1976) and Harpstead (1973) in Nigeria, Flores et al. (1978) in Peru, and De Alwis and Pluth (1976) in Sri Lanka. Some of these were studies of specific pedons for which data should have been adequate for classification at the family level. Many authors appear to prefer to use subgroups, perhaps because they are taxa of the lowest category in which the criteria have strong genetic bias. Soil families are identified in some studies involving practical interpretations and transfer of technology. The criteria for distinctions at the family level were selected specifically to enhance interpretive value for applied purposes. Uehara (1978) has described the potential of phases of soil families for transfer of technology internationally. North Carolina State University (Soil Science Department, 1978) has used soil families for quantitative identification of soils at experimental sites in the tropics as a basis for transfer of experimental results. Soil families are also used by some authors to identify soils used in studies of morphology and genesis. Family criteria below the subgroup level may be relevant to such studies, as in those by Snedden et al. (1972) and Hakimian (1977). C. COUNTRIES WHERESOILTAXONOMY Is USEDINFREQUENTLY In most of the 20 countries listed for this group in Table I, some individuals have at least studied Soil Taxonomy. Australian soil scientists have studied the system in detail as noted in Section 11, but most do not use it. Yet R. F. Isbell (PC) reports that he uses it as a secondary system. Soil Taxonomy is not known generally in the People’s Republic of China, but Lien Chieh Li (PC) writes that he has a copy and would welcome assistance in using it. The system is well known by leading soil scientists of France. The new French classification (Fauck et al., 1979) uses some of its principles, concepts, and elements of nomenclature, though the taxa and their organization are markedly different. E. Muckenhausen (PC) writes that the system is used little in the Federal Republic of Germany, though it is well known by some individuals. Guerrero’s unpublished study of Soil Taxonomy in the Tropics reveals that Guatemala and Panama use no taxonomic system and that Mexico and Haiti use other schemes exclusively. Even secondary use is unlikely in these countries. C. Charreau (PC) writes that the French-speaking soil scientists of Mali, Mauritania, Niger, Senegal, and Upper Volta use the French system in which they have been trained. From time to time, they do attempt correlations with Soil Taxonomy for surveys sponsored by international agencies. Responses from

202

MARLIN G.CLINE

individuals contacted in the Netherlands, Scotland, South Africa, and Zimbabwe indicate interest in the system by the leaders in soil science, though it is not used significantly. There has been no indication that Soil Taxonomy is used in any way in Hungary or the USSR, but the Russian literature on the subject is ample evidence of awareness. It should be noted that soil scientists of Iran may not use Soil Taxonomy currently. Iran is listed under countries using Soil Taxonomy as a secondary system in Table 1 on the basis of evidence in the older literature.

IV. PROBLEMS FOR USERS OF SOIL TAXONOMY A COMPLEXITY, PRESENTATION, A N D BACKGROUND

With few exceptions, the 53 foreign respondents to the author’s inquiry identified the complexity of the system, its presentation, or both as obstacles to its use. Beinroth (1975), Butler et al. (1961), Cook (1975), Ragg and Clayden (1973), Stephens (1963), and Webster (1968a) have published the same criticisms. The numbers of taxa and the quantitative precision of definitive criteria require complicated definitions and keys, which demand intense concentration and substantial competence if they are to be applied precisely. The difficulty is magnified by the manner in which these are presented in publishing documents. In many instances, precise interpretation of definitions and keys depends on the presence or absence of a comma or the use of the conjunction “and” or “or. A number of individuals with whom the author has discussed the subject assert that the manner of presentation makes application of the system unnecessarily complicated. These attributes of the system created problems for field scientists in the United States when Soil Taxonomy was first adopted. Only part of their difficulties could be attributed to the system and its presentation; a significant part was due to their background of qualitative or semiquantitative field methods and thought. This is probably true in other countries where Soil Taxonomy has been used. In the United States, difficulties of applying criteria largely disppaeared for individuals working with a limited range of soils as they gained experience with that part of the system which was relevant to their work. Experience was fortified by special training not only in use of the system but also through the example of supervisory personnel in the methods of quantitative science. Temporary difficulties in application of the system commonly reappear, though to a lesser degree, when even experienced field scientists encounter soils far removed from those parts of Soil Taxonomy with which they are familiar. These kinds of problems are major obstacles for individuals of countries where Soil Taxonomy is relatively new or is not used regularly. They are especially troublesome where English is not the common language. Respondents to the ”

SOIL TAXONOMY OF THE UNITED STATES

203

author’s questionnaire from 21 countries, including 6 where English is the common language, report that soil scientists classify soils incorrectly in a significant proportion of their placements because of misinterpretation of the keys or definitions. Respondents of only two countries reported that this is not a problem; the author questions their appraisal. Respondents of only three countries where English is not the common language report that translation is a source of error. Others familiar with the work, however, report that translation is a major problem. Buol (PC), responding from his perspective of the work in Latin America, observes that workers who are not fluent in English tend to use narrative descriptions rather than the keys. Consequently, they overlook the order of precedence of criteria, which is critical for classification in the system. He also notes that such problems are “surprisingly few” considering the handicaps under which Latin American soil scientists work. The author is convinced that the paucity of published information about the evolution of the ideas that shaped Soil Taxonomy and the reasons for seemingly arbitrary decisions about definitions of criteria and distinctions between taxa create problems for those who apply the system. Field personnel he supervised were commonly uncertain about the correct interpretation of keys and skeptical of distinctions until they understood the intent and reasons for it. General principles that control the system have been published in the Seventh Approximation (Soil Survey Staff, 1960), by Smith (1963, 1965). and in the “complete” publication (Soil Survey Staff, 1975). To understand the underlying rationale of the system and the bases of decisions it reflects, however, it is necessary to know relevant detail for an enormous amount of data assembled for thousands of individual soils, the results of testing each of six approximations against these data, the conflicts between theory and fact this testing revealed, and the conclusions from debate about ways to accommodate newly discovered facts. Nowhere is that information generally available, and no individual could assemble it in detail. Nor are the underlying precepts of soil genesis which shaped the system obvious in the literature. The system is presented as a device suitable for application empirically, which the inquiring mind finds both unsatisfying and, in parts, unreasonable. To rectify to a limited degree this enormous gap in background information, the author has recently compiled that part available to him (Cline, 1979), but much more is needed from those more intimately involved in development of the system.

B. APPLICATION OF CRITERIA 1 . Field Criteria

The application of diagnostic features of the argillic horizon was identified as a problem most frequently by respondents to the author’s questionnaire. It has also

204

MARLIN G. CLINE

been reported as a problem by Kesseba et al. (1972), Murthy et al. (1977), Nettleton et al. (1969), Ode11 et al. (1974), and Van Wambeke (1967). In 1977 Isbell presented a comprehensive summary of difficulties in identification of argillic horizons (scheduled for publication in the transactions of the meeting of Commissions IV and V of the International Society of Soil Science at Kuala Lumpur, Malaysia). The problems focus on identification of clay skins, estimation of their relative volumes, and conflicting evidence for genetic origin of both clay increases with depth and of clay skins. A second paper by Isbell at the same conference, currently unpublished, discusses the problem in relation to soils of Queensland, Hawaii, Brazil, Natal, and Mauritius. The criteria for plinthite require predicting whether or not the material will harden on exposure. Harpstead (1973) has described assumptions made in such predictions for Nigeria. Daniels et al. (1978) have described the problem and have suggested solutions. The criteria for identifying fragipans are more descriptive than definitive. Deckers (1966) has described some of the problems encountered in the field. From personal experience, the author knows that identification of fragipans is not consistent among workers in the United States. 2 . Laboratory Data Inadequate laboratory data is identified as a major problem by respondents from 22 nations. The cost of obtaining the data is considered limiting by respondents from most developing nations. Articles by Beinroth (1975), Butler et al. (1961), and Cook (1975) express similar concern. Antoine (1977) noted that for northern and western Africa lack of laboratory data is a serious limitation that can be corrected only by a massive laboratory program, for which resources are limited. Deficiencies of laboratory data were problems in the United States while Soil Taxonomy was being developed, and some individuals were critical of criteria that require them. With the resources of the United States, the deficiency was largely corrected, and soil science advanced enormously as a result. Responses to the author’s questionnaire show that supporting laboratory work has increased substantially in several developing nations as a consequence of adoption of Soil Taxonomy. Presumably soil science has advanced in those countries accordingly. Nevertheless, lack of adequate laboratory data is a problem in countries where resources are limited and will continue to be an obstacle to precise use of Soil Taxonomy for the foreseeable future. There does appear to be some misconception of the amount and sophistication of laboratory data that are essential. Obviously, not every pedon examined can be sampled and analyzed, nor is highly intensive sampling for laboratory work essential for realistic use of Soil Taxonomy. Enough data are needed to establish norms and to identify field clues that will permit reasonable estimates. This is the

SOIL TAXONOMY OF THE UNITED STATES

205

pattern followed in the United States. Murthy et al. (1977) have noted that reasonable approximations of many criteria can be made without resorting to sophisticated procedures. Cook ( 1975) reported that realistic estimates of criteria for classification were made in the Sudan even when laboratory data were inadequate for high precision. 3 . Inferred Criteria

Lack of data to establish soil temperature and moisture regimes is identified as a major problem by a large proportion of the respondents to the author’s queries. Ideally, direct measurement of these properties over time is preferred, but the time required, especially for soil moisture regimes, discourages workers in many countries. Consequently, these attributes are commonly inferred from other information. Uncertainty about estimates of soil moisture regime has been reported much more frequently than about estimates of soil temperature. Cook (1975) reported that estimating soil moisture regimes is a major problem for the Sudan; Antoine (1977), for northern and western Africa; Deckers (1966), for the Belgian Ardenne; Kesseba et al. (1972), for Tanzania; Isbell and Field (1977), for Australia and Brazil; and Murthy et al. (1977), for India. Antoine (1977) has emphasized that better correlations are needed between soil moisture regimes and both atmospheric climate and plant-water relations for reliable estimates. Respondents to the author’s questionnaire from the U.S. Forest Service and from Canada report that estimates of both soil moisture and soil temperature regimes are unreliable in vertically zoned mountainous areas. Establishing the necessary data base in these areas would be costly and time-consuming. Guidelines for estimating soil temperature regime (Smith et al., 1964) provide bases for reasonable approximations of soil temperature regimes in many areas if air temperature data are available. The guidelines need modification for regions where climates and vegetative cover differ markedly from those where the relationships were established. C. TRAINING A N D BIASOF PERSONNEL

Precise use of Soil Taxonomy demands expertise in field techniques, diligence in their application, and subordination of preconceived ideas about classification of soils. In the United States, Soil Taxonomy forced a major transformation of most of the field staff from qualitative observers in an atmosphere of historically derived concepts to quantitative investigators in an environment of detached reasoning. The transformation was not made easily or quickly, nor is it complete. It has involved intensive training, both formally and informally, which continues

206

MARLIN G . CLINE

as new personnel enter service. The system is no less demanding in other countries. Respondents from 10 foreign countries, including 3 of the developed nations, report that field workers commonly overlook definitive properties or fail to determine them precisely. Cook (1975) reported that inexperienced personnel prepared inadequate soil descriptions and mixed critical horizons when sampling for laboratory studies to determine criteria of the system. Personal communication with individuals who have observed field operations in several countries indicates that errors of these kinds are more widespread than the responses suggest. Buol (PC) comments that essentially all of the leading soil scientists of Latin America are trying diligently to use the system correctly but with too little direction in more than empirical application. These are problems that only experience and training can correct. Errors in application of Soil Taxonomy are not confined to field personnel. One response to the author’s questionnaire, for example, reports need for an “andic” subgroup of Ustox on the grounds that the soil is dominated by clay of high activity. This individual apparently does not understand that the Oxisol order, of which such a subgroup would be a member, is defined specifically to include only those soils which are so highly weathered that clays of high activity would not be present. He did not question the concept of Oxisols. Personal bias is, and probably always will be, a problem of soil taxonomists. (The term bias is not used in a derogatory sense.) It is exhibited among users of Soil Taxonomy in many ways, ranging from sincere disagreement with the system’s own bias to intellectual dishonesty in its application. Among the former, a number of respondents to the author’s inquiries report that Soil Taxonomy classifies some soils “incorrectly ” because it separates them from other soils that are considered members of the same taxon in another system. The most common “problem” of this kind is the complaint that many, if not most, Latosols (Laterite, Lateritic soils) are not Oxisols by criteria of Soil Taxonomy. The complaint may or may not have merit, but the real problem is that individuals base their criticisms on preconceived ideas rather than on the merits of the alternatives. Respondents from three countries say that some individuals deliberately classify soils incorrectly by criteria of Soil Taxonomy to preserve groupings of another system they have used. Six others report that some deliberately classify incorrectly to show similarities of soil potentials for use. Other misrepresentations of soils in the system arise from the mistaken idea that the system is infallible. Some workers distort or overlook factual evidence to make soils fit the system (Cline, 1977). Wilding (PC) mentions failure to report observations of facts that do not fit the system. These kinds of omissions and distortions conceal errors in the system, which was designed deliberately to expose them. Most countries that use Soil Taxonomy as a primary system have some provi-

SOIL TAXONOMY OF THE UNITED STATES

207

sion for training, if only informally by supervisory personnel. Buol (PC) indicates that this too commonly consists only of instruction in empirical application. India provides one of the more intensive training programs (Murthy, PC). Antoine (1977) has emphasized that in North Africa, where Soil Taxonomy is used little, acceptance of the system would depend on training prior to attempts to have field workers use it. Guerrero’s unpublished survey of Soil Taxonomy in the Tropics rates teaching of the system in Latin American universities as generally weak. Teaching and training are likely to remain problems in countries where resources are limited. The needs for training in many countries include instruction not only in use of the system but also in basic soil science and in scientific attitudes and ethics.

V. TAXONOMIC PROBLEMS Well over 100 “problems” are reported by respondents to the author’s questionnaire. Many are local problems for which solutions proposed would adversely affect classification of soils of other areas. Others are problems reflecting the personal bias and incorrect use of criteria discussed in the preceding section. Some are clearly deficiencies of the system and affect classification of soils in the world scene. The major ones of these are discussed here. A. NEEDFOR ADDITIONAL TAXA

This is the taxonomic problem mentioned most frequently by respondents. G.

D. Smith (PC) notes that lack of an appropriate taxon to accommodate unique kinds of soils is a problem he has encountered frequently in his extensive travels. It results in classification of unlike soils in the same taxon. The system was constructed to permit additions (Flach, 1963), so incomplete classification can be corrected. Uninhibited additions of taxa could, however, result in a chaotic mklange replete with contradictions. Some device is needed to weigh the merits of proposed new taxa in terms of the need for them, impact on other parts of the system, and appropriate criteria to define them. The main proposals for additional taxa which have come to this author’s attention are reported here, mainly with little comment for lack of bases to appraise them. Guthrie (PC) reports need for five great groups of Arents based on soil moisture regimes for the Southeastern United States. He also suggests alfic, ultic, spodic, ochreptic, and oxic subgroups of one or more of them. Arents are currently unclassified below the suborder level. Isbell (PC) notes the inadequacy of current great groups of Ultisols for Australia. Coover et al. (1975) have

208

MARLIN G. CLINE

proposed Halaqualf and Halaquept great groups with typic, salic, and histic subgroups for 1.5 million hectares of coastal marshlands of the United States. Their analysis of the problem has been reinforced by Coultas and Gross (1975). Westin er al. (1968) have reported need for a “Tropoll” suborder for Mollisols of ustic isohyperthermic regimes to distinguish them from Ustolls of temperate regions. The “Trop-” designation may be inconsistent with this use. Most of the proposals are for additional subgroups. Need for andic subgroups of a wide range of great groups is reported by respondents from regions having volcanic deposits: Alvarado (Costa Rica), Muchena (Kenya), Rijkse (New Zealand), Cline ( U S . Forest Service). These problems are related to those of Andepts discussed later. Need for additional halic, salic, and natric subgroups are also noted by several respondents and authors: Salk Udifluvent (India), Salic Natrargid (Sudan), Natric Torrifluvent (Pakistan), Halic Hydraquent (Coover er al., 1975), salic and salic-natric subgroups of Calciorthids, Camborthids, and Haplustalfs for the Indo-Gangetic Plain (Sehgal et al., 1975). Langohr (PC) describes a polysequum having a thick cambiclike horizon over an argillic horizon in Belgium. He considers these soils members of an unnamed subgroup of Hapludalfs. Ragg and Clayden ( 1973) reported unrecognized varieties of Placaquods. Leamy (PC) reports need for an ustic subgroup of Fragiochrepts. Two representatives of the U.S. Forest Service (PC) describe soils that fall in the Paleboralf great group because of their depths to the argillic horizon, yet have a fragipan. These are not accommodated well at present and may justify a new subgroup. Lewis (1977) has described an Argiudoll with vertic properties and near-ustic moisture regime, which he would classify as an Ustertic Argiudoll. Luzio and Menis (1975) described evidence of illuvial clay in lower horizons of Vertisols, and Luzio (1978) found argillans in some horizons of Pelloxererts, Durandepts, and Xerofluvents. Although potential alfic subgroups may be inferred, this is part of the argillic horizon problem described later. Kesseba et al. (1972) have reported need for an alfic intergrade to Ultisols. Guerrero (PC) reports need for additional subgroups of Ustalfs, Ustolls, and Oxisols in Colombia without specifying kinds. Sherrell (PC) suggests additional subgroups of Durumbrepts and Haplohumults in areas mapped by the U.S.Forest Service. Tan et al. (1970) proposed a new subgroup for some Spodosols of the tropics.

B. SOILSOF THE TROPICS Soil Taxonomy (Soil Survey Staff, 1975) calls attention to the fact that the classification of soils of the tropics remained to be tested as the manuscript went

SOIL TAXONOMY OF THE UNITED STATES

209

to press. That of Oxisols in particular was described as a compromise of conflicting judgments which might well satisfy no one. Problems that have been found are by no means confined to the Oxisols. The definition of the Oxisol order, including the range of soils it should encompass, remains a matter of controversy. An international committee has been charged with recommending solutions. Progress by the committee has been slow. One of the most troublesome problems involves relationships between the Oxisol order and the orders of Alfisols and Ultisols. Soils of extensive areas of the tropics are classified as Alfisols and Ultisols on the basis of presence of an argillic horizon. Many of these soils are characterized by properties associated with clays of low activity not strikingly different from those of the Oxic horizon. The contrast between those soils and their counterparts of the Alfisol and Ultisol orders in temperate zones is striking. The problem is widespread. Camargo (Brazil), Murthy (India), Buol (Latin America), De Alwis (Sri Lanka), Comerma (Venezuela), MacVicar (South Africa), and Ali (Sudan) identify it in responses to the author’s questionnaire. Gowaiker (1972), Murthy et al. (1977), and Rengasamy et al. (1978) have published reports about this problem for India. Van Wambeke (1967) identified the problem early in the development of the system. Sanchez and Buol (1974) have discussed it for soils of the upper Amazon basin; Lepsch and Buol (1974), for Sao Paulo State of Brazil. Isbell and Field (1977) were unable to classify certain soils of Brazil and Australia with confidence in either the Oxisol or Ultisol order because of uncertainty about presence or absence of an argillic horizon. Isbell has described the problem in detail (for the Proceedings of the 1977 International Workshop on Soil Classification held in Malaysia and Thailand). He suggested broadening the definition of Oxisols to resolve it. An international committee has been working on the problem over a period of several years. It has considered a number of options, including definition of special “kandi” great groups of Alfisol and Ultisol orders to accommodate these soils. Firm decisions have not been reached. Soil moisture regimes for the tropics also remain a controversial issue. Isbell has commented on it in his unpublished paper cited above. An international committee has been formed to recommend a solution, but its conclusions have not been reported. Reconsideration of great groups for soils of the tropics is also needed (Smith et al., 1975) not only for Oxisols but also for some Alfisols and Ultisols. Among respondents to the questionnaire, Sombroek (East Africa), Murthy (India), Muchena (Kenya), and De Alwis (Sri Lanka) variously report problems differentiating between Pale-, Rhod-, and Hapl- great groups of Ustalfs. Similarly, Buol (Latin America) and Sombroek (East Africa) identify problems distinguishing between Pale-, Rhod-, and Hapl- great groups of Udults, Ustults, and

210

MARLIN G. CLINE

Aquults. Great groups of soils of the tropics are listed as problems needing attention by McClelland (Soil Conservation Service), but no international committee is known to be assigned the problem at this time. C. ANDEPTS

The range of properties of the Andept suborder is very broad and is not adequately subdivided in lower categories. Bases can range 100-fold; soil moisture regimes include udic, ustic, and xeric; soil temperature regimes are even more varied. Problems related to this variation have been communicated by G. D. Smith, R. Cline and Warrington (US.Forest Service), Williams (Soil Conservation Service), Salgado (Chile), Alvarado (Costa Rica), and Rijkse and Leamy (New Zealand). Identification is an additional problem. The determination of thixotropy, for example, is highly subjective. Cortez and Franzmeier (1972) have questioned the limits for volcanic glass. These problems extend to Andic subgroups of a number of other suborders. The classification of Andepts is being studied by an international committee. One proposal would elevate the Andepts to order status, which would allow more options for distinguishing varieties by criteria consistent with those of other orders of the system. Proponents argue that this would recognize the character of the group at a level more nearly consistent with its uniqueness. As of this writing, the committee has not made recommendations.

D. PERGELIC SOILS Few respondents to the questionnaire mentioned the Pergelic soils, and the literature on applications of Soil Taxonomy to them is limited. Pettapiece (1975), however, has identified major problems in Canada. Currently, these soils are identified largely as pergelic subgroups of a variety of great groups, permafrost being treated as an “extragrade” property. Pettapiece and a number of authors whom he cites have documented extreme physical processes of disruption and mixing associated with cryoturbation, especially in wet soils. The resulting hummocky microrelief marks an abrupt break in soil character horizontally. Pettapiece finds genetic interrelationships between soils of the depressions and of the hummocks. The cyclic character of the resulting soils has led Pettapiece to the conclusion that the classification of these soils should be analogous to that of Vertisols, perhaps involving a sampling unit that includes depression and hummock as a unit. Pettapiece reported that a committee of Canadian Soil Survey has proposed a “Cryosolic” order in their system, paralleling the Vertisol order of Soil Taxonomy. In addition to Pettapiece, Tarnocai of the Canadian Land Resource Institute

SOIL TAXONOMY OF THE UNITED STATES

21 1

(PC) deplores the low level in Soil Taxonomy at which soils with pergelic temperature regimes are identified and the lack of good devices to separate cryoturbated soils. McClelland (Soil Conservation Service) lists the classification of these soils as problems demanding attention. E. VERTISOLS

Ali (1972) has reported that many of the cracking clays of the Sudan, commonly considered good examples, fail to meet criteria defined for Vertisols. Slickensides may not intersect; gilgai relief may be minimal; and aggregates may be platy instead of parallelepiped in a given soil unit. He also found that the distinction between Chromusterts and Pellusterts on the basis of chroma and the differentiation of subgroups on the basis of color value do not make the most meaningful distinctions. He reports that hue is useful in some instances. G. D. Smith (PC) lists Pellusterts and Chromusterts as groups presenting serious problems due to low correlation of chroma with drainage or erosion. He also cites failure to use acidity as a criterion for soil families as a deficiency. Murthy (PC) reports similar conclusions about use of chroma for India, as does Comerma for Venezuela. Comerma (PC) suggests also that a suborder of Aquerts is needed. Luzio (PC) reports a need for salic and natric subgroups of Xererts. Luzio (1978) and Luzio and Menis (1975) have described weak argillans in deep horizons of soils classified as Vertisols; Comerma (1971) reported uncertainty about distinguishing between Vertisols and vertic varieties of Alfisols, which may be related to Luzio’s findings. F. DIAGNOSTIC CRITERIA

The weighting, definition, and role of the argillic horizon are subjects of much controversy. The horizon is used in Soil Taxonomy as a mark of the illuviation of clay, which is weighted heavily in the system on the belief that it is an extremely important genetic process. Its definition is based on properties believed to be evidence of that process and allows for limits of observation and measurement. Problems of identification and those involving the horizon in Alfisols and U1tisols having low-activity clays have been discussed in preceding sections. Isbell (paper for publication in the 1977 transactions of Commissions IV and V, International Society of Soil Science, Kuala Lumpur, Malaysia) has summarized much of the criticism. He questions the genetic implications attributed to the horizon, as does Langhor (PC). MacVicar (PC), Isbell, and others believe that Soil Taxonomy weights illuviation of clay too heavily by using the argillic horizon as a criterion at the highest level in the system. Much of the problem with

212

MARLIN G . CLINE

Alfisols and Ultisols having low-activity clays can be attributed to its use at that level. Isbell as well as Nettleton et al. (1969), Gile and Grossman (1968), and Beinroth et al. (1974) have cited evidence that the presence or absence of argillans is not infallible proof that illuviation has, or has not, been significant. Oertel ( 1 968) contended that the ratio of fine to total clay is not a trustworthy criterion. Bullock (PC) questions the universal validity of both criteria. Few soil scientists appear to question the significance of argillic horizons at some level in the system, but many are critical of its definitive criteria and the weight given to it. It has been this author’s observation that field personnel commonly “stretch” the defined limits in their enthusiasm to find the horizon. Problems with criteria of the spodic horizon have persisted throughout the development of Soil Taxonomy and to the present. Avery et al. (1978) found that chemical criteria exclude from Spodosols many soils which are “spodic” by morphology. Witty (Northeastern States), Stout (Midwestern States), Gardiner (Ireland), Schelling (Netherlands), Lag (Norway), Ragg (Scotland), and R. Cline (Forest Service) support that observation in personal communications. McClelland (PC) lists criteria for the spodic horizon as a problem the Soil Conservation Service is studying. Criteria of wetness continue to cause problems in some soils. Pilgrim and Harter (1977) have reported that masking of low chroma mottles by iron and humus of the spodic horizon results in inclusion of a high proportion of wet Spodosols in Typic Haplorthods. Lavkulich (PC) also identifies this as a serious problem. Buol (PC) has difficulty with soils that classify as Typic Paleustults because chroma is not less than three, though the water table may be within the solum for long periods. Avery (PC) also describes water table studies which show that wet soils do not necessarily have mottles of two chroma or less. He attributed significance to ferruginous mottles in some soils. Sehgal and Sys (1970) have reported difficulties with criteria of wetness in the Punjab. G . MISCELLANEOUS PROBLEMS

In addition to the major problems discussed above, a number of relatively less widespread difficulties were reported by respondents to the author’s inquiries or were found in the literature. Examples of these are described here, excluding those which were clearly related to misinterpretation of criteria or to personal bias. De Bakker (1971) reported that stratification in soils of Dutch fluviatile and marine sediments appears to occur capriciously locally. As stratification is the critical criterion determining presence or absence of gleyed cambic horizons in these soils, Aquents and Aquepts are intimately intermingled with little apparent relation to landscape units or differences in genesis. Both Fenton and Stout (PC) are concerned with application of criteria for the

SOIL TAXONOMY OF THE UNITED STATES

213

mollic epipedon to eroded former Mollisols in the midwestern States. Technically, the eroded soils are no longer Mollisols, but they are intimately intermingled with Mollisols in the landscape. Both Avery and Clayden (PC) report that base status as a definitive criterion separates Eutrochrepts and Dystrochrepts in an intricate rectangular pattern where a long history of liming has changed the base saturation of soils in some fields but not in others. Avery also reports that man-made mollic epipedons create analogous complex patterns of soils distinguished at the order level. The foregoing are mentioned because they all represent mapping problems as artifacts of the system. Devices exist for handling such geographic mixtures in legends, but they are commonly awkward. As some respondents have indicated, changes to accommodate such situations may well create others elsewhere. Some difficulties have been reported for the clay mineralogy criteria at the family level. Witty (PC) reports that the definition of oxidic mineralogy classifies many soil series of New England in oxidic families. These are soils on very young glacial deposits. De Alwis (PC) reports problems distinguishing between oxidic and kaolinitic mineralogy in Sri Lanka. Fenton (PC) finds that failure to differentiate montmorillonitic and illitic mineralogy in fine loamy and fine silty families detracts seriously from the interpretive value of soil families in Iowa. Comerma (PC) has difficulty rationalizing the classification of those Haplustalfs that lack zones of carbonate accumulation in the udic subgroup when they clearly have an ustic moisture regime. Both Murthy and Nanda (PC) report the same problem with Udic Haplustalfs and Udic Paleustalfs in India. Scoppa (PC) is faced with a "pale-" great group of Mollisols having a udic moisture regime but with petrocalcic horizons. The key classifies these as Paleustolls. Sehgal et al. (1975) have reported that soils having salinity or sodicity limitations on the Indo-Gangetic Plain do not key out of typic or aquic subgroups of Calciorthids, Camborthids, and Haplustalfs. They have proposed salic and salic-natric subgroups. Both Leamy and Williams (PC) write that bulk density requirements keep some soils having other properties of andic subgroups out of those taxa, with serious consequences for interpretation. R. Cline (PC) reports from the Forest Service that some soils are classified as Andic Dystrochrepts by methods defined for base saturation, though research indicates that the soils are nearly base saturated at the pH of soils in the field.

VI. IMPACT OF SOIL TAXONOMY INTERNATIONALLY A . SOILCLASSIFICATION

The impact of Soil Taxonomy on soil classification internationally varies from adoption of a few concepts or terms in some countries to use as a primary system

214

MARLIN G. CLINE

in others. The evidence for impact in countries that have not adopted the system consists mainly of similarities between the innovative aspects of Soil Taxonomy and those of other schemes. Such evidence must be appraised cautiously. Some similarities might have developed independently. Others are known to be the result of ideas that originated with foreign soil scientists and were incorporated in Soil Taxonomy. The author has tried to confine his reporting to those aspects that would probably have taken somewhat different form in foreign countries had Soil Taxonomy not been available and to those that evolved through mutual exchange of ideas internationally at the initiative of Dr. Smith and his associates. The Canadian soil classification (Clayton er al., 1977) is most nearly like Soil Taxonomy among systems in general use. Although it differs from Soil Taxonomy somewhat in organization and uses different names, the greatest difference in principle is probably its use of diagnostic features that may easily be altered by man. The two systems were developed with frequent interchange of ideas, though they were not the results of a joint effort. The soil classification for the Soil Survey of England and Wales (Avery, 1973) departs substantially from Soil Taxonomy but uses quantitatively defined criteria, a three-dimensional “profile” comparable to the pedon, and some criteria and terms of Soil Taxonomy. Ragg and Clayden (1973) noted that Soil Taxonomy had prompted scrutiny of traditional British soil classification. The new French taxonomy (Fauck et al., 1979), like Soil Taxonomy, uses genetic theory with a strong bias for practical application, though the weighting of elements of genesis is greatly different. It uses the pedon and polypedon concepts. The nomenclature is similar in construction, though the terms are mainly different. The FA0 legend for the soil map of the world (FAO-UNESCO, 1974) is used as a primary or secondary classification in many countries. Although far from similar in outline and detail, it uses many of the devices of Soil Taxonomy. The nomenclature of some parts is similar in construction. Diagnostic horizons defined quantitatively are used as criteria; many are identified by terms used in Soil Taxonomy and are defined similarly. The correlations of the map units with taxa of Soil Taxonomy, however, are far from perfect (Beinroth, 1975). Rauta (PC) reports that Soil Taxonomy has influenced soil classification in Romania through increased precision of definitions and use of some of its diagnostic criteria. Some soil orders are the same. Thailand (Dent and Changprai, 1973) classifies soils by Soil Taxonomy, though it uses a national system for local use. Gardiner (PC) reports introduction of quantitative criteria from Soil Taxonomy into the system used in Ireland. Laker (PC) writes that the Seventh Approximation had a revolutionary impact on perspective of soils in South Africa, though a greatly different classification is used. Oyama (PC) states that “without the Seventh Approximation and Soil Taxonomy, soil science in Japan would be far from science. In that sense, the

SOIL TAXONOMY OF THE UNITED STATES

215

influence of Soil Taxonomy in Japan cannot be written.” This kind of impact is probably more important to soil classification world-wide than the adoption of specific concepts and devices in individual countries. B. SOILSURVEYS 1 . Soil Survey Operations

In the United States, Soil Taxonomy brought about an enormous increase in the precision and detail of field descriptions and a large increase in both kinds and numbers of supporting laboratory determinations. More importantly, it transformed the field force by necessity from qualitatively oriented observers to quantitatively inclined investigators. It had relatively little impact on the field methods of surveys, but it substituted quantitative criteria for much of the personal judgment that had dominated soil correlation and the development of legends. In the other 11 countries that use Soil Taxonomy as a primary system, the impact has been much less. Individuals from 9 of the 11 countries responded to the authors inquiries about the impact on soil surveys. All except the respondent from Pakistan report an increase in both the detail and the precision of field descriptions. Respondents from Argentina, Chile, Colombia, India, New Zealand, and Venezuela report an increase in laboratory support, mainly by the addition of determinations required for diagnostic criteria. The total input of laboratory technician time, however, ranges from a high of 14 man-years to a low of 0.25 man-year annually. Correlation in the soil surveys of developing nations of this group is notoriously weak. Responses to the author’s inquiries show that only New Zealand, Venezuela, and Colombia use the criteria of Soil Taxonomy both to define mapping units in field legends and to correlate mapping units when the field work is complete. In Argentina, India, and the Sudan, the criteria of Soil Taxonomy are apparently used mainly a f e r field work is complete to combine some units for publication and to correlate the published units with taxa of Soil Taxonomy. Murthy (PC), for example, reports that correlation of the 700 soil series established for mapping in India is only beginning. The situation in India is analogous to that in the United States in the 1950s, though the number of established series is much less. Soil Taxonomy required major revisions of existing soil series definitions before its criteria could be applied consistently in the development of legends for detailed surveys. Similar adjustments will be a major undertaking in countries that adopt the system. Some appreciation of the size of that task, including the necessary training, may be obtained from Cook’s (1975) report on use of Soil Taxonomy in the Sudan.

216

MARLIN G. CLINE

Indirect impact of Soil Taxonomy on soil survey operations is reported by respondents from Canada, Kenya, Sri Lanka, Trinidad, Romania, Ireland, Brazil, England, Japan, Belgium, and Costa Rica among countries that use the system only in a secondary role. They indicate that standards of precision, detail, or both have increased to some degree for field descriptions. Six of the 11 also report addition of new laboratory determinations for criteria used in classification. One respondent from Australia, which uses the system little, mentions some effect on field descriptions and addition of a few laboratory determinations.

2. Soil Maps and Mapping a . Soil Boundaries. Orvedal and Austin (1963) predicted little change in the number and location of soil boundaries in detailed soil surveys of the United States as a consequence of Soil Taxonomy. The boundaries of most delineations are based on the visible landscape. The identity of some mapping units did change as soil series were redefined and new series were established, but soil scientists contacted in the United States confirm that the effects on mapping were small. Six individuals report that “some” additional boundaries are drawn to accommodate criteria of Soil Taxonomy. Only two say that ‘‘some” boundaries that would have been drawn previously are located differently. Orvedal and Austin (1963) also compared compiled maps using the 1938 system with maps based on Soil Taxonomy for scales of 1:5,000,000 and 1 :20,000,000. They found that some adjustment of boundaries was necessary, mainly to accommodate the new soil temperature and soil moisture criteria that Soil Taxonomy introduced. Respondents from five foreign countries that use Soil Taxonomy as a primary system report that mappers use its criteria to make distinctions in the field. Only three of the five, however, report that the criteria are incorporated in legends; one may assume that use of the criteria is minimal for the other two. Four report that “some” boundaries that would not otherwise have been recognized are drawn; only two report that the location of other boundaries is affected. All of those reporting a significant impact on mapping are from countries where mapping intensity is less than for detailed soil surveys of the United States and where quality control has historically been poor. Respondents from eight countries where the system is used only in a secondary role report that soil mappers do use some of its criteria in the field. Respondents from Belgium, Brazil, Romania, and Sri Lanka report that some new boundaries are drawn as a result. None consider that location of other boundaries is affected significantly. b. Purity of Mapping Units. If, as reported, the location and number of soil boundaries drawn in the field is affected little by using Soil Taxonomy, it follows that the change of soil variation within delineations should be equally small.

SOIL TAXONOMY OF THE UNITED STATES

217

Assuming that any relocation or addition of boundaries accurately reflects adjustment to conform with new criteria, any change in purity should be positive when measured against composition according to taxa of the new system. Experience in the United States shows that the apparent purity of mapping units of detailed soil surveys may decrease when the criteria of Soil Taxonomy are applied as the basis of measurement. In an extreme instance, McCormack and Wilding (1969) studied the composition of mapping units in an area having intricate patterns of soils in stratified lake sediments. They measured the composition of delineated areas in terms of soil series present, first using concepts of series that predated Soil Taxonomy and then using series defined by the limits of criteria of the new system. They found that the proportion of the total area outside the ranges of series named increased by about one-third. Although the areas were not remapped using criteria of Soil Taxonomy, it has been this author’s experience in similar areas that mapping would have changed little. The apparent increase in inclusions is mainly an artifact of the quantitatively defined limits of taxa. The limits of criteria of all categories accumulate as limits of soil series. Thus, a pedon selected from a polypedon identified as a series of a coarse loamy family is outside the range of its series if it contains 19% clay-I% above the upper family limit. The accumulated limits at the series level cleave conceptual “splinters” of soil of another potential series from soil bodies that are identifiable as natural units. Commonly, these differ in only a minor degree of one or two properties from the series identified in mapping. Many are unclassified at the series level. The term “taxadjunct” has been coined to characterize unclassified soils that differ from established series to such a minor degree that soil potential for use is substantially the same. These are treated as if they were members of the parent series for practical purposes. c . Quantitative Limits. The discussion of the preceding section leads to reconsideration of quantitative limits and their application to bodies of soil in the field. The quantitatively defined criteria of Soil Taxonomy have been identified as one of the system’s most laudable attributes by 22 of the respondents to the author’s inquiries and by authors such as Antoine (1977), Beinroth (1975), Butler et al. (1961), and Ragg and Clayden (1973). It is the characteristic of the system that has had the greatest impact on soil science in general, soil classification in particular, and the scientific attitudes of soil survey workers. Webster (1968a, 1968b) has emphasized, however, that the absolute limits of Soil Taxonomy are not consistent with the errors of estimates of criteria in practice. Wilding (PC) expresses similar concern and notes that too few field workers appreciate the errors inherent not only in observation and measurement but also in field sampling. It is apparently not commonly appreciated that precise definition of limits for differentiating criteria does not a priori imply unyielding application of those limits to the conceptual boundaries of taxa. Wilding (PC) contends that although

218

MARLIN G. CLINE

limits of criteria should be defined in terms of single values, as at present, they should be applied to boundaries of taxa with latitude for variation consistent with the probable error of the estimate. To carry that idea one step further, not only could the degree of allowable departure from defined “limits” be prescribed but also how many criteria in combination might be permitted to violate their “limits” to that degree for determining the boundaries of taxa. This is effectively what is done in the definition of taxadjuncts, most of which would become members of their parent series in concept as well as in practice if such a device were adopted as general procedure. Such a convention would require substantial study and testing, and it would probably result in some loss of consistency of classification among individuals. It might, however, provide solutions to a number of the “problems” reported by respondents to the author’s questionnaire. C. INTERPRETATIONSFOR APPLIED OBJECTIVES

A sharp distinction is made in this section between the evaluation of soil taxa for applied objectives and predictions about potentials, limitations, and management needs of the real bodies of soil that are identified by the names of taxa on soil maps.

I , Evaluation of Taxa Bartelli (1978) has listed 11 objectives for which soil surveys are routinely interpreted in the United States. Although these interpretations are presented as predictions for mapping units that represent real soil bodies, they are, in fact, based mainly on the defined attributes of the phases of soil series used to name the mapping units. They do not usually evaluate the effects of the inclusions that are normally present. Phases of soil series are not new to Soil Taxonomy; comparable units would have been identified and interpreted if Soil Taxonomy had not been developed. Respondents of the Soil Conservation Service, the Forest Service, and several states report, however, that Soil Taxonomy has resulted in significantly more precise interpretation because of the quantitative limits it has imposed on the range of properties of soil series. It should be understood, however, that this increase of precision applies mainly to conceptual units of classification, not necessarily to the real bodies of soil that are identified by their names. Respondents from Argentina, Chile, India, New Zealand, the Sudan, and Venezuela report that adoption of Soil Taxonomy has increased both the precision and number of soil interpretations. Respondents from 4 of the 10 countries where Soil Taxonomy is used only as a secondary system report some increase in

SOIL TAXONOMY OF THE UNITED STATES

219

precision. The author would judge that the impact has been relatively small in all of these countries, which have either adopted Soil Taxonomy very recently or have limited programs for soil interpretation. As in the United States, increased precision applies primarily to taxonomic units, not mapping units. The foregoing refers mainly to interpretations for phases of soil series or their equivalents. Orvedal (1977) has stressed the sacrifice of information on which interpretations can be based as the taxa increase in heterogeneity from soil series to higher categories. Westin (1974) showed that the coefficient of variability of the selling price of land approximately doubled from strata representing soil series to strata representing orders. Chan (1978) found that soil series are correlated with productivity of Hevea in peninsular Malaysia; he could assign subgroups to five classes of productivity but the range of variability within subgroups was large. He found that the range of productivity was very large within great groups, though the best clones of Hevea realized their full potential only on soils of a few specific great groups. While the range of soil properties within taxa increases from series to families, phases of soil families are believed by a number of workers to be homogeneous enough to serve as basic units for transfer of technology on a broad geographic scale. Uehara (1978) has described the rationale for this function of the family. Preliminary results of research to test the hypothesis have been published (Anonymous, 1978). North Carolina State University uses soil families to provide unequivocal identity of soils at experimental sites in the tropics (Soil Science Department, 1978). Respondents to the author’s questionnaire suggest that soil families are suitable taxa for characterizing the soil factor in some Forest Service land inventories for predicting forest productivity, revegetation potential, soil stability, and similar elements of forestry operations. Respondents from India, New Zealand, the Sudan, and Venezuela report that soiI families have been tried and have proved useful for some applied interpretations, but their experience is limited. At successively higher categorical levels, the range of properties within taxa obviously restricts soil interpretation to increasingly broad objectives. Nevertheless, respondents from five countries report interpretive value of taxa at levels as high as the suborder. Ten cited the introduction of soil climatic parameters at a high level as an important element for enhancement of the interpretation potential of taxa of the higher categories. Ikawa (1 978) has discussed this in some detail. It is possible, for example, to identify soils in which wetness is limiting at the suborder level. Specific examples of interpretations for taxa above the family level are rare, however. Others report deficiencies of Soil Taxonomy as a base from which applied interpretations can be made. Stephens (1963) criticized the system because it separates similar soils in Australia and reduces the correlation of taxa with land use potential. This is not a unique attribute, nor is it limiting. Soils of similar

220

MARLIN G. CLINE

potential are classified in different taxa in most, if not all, taxonomies and are routinely grouped in technical classifications for applied uses (Bartelli, 1978). Problems are more serious if Soil Taxonomy fails to distinguish between soils of unlike potential. Instances of this kind reported by respondents include the following: 1. The range of soil temperature regimes is too wide. 2. Soils having fragipans are not distinguished from “pale-” great groups. 3. Color criteria of Vertisols do not make meaningful separations. 4. Some soils with pronounced ustic moisture regimes are classified as udic subgroups.

Buol (PC) reports that failure of the system to make distinctions on the basis of properties of surface horizons leaves soils having significantly different potentials for primitive farming in the same taxon at the family level. Buol’s concern is with the loss of confidence of soil scientists who expect too much of the system-not with the omission of properties of surface horizons. The examples given above are, of course, appropriate criteria for soil series or for phases of families. None of the deficiences reported is unique in principle to Soil Taxonomy nor is insurmountable if the principles and techniques of soil interpretation are understood. Low levels of technological understanding are more critical than deficiencies of the system. Guerrero’s survey of Soil Taxonomy in the Tropics, for example, reports vigorous criticism because (1) the taxa are not interpretive groups and (2) the interpretive potential decreases from low to high categories. Both of these attributes are characteristic of any hierarchal taxonomic system. It is quite possible that the subtitle, “A Basic System of Soil Classification for Making and Interpreting Soil Surveys, has unwittingly inspired expectations of direct interpretive meaning of the taxa for workers having weak technical backgrounds. ”

2 . Predictions f o r Map Units These entail predictions that encompass soil potentials of not only the taxa identified in map unit names but also the inclusions and, in addition, the effects of interactions between the two on use of land areas. Almost all interpretations of detailed soil surveys consider only the taxa identified in names. Soil Taxonomy has had essentially no effect on this practice. The system has increased the precision of predictions for the named taxa, but if as indicated in Section VI,B,2 it has had little effect on mapping, the net error of the predictions for delineated areas should have increased. From personal observation, the author believes that the “purity” of mapping units in terms of soil potentials for use has increased substantially in detailed soil

SOIL TAXONOMY OF THE UNITED STATES

22 1

surveys of the United States over the past 15 years, but not as an effect of Soil Taxonomy. During that period the pressures for both greater variety and greater accuracy of predictions about the potentials and limitations of soil mapping units have increased dramatically. These pressures have been translated into legends tailored for predictive value and into greater diligence on the part of mappers. Mappers strive to exclude from their delineations small areas of soil having markedly lower potential than the taxa identified in the names. They are less concerned with small inclusions of soil having higher potential. The fact that mappers may still be located where they can be held accountable when some users apply the predictions is a potent compulsion for diligence. These elements are largely lacking in other countries that use Soil Taxonomy. Interpretation of small-scale maps is a neglected subject in soil science. Most of those that have been made are primitive. Orvedal (1977) has emphasized the loss of geographic information and its precision as map scale decreases. Interpretations that can be made must be correspondingly general in purpose and in geographic detail. Small-scale maps based on Soil Taxonomy or any other taxonomic system can be interpreted only within these limitations. It is common misconception that units of small-scale maps can be identified with precision in terms of taxa of higher categories. At small scales and correspondingly large minimum size of mappable areas, soil heterogeneity within delineations necessarily is large. No taxonomic device can alter that fact. When map units that represent large land areas are identified by taxa of high categories, part of that heterogeneity is represented by the broad ranges of the taxa; the remainder can be identified in terms of the contrasting taxa present. If the taxa have strong geographic bias, the latter can be minimized by appropriate location of boundaries. There is strong geographic bias in the Alfisol, Aridisol, Mollisol, Oxisol, Spodosol, and Ultisol orders of Soil Taxonomy. Sehgal and Sys (1970) have correctly noted, however, that the Entisol, Inceptisol, Histosol, and Vertisol orders have much more limited geographic connotations. Areas of Entisols and Inceptisols ranging from a fraction of a hectare to a few thousand hectares, for example, are commonly intimately intermingled with areas of one or more of the six orders that have stronger general geographic connotations. Consequently, soil map units at scales so small that minimum delineations represent hundreds of square kilometers are typically geographic mixtures of contrasting soil orders. At the suborder level, soils having aquic moisture regimes are separated from their nonaquic counterparts of the same order. These contrasting taxa at the suborder level are typically even more intricately intermingled in the landscape and cannot be delineated separately on small-scale maps. Thus, prediction of the soil potentials of the large areas mappable at small scales usually involves appraisal of geographic mixtures of contrasting soils, even though they may be identified at the highest levels of Soil Taxonomy.

222

MARLIN G. CLINE

Small-scale maps, like those of Aubert and Tavernier (1972) and Moorman and van Breemen (1978), identify map units in terms of one or a few dominant taxa at the highest levels of Soil Taxonomy, but it is understood that the areas are likely to contain major proportions of soils that may be highly contrasting. The maps at larger scales produced by the low-intensity surveys common in developing nations also delineate intimate mixtures of soils that not only are members of different taxa of Soil Taxonomy but also have markedly different potentials for use. Many maps of this kind identify map units in terms of a single dominant taxon and thereby lose much of their potential interpretive value. The kinds, amounts, and patterns of associated contrasting soils are commonly not recorded. Buol (PC) reports that many workers in Latin America are disappointed and critical when they discover that such map units cannot be interpreted with confidence for the site-specific requirements of small farms. Even if the composition of map units is known, relatively few soil scientists have the understanding and experience with geographic interpretations to take advantage of the potential such soil maps present. Most predictions for them stop with generalizations based on the dominant taxa. While the data do not permit site-specific predictions, they do provide the information necessary for quantitative estimates of the relative proportions of map units having different soil potentials. They also provide the information necessary for estimates of probability that soil of a given potential will be found at a given site. Cline and Marshall (1977) have published an interpretation of this kind.

VII. SUMMARY Response to international distribution of Soil Taxonomy has ranged from acceptance as an official system to absolute rejection. The scheme is used as the primary classification for national soils programs in 12 countries. It is used commonly by soil scientists of 19 other countries for international communication, though the national programs of these countries depend on other systems. It is used infrequently in 20 others for which information is available. In a few of the latter, it is rejected almost completely. The system addresses the world-wide range of soils in detail. Its scope is praised by advocates; the complexity which this attribute necessarily entails is deplored by critics. Among its innovations, the use of quantitative criteria and concepts of diagnostic horizons are acclaimed by its supporters; the complicated definitions and keys that result are cited as defects by many. The innovative nomenclature is welcomed by some but condemned by others. Problems in applying the new approaches and criteria developed for Soil Taxonomy were inevitable. Some have been problems inherent in its potential users, such as personal bias and lack of training and experience. Deficiencies in published explanations of reasons for the choice and application of criteria have

SOIL TAXONOMY OF THE UNITED STATES

223

created an appearance of empirical or capricious decisions and uncertainties among some of its potential users. Some attributes of the system itself are obstacles to users. The system is highly demanding, and the manner in which it is presented in published form magnifies the problem. Some field criteria, such as those for argillic horizons, plinthite, and fragipans, are difficult to apply with precision. Lack of laboratory support for determination of some criteria and lack of data from which soil moisture and temperature regimes can be estimated have been cited by many as obstacles. The system is still in a process of evolution as data about soils of the world accumulate and expose taxonomic problems. Need for additional taxa, which can be accommodated, has been reported by many users. The system is incomplete for soils of the tropics. International committees have been appointed to consider the definition of Oxisols, the problems with soils having low-activity clays and argillic horizons in the tropics, and the questions about soil moisture regimes for these regions. A committee is studying a more nearly adequate subdivision of the Andept suborder, including the possibility of elevating the group to order status. Workers in regions of cold climates find classification of cryoturbated soils associated with permafrost inadequate. The criteria for subdivision of Vertisols, and for distinguishing the Vertisol order, do not everywhere make the distinctions anticipated. Taxonomic problems with some diagnostic horizons and several criteria have been found. Among these, the validity of the heavy weight given the argillic horizon in the system is questioned by many. Criteria for the spodic horizon continue to be troublesome. The criteria of wetness appear to need adjustment for some soils. The impact of Soil Taxonomy on soil classification world-wide has been far greater than that of any other development in the discipline during the past 50 years. In addition to its use as a primary system in 12 countries, elements of its principles, concepts, and devices have been incorporated to varying extent in both general and national schemes used in many countries. More importantly, it has prompted a changed perspective of soils and the beginnings of transformation from a qualitative to a quantitative approach to soil classification in many countries. The impact on soil surveys and soil interpretations has been much less. Major improvements remain to be made. Among these, devices that will adapt quantitative limits of criteria more realistically to soil variation in the field are needed to reconcile the conceptual framework of taxonomy with the reality of soils in nature. REFERENCES Ahmad. M., Ryan, J . , and Paeth, R. C. 1977. Soil Sci. SOC.Am. J . 41, 1162-1 166. Ali, M. A. 1972. “Classification of Sudan Soils.’’ Soil Survey Dept., Democratic Republic of Sudan, Wad Medani. Sudan.

224

MARLIN G. CLINE

Altaie, F. H., Sys, C., and Stoops, G. 1969. Pedologie 19, 65-148. Anonymous. 1978. “Research on Agrotechnology Transfer in the Tropics Based on the Soil Family.” Prog. Rep. No. I , Benchmark Soils Project. Univ. Hawaii, Honolulu. Antoine, P. 1977. I n “Soil Resource Inventories,” pp. 189-194. Agron. Mimeo 77-23. Cornell Univ., Ithaca, New York. Aubert, G., and Tavernier, R. 1972. I n “Soils of the Humid Tropics.” Natl. Acad. Sci., Washington, D.C. Avery, B. W . 1968. Trans. Int. Congr. Soil Sci., 9th, Adelaide 4, 169-175. Avery, B. W. 1973. J. SoilSci. 24, 324-338. Avery, B. W., Clayden, B., and Ragg, J. M. 1978. Soil Sci. 123, 306-318. Bartelli, L. J . 1978. Adv. Agron. 30, 247-289. Beinroth, F. H. 1975. I n “Soil Management in Tropical America” (E. Bornemisza and A. Alvarado, eds.), pp. 92-108. North Carolina State Univ., Raleigh. Beinroth, F. H., Uehara, G.,and Ikawa, H. 1974. SoilSci. SOC. Am. Proc. 38, 128-131. Bunting, B. T. 1969. Geogr. J. 135, 646-647. Butler, B. E.. Downes, R. G.,Hubbel, G.D., Nicols, K. D., andTeakle, L. J. H. 1961. Ausr. Soc. Soil Sci.,Publ. 1. Camargo, M. N., and Falesi, 1. C. 1975. In “Soil Management in Tropical America” (E. Bornemisza and A. Alvarado, eds.), pp. 25-45. North Carolina State Univ., Raleigh. Chan, H. Y . 1978. In ‘‘Soil Resource Data for Agricultural Development” (L. D. Swindale, ed.), pp. 41-66. Hawaii Agr. Exp. Stn., Honolulu. Clayton, J. S., Ehrlich, W. A., Cann. D. B., Day, J. H.,and Marshall, I. B. 1977. “Soils of Canada,” Vol. I . Canada Dept. Agr., Ottawa. Cline, M. G. 1977. Soil Sci. Soc. Am. J. 41, 250-254. Cline, M. G.1979. “Soil Classification in the United States.” Agron. Mimeo 79-12. Cornell Univ., Ithaca, New York. Cline, M. G.,and Marshall, R. L. 1977. Cornell Univ. Coop. Exr. Inf. Bull. 119. Comerma, J. A. 1971. Agron. Trop. (Maracay, Venezuela) 21, 365-377. Cook, T. D. 1975. “Final Report on the American System of Soil Classification as Applied to the Soils of Sudan.” FA0 Proj. SUD/71/553. Rome. Coover, J. R., Bartelli, L. J., and Lynn, W. C. 1975. Soil Sci. SOC.Am. Proc. 39, 703-706. Cortes, A., and Franzmeier, D. P. 1972. Soil Sci. SOC. Am. Proc. 36, 653-659. Coultas, C. L., and Gross, E. R. 1975. Soil Sci. SOC. Am. Proc. 39, 914-919. Daniels, R. B., Perkins, H. F., Hajek, B. F., and Gamble, E. E. 1978. Soil Sci. Sac. Am. J . 42, 944-949. De Alwis, K. A., and Pluth, D. J. 1976. Soil Sci. SOC. Am. J . 40, 912-920. De Bakker, H. 1971. Geoderma 5, 169-177. Deckers, J . 1966. Pedolog. Mem. 6. De Coninck, F., Vasu, Al., and Rapaport, C. 1976. Pedologie 26, 255-279. Dent, F. J . , and Changprai, C. 1973. “Soil Survey Handbook for Thailand,” pp. 70-94. Ministry of Agr. and Coop., Bangkok. Deshpande, S. B., Fehrenbacher, J. B., and Ray, B. W. 1971. Geoderma 6, 195-201. Dewan, M. L., and Famouri, J. 1964. “The Soils of Iran.” FAO, Rome. D’Hoore, J. L. 1968. In “The Soil Resources of Tropical Africa” (R. P. Moss, ed.), pp. 7-28. Cambridge Univ. Press, London and New York. Dijkerman, J. C. 1969. Afr. Soils 14, 185-206. Duchaufour, P. 1963. J . Soil Sci. 14, 149-155. FAO-UNESCO. 1974. “Soil Map of the World,” Vol. I , Legend. UNESCO, Paris. Fauck, R., Lamouroux, M., Perraud, A., Quantin, P., Roederer, P., Viellefon. J., and Segelin, P. 1979. “Projet de Classification des Sols.” ORSTOM (Off. Rech. Sci. Tech. Outre-Mer), Paris.

SOIL TAXONOMY OF THE UNITED STATES

225

Flach, K. W. 1963 Soil Sci. SOC.Am. Proc. 27, 226-228. Flores, P. S . , Alvarado, A., and Bornemisza, E. 1978. Turrialba 28, 99-103. Fridland, V. M. 1964. Sov. SoilSci. pp. 576-577. Gerasimov, 1. P. 1962. Sov. Soil Sci. pp. 601-609. Gerasimov, I. P. 1964. Sov. Soil Sci. p. 603. Gerasimov, 1. P. 1969. Sov. Soil Sci. pp. 5 I 1-524. Gerasimov. 1. P. 1978. Pochvovedenie ( I ) , 152-157. Gile, L. H., and Grossman, R. B. 1968. Soil Sci. 106, 6-15. Gowaiker, A. S. 1972. J. Indian Soc. Soil Sci. 20, 59-66. Hakimian, M. 1977. SoilSci. Soc. Am. J . 41, 1155-1161. Harpstead, M. I. 1973. Soil Sci. 116, 437-443. Harris, S. A., Neumann, A. M., and Stouse, P. A. D., Jr. 1971. Soil Sci. 112, 439-447. Hendershot, W. H., and Lavkulich, L. M. 1978. Soil Sci. SOC. Am. J . 42, 468-472. Ikawa, H . 1978. In “Resource Data for Agricultural Development” (L. D. Swindale, ed.), pp. 20-27. Hawaii Agr. Exp. Stn., Honolulu. Isbell, R. F., and Field, J. B. F. 1977. Geodermu 18, 155-175. Kesseba, A , , Pitblado, J. R., and Uriyo, A. P. 1972. J. Soil Sci. 23, 235-247. Leahey, A. 1963. Soil Sci. Soc. Am. Proc. 27, 224-225. Lepsch, I . F., and Buol, S . W. 1974. Soil Sci. Soc. Am. Proc. 38, 491-501. Lepsch, 1. F., Buol, S. W . , and Daniels, R. B. 1977. Soil Sci. SOC. Am. J . 41, 109-115. Lewis, D. T. 1977. Soil Sci. Soc. Am. J . 41, 940-945. Luzio, L., W. 1978. Cienc. Invest. Agr. (Chile) 5, 181-186. Luzio L. W., and Menis M. M. 1975 Turrialba 25, 271-282. McCormack, D. E . , and Wilding, L. P. 1969. SoilSci. Soc. A m . Proc. 33, 587-593. MacVicar, C. N., DeVilliers, J. M., Loxton, R. F., Verster, E., Lambrechts, J. J . N., Merryweather, F. R., LeRoux, J., Van Rooyen, T. H., and Von M. Harmse, H. J. 1977. S. Afr. Dep. Tech. Serv. Bull. 390. Miller, R. B. 1978. N . Z. Soil News 20, 176-179. Mitchell, C. W. 1973. J. Soil Sci. 24, 41 1-420. Moore, A. W. 1978. In “Soil Resource Data for Agricultural Development” (L. D. Swindale, ed.), pp. 193-203. Hawaii Agr. Exp. Stn., Honolulu. Moorman, F. R., and van Breemen, N. 1978. “Rice: Soil, Water, Land,”pp. 51-82. Int. Rice Res. Inst., Los Banos, Philippines. Muir, J. W. 1962. J . Soil Sci. 13, 22-30. Mulcahy, M. J . , and Humphries, A. W. 1967. Soils Fert. 30, 1-8. Murthy, R. S . , ed. 1979. “Detailed Soil Survey and Land Use Planning, UnionTerritory of Delhi.” Natl. Bur. Soil Survey and Land Use Planning, Nagpur, India. Murthy, R. S.. Shankaranarayana, H.S., and Hirekerur, L. R. 1977. J . Indian Soc. Soil Sci. 25, 284-287. Nettleton, W. D., Flach, K . W., and Brasher, B. R. 1969. Soil Sci. Soc. Am. Proc. 33, 121-125. Odell, R. T . , Dijkerman, J. C., vanvuure, W., and Melsted, S. W. 1974. Llniv. Illinois Coll. Agr. Bull. 784. Oertel, A. C. 1968. Truns. Inr. Congr. Soil Sci.. 9rh, Adelaide 4, 481-486. Ojanuga, A. G., Lee, G. B., and Folster. H. 1976. Soil Sci. SOC. Am. J . 40, 287-292. Orvedal, A. C. 1977. In “Soil Resource Inventories.” Agron. Mimeo 77-23, pp. 19-24. Cornell Univ., Ithaca, New York. Orvedal, A. C., and Austin, M. E. 1963. Soil Sci. Soc. Am. Proc. 27, 228-231. Pettapiece, W. W. 1975. Arctic 28, 35-53. Pilgrim, S. A. L., and Harter, R. D. 1977. New Hampshire Agr. Exp. Stn. Bull. 507. Ragg, J. M., and Clayden. B. 1973. Rothamsred Exp. Srn. Monogr. 3.

226

MARLIN G. CLINE

Rengasamy, P., Sarma, V. A. K.,Murthy, R. S., and Krishna Mwti, G.S. R. 1978. J . Soil Sci. 29, 43 1 - 4 5 . Sanchez, P. A,, and Buol, S. W. 1974. Soil Sci. Sor. Am. Proc. 38, 1 17-12. Sehgal, J. L., and Sys, C. 1970. Pedologie 20, 244-267. Sehgal, J . L., Hall, G.F., and Bhargava, G.P. 1975. Geoderma 14, 75-91. Smith, G. D. 1963. Soil Sci. 96, 6-16. Smith, G. D. 1965. Pedologie Special No. 4. Smith, G. D., Newhall, F., Robinson, L. H., and Swanson, D. 1964. “Soil Temperature Regimes-Their Characteristics and Predictability.” U.S. Dept. Agr., Soil Conserv. Serv. TP-144., Washington, D.C. Smith, G. D., Sys, C., and Van Wambeke, A. 1975. Pedologie Special No. 5. Snedden, J. I., Lavkulich, L. M., and Farstad, L. 1972. Soil Sci. SOC. Am. Proc. 36, 100-104. Soil Science Department. 1978. “Research Program on Soils of the Tropics, 1976-77 Annual Report.” North Carolina State Univ., Raleigh. Soil Survey Staff. 1960. “Soil Classification. A Comprehensive System, 7th Approximation.” U.S. Dept. Agr., Soil Conserv. Serv., Washington, D.C. Soil Survey Staff. 1975. “Soil Taxonomy. A Basic System for Making and Interpreting Soil Surveys. ” Agric. Handb. No. 436. U.S. Govt. Printing Office, Washington, D.C. Stephens, C. G. 1963. Soil Sci. 96, 40-48. Tan, K. H., Perkins, H. F., and McCreery, R. A. 1970. Soil Sci. SOC.Am. Proc. 34, 775-779. Uehara, G.1978. In “Soil Resource Data for Agricultural Development” (L. D. Swindle, ed.), pp. 204-209. Hawaii Agr. Exp. Stn., Honolulu. Van Wambeke, A. 1967. Soil Sci. 104, 309-313. Walmsley. M. E., and Lavkulich, L. M. 1975. Soil Sci. Soc. Am. Proc. 39, 84-88. Webster, R. 1968a. J. Soil Sci. 19, 354-366. Webster, R. 1968b. Geogr. J . 134, 394-396. Westin, F. C. 1963. SoilSci. SOC.Am. Proc. 27, 222-223. Westin, F. C. 1974. Soil Sci. SOC.Am. Proc. 38, 804-807. Westin, F. C., Avilon, J., Bustamante, A., and Marino, M. 1968. Soil Sci. 105, 92-102. Yerokhina, A. A., and Sokolova, T. A. 1964. Sov. Soil Sci. pp. 579-581. Zamora, C. 1975. In “Soil Management in Tropical America” (E. Bornemisza and A. Alvarado, eds.), pp. 46-60.North Carolina State Univ., Raleigh.

ADVANCES IN AGRONOMY, VOL. 33

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION R. J. Haynes Department of Soil Science, Lincoln College, Canterbury, New Zealand

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 11. Competition in the Pasture Community . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

A. Nature of Competition.. . . . . . . . . . . . . . . . . . . . , . , . , . . . . . . . . . . . . . . . . , . . . , . , B. Evolution of the Grasses and Legumes . . . . . . . . . . . . . . , . . . . . . C. Allelopathy.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . 111. Physiological Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . A. Rhizobium Symbiosis and Nitrogen Transfer.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Vesicular-Arbuscular Mycorrhizae . . . C. Period and Rate of Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Light Requirement

229 229 230 23 1 232 233 235 237 238 F. Root Catio 238 IV. Morphological .............................................. 239 239 ................. ........................ 24 1 .......... . 247 C. Root Morphology ..................................................... 248 V. Competition for Environmental Factors A. Light.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 250 B. Nutrients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 C. Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 References . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

I. INTRODUCTION The concept of culturing legumes and nonlegumes as a pasture has been an important feature of agriculture from the early days, and historical reviews have been compiled by Nicol (1935) and Wilson (1940). The earliest records of the use of clover pastures appears to be in the early sixteenth century (Fussell, 1964). A grass-legume association has been used in many countries of the world because a greater total herbage yield may be obtained by growing a grass and a legume in association, rather than in individual swards, where no fertilizer N is applied. 227

Copyright @ 1980 by Academic Ress. Inc. All rights of repmduaion in any form reserved.

ISBN 0-12-ooO733-9

228

R. J. HAYNES

Studying the grass-legume association, de Wit et a l . (1966) found no net benefit from grass-legume associations where N fixation was precluded; but where N fixation was permitted, results indicated a net benefit from the inclusion of the legume, irrespective of its inherent lower yielding ability. The use of legumes in pastures may also result in increased N content and digestibility and a high well-balanced mineral content of herbage, all of which are of importance in animal nutrition (Chestnutt and Lowe, 1970). White clover (Trifolium repens L.) is the most important pasture legume in many parts of the temperate zone (Leffel and Gibson, 1974). The ryegrass-white clover association dominates New Zealand pasture (Brougham et al., 1978) and it is also prevalent in northwestern Europe. In such pastures in New Zealand, nitrogen is supplied almost entirely from the clovers (Scott, 1972). Alfalfa or lucerne (Medicago saliva L.) is the other very important legume in the temperate region, although it is well adapted to a wide range of climatic and soil conditions and has worldwide distribution (Hanson and Barnes, 1974). Alfalfa is the mainstay of agriculture in the United States (Bolten et a l . , 1972), where it is used often in association with grasses as a grazing pasture, or it is cut, dehydrated, and fed to stock as alfalfa meal. In Britain (Spedding and Diekmahns, 1972) and New Zealand (Langer, 1973c), lucerne, often in association with grasses, is frequently used to make hay and silage. Other legumes such as red clover (Trifolium pratense), other species of Trifolium, sainfoin (Onobrychis viciaefolia), the birdsfoot trefoils (Lotus spp.), and the vetches (Vicia spp.) are components of temperate pastures in many parts of the world. In tropical and subtropical pastures, Siratro (Macroptilium atrupurpureum), Townsville stylo (Stylosanfhes humilis) , and other legumes are becoming increasingly important components in herbage production. It is therefore evident that interferences and competition between legumes and associated grasses are of great importance in world pasture production. Such interspecific interferences are likely to be a dominating factor in the nonequilibrium, man-made and -managed pasture vegetation. It is generally accepted that grasses normally have a competitive advantage over legumes and therefore tend to dominate pastures, but that in order to maintain high pasture productivity, a balance between grasses and legumes is desirable. A large number of factors (both plant and environmental) can influence the delicate balance in a nonequilibrium pasture mixture, and through an appreciation of the way in which these factors operate both separately and in combination, a desirable balance between grass and legume can be maintained. This review therefore examines the competitive relationships between grasses and legumes and also considers the way in which pasture management techniques alter the balance of legumes and grasses by manipulating She environment, which results in a competitive advantage to the legume component. Examples in this review are taken principally from temperate pastures, although it is recognized that the use of legumes is becoming increasingly important in tropical grasslands.

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

229

II. COMPETITION IN THE PASTURE COMMUNITY A. NATUREOF COMPETITION

In the exact sense, two plants do not compete with each other as long as the water supply, nutrient supply, oxygen and C 0 2 . and light and heat are in excess of the needs of both. Harper (1964) pointed out that competition itself is only one facet of interference between plants, although at times it may be a very dominating one. Noncompetitive interferences may be the direct stimulation of one species by another (e.g., the nitrogen fixed by a legume becoming available to a nonlegume). Competition among plants most commonly occurs for light, water, and nutrients (Donald, 1963; Rhodes, 1970), although competition for oxygen, carbon dioxide, and space may also occur. It is generally accepted that the more similar the needs of the two organisms, the more intense the contest. Thus intraspecific competition is more intense than interspecific competition. Competition and other interferences between plants play an integral role in the ecological principle of succession in which assemblages of species succeed one another until a steady state is reached. In the process of succession, plant species in an area modify each other’s environment in such a way that they progressively replace one another. The species involved in the process of succession are loosely referred to as seral species. The ecology of a seral species is therefore critically defined by its reaction to the presence of others-those it ousts and those which in turn oust it (Harper, 1964). When the steady state is reached, the vegetation is in equilibrium with its environment and is known as climax vegetation. A mixture of agricultural plants such as a pasture does not often behave as a stable diverse climax ecosystem, since it is artificial and man-made. Tothill (1978) states that “most pasture vegetation, as opposed to natural grassland vegetation, may be considered to be in a seral nonequilibrium condition.” A pasture may be in equilibrium with the environment under a certain management practice. If pasture management procedures are altered, then the pasture begins to revert to some other form of vegetation. Hence pasture management techniques (e.g., choice of species, date of seeding, density of seeding, fertilization, irrigation, and time and height of defoliation) are utilized in order to secure a desirable balance of grasses and legumes within the pasture. This is the basis of pasture science. B. EVOLUTION OF

THE

GRASSES A N D LEGUMES

In order to understand the nature and growth of grasses and legumes some knowledge of their origin and evolution is desirable. Both have their origins in the tropical forests; the grass.es in woody bamboo forms and the legumes as trees

230

R. J . HAYNES

and shrubs. The evolution of the grasses and legumes of grasslands has been adequately reviewed elsewhere (Harlan, 1956; Tothill, 1978) and will not be discussed in detail here. The grasses belong to the family Gramineae which comprises about 620 genera and 10,OOO species. The family appears to be uniformly spread throughout the world, being found from subpolar regions to the tropics, and from seashores to mountain tops. Their adaptability to such diverse environments is largely due to the fact that most species have their growing points close to the soil where they are fairly safe from damage by climatic conditions, fire, or grazing. Few, if any, of the important cultivated pasture grasses are constituents of the major grasslands of the world. The typical agricultural grasses, species of Poa, Festucce, Lolium, and Dactylis, are allied to and probably derived from grasses of woodland and forest margins (Hartley and Williams, 1956) where rainfall and soil fertility is generally relatively favorable (Tothill, 1978). The grasses of the natural climax grassland communities have generally developed in regions of lower rainfall and lower soil fertility (Hartley and Williams, 1956). The primitive members of the family Leguminosae were believed to be tropical and subtropical trees and shrubs. Tropical rainforest soils are generally of poor fertility and thus Rhizobiurn symbiosis would be advantageous in such habitats (Tothill, 1978). A small part of one of the three subfamilies, the Pipilionateae, has evolved a herbaceous habit. Three tribes of this family, Trifolieae, Loteae, and Vinceae, are found entirely as herbaceous species of the Eurasian region. Some legumes became so intimately adapted to the grasslands of the world and the grazing herbivores that evolved on them that their chief mode of seed dispersal is through grazing animals. For example, the Trifolieae are distributed in nature almost exclusively by animals that eat the seed. In nature the legumes seldom, if ever, dominate the landscape as the grasses do in the grassland regions of the world. A great number of the Trifolium species originated in the natural grasslands of those countries that border the Mediterranean and Red Seas (Fergus and Hollowell, 1960). C . ALLELOPATHY

Before discussing competition between grasses and legumes in detail, the possibility of noncompetitive allelopathy must be considered. The influence of one plant upon another by chemical means has been studied extensively in recent times and there is now strong evidence that this type of interaction does occur (Whittaker and Freeny, 1971; Rice, 1974). Allelopathy refers to the liberation of chemical retardants from the roots and/or shoots of a plant into the environment from where they have a deleterious effect on another plant. Critical proof of the ecological importance of allelopathy is still ill-defined (Etherington, 1976). Impressive in vitro effects often disappear when experiments are carried out in soil.

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

23 1

It is relatively easy to extract metabolic products of one plant than can inhibit the growth of another. However, as Etherington (1976) has pointed out, it is extremely difficult to prove that the inhibitory growth effect of one plant upon another is related to secretion of an inhibitory metabolite into the soil and not to competitive interactions. Thus in plant communities it is likely that a combination of competitive and allelochemical effects may occur (Trenbath, 1974). Since crop plants are normally grown at densities high enough for intense competition for resources to occur, any allelochemical effect would be either exaggerated or reversed by competition (Trenbath, 1974). Newman and Rovira (1975) who compared individual receiver species to different plant leachates showed that some species [e.g., perennial ryegrass (Loliurn perenne L.) and white clover] grew more slowly when receiving leachate from their own species than they did from other species, whereas other species [e.g., Yorkshire fog (Holcus lanatus L.)] showed the opposite response, growing faster with leachate from their own species than from the others. Newman and Rovira (1975) claimed these results may have some bearing on the composition of pastures and that they could explain why some species are strongly dominant in grasslands whereas others are found interspersed with other species. In a subsequent study in sand culture, Newman and Miller (1977) showed that application of different plant leachates to a range of plant species resulted in significant differences in the uptake of phosphorus, sometimes being greater and sometimes less than the control. Perennial ryegrass exudate generally appeared to decrease P uptake by white clover, while the exudate of white clover had little effect on P uptake by perennial ryegrass. It has been demonstrated that certain crop plants contain water-soluble materials that inhibit seed germination and seedling growth of several crop plants (Rice, 1974). Aqueous extracts of both lucerne tops and roots appear to be particularly detrimental to the germination and subsequent root growth of other pasture grasses and legumes (Grant and Sellans, 1964). Exudates of some native tussock grasses of New Zealand have allelochemical interactions with the germination of oversown seed (Janson and White, 1971; Scott, 1975). Exudates generally depress germination of grass seed but either promote or depress germination of legume seed. Thus it is known that allelochemical effects do occur between pasture legumes and grasses under laboratory conditions. The relevance of such interactions under field conditions is unknown, and more research is needed in this area of pasture-plant interactions.

111. PHYSIOLOGICAL CONSIDERATIONS Basic physiological and morphological differences between grasses and legumes materially affect the nature of competition that develops between

232

R. 1. HAYNES

species in various mixtures. It is difficult and artificial to separate and categorize these differences as they are interrelated and they interact in a complex manner and in different ways in different situations. It is recognized that the categories used in this review are artificial and are made in order to gain a better understanding of the grass-legume association. It must also be remembered that both interand intraspecific competition are occurring within the grass and legume components of a pasture community. A. Rhizobiurn SYMBIOSIS AND NITROGEN TRANSFER

Pasture plants obtain most of their nitrogen (N) either from mineral N in the soil arising from mineralization of organic N or from N fixed from the air by symbiotic rhizobia of nodulated legumes. Nitrogen relationships in grass-legume mixtures have been reviewed recently (Vallis, 1978) and the physiological processes involved have also been reviewed (Stiefel, 1977). The effect of N fixation on the grass-legume balance is, however, of relevance to this review. The competition for N between grass and legume represents a unique situation because the N uptake of the grass can be influenced by the legume by two opposing processes. The legume may increase the supply of available N in the root medium by N fixation, but it may also compete for mineral soil N (Simpson, 1965). The balance between competition and transfer is not constant with time but changes with the growth cycles of the species in the sward (see Vallis, 1978). The percentage of N fixed by the legume component of a mixed sward that is transferred to the grass component may vary from nil to 75% (Whitehead, 1970). Temperate pasture legumes appear to be less effective than grasses in taking up mineral N (Walker et al., 1956; Akatsuka and Sugihara, 1973). It is usually assumed that legumes in mixed pastures obtain only a small proportion of the available mineral N, although in some cases the legume has been shown to compete with the grass for’available soil N (Davies, 1964; Simpson, 1965). Growth of the legume and N fixation is often increased by small doses of N especially if added N is available during the time between exhaustion of the seed reserves and establishment of an effective N-fixing system (Pate and Dart, 1961; Diatloff, 1974), however as the level of N applied is increased the number of effective nodules is reduced (Allos and Bartholomew, 1959; McAuliffe et al., 1958). The reduction in fixation caused by high levels of mineral N is less in fast-growing species than in slow-growing ones (Allos and Bartholomew. 1959; Dart and Wildon, 1970). Direct measures of N relations in grass-legume mixtures can be made by application of small amounts of 15N-labeled mineral N (e.g., Walker et al., 1956; Vallis et al., 1967; Henzell et al., 1968; Akatsuka and Sugihara, 1973). Walker et al. (1956) using 15N in a pot trial found that when Italian ryegrass (Lolium multiforium Lam.) and white clover were grown together, the grass took up practically all the N (over 90%); under these conditions

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

233

the clover became almost fully dependent on symbiotic N (100% at low N, 85% at high N). Under most conditions N transfer occurs through cycling by grazing animals and decay of legume herbage and roots and/or nodules (Sears, 1953; Bakhuis and Kleter, 1965; Whitehead, 1970). Under some conditions legumes may excrete soluble amino compounds, but this would rarely occur in the field (Butler and Bathurst, 1956). Many agronomic experiments have shown that the amount of N transferred from legumes to associated grasses varies with species of legume, species of grass, percentage legume in the sward, age of sward, and type of management (Chestnutt and Lowe, 1970; Whitehead, 1970). The white clovergrass sward has been extensively studied in the United Kingdom (Chestnutt and Lowe, 1970; Chestnutt, 1972). Generally those grass species which are high yielding in pure swards (e.g., cocksfoot, Dactylis glomerata L.) are also high yielding when grown with clover, and are those which may be less compatible with clover, particularly when environmental conditions favor grass growth (Cowling and Lockyer, 1967). Thus these more vigorous grasses shade the clovers and hence suppress clover growth. This leads to decay of above- and below-ground organs of clovers and the release of N which may considerably increase grass yields. Cowling and Lockyer ( 1965, 1967) have shown evidence of an inverse relationship between yield of clover and yield of associated grasses. Perennial ryegrass is highly compatible with clover (Chestnutt and Lowe, 1970). Hence Reid and Castle (1965), in a 3-year cutting trial, estimated that although the total amount of N fixed was less in a cocksfoot-clover sward than in a ryegrass-clover sward ( 121 and 184 kg N/ha, respectively), a greater proportion of the fixed N was transferred to cocksfoot (5 1 %) than to ryegrass (36%). The species of legume is often an important factor in determining the extent and nature of N transfer. Simpson (1965) found that alfalfa gradually released N over 12-18 months, white clover was competitive for N until the autumn-winter period. Dilz and Mulder (1962) demonstrated that the amounts of fixed N transferred to ryegrass from alfalfa, white clover, and red clover were 16, 36, and 6%, respectively, of N in the legume tops. Among temperate perennial legumes, white clover appears to transfer N more efficiently than others. This may be because of a faster turnover of root and nodule tissue (Vallis, 1978). The senescence of an annual legume [e.g., subterranean clover (Trifoliumsubterraneum)] grown with a perennial grass may enhance the efficiency of N transfer (Vallis, 1978). B . VESICULAR-ARBUSCULAR MYCORRHIZAE

The fungal group of vesicular-arbuscular mycorrhizae (VAM) are very widespread in nature and are associated with a great variety of plants of different taxonomic groups (Nicolson, 1975). Although the relationship between host

234

R. 1. HAYNES

plant and endophyte is obligate as far as the endophyte is concerned, endophytes display a remarkable lack of host-endophyte specificity (Nicolson, 1975). Within the root cells, the fungal hyphae may take two forms, many-branched structures called arbuscules and oil-filled terminal swellings called vesicles. External to the root, the loose network of hyphae may extend a considerable distance into the soil. The ability of these fungal hyphae to explore more soil than nonmycorrhizal roots enables infected plants to use effectively soil nutrients more efficiently by exploring a greater volume of soil than nonmycorrhizalplants (Nye and Tinker, 1977). This is particularly important when considering the nutrition of plants for immobile nutrients such as phosphorus. It is known that white clover is much more reliant on the formation of VAM than perennial ryegrass is (Crush, 1974; Hall, 1978). It has been suggested that mycorrhizal formation is only beneficial to a host plant when it has few or no root hairs (Baylis, 1970, 1975). Therefore, because grasses generally possess a fine, fibrous root system with many root hairs, the benefits of the mycorrhizal association may be limited (Sparling and Tinker, 1975). However, in soils low in available phosphorus, various workers have shown significant growth responses in VAM-infected grasses (Crush, 1973, 1975; Fitter, 1977). Powell (1977~) showed that with the use of more efficient strains of VAM such as Gigaspora margarita, significant growth responses by ryegrass could be achieved in sterilized soils of moderate phosphorus fertility. Legumes, particularly white clover with its less extensive root system and characteristically less efficient nutrient absorption than grasses, consistently show large responses to VAM infection (Crush, 1976; Powell, 1976; Abbot and Robson, 1977). In many infertile soils, white clover is highly dependent on infection with indigenous mycorrhizae (Powell, 1976, 1977a) and growth responses may be greatly increased by infection with more efficient species of VAM (e.g., Glomus mosseae) (Mosse, 1975, 1977; Powell, 1977b). The differential ability of white clover cultivars to become infected by VAM (Hall et al., 1977) may be an important consideration in their use in the future. Observations (Daft and El Giahmi, 1975, 1976; Mosse et al., 1976; Smith and Daft, 1977) have shown that mycorrhizal infection increases nodulation and nitrogenase activity of legumes as well as improving phosphate nutrition. The effect appears very early in the life of the plant so that, within 1-2 weeks of inoculation with Rhizobium meliloti and Glomus mosseae, Smith and Daft (1977) showed an enhancement of nodulation and nitrogenase activity in dually infected alfalfa. Since legumes require high levels of phosphate for growth and effective nodulation and it is known that N2 fixation has a high phosphate requirement (Munns, 1977), improved phosphorus nutrition, by way of the mycorrhizal association, may therefore have the secondary effect of stimulating nodulation and N 2 fixation by Rhizobium spp. bacteria (Mosse, 1977; Hall, 1978). Generally, mycorrhizal infection of the grass-clover association therefore confers a competitive advantage on the clover. Crush (1974) showed that when

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

235

ryegrass and clover were grown together, mycorrhizal clover competed more effectively for P than nonmycorrhizal clover. Hall (1978) showed that at low soil P levels VAM produced a 5-fold increase in total clover dry weight when grown alone, and a 40-fold increase when grown with ryegrass. The clover appeared to be infected in preference to the grass, and apparent VAM-induced depressions in yield were observed in the grass. Thus it may be concluded that VAM inoculation gives infected clover an advantage over noninfected clover when in competition with grasses by increasing the volume of soil explored by the clover root system. However, most research, through necessity, has been carried out in sterilized soil in order to destroy indigenous mycorrhizal fungi. In the field both clovers and grasses are generally already infected with indigenous VAM, hence inoculation with improved VAM strains may not have a very large effect on the grass-clover balance. The effect of VAM inoculation on the competitive ability of legumes other than white clover does not appear to have been studied yet. However, Crush (1976) found that although VAM- (Acaulospora laevis) infected red and white clover grew better than nonmycorrhizal plants in a range of soils, the same fungus was parasitic on alsike clover (Trifoliurn hybridurn L.) and alfalfa, reducing growth by 3-16%. It has also been suggested that the host-endophyte relationship in the VAM symbiosis may change from mutualism to parasitism as soil P availability increases (Crush, 1975). Thus, substantially more research is needed in this apparently important area of grass-legume relations. C. PERIODA N D RATEOF GROWTH

The stability of a grass-legume pasture may, in part, depend upon differences in seasonal growth rates. Brougham (1959) showed that at Palmerston North, New Zealand, the annual growth curves of white clover and short rotation ryegrass in a mixture were complementary (Fig. l), white clover contributing most to total herbage production in late summer and autumn. This occurred as a consequence of different temperature optima for growth of ryegrass (18-21°C) and white clover (24"C), postflowering depression of ryegrass growth, and susceptibility of ryegrass to fungal disease in summer (Harris and Thomas, 1973). Thus during midsummer and early autumn, when light and temperature were nearer optimal for white clover growth and when ryegrass was showing postreproductive or midsummer depression of growth, white clover was able to spread through the swards by stoloniferous growth. The greater growth rates of ryegrass under the low temperatures and light regimes of late autumn, winter, and spring, which are more suboptimal for white clover than for ryegrasses, results in ryegrass dominance during this period. This ryegrass dominance is also partially explained by the balance between N fixation by the clover and the response of ryegrass to this N with consequent suppression of clover by ryegrass. Hence,

236

R. J. HAYNES

C

\

a

s

--

1.1

Ryegrass

White clover

I-

z Y E

AUG

OCT

DEC

FEE

APR

JUNE

AUG

OCT

FIG. 1. The relationship between the seasonal growth rates of white clover and short rotational ryegrass in Palmerston North, New Zealand. (Redrawn from Brougham, 1959.)

maximum yield contributions of ryegrass and clover occur at different times of the year resulting in an advantageous effect of equalizing dry matter production throughout the year. Approximately 30% clover content on average throughout the year results in an equilibrium being attained (Martin, 1960; Harris and Thomas, 1973). Similarly, in comparison with white clover, the grass Phaluris tuberosu has a high production potential in late autumn, winter, and early spring, which enables Phalaris to cohabit with white clover in the pastures of Armidale, Australia (Lazenby and Swain, 1972). Pineiro and Harris (1978) studied seasonal production in establishment year of three cultivars of red clover grown in association with two cultivars of prairie grass (Bromus cuthurticus Vahl) and ryegrass. All mixtures showed a marked shift of dominance from grass in spring to red clover in summer to give complementary growth of sown grass and sown legume, resulting in continuously high total production from early spring to late autumn. This balance may be upset by breeding for increased grass growth in midsummer to late autumn or for increased clover growth in winter and spring. For instance, the greater reduction of ryegrass yield in association with Pitau white clover (bred for winter cool-season production) than when associated with Hui white clover during winter indicates this effect (Brock, 1974). Growth vigor and yielding ability of components is another factor affecting the botanical composition of a pasture. Cowling and Lockyer (1967) observed that those grass species that are high yielding as pure swards (e.g., cocksfoot) are also high yielding when grown with clover and are those which are less compatible with white clover, particularly when environmental conditions favored grass growth. The vigorous cocksfoot is also known to be very competitive toward

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

237

sainfoin and alfalfa in comparison with other grasses such as timothy and meadow fescue (Spedding and Diekmahns, 1972). Donald (1978) has observed that the breeding of high-yielding grasses results in species that are highly competitive toward legumes. The rate of emergence and growth from seeds is also important. Blaser et al. (1956) have shown that lucerne, relative to many legumes and grasses, is an aggressive species in the seedling stage. Using 100 as a seedling vigor for alfalfa, red clover and ladino clover had values of 63 and 38, respectively, whereas cocksfoot, tall fescue, bromegrass, and timothy had values of 42,45, 52, and 17, respectively. It is well known that cocksfoot develops slowly from seed and that seedling growth is much slower than that of ryegrass, but once established cocksfoot is more aggressive. Laskey and Wakefield (1978) observed that reduced stands of birdsfoot trefoil, when grown with perennial ryegrass, appeared to be due to rapid seedling emergence of the grass at a critical stage in the early growth and development of the legume. The slower growing grasses, Kentucky bluegrass (Poa pratensis L.) and red fescue (Festuca rubra L.), did not inhibit growth of birdsfoot trefoil. D. LIGHTREQUIREMENT

A great deal of experimental work has shown that clovers are light-demanding plants (Langer, 1973a). Most research has centered on the canopy architecture of grasses and clovers and how this influences light use, which is discussed later. Photosynthetic rate and efficiency of temperate legumes and grasses does not appear to have been studied in any detail. Blackman and Templeman (1938) found that white clover grown with grasses was greatly depressed by reducing daylight to 60%. Grasses appear to adapt more readily to shaded conditions and are less affected by reduced light intensities than clovers (Langer, 1973a). Blackman and Black ( 1 959) estimated the light intensities necessary to obtain maximum growth rates of seedlings of various herbage species (Table I). These Table I Optimal Light Levels for Maximal Relative Growth Rate of Different Pasture Species"

Species

Light level (daylight = 1)

Lolium multiflorum Trifolium repens T . pratense T . hybridum Medicago sativa

0.71 1.85 1.oo 1.48 2.51

Data from Blackman and Black (1959).

238

R. J. HAYNES

data c o n f m that, in general, legumes are more light-demanding than the grass Italian ryegrass. The most marked response of white clover to light intensity is a reduction in the formation of stolons from axillary buds (Beinhart, 1963). Mitchell (1956) showed that the deleterious effect of shading on white and subterranean clover and on Lotus are particularly severe at high temperatures. E. EFFICIENCY OF WATER USE

Tantrum and Mitchell (1972) found that under conditions comparable to warm sunny days the water loss (grams per hour per gradleaf blade dry weight) by lucerne and white clover was significantly higher than that from the temperate grasses Phalaris, Tama ryegrass, and Raunui ryegrass. Other workers have also reported that legumes are the least efficient in water use of a range of crop and pasture plants (Leach, 1978). Burch and Johns (1978) found that white clover was unable to reduce either its relative canopy conductance or its rate of leaf transpiration as effectively as the grass tall fescue (Festuca arundinacea Schreb.). Poor control of leaf transpiration in clover resulted in low leaf water potentials and increased leaf senescence. In contrast the better stornatal control and higher leaf water potentials in fescue prolonged its period of active growth. Results indicated that clover leaf senescence was the main mechanism for balancing transpiration with water uptake as the soil dried out. The high water loss by alfalfa has also been attributed to poor stomatal control. The stomata tend to remain fully open throughout the day, thus exerting minimal control over water loss (Leach, 1978). The limited data available indicate that grasses may generally make more efficient use of soil water than legumes do. F. ROOT CATION EXCHANGE CAPACITY

It is well known that plant roots have a cation exchange capacity (CEC), because negatively charged points located in the cell wall matrix result in the free space of roots having a negative charge (Lauchli, 1976). The role of root CEC in ion accumulation by plants is still controversial and has been questioned (Nissen, 1974; Nye and Tinker, 1977). It is generally accepted (e.g., Volz and Jacobson, 1977) that the root CEC of legumes is about twice that of grasses as the results of Asher and Ozanne (1961) show (Table 11). It has been suggested that this may be an explanation of why legumes are such poor competitors for K when grown in association with grass (Gray et al., 1953; Mouat and Walker, 1959b). A plant root surface having a high CEC might absorb relatively more divalent cations (such as Ca) than a plant root, such as grasses, with a low CEC. The work of

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

239

Table I1 Root Cation Exchange Capacities of Some Monocots and Legumes Grown under L o w Nitrogen Supply“

Monocots

(meqI100 g)

Root CECb

Legumes

Root CECb (meq/100 9)

Lolium rigidum Bromus rigidus Hordeum vulgare Avena sativa

12.0 13.2 7.7 9.5

Trifolium subterraneum Medicago tribuloides Pisum arvense Vicia sativa

25.6 32.3 33.9 31.6

Data from Asher and Ozanne (1961). Electrodialysis carried out prior to oven drying and subsequent titration with KOH.

Drake et al. (1951) showed that grasses were generally less capable of cation absorption, but competed better for K than associated legumes in a soil low in available K. Mouat and Walker (1959b) also postulated that the poor competitive ability of clover to accumulate P when grown in association with grasses may also be as a result of the high CEC of clover roots. The poor competitive ability of legumes for P and K can also be explained in terms of root morphology (see Section IV,C).

IV. MORPHOLOGICAL CONSIDERATIONS A. FOLIAGE ARCHITECTURE

Most of the radiation intercepted by a field crop is absorbed by leaves, more specifically by leaf laminae. The architecture of the leaf canopy is therefore fundamental to light interception by plants. Probably the most important feature of plants that determines their competitive ability for light is height. Trenbath (1974) concluded that the component in a mixture with its leaf area higher in the canopy is at a general advantage. In most cases in grass-legume pastures the grasses are taller and more vigorously growing plants than the legumes, and this often leads to shading of the legumes particularly if the sward is not well managed. The angles of leaves to the horizontal and the area of the leaves are also important adjuncts to canopy height in determining the competitive ability of plants for light. Full light only reaches those levels or parts of leaves in the top layer of a plant stand. Thus, light intensity decreases by absorption and reflection down the various leaf layers to the soil surface. This gradient may be expressed as the “light extinction coefficient” (K) (see Geyger, 1978). Variations in K have been related to variations in canopy structure, especially to angle of leaf display

240

R. J. HAYNES

(Takeda, 1961; Loomis er al., 1968). The broad horizontally positioned (planophile)leaves of dicotyledonousplants such as clover absorb a great deal of light with only a few layers of leaves; the extinction coefficient is high (K > 1). In plants with vertically inclined leaves, such as grasses (erectophile), the light is distributed more evenly throughout the leaf canopy and K is consequently low (K = 0.3-0.7) (Geyger, 1978). Extinction coefficients calculated by Loomis and Williams (1969) were 0.60 for clover and 0.25 for grass. The amount of foliage present is an important parameter in studying light penetration and usage by pasture. It is usually expressed in terms of leaf area index (LAI), the ratio of total area of leaf to total area of gound beneath the canopy. Without high LA1 values, the useful light cannot be intercepted as efficiently at low levels of illumination; however with flat horizontal leaves a high LA1 is a disadvantage because of excessive self-shading. When net dry matter production is related to LAI, two types of relationships have been found. In the first relationship, yield increases as LA1 increases up to some optimum value (optimal LAI) and then declines. For instance Trifoliurn repens and T . subrerraneum reach an optimal LAI at which an equilibrium is established between leaf formation and leaf senescence (Brougham, 1958; Davidson and Donald, 1958). Blackman (1962) has postulated that the optimal LA1 for slightly inclined planophile leaves would be about 4-5 and that for sharply inclined erectophile leaves would be about 7. The results of Brougham (1957) show good agreement with this with white clover having an optimal LA1 of about 3.5 and that of perennial ryegrass being about 7. However, as the LA1 increases and the level of light interception rises, the index at which interception by the canopy is complete depends upon leaf shape and arrangement as well as the angle of the leaf. Furthermore, environmental factors such as light intensity, temperature, and plant density also effect the optimal LA1 (Black, 1963; Ludwig etal., 1965). In other cases, a plateau response has been found with yield remaining constant while LA1 continues to increase. Alfalfa is an example of this case; LA1 continues to increase after the index at which interception of light by the canopy is complete (Brown and Blaser, 1968). King and Evans (1967) suggested that lucerne has an advantageous canopy structure which enables adequate light penetration to lower levels, leading to more efficient light energy conservation by the whole canopy. The LA1 of alfalfa may reach 10 or 1 1 (e.g., Smith er al., 1964). Langer (1973a) has stated that because many clovers have horizontally orientated leaves, they reach a critical LA1 more quickly than grasses, and from this time onward light within the canopy becomes critical factor. Because of the complex nature of LA1 only general statements such as this can be made. However, the previous discussion has shown that legumes generally possess horizontally inclined leaves and absorb light from only a few layers of leaves, while in grasses light is distributed more evenly throughout the canopy because of their

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

24 1

more upright leaves. Hence legumes are generally more prone to be shaded by competitors than are grasses, and therefore are poor competitors for light. The position is exacerbated by the fact that the grasses tend to be taller and their growth rate is generally greater than that of the associated pasture legumes. However, in a pasture, defoliation is a common occurrence either through cutting for fodder or by the grazing of animals, and hence the nature of competition between species is modified. This aspect is discussed in the next section. B. GROWTH HABITA N D DEFOLIATION

The growth habit of pasture plants is extremely important because it is one of the principal factors governing the response of a plant to defoliation. The situation is however complex, since an interaction between grazing and environmental factors exists. The success and wide distribution of grasses as pasture plants is attributable to their structure and growth habit which is extremely well adapted to being grazed (Langer, 1973b). The key morphological feature is the location of the shoot meristems which lie close to the soil surface, well below the level normally reached by the grazing animal or machine. Both tillers and expanding leaf blades are able to continue growth after defoliation. The leaf forming process is thus unaffected by defoliation. Regrowth arises from initiation of new leaves from axillary buds, and also each leaf normally has its own meristematic tissues low enough to escape permanent damage, hence regrowth may occur. Culmed grasses, which may extend their internodes even in the absence of flowering, have their apexes elevated and thus vulnerable. Following defoliation, new shoot growth depends upon initiation of activity in axillary buds and regrowth may be slower than that of culmless grasses. Continuous grazing is therefore appropriate for culmless vegetative grasses, whereas rotational grazing is better suited for culmed grasses. Once flowering occurs, the stem apex of grasses becomes elevated above the grazing level. Defoliation during flowering may be particularly detrimental to regrowth. The poor persistence of Lolium x hybridum cv. MANAWA in New Zealand grasslands has been attributed in part to hard grazing during summer when tillers are flowering, and apexes are elevated (Harris, 1978). Defoliation of grasses results in variable responses in tillering. This is due to interactions of many different factors such as nutrients and water supply, extent of removal of stem apexes, supply of photosynthates for growth, and shading effects (Alberda, 1957; Harris, 1978). One of the reasons for the superiority and usefulness of the ryegrasses is that its yield is not depressed nearly as much as the majority . of grasses by continuous stocking compared to intermittent grazing or cutting (Smetham, 1972). It has been observed that ryegrass tillers more perfusely under

242

R. J. HAYNES

heavy stocking (Suckling, 1975). The grazing animal removes a relatively greater proportion of each tiller or stem with an upright plant in comparison with a plant of prostrate growth habit; and hence in the latter case relatively more green leaf, stem, and sheath is left after grazing to facilitate regrowth (Fig. 2). The grasses with upright growth habits (e.g., timothy, prairie grass, and cocksfoot) are slow to recover from defoliation. For example, Davies (1960) found excessively hard grazing soon reduced vigor of cocksfoot and extinguished it, but if grazing was too lenient it was apt to grow luxuriantly and gain dominance within the sward. Morphological differences exist in hybrid ryegrass (Loliurn x hybridurn), for instance the more prostrate cv. RUANUI is higher producing under close continuous grazing than the more upright cv. MANAWA (Smetham, 1972). Similarly, the suitability of legumes for grazing is primarily governed by their growth habit. Those with creeping stems tend to escape serious damage by grazing animals, while plants with an upright habit are more susceptible to damage. Adaptation to grazing in those species with an upright growth habit depends on the position of the crown. If a high proportion of shoots and their associated axillary buds are accessible to the animal, then the plant will require a prolonged period of regeneration. Stoloniferous legumes especially if they are rooted at frequent intervals (e.g., white clover) are well adapted to grazing (Fig. 2). The main temperate legumes of agricultural significance appear to be able to adapt their growth habit following frequent defoliation. For example Davidson Alfalfa

Ryegrass

Cocksfoot

Clover

FIG. 2. Contrasting growth habits of upright and prostrate legumes and grasses and their influence on the relative amount of shoot material left after defoliation. Dotted line represents a height of defoliation.

243

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

and Birch (1972) found that subterranean clover cut at weekly intervals covered the soil with a network of low-growing stolons. The development of this prostrate habit resulted in a large photosynthetic area and a high density of leaf production sites following defoliation. The wide diversity of growth forms of T. repens is a major factor in its success in temperate agriculture (Brougham et al., 1978). Its growth habit varies from erect, lax, large-leaved plants in the Mediterranean regions (of Ladino forms) to the small-leaved many-branched prostrate habit in northern Europe (e.g., Kentish wild forms). The small-leaved varieties are also more winter hardy. Growth and consequent survival of white clover depends upon stolon proliferation. The more stoloniferous types are strongly perennial. The growth form of a plant is of great importance when grown in a pasture. The red clovers (T. pratense) have an upright growth habit and have an immediate advantage over the prostrate white clover whenever the grazing interval is long enough to allow the grasses to shade the clovers. However, because of its prostrate habit, white clover is able to recover very quickly from cutting or grazing in comparison with red clover (Butler et al., 1959). Studies of the effect of cutting height and frequency on the growth of ladino white clover grown in association with three different grass species by Hunt and Wagner (1963a) showed that frequent cutting at the lower height resulted in the highest percentage clover in the harvested forage (Table 111). Infrequent cutting to 5 cm was more advantageous to clover growth than either frequent or infrequent cutting to the 10 cm height. It has been suggested that white clover also has a great ability to adapt to changes in environmental conditions. For example, plants that normally have medium-sized leaves may become very small-leaved when place under heavy grazing pressure (Brougham et al., 1978). Overgrazing of pastures has been reported to result in a close mat of white clover stolons bearing small leaves close to the ground (Kydd, 1957; Suckling, 1975). It has been found that with the birdsfoot trefoils, the more prostrate EmpireTable I11 Effects of Management Practices on the Average Percent Clover in Harvested Forage over One Year" Average percent clover for indicated frequency of cuttingb

Height of cut (cm) 5 10

Bromegrass-Wino clover

Cocksfoot-Ladino clover

Tall fescue-ladino clover

Frequent

Infrequent

Frequent

Infrequent

Frequent

Infrequent

I

4

11

2

18 1

3

1

1

4

6 2

Data from Hunt and Wagner, Agron. J. 55, 13-16 (1963a). bAverage percent clover (3 cuttings per year) for 1955.

244

R. J. HAYNES

type cultivars persist better in pastures than the more upright erect European types such as Viking and Mansfield (Seaney , 1974). The European cultivars can be used in pastures but must be rotationally grazed. The more prostrate Empire types, even when closely grazed, can flower and set seed on stems near the ground, thus facilitating reseeding. The alfalfa plant has a very different growth habit from that of the clovers (Fig. 2). There may be 5-25 or more erect stems per plant arising from a woody crown from which new stems grow when older ones are cut. The apical meristem, the source of new leaves, is elevated by stem elongation; hence alfalfa loses the growing points upon defoliation. In general, Langer (1973~)considers that there is a rhythmic sequence in the activity of buds whereby active extension growth of new shoots begins at the base of the plant when the previous crop of shoots has reached a certain stage of maturity, normally coinciding with the first appearance of young flowers. Buds at or close to the crown are the most important centers for regeneration after grazing (Langer, 1973c; Smith, 1972). Defoliation when the plant reaches the early flowering stage will give the highest yield of dry matter and preserve the longevity of the stand. Due to its nature of regrowth, alfalfa is often more delayed in making regrowth after defoliation than grasses, which may result in a grass-dominated stand (O’Connor, 1967; Spedding and Diekmahns, 1972). Hence, alfalfa will not persist under close defoliation, and continuous sheep grazing is known to practically eliminate alfalfa from a mixed pasture (Van Keuren and Marten, 1972). Generally, alfalfa-grass mixtures (Rohweder and Thompson, 1974) should be permitted to reach 20-45 cm before grazing, and a rest period of 5-6 weeks is needed before regrazing. Some of the aspects just discussed are illustrated by the results of Dobson et al. (1976) who compared a range of legumes when grown in association with tall fescue (Table IV). The white clovers had a markedly higher percentage legume coverage than the other more upright vetches, trefoils, and red clovers, particularly at the lower cutting height. With the latter three legumes, the percentage coverage increased with increased cutting height, while with the more prostrate white clovers, increased percentage coverage was achieved at the lower cutting height. Forage yield was, however, generally greater from the 5 cm than the 10 cm cutting height irrespective of the percentage legume coverage. The significance of the interaction between defoliation intensity and growth habit on influencing the botanical composition of pastures is highlighted in a recent review of New Zealand research (Harris, 1978). It was found that rotational defoliation with a high level of pasture utilization led to a simple pasture mixture of L . perenne and T . repens. A low level of utilization allowed ingress of taller species like D . glomerata, Holcus lanatus, and Bromus cathariticus. Close continuous defoliation led to a more species-rich association dominated by prostrate rhizomatous, stoloniferous or basal rosette habits such as Agrostis tenuis, Poa trivalis, Oxalis corniculata, Hydrocotyle moschuta, Sagina procum-

245

COMPETITIVE ASPECTS O F THE GRASS-LEGUME ASSOCIATION

Table IV Effect of Clipping Height on Forage Yield and Area Covered by Legumes at Falln for Selected Legume/Fescue Swardsb Forage yield (metric tons/ha)

Area covered by legumes (9%)

Sward

5 cm

10 cm

5 cm

10 cm

Fescue-milkvetch Fescue-Penngift crown vetch Fescue-Chemung crown vetch Fescue-Empire trefoil Fescue-Viking trefoil Fescue-Kenland red clover Fescue-Kenstar red clover Fescue-Regal white clover Fescue-Tillman white clover Fescue-Ladino white clover

4.3 8.5 9.3 7.7 10.7 9.1 10.4 8.3 9.5 7.0

4.4 7.8

27 41 45 38 39 47 37 68 70 52

32 51 50

8.1

6.8 9.4 9.0

10.5 6.0 8.1

5.9

46 55 50 42 61 61 47

aData from Dobson er al., Agron. J . 68, 123-125 (1976). bThree-year average results for 1971, 1972, and 1973.

bens, and Taraxacum oflcinale. Similarly, Cooper and Moms (1973) found that hard grazing of a mixed pasture with sheep led to a virtually pure ryegrass-white clover sward. On the other hand, repeated annual cutting at an advanced stage of growth, or even continued lax grazing, resulted in cocksfoot dominance and the almost complete disappearance of clover. Defoliation intensity may have a secondary effect on the grass-legume balance. By defoliation, the principal source of photosynthate necessary for continued nodule development and function is removed, and nodule decay and “shedding” is often observed following defoliation (Vallis, 1978). Thus it has been suggested that under more frequent defoliation there may be a greater transfer of N from legume to grass (Chestnutt and Lowe, 1970). However, experimental results are contradictory, and it may be that much of the N in senescing nodules is redistributed within the plant before the nodules slough off. So far, the general effects of defoliation have been discussed; however, grazing by livestock may have specific influences on the botanical composition of a pasture (Watkin and Clements, 1974). Legumes tend to be dominated by grasses when the two are present in mixtures partly because the legumes are selectively grazed by livestock (Spedding and Diekmahns, 1972; Watkin and Clements, 1974). Sheep and cattle may express differential selection patterns while grazing the same forage mixture (Bennett et al., 1970), but generally sheep are more selective than cattle (Watkin and Clements, 1974). In New Zealand conditions, observations indicate that at most times of the year the grazing animal shows

246

R. J. HAYNES

some preference for the white clover component of mixed pastures (Brougham, 1966). Other workers have also shown sheep have a significant preference for white clover over other grasses in a mixed sward (Grimes et al., 1967; Hodge and Doyle, 1967; Spedding and Diekmahns, 1972). Under intensive grazing systems this can lead to overgrazing of the clover resulting in regrowth characterized by small leaves with short petioles (Brougham, 1966) thus giving the grass component of the pasture a competitive advantage. In hill country farming, continuous selective grazing by dispersed flocks and herds can result in very low white clover yields and consequent poor nitrogen economy (Watkin and Clements, 1974). Mixtures of alfalfa and grass are known to be more unstable if the companion grass is one of the less palatable species such as cocksfoot or tall fescue, because the relatively palatable alfalfa is liable to suffer excess defoliation (Spedding and Diekmahns, 1972). Selective grazing patterns are not always straightforward; for instance, Bedell (1968) found that sheep selected consistently high amounts of subterranean clover in both perennial ryegrass-subclover and tall fescue-subclover pasture mixtures during spring, but in summer sheep preferred tall fescue to subclover, and on ryegrass mixtures they retained or increased their dietary preference for subclover. Treading damage to pasture plants which can be inflicted by both sheep and cattle (Edmond, 1970) is another factor that may influence the grass-legume balance in pastures (Brougham, 1966). The size of leaves, the height of the growing point, the physical strength of the leaf, and the ability to assume a rhizomatous growth habit have all been advanced as explanations for differential treading damage to pasture plants (Mitchell, 1960; Edmond, 1966, 1974; Evans, 1967). The direct effects are attributable to differential damage to stem and leaf species (Brougham, 1966). Total herbage yield of pastures is often significantly reduced through treading damage as stocking rates increase (Bryant et al., 1972), particularly under moist soil conditions. Notwithstanding the importance of legumes in pasture production, little is known about the differential reactions of grasses and legumes to treading and the underlying mechanisms involved. Edmond (1964) found that under heavy grazing, plant tolerances to treading were in the order from most to least: perennial ryegrass, Kentucky bluegrass, Poa trivulis, short rotational ryegrass, browntop, white clover, timothy, cocksfoot, red clover, and Yorkshire fog. Generally, British results have shown that trampling can adversely affect dry matter yield of cocksfoot and sainfoin, but can encourage the growth of perennial ryegrass (Spedding and Diekmahns, 1972). It was demonstrated by Edmond (1963) that in both summer and particularly winter, white clover was more affected by treading than perennial ryegrass thus giving the grass a competitive advantage. Growth habit may also affect competition for space on the soil surface. For example, Harris and Thomas (1973) suggested that a white clover-ryegrass mixture provides a more stable association than a grass monoculture because of the

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

247

ability of white clover to spread vegetatively by stolons into areas of bare ground, enabling a more efficient use of horizontal space. Lieth (1960) also showed that in a T. repens-L. perenne mixture the two species formed a mobile mosaic in which low grass density areas are invaded by clover and vice versa. Competition between white clover and low-yielding stoloniferous grasses (e.g., browntop and Poa trivafis) for the inter-ryegrass spaces is important in maintaining highyielding pastures (Harris and Thomas, 1972; Harris, 1974). Under frequent defoliation browntop may impose a shade effect on white clover stolons and thus restrict stolon development (Jackman and Mouat, 1972a), and browntop stolons may also impede spread and rooting down of white clover stolons. Increased intervals between defoliation allow white clover and ryegrass to shade and suppress browntop (Harris, 1974). It is apparent from the previous discussion that growth habit is a major factor involved in the competition between grasses and legumes for light and space. Competitive relationships are altered enormously by light and frequency and method of defoliation. Thus, frequency and intensity of defoliation must be planned in accordance with the growth habit and regrowth characteristics of individual components of a pasture, particularly in relation to the maintenance of the legume component. C. ROOT MORPHOLOGY

The study of the root system of pasture plants has lagged behind studies of shoots and leaves because of the difficulties involved in extracting and measuring the characteristics of roots. Generally for one component of a pasture to gain an advantage over the other in the early competition for nutrients and water, a faster growing root system (generally a greater root length) is required (Trenbath, 1974). Snaydon (1971) showed that root competition had a greater effect and probably began earlier than shoot competition. Donald (1958) showed that the proportionate yield reduction due to shoot competition was greater when root competition was occurring than when it was absent. Thus, an interaction between root and top competition occurs (Milthorpe, 1961). The results of Evans (1977, 1978) indicate that root morphology may be a very important factor in competition for water and nutrients within the pasture. Evans (1977) found that, in general, grasses had longer, thinner, more finely branched roots than clovers, although they have similar root surface areas per unit dry weight (Table V). Because root hairs of grasses were longer and more frequent, the calculated surface of the root hair cylinder and the volume of soil within the root hair cylinder were several times greater than those for clovers. Hence, on theoretical grounds, Evans (1977) postulated that in a grass-clover pasture most of the clover roots would be competing with grass roots for available nutrients, but only a

248

R. J . HAYNES Table V Characteristics of the Roots of Some Pasture Legumes and Grasses"

Species

Length per unit weight (cdmg)

Mean root diameter (mm)

Percentage of roots with root hairs

Root hair length (mm)

Volume within root hair cylindefl (mm?

Trifolium repens T . pratense T . subterraneum Lolium perenne Dactylis glomerata Phleum pratense

27.6 26.8 22.3 30.6 44.4 36.7

0.26 0.27 0.28 0.19 0.16 0.17

68 58 57 95 95 97

0.23 0.20 0.22 0.51 0.61 0.77

68 49 49 41 1 623 853

Data from Evans (1977). *Root hair cylinder defined as the cylinder described by the tips of the root hairs.

a

small proportion of grass roots would be in competition with those of clover. These differences could give the grasses a strong competitive advantage over clovers in nutrient and water uptake, especially of immobile nutrients such as phosphate. The inability of legumes in general to compete for P, K , and S has also been attributed by Rabotnov (1977) to their less ramified root systems compared to grasses. The distribution of roots within the soil profile often differs markedly among pasture plants. In a recent study, Evans (1978) showed that for white clover, ryegrass, cocksfoot, and alfalfa growing in a deep free-draining soil, most of the roots of all species occurred in the top 20 cm of soil, root numbers decreasing down to 100 cm. However, cocksfoot roots continued to be more numerous down to 140 cm, while the taproot of alfalfa reached at least 210 cm. It is well known that grasses commonly grown with alfalfa have a much higher proportion of the root system in the upper 30 cm of soil than alfalfa since the latter possesses a long taproot (Chamblee, 1972) (Fig. 1). Rumball (1977) observed that alfalfa roots reached 72 cm after 10 weeks during which the water table was progressively falling. Under favorable soil and moisture conditions, the roots of alfalfa may penetrate to a depth of 10 m and penetration of active roots to a depth of 3-4 m is common (Peterson, 1972). Under drought conditions, alfalfa obviously has a competitive advantage over other pasture plants.

V. COMPETITION FOR ENVIRONMENTAL FACTORS The main physiological and morphological characteristics of grasses and legumes have been discussed in relation to their potential competitive ability in a

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

249

pasture. Competition between grasses and legumes for the three main environmental factors (light, nutrients, and water) are now summarized in relation to these characteristics. A. LIGHT

Competition for light is unique; there is no common pool from which plants continue to draw their supplies until it is depleted. Incoming light energy is instantaneously available, it must be used or lost. Competition for light may occur whenever one plant casts a shadow on another or when one leaf shades another (Donald, 1963). Competition for light is between individual leaves, rather than between plants, because if leaves remain below compensation point for long periods, they are not supported by export of assimilates from other parts of the plant, and quickly die (Etherington, 1976). In a pasture, foliage of each plant will be intermingled with that of several of its neighbors. The successful plant is not necessarily the plant with more foliage (Etherington, 1976), but the plant which has its foliage in an advantageous position, relative to the foliage of its competitors for light interception. Thus the physiological characteristics, canopy height and architecture, and whole plant morphological characteristics, previously discussed, determine peak photosynthetic rates and light-competitive abilities of plants In practice, competition for light in a grass-legume pasture involves strong interactions with water and nutrient availability, notably mineral N. Interactions are modified by management practices applied to the pasture, particularly cutting height and frequency. Competition for many factors may ultimately operate through modified light relationships (Jennings and Aquino, 1968). Competition for light may influence the rate of N transfer from legume to grass. Shading of legumes restricts the supply of carbohydrates to the root system, thus causing death of nodule tissue and possibly an increase in the rate of N transfer to the competing grasses (Chestnutt and Lowe, 1970). Shading is known to induce rapid root nodule senescence in white clover and birdsfoot trefoil, although alfalfa and red clover are more tolerant of this treatment (Butler er al., 1959; McKee, 1962). The increase in height and closeness of swards as a result of fertilizer application creates, for low-growing plants, unfavorable conditions of light supply; and shading reduces the growth of these species and their ability to compete for nutrients (Donald, 1963). Therefore, many low-growing species of low shade resistance respond negatively to fertilizers, especially N. This is a particularly severe effect on low-growing legumes since a far greater share of applied N is taken up by the fast-growing tall grasses and that taken up by the legumes has little effect on yield, as N fixation is simultaneously depressed. Low-growing species may respond positively to fertilizers where frequent defoliation prevents a long-term shading (Rabotnov, 1977). For instance, Harris and

250

R. J. HAYNES

Thomas (1973) found that initial dominance by ryegrass in a perennial ryegrasdwhite clover pasture was a function of relatively high N availability and the greater growth rates of ryegrass under low light and low temperature regimes of winter resulting in the shading of clover. This shading was increased by increasing interval between defoliations. Defoliation frequency and intensity as a determinant of pasture composition has been reviewed recently (Harris, 1978). Generally, dominance or suppression can largely be explained by defoliation, which allows light to penetrate to levels where prostrate species dispose their leaf canopies, or by restriction of the ability of taller growing species to elevate their leaves to shade prostrate species. The proportion of stoloniferous species such as white clover in pastures is fostered by more frequent and intensive grazing systems, whereas the more erect-growing species are fostered by less frequent and less intensive grazing (Brougham, 1959; Harkness et al., 1970). Hence, whereas the prostrate Huia white clover is suppressed by an infrequent, lax grazing system (Brougham, 1959), the largerleafed, more erect open habit of Pitau white clover is best suited to that type of grazing system (Brock, 1974). Conversely, the open habit of Pitau makes it more susceptible than Huia to hard grazing. The alfalfa cultivars utilized in North America tend to dominate their companion grasses (Chamblee, 1972). One reason is possibly that alfalfa has an advantageous canopy structure which enables adequate light penetration to lower levels, leading to more efficient light energy usage by the whole canopy (Leach, 1978). Some workers (e.g., Chamblee, 1958) have postulated that alfalfa may produce a favorable “light” environment for the shade-tolerant grasses such as cocksfoot by shading them. However, in Europe and Australasia, grasses often have a competitive advantage for light over alfalfa (O’Connor, 1967; Spedding and Diekmahns, 1972). Grasses offer competition to alfalfa as the new growth is arising from the crown buds. The crown buds may be completely shaded by the grass Festuca arundinucea Schreb., resulting in many buds failing to develop (Chamblee and Lovvorn, 1953). B. NUTRIENTS

Some recent studies of interference between plant species have led to the conclusion that competition for nutrients is of greater importance than competition for light (e.g., Snaydon, 1971; Eagles, 1972; Hall, 1974). However, there is a large interaction between the two factors and competition for light is often causally related to competition for nutrients. Thus Donald (1963) has noted that a plant’s success in gaining a greater share of the limiting nutrient may cause such an increase in growth that a competing species may be suppressed secondarily by shading. However, in some situations where competition for light is low in a

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

25 1

nutrient-poor soil, the smaller plant may be at a competitive advantage (Newberry and Newman, 1978). The smaller plant may be able to grow relatively faster than the larger, so that despite its initial disadvantage it would not be suppressed further, and the smaller plant might increase its contribution to the total biomass of the mixture. It has already been noted that most legumes are poor competitors with grass for nitrogen. Sears (1962) has observed that white clover is not a pioneer legume but that it is adapted to high soil fertility conditions, although many ecotypes with different nutritional preferences occur (Snaydon, 1962; Snaydon and Bradshaw, 1962). In association with grasses, white clover is also a poor competitor for P (Mouat and Walker, 1959a; Jackman and Mouat, 1972a,b), K (McNaught, 1958; Mouat and Walker, 1959a), and S (Walker and Adams, 1958; Neller, 1960). The poor competitive ability of white clover for nutrients is probably related to differences in root morphology (Evans, 1977) and/or root CEC (Blaser and Brady, 1950; Mouat and Walker, 1959b). There are many examples in the literature that reflect the poor nutrient competitive ability of legumes. Generally the response of applied P to established grass-legume pastures is an increase in total yield as well as an increase in the percentage of the legume component in the sward (Baylor, 1974). The differing response curves of a legume and a grass to applied P in a pot experiment (Jackman and Mouat, 1972b) are shown in Fig. 3. Similar response curves which show the lower efficiency to applied P of clovers in comparison with grasses have been found by other workers (Ozanne et al., 1969; Barrow, 1975). The depressive effect of a relatively constant amount of browntop on clover growth in a field experiment, particularly at low levels of applied superphosphate, is clearly illustrated in Fig. 3. 1000 h

m

c

\

m

1 b

n 4

y1

> I 500 W

t I-

a 5 t

Browntop grown w i t h clover

0:

n 0

0 APPLIED

100 P Cppm)

200

0

500 1000 SUPERPHOSPHATE ( k g / h a )

FIG.3. The relative growth response of browntop and white clover to applied P in a pot experiment (left) and the dry matter yield of white clover grown alone and white clover and browntop grown together in a field experiment (right). (Compiled from Jackman and Mouat, 1972a. b.)

252

R. J . HAYNES

Jones (1974) found that in grass-clover mixtures in California and Oregon, yields of clover more than doubled with 80 kg S/ha, but those of grasses only increased by about 50%. The addition of P and K to pastures was shown by Rabotnov (1977) to increase both the yield and percent participation of legumes. Not all species of legumes responded identically to P and K; Trifolium repens was one of the species most responsive to P and K. Trifolium praiense and T . hybridum responded to P and K more favorably than Laihyrus praiensis and Vicia cracca. The capability of the grass species to compete for a particular nutrient is also important. Baylor (1974) found that competition for K by legumes was less severe with bromegrass (Bromus inermis) than with Kentucky bluegrass and most severe with Agrostis spp. O’Connor (1967) reviewed results of investigations that show that alfalfa suffers from competition with associated grasses for P, K , and S. Competition for K appears to be particularly important; and researchers have found that when K is present at high levels alfalfa dominates, while grass dominates when K is in limited supply (Hunt and Wagner, 1963a; MacLeod and Bradfield, 1963). The deep taproot system of alfalfa may give it some advantage where nutrients such as P and S are leached below the depth from which shallow-rooted companion crops can absorb them (O’Connor, 1967; Jones, 1970). In dry regions, surface-applied nutrients may be less available to alfalfa than to surface-rooting species (Brownlee et a l . , 1975). The subject of grass-legume sward nutrition is made more complex by the requirements of individual nutrients by the symbiotic Rhizobiurn spp. on the legume roots. Nodulated legumes have been shown to require increased amounts of Mo, P, and Cu and to have a unique requirement for Co (Bergersen, 1971). The mineral requirements of legume symbiosis has been reviewed recently (Munns, 1977). An important source of nutrients in pasture production comes from the recycling of nutrients by livestock (Wolton, 1963; Mott, 1974). The grazing animal removes only a small quantity of nutrients from the pasture, the remainder being excreted in dung and urine. The major nutrients in urine are N and K and sometimes S, while most of the P, Ca, and Mg ingested is excreted in dung as well as appreciable quantities of N and K (Watkin and Clements, 1974). Thus legume growth is strongly stimulated by dung where soil P is deficient, whereas the N in urine strongly stimulates grass growth (Brougham et al., 1978). Excretion in dairy cows is such that only two-fifths of the total daily amount of nutrients is excreted during the day and the remainder during the evening (Goodall, 1951). Thus in dairy farms there is a marked transfer in fertility from day paddocks to night paddocks; consequently the night paddocks become strongly grass dominant through high N returns and, conversely, the day paddocks become much more clover dominant because of lowering of soil N (Brougham, 1966). If urine scorches or kills pasture plants in the excreted area, deterioration in

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

253

botanical composition of the pasture may begin through ingress of weed species (Richards and Wolton, 1975). Similarly the common cattle dry pat can effect a major change through the death of the covered plants and the ingress of volunteer species (Watkin and Clements, 1974). Although the significance of such processes to the deterioration in botanical composition of pastures is probably very small, they appear to contribute to decline, which is often characterized by the ingress of volunteer species, particularly at the expense of the legume component. C . WATER

Success of any plant or species in competition for water will depend on the rate and completeness with which it can make use of the soil water supply. Because water is so often physiologically limiting, the rate and extent of exploitation of the soil space is important (Etherington, 1976). The use of different parts of the soil by the root systems of different species is one of the most important factors affecting competition for water. The other important factor is water use efficiency, which is related to the plant’s ability to regulate water loss. Competition for water usually occurs together with other forms of competition, especially for light and nitrogen (Donald, 1963). The water use efficiency of pasture plants is usually increased by application of fertilizers (Keller and Carlson, 1967; Rabotnov, 1977) since if the soil is kept moist anything that increases yield also increases water use efficiency. Both root morphology and water use efficiency have been outlined previously and will not be discussed in detail here. Generally, temperate grasses are less affected by dry conditions than is the production of temperate pasture legumes, with the exception of lucerne (Johns, 1972). Irrigation under drought conditions usually results in a greater yield response by red and white clover than by temperate grasses, resulting in an increased proportion of clover in the sward (Spedding and Diekmahns, 1972; Johns, 1972). The reason for this greater response to irrigation appears not to have been studied in detail. However, it is known that although grasses and clovers explore approximately the top 140 cm of soil, the roots of clovers are less ramified and hence the volume of soil explored is less than that for grasses (Evans, 1977, 1978). This may result in grasses having a competitive advantage over clovers for water uptake. Although cocksfoot shows a response similar to that of other grasses to irrigation, it is reputed to be particularly drought resistant (Spedding and Diekmahns, 1972). Evans (1978) concluded that cocksfoot made more efficient use of the water available in the top 140 cm than did perennial ryegrass and white clover. Depth of rooting is an important factor in competition between plants for water. For example, Burch and Johns (1978) found that fescue had more deep

254

R. J. HAYNES

roots than clover and extracted more water from deep soil layers, finally resulting in a drier soil profile. It is well known that alfalfa has a favorable competitive performance when grown with grasses when water is limiting (Chamblee, 1972; Smith and Stiefel, 1977). The long taproot of alfalfa enables it to obtain water from lower depths of the soil than grasses. Furthermore, Chamblee (1958) concluded that alfalfa is equally competitive with some grasses (e.g., cocksfoot) for soil moisture in the upper levels of the soil. The deep-rooted habit of Phaluris tuberosu enables it to persist in low rainfall areas. This makes it an important grass in the improved pastures of temperate Australia (Lazenby and Swain, 1972). The adaptability of a plant's growth characteristics is also important in competition for moisture. Many perennial grasses are able to exist in a dormant state throughout long periods of water stress, while annual grasses and temperate legumes (except alfalfa) may die fairly quickly (Johns, 1972).

VI. CONCLUSIONS The beneficial association of legumes and grasses is mainly related to the fixation of nitrogen by rhizobia, allowing the clover to become independent of the soil nitrogen, and in time this fixed nitrogen may become available to the grass. Thus competition for N is reduced within a grass-legume pasture where available soil N is at a low level initially. Basic morphological and physiological differences between pasture plants materially affect the nature of competition that develops between species in various mixtures. These are summarized in Table VI as above- and below-ground effects. Differences in seasonal growth patterns may minimize competition for environmental factors and hence confer greater a degree of stability on a mixture. Differences in growth rate of seedlings or of established plants during regrowth following defoliation also affect the competitive ability of plants. Generally, legumes are poor competitors with grasses for light nutrients and water. The legumes generally possess planophile leaves, while the leaves of grasses are generally erectophile. For this reason, legumes generally absorb a great deal of light with only a few layers of leaves (with the notable exception of alfalfa); while with grasses, light is distributed more evenly throughout the leaf canopy. Legumes also appear to have a physiological need for higher light intensities than grasses for maximum growth rates. The competitive disadvantage of legumes is exaggerated more by the fact that pasture grasses are generally taller and often have a faster growth rate and can thus overtop and shade the legumes. Defoliation frequency and intensity are therefore important in order to maintain the legume component of a pasture.

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

255

Table VI Characteristics of Plants Mecting the Interactions between the Environment and the Plants of a Grass-Legume Association Plant characteristic Aboveground Growth rate and period Growth habit Leaf architecture Light requirement Water use efficiency Below-ground Root morphology Root CEC Symbiotic N Fixation VA mycorrhizae

Environmental parameter

Light, space Light, space Light, space Light Water Water, nutrients, space Nutrients Nutrients Nutrients

The suitability of pasture plants for grazing is primarily governed by their growth habit. Grasses generally have their shoot meristems close to the ground surface, below the level normally reached by grazing animals. With the legumes, differences in growth habit are important since those with creeping rhizomatous stems tend to escape serious damage by grazing (defoliation and treading), while those with more upright habits are more susceptible to damage and should be rotationally grazed. However, those with upright habits have a better ability to compete for light with taller growing grasses, whereas those with more prostrate habits may be shaded out unless frequent defoliation to a low cutting height is practiced in order to reduce the competitive ability of the grasses for light. However, overgrazing can be detrimental since legumes are generally selectively grazed by livestock. Competition for many environmental factors may ultimately operate through modified light relationships. Grasses generally have longer, thinner, more finely branched roots than clovers; thus the former explore a greater volume of soil. This may give grasses a competitive advantage over clovers in nutrient (particularly P, K, and S) and water uptake when they are in short supply. The inoculation of clovers with improved strains of VAM can increase the volume of soil explored by the white clover roots by the formation of a loose network of external hyphae. The deep taproot of alfalfa enables it to gain a competitive advantage over shallow-rooted grasses when moisture andor nutrients are limiting in the upper part of the soil profile. The roots of legumes generally have a root CEC about twice that of grass roots. This may be a partial explanation for the poor ability of the legumes to compete with the grasses for P, K, and S. The application of P, K,or S is generally likely to favor legume growth at the expense of the grass component. How-

256

R. J. HAYNES

ever, N application is likely to be extremely detrimental to the legumes through luxuriant grass growth. An appreciation of the characteristics of individual pasture plants that give them a competitive advantage over others for one or more environmental factor in a particular situation is of paramount importance in the choice of pasture management techniques. It is particularly important in maintaining a desirable balance between legumes and grasses within a pasture. Consideration of these characteristics is particularly important in the breeding of new cultivars of legumes and grasses and their subsequent incorporation into contemporary pastures. Simply breeding for high yield of individuals is no guarantee of high total mixed pasture production or of stability in the botanical composition of the pasture. ACKNOWLEDGMENTS I wish to thank my colleagues Miss P. N. Crisp and Mr. J. G. Buwalda, and Professor T. W. Walker for stimulating my interest in pasture plant relations. I am also indebted to the University Grants Committee and the Public Trust for financial assistance.

REFERENCES Abbot, L. K . , and Robson, A. D. 1977. Aust. J . Agr. Res. 28,639-649. Akatsuka, K., and Sugihara, S. 1973. J . Jap. Soc. Grassl. Sci. 19, 215-221. Alberda, Th. 1957. Plant Soil 8, 199-230. Allos, H. F., and Bartholomew, M. V. 1959. Soil Sci. 19, 61-66. Asher, C. J., and Ozanne, P. G. 1961. Aust. J . Agr. Res. 12, 755-766. Bakhuis, J. A., and Kleter, H. 1965. Neth. J . Agr. Sci. 13, 280-310. Barrow, N. J. 1975. Aust. J. Agr. Res. 226, 137-143. Baylis, G . T. S. 1970. Plant Soil 33, 713-716. Baylis, G . T. S. 1975. In “Endomycorrhizas” (F. E. Sanders, B. Moses, and P. B. Tinker, eds.), pp. 373-389. Academic Press, New York. Baylor, J. E. 1974. In “Forage Fertilization” (D. A. Mays, ed.), pp. 171-188. Am. SOC. Agron., Madison, Wisconsin. Bedell. T. E. 1968. J . Range Manage. 21, 291-297. Beinhart, G. 1963. Crop Sci. 3, 209-213. Bennett, D., Morley, F. H. W., Clark, K. W., and Dudzinski, M. L. 1970. Ausr. J . Exp. Agric. Anim. Husb. 10, 694-709. Bergersen, F. J. 1971, Annu. Rev. Plant Physiol. 22, 121-140. Black, J. N. 1963. Ausi. J . Agr. Res. 14, 20-38. Blackman, G. E. 1962. Annu. Rep. Easr Mailing Srn. 1961. pp. 39-50. Blackman, G. E., and Black, J. N. 1959. Ann. Eor. 23, 51-63. Blackman, G. E., and Templeman, W.G. 1938. Ann. Bor. 2, 765-791. Blaser, R. E., and Brady, N. C. 1950. Agron J . 42, 128-135. Blaser, R. E., Taylor, T., Griffeth, W., and Skrdla, W. H. 1956. Agron. J . 48, 1-6. Bolton, J. R.,Goplen, 8. P.,and Baenziger, H. 1972. In “Alfalfa Science and Technology” (C. H. Hanson, ed.), pp. 1-34. Am. SOC.Agron., Madison, Wisconsin.

COMPETITIVE ASPECTS OF THE GRASS-LEGUME ASSOCIATION

257

Brock, J. L. 1974. N.Z. J. Exp. Agr. 2, 365-369. Brougham, R. W. 1957. Proc. N.Z. Grassl. Assoc. 19, 109-116. Brougham, R. W. 1958. Aust. J. Agr. Res. 9, 35-52. Brougham, R. W. 1959. N . Z . J. Agr. Res. 2, 1232-1248. Brougham, R. W. 1966. Proc. N.Z. Ecol. Soc. 13, 58-65. Brougham, R. W., Ball, P. R., and Williams, W. M. 1978. In “Plant Relations in Pastures’’ (J. R. Wilson, ed.), pp. 309-324. CSIRO, Canberra. Australia. Brown, R. H., and Blaser, R. E. 1968. Herb. Abstr. 38, 1-9. Brownlee, H., Scott, B. J., Kearins, R. D., and Bradley, J. 1975. Am:. J . Exp. Agr. Anim. Husb. 15, 475-483. Bryant, H. T., Blaser, R. E., and Peterson, J. R. 1972. Agron. J. 64, 331-334. Burch, G. J., and Johns, G. G. 1978. Aust. J. Plant Physiol. 5, 859-871. Butler, C. W., and Bathurst, N. 0. 1956. Proc. Int. Grassl. Congr. Palmerston North, VIIth, pp. 168- 178. Butler, G. W., Greenwood, R. W., and Soper, K. 1959. N . Z . J. Agr. Res. 2, 415-426. Chamblee, D. S. 1958. Agron. J . 50, 434-437. Chamblee, D. S. 1972. In “Alfalfa Science and Technology” (C. H. Hanson, ed.), pp. 21 1-228. Am. SOC.Agron., Madison, Wisconsin. Chamblee, D. S . , and Lovvom, R. L. 1953. Agron. J. 45, 192-196. Chestnutt, D. M. B. 1972. J. Br. GrasslandSoc. 27, 211-216. Chestnutt, D. M. B., and Lowe, J. 1970. In “White Clover Research” (J. Lowe, ed.), pp. 121-213. Br. Grassl. Soc. Occ. Symp. No. 6. Cooper,M. McG., and Moms, D. W. 1973. “Grass Farming.” Farming Press, Ipswich, England. Cowling, D. W., and Lockyer, D. R. 1965. J. Br. Grassl. SOC. 20, 197-204. Cowling, D. W., and Lockyer, D. R. 1967. J. Er. Grassl. Soc. 22, 53-61. Crush, J. R. 1973. New Phytol. 72, 965-973. Crush, J. R. 1974. New Phytol. 73, 743-749. Crush, J . R. 1975. N.Z. J. Agr. Res. 18, 361-364. Crush, J. R. 1976. N . Z . J. Agr. Res. 19, 473-476. Daft, M. J., and El Giahmi, A. A. 1975. In “Endomycorrhizas” (F. E. Sanders, B. Mosse, and P. B. Tinker, eds.), pp. 581-592. Academic Press, New York. Daft, M. J., and El Giahmi, A. A. 1976. Ann. Appl. Biol. 83, 273-276. Dart, P. J . , and Wildon, D. C. 1970. Ausr. J. Agr. Res. 2 1 , 4 5 5 6 . Davidson, I. L., and Birch, J. W. 1972. Aust. J. Agr. Res. 23, 981-993. Davidson, J . L., and Donald, C. M. 1958. Aust. J. Agr. Res. 9, 53-72. Davies, W. 1960. “The Grass Crop.” Spon, London. Davies, W. E. 1964. J. Br. Grussl. Soc. 19, 263-270. Diatloff, A. 1974. Aust. J. Agr. Res. 25, 577-581. Dilz, K., and Mulder. E. G. 1962. Plant Soil 16, 229-237. Dobson, J. W., Fisher, C. D., and Beaty, E. R. 1976. Agron. J. 68, 123-125. Donald, C. M. 1958. Aust. J. Agr. Res. 9, 421-435. Donald, C. M. 1963. Adv. Agron. 15, 1-1 17. Donald, C. M. 1978. In “Plant Relations in Pastures’’ (J. R. Wilson, ed.), pp. 41 1-420. CSIRO, Canberra, Australia. Drake, M., Vengris, J., and Colby, W. C. 1951. Soil Sci. 72, 139-147. Eagles, C . F. 1972. J. Appl. Ecol. 9, 141-151. Edmond, D. B. 1963. N . Z . J. Agr. Res. 6, 265-276. Edmond, D. B. 1964. N . Z . J. Agr. Res. 7, 1-16. Edmond, D. B. 1966. Proc. In:. Grassl. Congr. Xth Helsinki. pp. 453-457. Edmond, D. B. 1970. Proc. Int. Grassl. Congr. Xlth Surfers Paradise, pp. 604-608.

258

R. J. HAYNES

Edmond, D. B. 1974. N.Z. J. Exp. Agr. 2, 39-45. Etherington, J. E. 1976. “Environment and Plant Ecology.” Wiley, New York. Evans, P. S. 1967. J . Agr. Sci. 69, 175-181. Evans, P. S. 1977. N.Z. J . Agr. Res. 20, 331-335. Evans, P. S. 1978. N.Z. J . Agr. Res. 21, 261-265. Fergus, E. N., and Hollowell, E. A. 1960. Adv. Agron. 12, 365-436. Fitter, A. H. 1977. New Phytol. 79, 119-125. Fussell, G. E. 1964. J . Br. Grassl. Soc. 19, 49-54. Geyger, E. 1978. In “Application of Vegetation Science to Grassland Husbandry” (W. Krause, ed.), pp. 499-520. Dr. W. Jung, The Hague. Goodall, V. C. 1951. Proc. N . Z . Grassl. Assoc. 13, 86-96. Grant, E. A , , and Sellans, W. G. 1964. J . Br. Grassl. Soc. 19, 191-197. Gray, B., Drake, M., and Colby, W. G. 1953. Soil Sci. Soc. Am. Proc. 17, 235-239. Grimes, R. C., Watkin, B. R., and Gallagher, J. R. 1967. J. Agric. Sci. 68, 11-21. Hall, I. R. 1978. N . Z . J . Agr. Res. 21, 509-515. Hall, I. R., Scott, R. S., and Johnstone, P. D. 1977. N.Z. J . Agr. Res. 20, 349-355. Hall, R. 1. 1974. Aust. J . Agr. Res. 25, 739-747. Hanson, C. H., and Barnes, D. K. 1974. In “Forages. The Science of Grassland Agriculture” (M. E. Heath, D. S. Metcalfe and R. F. Barnes, eds.), pp. 136-147. Iowa State Univ. Press, Ames. Harkness, R. D., Hunt, 1. V., and Frame, J. 1970. In “White Clover Research” (J. Lowe, ed.), pp. 175-186. Br. Grassl. Soc. Occ. Symp. No. 6. Harlan, J. R. 1956. “Theory and Dynamics of Grassland Agriculture. ” Van Nostrand-Reinhold, Princeton, New Jersey. Harper, J. L. 1964. J . Ecol. 5 2 (Suppl.) 149-158. Harper, J. L. 1967. J. Ecol. 55, 247-270. Harris, W. 1974. N . Z . J . Agr. Res. 17, 251-256. Harris, W. 1978. In “Plant Relations in Pastures” (J. R. Wilson, ed.), pp. 67-85. CSIRO, Canberra, Australia. Harris, W., and Thomas, V. J. 1972. N . Z . J. Agr. Res. 15, 19-32. Harris, W., and Thomas, V. J. 1973. N . Z . J . Agr. Res. 16, 49-58. Hartley, W., and Williams, R. J. 1956. Proc. Int. Grassl. Congr. Vllth Palmerston North, pp. 190- 199.

Henzell, E. F., Martin, A. E., Ross, P. J., and Haydock, K. P. 1968. Aust. J . Agr. Res. 19,65-77. Hodge, R. W., and Doyle, J. J. 1967. Aust. J . Exp. Agr. Anim. Husb. 7, 141-143. Hunt, 0. J., and Wagner, R. E. 1963a. Agron. J . 55, 13-16. Hunt, 0. J., and Wagner, R. E. 1963b. Agron. J . 55, 16-19. Jackman, R. H., and Mouat, M. C. H. 1972a. N . Z . J . Agr. Res. 15, 653-666. Jackman, R. H.. and Mouat, M. C. H. 1972b. N.Z. J. Agr. Res. 15, 667-675. Janson, C. G., and White, J. G. H. 1971. N.Z. J. Agr. Res. 14. Jennings, P. R., and Aquino, R. C. 1968. Evolution (Loncaster) 2 2 , 529-542. Johns, G. G. 1972. In “Intensive Pasture Production” (A. Lazenby and F. G . Swain, eds.), pp. 160-175. Angus & Robertson, Sydney, Australia. Jones, M. B. 1974. In “Forage Fertilization” (D. A. Mays, ed.), pp. 171-188. Am. Soc. Agron., Madison, Wisconsin. Keller, W., and Carlson, C. W. 1967. In “Irrigation of Agricultural Lands” (R. M. Hagan, H. R. Haise, and T. W. Edminster, eds.), pp. 607-621. Am. SOC.Agron., Madison, Wisconsin. Keuren, R. W., Van, and Marten, G. C. 1972. In “Alfalfa Science and Technology” (C. H. Hanson, ed.), pp. 641-658. Am. Soc. Agron., Madison, Wisconsin. King, R. W., and Evans, L. T. 1967. Aust. J . Biol. Sci. 20, 623-635. Kydd, D. D. 1957. Agric. Rev. 2 , 25-30.

COMPETITIVE ASPECTS O F THE GRASS-LEGUME ASSOCIATION

259

Langer, R. H. M. 1973a. I n “Pastures and Pasture Plants” (R. H. M. Langer, ed.), pp. 41-63. Reed, Wellington, New Zealand. Langer, R. H. M. 1973b. In “Pastures and Pasture Plants” (R. H. M. Langer, ed.), pp. 65-84. Reed, Wellington, New Zealand. Langer, R. H. M. 1973~.In “Pastures and Pasture Plants” (R. H. M. Langer, ed.), pp. 347-363. Reed, Wellington, New Zealand. Laskey, B. C., and Wakefield, R. C. 1978. Agron. J . 70, 146-148. Lauchli, A. 1976. I n “Encyclopedia of Plant Physiology, Transport in Plants 11. Part B. Tissues and Organs” (U. Luttge and M. G. Pitman, eds.), pp. 3-34. Springer-Verlag, Berlin and New York. Lazenby, A,, and Swain, F. G. 1972. In “Intensive Pasture Production” (A. Lazenby, and F. G. Swain, eds.), pp. 67-96. Angus & Robertson, Sydney, Australia. Leach, G. J. 1978. In “Plant Relations in Pastures” (J. R. Wilson, ed.), pp. 290-308. CSIRO, Canberra, Australia. Leffel, R. C., and Gibson, P. B. 1974. In “Forages. The Science of Grassland Agriculture” (M. E. Heath, D. S. Metcalfe, and R. F. Barnes, eds.), pp. 167-176. Iowa State Univ. Press, Ames. Lieth, H. 1960. I n “The Biology of Weeds,” pp. 27-39. Blackwell, Oxford. Loomis, R. S., and Williams, W. A. 1969. In “Physiological Aspects of Crop Yield” (R. C. Dinauer, ed.), pp. 27-47. Am. SOC.Agron., Madison, Wisconsin. Loomis, R. S . , William, W. A., Duncan, W. G., Dovrat, A,, and Nunez, F. 1968. Crop Sci. 8, 352-356. Ludwig, L. C., Saeki, T., and Evans, L. T. 1965. Ausr. J . Eiol. Sci. 18, 1103-1118. McAuliffe, C., Chamblee, D. S., Uribe-Arango, H., and Woodhouse, W. W. 1958. Agron. J. 50, 334-337. McKee, G. W. 1962. P a . Sfate Univ. Coll. Agric. Bull. No. 689. MacLeod, L. B., and Bradfield, R. 1963. Agron. J. 55, 435-439. McNaught, K . J . 1958. N . Z . J . Agr. Res. 1, 148-181. Martin, T. W. 1960. Herb. Absrr. 30, 159-164. Milthorpe, F. L. 1961. In “Mechanisms in Biological Competition,” pp. 330-355. Cambridge Univ. Press, London and New York. Mitchell, K. J . 1956. N . Z . J . Sci. Technol. A . 38, 203-215. Mitchell, K. J . 1960. Proc. N . Z . SOC. Anim. Prod. 20, 82-92. Mosse, B. 1975. In “Endomycorrhizas” (F. E . Sanders, B. Mosse, and P. B. Tinker, eds.), pp. 467-484. Academic Press, New York. Mosse, B. 1977. New Phytol. 78, 277-288. Mosse, B., Powell, C. LI., and Hayman, D. S. 1976. New Phyrol. 76, 331-342. Mott, G. 0. 1974. In “Forage Fertilization” (D. A. Mays, ed.), pp. 323-339. Am. Soc. Agron., Madison, Wisconsin. Mouat, M. C. H., and Walker, T. W. 1959a. PIanr Soil 11, 30-40. Mouat, M. C. H., and Walker, T. W. 1959b. PIanr Soil 11, 41-52. Munns, D. N. 1977. In “A Treatise on Dinitrogen Fixation. Section IV: Agronomy and Ecology” (R. W. F. Hardy and A. H. Gibson, eds.), pp. 353-391. Wiley, New York. Neller, J. R. 1960. Proc. Inr. Grassl. Congr. 8th Reading pp. 90-93. Newberry, D. McC., and Newman. E. 1. 1978. Oecologia, 33, 361-380. Newman, E. I . , and Miller, M. H. 1977. J . Ecol. 65, 399-411. Newman, E. I . , and Rovira, A. D. 1975. J. Ecol. 63, 727-737. Nicol, H. 1935. Emp. J. Exp. Agric. 3, 189-195. Nicolson, T. H. 1975. I n “Endomycorrhizas” (F. E. Sanders, B. Mosse, and P. B. Tinker, eds.), pp. 25-34. Academic Press, New York. Nissen, P. 1974. Annu. Rev. Planr Physiol. 25, 53-79.

260

R. J . HAYNES

Nye, P. H., and Tinker, P. B. 1977. “Solute Movement in the Soil-Root System.” Blackwell, Oxford. O’Connor, K. F. 1967. I n “The Lucerne Crop” (R. H. M. Langer, ed.), pp. 47-61. Reed, Wellington, New Zealand. Ozanne, P. G., Keay. J., and Biddiscombe, E. F. 1969. Ausf. J . Agr. Res. 20, 809-818. Pate, J. S., and Dart, P. J. 1961. Plant Soil 15, 329-346. Peterson, H. B. 1972. I n “Alfalfa Science and Technology” (C. H. Hanson, ed.), pp. 469-480. Am. Soc. Agron., Madison, Wisconsin. Pineiro, J., and Hanis, W. 1978. N . Z . J . Agr. Res. 21, 83-92. Powell, C. L1. 1976. Nature (Landon) 264, 53-57. Powell, C. L1. 1977a. N . Z . J . Agr. Res. 20, 59-62. Powell, C. L1. 1977b. N . Z . J. Agr. Res. 20, 343-348. Powell, C. L1. 1977c. N . Z . J . Agr. Res. 20, 495-502. Rabotnov, T. A. 1977. I n “Application of Vegetation Science to Grassland Husbandry” (W. Krause, ed.), pp. 459-497. Dr. W. Jung, The Hague. Reid, D.,an.‘ Castle, M. E. 1965. J . Agr. Sci. 64, 184-194. Rhodes, I. l Y i c i Herb. Abstr. 40, 115-121. Rice, E. L. 1974. “Allelopathy.” Academic Press, New York. Richards, E. R., and Wolton, K. M. 1975. J . Er. Grassl. Soc. 30, 187-194. Rohweder, D. A., and Thompson, W. C. 1974. I n “Forages. The Science of Grassland Agriculture” (M. E. Heath, D. S. Metcalfe and R. F. Barnes, eds.), pp. 596-606. Iowa State Univ. Press, Ames. Rumball, P. J. 1977. Proc. N . Z . Grassl. Assoc. 39, 53-60. Scott, D. 1975. N . Z . J. Exp. Agr. 3, 135-141. Scott, W . R. 1972. I n “Pastures and Pasture Plants” (R. H. M. Langer, ed.), pp. 159-178. Reed, Wellington, New Zealand. Seaney, R. R. 1974. I n “Forages. The Science of Grassland Agriculture” (M. E. Heath, D. S. Metcalfe and R. F. Barnes, eds.), pp. 177-188. Iowa State Univ. Press, Ames. Sears, P. D. 1962. N . Z . J . Sci. Technol. A . 35, 221-236. Sears, P. D. 1962. N . Z . Ecol. Soc. Proc. 9, 57-63. Simpson, J. R. 1965. Ausr. J . Agric. Res. 16, 915-926. Smetham, M. L. 1972. In “Pastures and Pasture Plants” (R. H. M. Langer, ed.), pp. 179-228. Reed, Wellington, New Zealand. Smith, D. 1972. I n “Alfalfa Science andTechnology” (C. H. Hanson, ed.), pp. 481-496. Am. SOC. Agron., Madison, Wisconsin. and Bula, R. J. 1964. Crop Sci. 4, 577-580. Smith, J. G., Mott, G. 0.. Smith, R. G., and Stiefel, W. 1977. Proc. N . Z . Grassl. Assoc. 39, 61-69. Smith, S. E., and Daft, M. J. 1977. Ausr. J . Plant Physiol. 4, 403-413. Snaydon, R. W. 1962. J. Ecol. 50, 439-447. Snaydon, R. W. 1971. J . Appl. Ecol. 8, 687-697. Snaydon, R. W., and Bradshaw, A. D. 1962. J . Ecol. 50, 439-447. Sparling, G. P., and Tinker, P. B. 1975. I n “Endomycorrhizas” (F. E. Sanders, B. Mosse, and P. B. Tinker, eds.), pp. 545-560. Academic Press, New York. Spedding, C. R. W., and Diekmahns, E. C. (eds.). 1972. “Grasses and Legumes in British Agriculture.” Commonw. Agric. Bur., Hurley, England. Stiefel, E. I. 1977. I n “Recent Developments in Nitrogen Fixation” (W. Newton, J. R. Postgate, and C. Rodriguez, eds.), pp. 69-108. Academic Press, New York. Suckling, F. E. T. 1975. N . Z . J . Exp. Agr. 3, 351-436. Takeda, T. 1961. Jap. J . Bot. 17, 403-437. Tantrum, I.. and Mitchell, K. J. 1972. N . Z . Agr. Sci. 6, 7-8. Tothill, 1. C. 1978. I n “Plant Relations in Pastures” (J. R. Wilson, ed.), pp. 385-401. CSIRO, Canberra, Australia.

COMPETITIVE ASPECTS O F THE GRASS-LEGUME ASSOCIATION

26 1

Trenbath, B. R. 1974. Adv. Agron. 26, 177-210. Vallis, 1. 1978. In “Plant Relations in Pastures” (J. R. Wilson, ed.), pp. 190-201. CSIRO, Canberra, Australia. Vallis, I., Haydock, K. P., Ross, P. J., and Henzell, E. F. 1967. Aust. J . Agr. Res. 18, 865-877. Volz, M. G . , and Jacobson, L. 1977. Plant Soil 46, 79-91. Walker, T. W., and Adams, A. F. R. 1958. Plant Soil 9, 353-366. Walker, T. W., Adams, A. F. R., and Orchiston, H. D. 1956. Soil Sci. 81, 339-351. Watkin, B. R.. and Clements, R. J . 1974. I n “Plant Relations in Pastures” (J. R. Wilson, ed.), pp. 273-289.

Whitehead, D. C. 1970. Commonw. Bur. Pastures Field Crops Hurley, Berkshire Bull. NO. 48. Whittaker, R. H., and Freeny, P. P. 1971. Science 171,757-770. Wilson, P. W. 1940. “The Biochemistry of Nitrogen Fixation.” Univ. Wisconsin Press, Madison. Wit, C. T., de, Tow, P. G., and Ennik, G. C. 1966. Agric. Res. Rep. Wugeningen No. 687. Wolton, K. M. 1963. J . Br. Grassl. Soc. 18, 213-219.

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY. VOL. 33

NITROGEN LOSSES FROM TOPS OF PLANTS R. Wetselaar* and G. D. Farquhart 'Division of Land Use Research, Commonwealth Scientific and Industrial Research Organization, Canberra, Australia and tDepartment of Environmental Biology, Research School of Biological Sciences, Australian National University, Canberra, Australia

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

263

11. Record of Observed Decreases of Nitrogen Content of Plant Tops . . . . . . . . . . . . . . . . 264 A. Grain Crops . . . . . . . . . . .................................. 264

B . Pasture Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Other Plants and Crops . .................................. D. Discussion ........... .................................. ID. Possible Pathways of Nitrogen Losses from T o p s . . ............................ A. Translocation to Roots and Soil ............................... B. Loss of Plant Materi .................................. C. Gaseous Losses fro IV. Associated Methodology Problems ......... A. Plant Sampling.. ..................................................... B. Fresh Sample Storage . . . . . . . . . .................................. C. Oven-Drying . . . . D. Grinding . . . . . . . . ......... E. Kjeldahl Nitrogen ............................... F. Root Nitrogen . . . . . . . . . . G. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Conclusions ........................................ . References . ......................................................

271 275 276 279 283 293 294 294 296 297 297 298 299

I. INTRODUCTION Recognition of the importance of nitrogen in increasing crop production has led, in the last few decades, to an increased demand for nitrogen fertilizers. These fertilizers are expensive, requiring large amounts of energy for their production, and may contribute to eutrophication and perturbations of the chemistry of the atmosphere. Thus the proportion of the applied nitrogen that appears in harvestable products is important, both economically and environmentally. We have been aware for some time of instances where the absolute amount of nitrogen in the tops (aboveground parts) of crops appears to have decreased 263

Copyright @ 1980 by Academic h u . Is. All righls of npmduction in any form reserved. ISBNO 12-000733-9

264

R. WETSELAAR AND G . D. FARQUHAR

before harvest, the maximum amount of nitrogen in the tops being reached nearer to anthesis. Informal discussions with other agronomists gained both support for, and opposition to, the notion that the period of maximum net accumulation of nitrogen may occur some time before maturity in well-fertilized crops. We reviewed the literature and found that the phenomenon was not uncommon; we then examined possible causes for it. This article is the result of those efforts and its format reflects it. We collate known cases of decrease with time in the amount of nitrogen aboveground, and assess the possible contributions of various causes, including those associated with methodology. The conclusion that these decreases are real has implications. To the extent that they represent losses from the soil-plant system, they are important in assessing the results of nitrogen balance studies. It is not inconceivable that losses from the system, directly from plant tops, occur continuously over the full period of plant growth, only occasionally becoming apparent in records of aboveground nitrogen, when the rates of loss exceed the rates of uptake from the roots. Even if decreases in aboveground nitrogen merely reflect a reallocation of potentially harvestable amino acids, they are sources of inefficiency for those engaged in primary production. We assess these decreases and their implications, and identify priority areas of need for further knowledge.

II. RECORD OF OBSERVED DECREASES OF NITROGEN CONTENT OF PLANT TOPS The curve representing the dependence on time of the absolute N content of the tops of plants and crops usually has a sigmoidal shape. However, in some cases, a maximum is reached well before maturity with a decline from then on until harvesting. In this section we review cases where the latter has been observed. We include factors relevant to possible mechanisms of loss, such as leaching and loss of plant material. We examine in turn grain crops, pasture species, and other plants and crops. The subsection on grain crops is further subdivided on the basis of species; within each species the findings are listed chronologically. By nitrogen content we always mean the absolute amount of nitrogen expressed as weight per unit ground area or weight per plant or pot. A. GRAINCROPS

1 . Wheat

Wilfarth et al. (1905) grew spring wheat (Triticurn aestivum) in the field in Germany in the presence of nitrogenous and phosphatic fertilizers. During the

265

NITROGEN LOSSES FROM TOPS OF PLANTS

last three weeks before maturity, the tops lost 20% and the stubble plus roots 29% of their N content, while the dry matter of the tops increased. In the state of Washington, Doneen (1934) grew eight different wheat varieties in the field with or without 92 kg N ha-', applied in November, March, or May. The total N contents in the tops at maturity were 43.8,44.1, and 26.0% less than their respective maxima, for the three times of nitrogen application (means of all varieties). For the unfertilized crops this reduction was either small or negative. Miller (1939) grew two winter wheat varieties in the field in Kansas, for four consecutive seasons between 1931 and 1935. During the later stages of growth, weekly changes in the N content of tops were recorded. Statistical analyses were not made, and the only decreases in nitrogen yield that were probably of significance were in 1933-1934 and 1934-1935 for one variety. These losses did not appear to be related to the amount of rainfall received during the loss period. In fact, the highest decline in nitrogen yield occurred during a dry period. Boatwright and Haas (1961) grew a field crop of spring wheat in North Dakota, without fertilizer, with 50 kg ha-' of P only, and with 50 kg ha-' of P plus 50 kg ha-' of N. The total rainfall during the growing period was 265 mm. The nitrogen maxima were reached at heading, soft dough, and maturity for NP, N, and nil fertilizer, respectively (Fig. 1A). The nitrogen decrease for the NP treatment was due to losses of nitrogen from leaves, stems, and chaff of 22, 11, and 11 kg ha-', respectively, with a partial compensation due to increases in grain nitrogen. The decreases were not accompanied by decreases in dry matter (Fig. 1B). In 1957-1958 wheat was grown on a heavy soil (Gasser, 1961) and winter wheat and spring wheat in 1958-1959 on a light soil (Gasser, 1962) at Rothamsted, England, with or without 112 kg ha-' of nitrogen applied in the nitrate or ammonium form, either in autumn or in spring. In the case of the heavy soil, with all types of nitrogen application, there was an average decline in nitrogen yield in

-

80 r

N50 '50

N50 '0

No Po

~

v 2

263

MAY

1324

JUNE

8

25

JULY

2

26 3

MAY

13 24

JUNE

8

25

JULY

FIG. 1. Time course of N content (A) and dry matter (B) of tops of spring wheat without fertilizer, with 50 kg N ha-', or with 50 kg N ha-' plus 50 kg P ha-', applied as ammonium nitrate and superphosphate (Boatwright and Haas, 1961). Redrawn by permission of the American Society of Agronomy.

266

R. WETSELAAR AND G. D. FARQUHAR

the tops of winter wheat between ear emergence (mid-June) and maturity (early September) of 20 kg ha-'. During the same period, dry matter increased markedly. On the light soil (Fig. 2) the nitrogen in the tops for both winter and spring wheat increased at the later stages of growth when the N content was relatively low. In the high nitrogen plants the N content declined slightly with winter wheat and markedly with spring wheat (maximum 36 kg N ha-') over the same period. Dry matter changes were always positive. In Australia, Stonier (1962) grew wheat at three sites differing in soil nitrogen status. No nitrogen fertilizer was applied. At the site with the highest amount of soil mineral nitrogen throughout the growing season, the amount of nitrogen in the tops decreased to a statistically significant extent after flowering. Over the same period, dry matter first increased markedly between flowering and dough stage, and then decreased sharply until maturity. In South Australia three wheat varieties were grown in the field by Barley and Naidu (1964) using four levels of nitrogen application. The total amounts in the

7 ~

-I

0

x

*Oo0

B

6000 -

* 0

4000

5c *

e

-

2000 -

0

0-

MONTHS

FIG.2. Time course of dry matter and N contents of tops of winter and spring wheat without fertilizer nitrogen or with I12 kg N ha-' as ammonium sulfate or calcium nitrate, applied in October or March, on a light soil (Gasser, 1962). A = Winter dry matter, B = spring wheat dry matter, C = winter wheat N content, D = spring wheat N content, 0 = all treatments, V = ammonium sulfate (October), = ammonium sulfate (March), W = no fertilizer-N, A = calcium nitrate (October), 0 = calcium nitrate (March). Redrawn by permission of Blackwell Scientific Publications, Ltd.

+

267

NITROGEN LOSSES FROM TOPS OF PLANTS 160 -

--

140

-

120

-

mi34

O

r

“1 O

1

0GABO

~

JAVELIN

malure eAr emergence

ear emergence

BENCUBBIN

mature

mature

ear emergence

PHYSIOLOGICAL STAGE OF CROP

FIG.3. N content of tops of wheat, at two growth stages, for the varieties GABO, JAVELIN, and BENCUBBIN with different rates of fertilizer-N application as ammonium sulfate (Barley and Naidu, 1964).

tops at ear emergence and maturity are given in Fig. 3. Except for GABO, the higher the amount of nitrogen at ear emergence, the greater was the nitrogen loss thereafter. For BENCUBBIN, this nitrogen loss was 40 kg ha-’ for the treatment with the highest rate of nitrogen application. In the U.S.S.R., Nikitischev (1974) grew winter wheat for 4 consecutive years after either a pasture or maize, with different rates of fertilizer-N application. In 10 of the 14 experiments the N content in the tops of the winter wheat decreased between flowering and maturity at the higher levels of plant nitrogen content at flowering. Similarly, Nikitishev and Krusser (1974) observed a decline of nitrogen in the tops of the same crop for virtually all treatments (mean for 3 consecutive years given only). The crop was grown after either bare fallow, pasture, or maize, at different levels of nitrogen application. On each occasion the phosphorus content of the tops either increased or remained the same in the same growth period. A high loss of nitrogen was observed by Daigger ef al. (1976) from winter wheat grown at different locations in western Nebraska, with different rates of

268

R. WETSELAAR AND G. D. FARQUHAR

gn &k H Z

a 30

60

90

150

120

APPLIED NITROGEN (kg ha-'

I

FIG. 4. Nitrogen loss from wheat plants (including roots) between anthesis and maturity, versus rate of fertilizer-N application (Daigger et al.. 1976). Redrawn by permission of the American Society of Agronomy.

nitrogen application. In general, dry matter and N content of tops plus roots cached a maximum at anthesis. Thereafter, dry matter declined by about lo%, while losses of nitrogen from the tops plus roots between anthesis and maturity were dependent on the amount of nitrogen applied (Fig. 4). There was no apparent translocation of nitrogen to the roots, of which the dry matter content changed only slightly. Campbell and Davidson (1979) grew spring wheat in 9-liter, soil-filled containers in growth chambers under simulated irrigation conditions at three rates of nitrogen application and at two different temperatures. The N contents in the tops declined after anthesis in those treatments combining high temperature and least water stress. The decrease was greater the greater the rate of nitrogen application. In all cases the N and dry matter contents in the roots decreased from or before anthesis onward. There was no decrease of dry matter in the tops.

2. Rice High losses have also been observed in rice (Oryza sativa) plants. According to Tanaka and Navasero (1964) such losses occurred under high levels of nitrogen application during the latter stages of growth in Orissa, India. In the Philippines, in the field during the wet season, they obtained a nitrogen decrease of 47 kg ha-' in N content of rice tops. This occurred between three weeks before flowering and maturity, for the highest nitrogen input. With a lower fertilizer-N dressing the loss was smaller and occurred later (Fig. 5). A similar effect of nitrogen fertilizer rate on plant nitrogen losses was observed by J. J. Basinski and D. R. Airey (unpublished data), who grew rice in the field near Darwin, Australia, during the wet season (Fig. 6). When a similar experiment was repeated during the dry season the same effect was obtained, but losses

269

NITROGEN LOSSES FROM TOPS OF PLANTS

ha-’)

Z

FIG.5. Time course of N content of tops of rice as affected by rate of fertilizer-N application (Tanaka and Navasero, 1964).

were less and occurred only at higher application rates (225 and 270 kg N ha-’). From the data in Fig. 6 the apparent fertilizer-N recoveries can be calculated (Fig. 7). This recovery is defined as 100 x (N content in tops of fertilizer-N treatments - N content in tops of control)/amount of N applied, and is generally regarded as a measure of efficiency of application. The implications of Fig. 7 will be discussed in Section II,D. maximum

1st

head

50%

mature

DAYS AFTER SOWING

FIG. 6. Time course of N content of tops of rice as affected by rate of fertilizer-N application at sowing (J. J . Basinski and D. Airey, unpublished data).

270

R. WETSELAAR AND G. D. FARQUHAR

0

L I

40

I

I

I

80

I

120

I

I

I

160

I

200

I

I

240

FERTILIZER - N APPLIED (kg

FIG. 7. Apparent fertilizer-N recoveries as affected by amount of nitrogen applied, calculated from Fig. 6, for 68 days after sowing and at maturity.

3 . Sorghum Herron et al. (1963) grew grain sorghum (Sorghum bicolor) under four different irrigation regimes, at four levels of nitrogen, for 3 years. In one year, at all nitrogen inputs, the nitrogen in the tops decreased after the soft dough stage, with the greatest decline of 48 kg ha-', or 27.5%,after an application of 80 kg ha-' of fertilizer-N. 4 . Cereal Rye

Sneva and Hyder (1963) observed nitrogen losses from the tops of cereal rye (Secale cereale) grown for hay under dryland conditions. This loss, greatest one week after anthesis, was continuous between early flowering and hard dough stage. When cropped biennially this loss was 21 kg ha-' (or 45%), and when cropped annually it was 8 kg ha-' (or 43%) during the three weeks following anthesis. For the same crop Rumburg and Sneva (1970) measured, in Oregon, a statistically highly significant loss of 7.9 kg ha-' of nitrogen (or 27.4%)from the tops during the two weeks after anthesis. Over the same growth period a loss of 35% was obtained from the tops of plants grown in polythene cylinders filled with soil, and placed in the same field. The loss was accompanied by a 40% nitrogen and 19% dry matter loss in the roots, while the dry matter in the tops increased slightly.

5 . Barley Considerable losses of nitrogen from the tops were found by Burd (1919) during head formation of barley (Hordeum vulgare) grown in large containers filled with soil. The plants were protected against birds, rodents, and leaching by rain. Power ef al. (1 970) determined the effect of soil temperature on spring barley growth and nutrition in a growth chamber. Seedlings were placed in cans filled

27 1

NITROGEN LOSSES FROM TOPS OF PLANTS Soil

0-3 leaf a.4

Temperature 9 "C ,?,, 15.5 "C I)

leaf

'I(

225

? P

5

u I-l Z

8 Z

2 0

P 0

20

40

60

80

0

20

40

60

80

DAYS AFTER TRANSPLANTING

FIG.8. Time course of N content of tops and roots of spring barley, grown in cans containing soil, as affected by soil temperature and P supply (Power et al., 1970). Redrawn by permission of the American Society of Agronomy.

with a sandy loam to which fertilizer was added. The results (Fig. 8) suggest that the higher the N content in the tops at heading stage the greater was the nitrogen decrease. There did not appear to be any increase in root nitrogen to account for this loss in the tops. The dry matter of tops and roots sometimes decreased, but these decreases were much smaller than those in N content.

6 . Maize Terman and Allen (1974) grew maize (Zea mays) in the greenhouse in pots filled with soil with different types and rates of nitrogen fertilizer applied before sowing. The N content of tops and tops plus roots was determined at 4, 6, and 9 weeks in one experiment and at 4, 6, and 8 weeks in another. With high rates of applied nitrogen, significant losses of nitrogen occurred from 6 to 8 or 9 weeks after sowing. When Terman and Allen used fertilizers with a low dissolution rate, under otherwise identical experimental conditions, decrease in N content in the tops was confined to the highest rate of application. B. PASTURE SPECIES

1 . Annuals

In South Australia, Richardson et al. (193 1) grew Wimmera ryegrass (Lolium subulutum) in large, soil-filled pots in the glasshouse and noted a decrease in P content and a 14.9% decrease in N content in the tops between seed setting and maturity.

272

R. WETSELAAR AND G. D.FARQUHAR end

flowering and seed setting

900

-

800

c

of

plants mature

-'&zlbndl

I-

dry

N 0D.W.0 N *-• D . W . * - - - - - - * Herbage

.O--. I'

\

I

10

\\ .

I

I

0

I

c

700

0 Q

L L

0) Q

-

30

600

-

0

5 5

0

500

e Ly

I-

Ly

8 Z

400

20

; c

> LL n

300

200

10

100

0

10

8

6

12

14

0

18

16

WEEKS FROM SOWING

FIG.9. Dry matter and N contents of Trifolium subferranium. The plants were grown under cover in pots containing soil with 15.5 g of superphosphate added (Lapins and Watson, 1970). Redrawn by permission of CSIRO, Australia.

I2 000

--

r

0

10000 -

1

I

2

I

8000 -

2oo

,. -

llSO

0

5 m 6000Y

E 4

z 4 000 > m 2ooo, n

I =

t_

O'

JUL

'

AUG

'

SEP

'

OCT

NOV

'

0

FIG.10. Dry matter and N contents of tops of native pasture, dominated by the annuals Dihereropogon hagerupii and Louderia rogoensis, in the Sahel area of western Africa (F. W. T. Penning de Vries, personal communication).

NITROGEN LOSSES FROM TOPS OF PLANTS

273

In order to study the pathways of nitrogen losses from plant tops Lapins and Watson ( 1 970) grew subterranean clover (Trifolium subterraneum) and soft brome (Bromus mollis) in the field or in soil-filled pots under cover in Western Australia. In all cases the N content in the tops decreased at the later stages of growth. These losses were not compensated by increases in root nitrogen. For subterranean clover, the dry matter changes were proportionally much smaller than the corresponding changes in top and root nitrogen (Fig. 9). In the Sahel area of west Africa, Penning de Vries (personal communication) observed marked decreases in the total N content of the tops in the native pasture dominated by Diheteropogon hagerupii and Loudetia togoensis. This decrease was paralleled by a decline in dry matter (Fig. 10). 2 . Perennials It is advantageous for perennial species to translocate nitrogen into underground organs for storage; the following data are consistent with this. In Kansas, Aldous (1930) observed that in the roots the percentage of nitrogen, total sugars, and starch decreased until about the time of flowering, after which major increases were recorded for Vernonia baldwinii, Andropogon scoparius, Symphoricarpos vulgaris, Rhus glabra, and Verbena stricta. For the last species Grandfield ( I 930) found in the same region an increase in the roots of 24 and 17%for total carbohydrates and N content, respectively, between budding and maturity. In a glasshouse study Richardson et al. (1932) noted a 36% decrease in nitrogen in the tops of Phalaris tuberosa during the last two months before maturity. The nitrogen decline was accompanied by a decrease in P content. The nitrogen and phosphorus losses in the tops were accounted for by corresponding gains in the N and P contents in the butts and roots of the plant. A similar compensation by roots for nitrogen loss in the tops was found by Weinmann (1940) in South Africa for Rhodes grass (Chloris gayana) grown in pots. The higher the N content in the tops the greater was the decline in nitrogen at the later stages of growth (Fig. 11). Bird ( 1 943) investigated the effects of different cutting times on the performance of bromegrass (Bromus inermis), timothy (Phleum pratense), red top (Agrostis alba), and Kentucky bluegrass (Poa pratensis) in Canada. One year after sowing the total N contents in the tops decreased 16, 9, 8, and 4%, respectively, between the beginning of flowering and end of seed formation. In the semiarid climate of the Oregon High Desert, Sneva et al. (1958) applied 0 , 1 1 , 22, 33, and 44 kg ha-' of fertilizer-N to an established stand of crested wheatgrass (Agropyron cristatum) which had never been grazed or fertilized. The nitrogen yields in the tops decreased 37, 34,43, 48, and 50%, respectively, between June 1 and August 1.

274

R. WETSELAAR AND G . D. FARQUHAR

-,"

0.4 -

0.6

-:2

n9 2

- *' 0.2-

L

El

n I

--\

-*

NPK

............... .................. NIL

!-

'

23 MAR'37 17 APR 37

5

0-

5

8

,*-\ , , ,

2 0.20

*............. *--

-

2 0.4 -

I8 M A Y 37 I

"* .-. *- -----• ...................

1

a\*'*

0.6

-.-*-*-a

N! P?

/

--

standard error 1

1

. . . . . . .

I

r 0

' 0.75 0.5

0.25 -

200 100 -

DEC 1

m

p 5

5 8 n

JAN 1 FEB IMAR IAPR [MAY I JUN IJUL

RG. 12. Rainfall, and dry matter and N contents of tops of native pasture, dominated by the perennials Sorghum plumosum, Themeda australis and Chrysopogon fallax. in northern Australia (Norman, 1963). Redrawn by permission of CSIRO, Australia.

NITROGEN LOSSES FROM TOPS OF PLANTS

275

Norman ( 1963) found marked decreases in dry matter and in N and P contents in the tops of a native pasture dominated by the perennials Sorghum plumosum, Themeda australis, and Chrysopogon fallax in tropical northern Australia (Fig. 12). Other measurements showed that to some degree the butts of all three species, and the rhizome of C. fallux, acted as storage sites for the nitrogen and phosphorus. However, since there was a net loss of both nutrients from the leaves plus butts, a translocation to the roots was suggested. C. OTHERPLANTS A N D CROPS

For annual fodder crops in tropical northern Australia, Norman and Wetselaar (1960) obtained decreases after 12 weeks' growth in nitrogen yields of tops of fodder crops in 1956-1957 but not in 1957-1958. The higher the nitrogen input the greater was the decrease. The fodder crops were grain sorghum, Italian fodder sorghum, bulrush millet (Pennisetum americanum), and cowpea ( Vigna unguiculata). The nitrogen decline was most pronounced for millet. In the same region, cotton (Gossypium hirsutum) was grown under furrow irrigation with 10 different nitrogen input regimes. The nitrogen yields decreased after first boll opening only in cases where all nitrogen was applied at sowing (Fig. 13) (Basinski et al., 1971). Anderson ( 1866) grew broad beans (Viciafuba) in the field in western Europe and observed a decline of 75 kg N ha-' in the tops at the later stage of growth without any compensation by the roots for this loss (Fig. 14). Fertilizer -N

240

2oo

E

F

1

I-

8

=

*o[ O.--//

40t/ 1 5k

;6

1:6

DAYS AFTER SOWING

FIG.13. Time course of N content of tops of cotton as affected by rate of fertilizer-N application at sowing (Basinskier al., 1971).

276

R. WETSELAAR AND G. D. FARQUHAR

r t

I

5 2

85 9,,1

=

401

i /

MAYOJUN, JUL ,AUG, S E P , OCT,NOV,

*\*--*

*-*-..

2 40 FIG. 14.

N content in tops and roots of Viciafaba versus time (Anderson, 1866).

In West Germany, Frank (1954) grew Arabidopsis thaliana and Oenothera biennis in small pots filled with soil in the greenhouse. With the former species the total N content in the whole plant (including roots) decreased by 30% during the last month of growth. Over the same period the total N content in the underground organs decreased by 38%. A similar, but much less pronounced trend, was observed for 0. biennis. Although both species are biennials, they were grown from seedling to seed production and maturity within one season during the experiment. For hemp (Cannabis sativa) Mothes and Engelbrecht ( 1 952) noted in western Europe 67 and 22% decreases in the N content of the tops of male and female plants, respectively, toward the later stages of growth; while the N content in the roots decreased by 50 and 42%, respectively. According to Penston (1935) losses of plant nutrients from tobacco occur when the main stem reaches maturity. Watson and Petrie (1940) grew tobacco, of which the inflorescences were retained or removed, under controlled conditions in the glasshouse under two phosphate regimes. In all four cases the total amount of nitrogen in the tops decreased at the later stage of growth. In two out of the four cases the N content in the roots declined in the same period, while in the other two cases there was some compensation by the roots. D. DISCUSSION

A decline in N content of tops is not a rare phenomenon. We are led to this conclusion by the weight of examples and the magnitudes of observed declines

277

NITROGEN LOSSES FROM TOPS OF PLANTS

Table I Maximum Nitrogen Losses in Tops and Changes in Dry Matter of Tops and in Root Nitrogen Content of Annual Plants and Crops Max. N loss

Change

Dry matter Species Wheat Wheat Wheat Wheat Wheat Wheat Wheat Wheat Wheat Wheat Rice Rice

(kg ha-')

N content of roots

17 38 16 40 23 21 70 47 48 48

Cotton Tobacco

42

21 8

75

Environment

Reference

Field Field Field Field Field Field Field Field Field Chamber Field Field

54

- 19%

-10%

Miller ( 1939) Doneen ( 1934) Gasser ( 1 962) Wilfarth ef al. (1905) Boatwright and Haas (1961) Barley and Naidu (1964) Nikitishev (1974) Nikitishev and Krusser (1974) Daigger e f al. (1976) Campbell et al. (1979) Tanaka and Navasero (1964) Basinski and Airey (pers. comm.) Field H e m n et al. ( 1 963) Field Norman and Wetselaar (1960) Field Sneva and Hyder (1963) Field Rumburg and Sneva (1970) Chamber Power et al. (1970) Green- Terman and Allen ( 1 974) house Basinski e f al. (1971) Field Green- Watson and Petrie (1940) house Norman and Wetselaar (1960) Field Field Norman and Wetselaar (1960) Field Mothes and Engelbrecht (1952) Field Mothes and Engelbrecht (1952) Field Norman and Wetselaar (1960) Field Anderson (1866) GreenFrank ( 1 954) house Richardson e f al. (1931) Greenhouse Cover Lapins and Watson ( 1 970)

12.5

- 10%

-43%

Cover

44

Grain sorghum Grain sorghum Cereal rye Cereal rye Barley Maize

Fodder sorghum Millet Hemp (male) Hemp (female) Cowpea Vicia faba Arabidopsis thaliana Lolium subulatum Trifolium subrerraneum Bromus mollis

(%)

in tops

37 20 17.5 25 19 16.5

+ +

-

0

29 28 27 27.5 38 45 27 34 32

0

-

+ -35%

-40%

-50%

17.5 6 19 35 67 22 14 34

+ + -50% -42% 0

0 -38%

-

15

Lapins and Watson (1970)

278

R. WETSELAAR A N D G. D . FARQUHAR

(Table I). The phenomenon manifests itself in many geographical areas, under a variety of environmental conditions, and in annual as well as perennial species. The decreases are not always associated with a decrease in dry matter (Table I). Loss of plant parts, such as leaves, does not provide an explanation for the observed nitrogen losses, although continued dry matter production must be taken into account. This aspect is discussed in greater detail in Section III,B. For perennials the decrease was always associated with a transfer of nitrogen and other nutrients to underground storage organs, presumably to be used for the following growth period. For annuals, in all cases where root nitrogen contents were recorded, these contents decreased during the period that a nitrogen decline in the tops was observed (Table I). Therefore, the possible pathways of nitrogen disappearance from the tops of annuals need further identification. This we shall do in Section 111. Virtually all nitrogen declines occur from anthesis onward and are greater the higher the N content in the tops at that growth stage (e.g., Figs. 1-6, 8, 11, and 13). We assume for the moment that these declines are not associated with problems of methodology, which are discussed in Section IV. This being so, the declines in N content between anthesis and maturity represent actual losses of nitrogen from the tops. The effects of nitrogen losses from plant tops on fertilizer-N recovery and nitrogen balance sheets are several. Virtually all fertilizer-N recoveries are calculated at the time the plant or crop is mature. In the case of the data shown in Fig. 7 this leads to a classical decrease in recovery with increasing rate of nitrogen application. However, since the N content of the tops actually reached a maximum near anthesis (Fig. 6), a more appropriate time for calculation of recovery is at 68 days after sowing. When this is done no decrease in recovery is observed (Fig. 7), with high apparent recoveries of about 63% being maintained even at an application rate as high as 225 kg N ha-’. In other words, the plants had access to and made use of the fertilizer-N in their tops in the earlier stages of growth, but lost some of this nitrogen at later stages. Most balance sheets drawn up at maturity result in an “unaccounted-for” loss (Allison, 1955, 1966). Denitrification is usually assumed to be the major cause of this. From the above discussion it is clear that this assumption may lead to a gross overestimation of the amount lost by this mechanism. It is possible that the decline in nitrogen in the tops reflects the limited capacity of the inflorescences to store nitrogen. In this context, the unpublished data of J. J. Basinski and D. R. Airey are relevant; they relate to 23 different nitrogen treatments, representing different rates and times of nitrogen application, for wet and dry season crops of rice. Wetselaar (1972) analyzed them and found in all cases that the ratio between the N content of grain and the total N content of the tops was always 0.65 at maturity, irrespective of whether the tops lost or gained nitrogen between head formation and maturity. Austin et al. (1977) termed this

279

NITROGEN LOSSES FROM TOPS O F PLANTS

Table I1 Rate of Decline of Nitrogen Content in Tops of Crops N decline (kdha)

Period of decline (days)

Rate of N decline (kg/ha/day )

Vicia faba Cotton Rice Rice Wheat Wheat Wheat Wheat Grain sorghum

75 42 40 47 16 38 70 40 48

69 19 22 59 31 83 40" 40" 42

1.09 2.21 1.82 0.80 0.52

Unweighted mean

46

45

1.20

Crop

0.46 1.75 1 .oo

1.14

Reference Anderson ( 1866) Basinski er al. (1971) Fig. 6 Tanaka and Navasero (1964) Boatwright and Haas (1961) Gasser (1962) Daigger er al. (1976) Barley and Naidu (1964) Heron et al. (1963)

Estimated.

ratio the nitrogen harvest index (NHI). They tested 47 genotypes of wheat and measured a mean NHI of 0.683, with a standard deviation of 0.0033. From the data of Bartholomew ( 1 972), we conclude that this index was also conservative for rice, maize, and wheat over a wide range of nitrogen inputs. The indices calculated from the data of Boatwright and Haas (1961) were 0.68, 0.69, and 0.7 1 for zero, N, and NP fertilization; nitrogen decreases occurred in the last two cases (see Fig. 1). We hypothesize that for annuals any nitrogen in excess of (I/NHI) times the capacity of the inflorescences may be lost. It is also possible that losses from plant tops occur continuously during all stages of plant growth, but only become apparent when they are greater than the rate of uptake of nitrogen via the roots. Continuous uptake of nitrogen may occur when fertilizers with low dissolution rates are applied. This could mask nitrogen losses from the tops. Such may have been the case in the experiments of Terman and Allen (1974) on maize. The possibility of continuous losses must be borne in mind at all times in Section 111. Table I1 was prepared to aid in assessing the significance of the various pathways of loss. It summarizes the decline in N content that occurred in the tops of annual field crops, the times taken, and the rates of net loss. For the cases cited the decline was 40-50 kg N ha-' at a mean rate of 1.2 kg N ha-' day-'. 111. POSSIBLE PATHWAYS OF NITROGEN LOSSES FROM TOPS In this section we examine mechanisms of plant nitrogen losses in order to determine the extent to which nitrogen losses given in Section I1 can be explained. The pathways are grouped in three different categories.

280

R. WETSELAAR AND G . D. FARQUHAR

A. TRANSLOCATION TO ROOTSA N D SOIL

I . Translocation to Roots

Nitrogen can be redistributed within the plant, either from older leaves to younger ones, or from leaves to inflorescences, mainly in the form of products of protein hydrolysis such as amino acids. These products move from the point of hydrolysis to sinks in other parts of the plant. According to Williams (1955) roots are such potential sinks, becoming operative only if nitrogen is not wholly taken up by other plant parts. It is therefore logical to postulate that the observed nitrogen losses from tops could be due to migration of nitrogenous products from the tops to the roots, especially since (1) the nitrogen losses occur during a period when leaves senesce, liberating mobile nitrogenous products; and (2) higher losses occur from tops that have a higher maximum nitrogen yield. In the latter case there is more likely to be a nitrogen “surplus” after the demand by the inflorescence has been met. For perennials there is a strong indication (Section II,B,2) that such a transfer of nitrogen and other nutrients is into underground organs, presumably for future use. On the other hand, in all cases where measurements were made on the roots of annuals the N content of the roots decreased during the period in which nitrogen declines in the tops were observed (Table I). Although we acknowledge the problems associated with the determination of root nitrogen content (Section IV,F), we must conclude that at present there is no evidence that roots of annuals act as termini for a nitrogen surplus in the tops. Whether or not annuals translocate nitrogen via the roots to the soil will be discussed in the next subsection.

2 . Root Exudates In relation to translocation of nitrogen from roots to soil, Rovira (1969b) differentiates between diffusible and nondiffusible exudates, the latter consisting of sloughed cells and mucilagenous material. For the former, Hale et al. (1971) calculated the amount of amino nitrogen exuded by different plant species such as wheat, peas (Pisurn sativum), alfalfa (Medicago sativa), sorghum, and sunhemp, based on experiments with seedlings grown in nutrient solutions. Extrapolating to the field situation and assuming a plant population of between 4 X lo4and 20 X lo4ha-’ (depending on the crop), we arrive at rates of exudation in the order of 0.01 g amino-N ha-’ day-’. Since it is generally agreed that the tip is the major site of exudation (Oades, 1978), one would have to assume an increase in number of root tips between seedling and flowering stage of about four to five orders of magnitude in order to account for the observed nitrogen losses under discussion if exudation alone were the main cause for these losses.

28 1

NITROGEN LOSSES FROM TOPS OF PLANTS

NUTRIENT CONTENl {mg per pot)

FIG. 15. Dry matter and N, P, and K contents of maize tops plus roots. The plants were grown in pots containing soil plus 100,500, or lo00 mg fertilizer-N and 300 mg P and K per pot (Terman and Allen, 1974). Redrawn by permission of the Soil Science Society of America.

However, as plants develop, exudation by diffusion declines and transfer of organic nutrients via decomposition of moribund tissues becomes more important (Rovira, 1965). Accepting this, and also accepting for the moment that the latter mechanism is the main cause for nitrogen losses from plant tops, one would expect this loss to be paralleled by increases in the total N content of soil roots. Terman and Allen (1974) obtained decreases in N contents in both tops and roots of maize (Section II,A,6). These decreases were paralleled by similar declines of potassium and phosphorus (Fig. 15). From these results they then concluded that the loss of mineral nutrients was through the roots into the soil, since leaching loss from the leaves by guttation, dew, or rain did not occur under their prevailing glasshouse conditions. Such an efflux of nitrate-N from the roots to a solution surrounding the roots has been demonstrated by Morgan et al. (1973). There does not appear to be any evidence, however, for an increase of nitrogen in the soil. Firth et al. (1973) applied 15N-labeledammonium sulfate to maize under furrow imgation in the Central Plain of Thailand and found a decrease in (soil + root) - N between 8 and 12 weeks from sowing (Fig. 16). Similar results were obtained with rice (Wetselaar, 1974), despite the fact that declines of I5N in the tops occurred. These findings need to be confirmed for other crops and environmental conditions, especially since with lowland rice the possibility of denitrification losses of nitrogen translocated to the soil cannot be excluded.

+

282

R. WETSELAAR AND G . D. FARQUHAR

"

0

4 8 WEEKS FROM SOWING

12

FIG. 16. Fate of 15N-labeledammonium sulfate applied at 100 kg ha-l, at sowing, to maize (Firth e r a / . , 1974). The numbers in circles represent the percentages for individual compartments. Redrawn by permission of Pergamon Press.

3 . Loss from the Soil of Nitrogen Transferred from Roots

It is possible that nitrogen from tops may be translocated from the tops to the roots, transferred to the soil, and then, after conversion to nitrate, be lost by leaching or denitrification. Translocation of nitrogen from tops to the roots of perennials occurs, as discussed earlier. In perennial ryegrass (Lolium perenne), at least, the nitrogen moves further. Dilz and Woldendorp (1960) and Huntjens (197 1) found with this species that I5N-labeled organic nitrogen compounds were translocated from the grass tops to the soil. The potential exists for subsequent denitrification. Woldendorp ( 1962) found that 20% of I5N-labeled nitrate-N and 7% of similarly labeled NK-N were lost when applied to perennial ryegrass sods with living roots, even though the soil was not wet (60% of field capacity). The losses were only 7% and zero, respectively, when the roots were killed. The same author confirmed the formation of N,O plus N2 in large Warburg vessels when nitrate was applied to 2-week-old perennial ryegrass or pea plants. In a similar system he observed an oxygen consumption that was 20 times higher for sods with live roots than sods in which the roots were killed. He inferred that the partial pressure of oxygen in the rhizosphere was low enough for denitrification to take place. Substantial translocation of nitrogen from the tops to the roots of annuals after anthesis has not been demonstrated. It is possible, but unlikely, that the nitrogen is transferred to the soil and lost sufficiently rapidly that no increase is observed in the N content of root soil. The main source of root exudates appears to be the decomposition of moribund tissues (Section III,A,2); sufficient energy for denitrification is possibly available from this same source. Lapins and Watson

+

NITROGEN LOSSES FROM TOPS OF PLANTS

283

(1970), who found substantial nitrogen losses from the tops, regarded losses via denitrification as unlikely in their case, in view of the dry condition of the soil during the period that the losses occurred. Whether conditions favorable for denitrification prevail in the period between flowering and maturity of an annual crop needs to be investigated. We doubt, however, that denitrification of nitrogen derived from translocation from the tops is a major source of loss in annual crops. B. Loss OF PLANTMATERIAL

Nitrogen in the tops of plants and crops can be lost due to abortion of different plant parts; removal by animals, microorganisms, and mechanical agitation; and by exudation and leaching. These will be discussed in the following subsections. I . Pollen, Flowers, and Fruits

Pollen, flowers, and fruits can abort spontaneously. Some species, such as maize, discharge considerable amounts of pollen into the atmosphere, and the nitrogen it contains must therefore be regarded as lost from the tops. This does occur during the period when nitrogen decreases in tops have been observed. Unfortunately, we are not aware of any quantitative measurements of the amount of nitrogen contained in airborne pollen. Rumburg and Sneva (1970) determined the nitrogen contained in stamens of cereal rye and estimated on that basis a possible loss of 16.1 kg N ha-', which was far in excess of the actually measured nitrogen decline in the tops of 7.9 kg ha-'. Further losses could occur through the shedding of anthers and filaments. Sneva (1967) estimated the nitrogen loss through the shedding of reproductive organs of crested wheatgrass (Agropyron desortorum) to be 0.03 kg N ha-' day-' during one summer month. Nitrogen losses through the shedding of flowers and fruits can be estimated easily in the field, and these losses should be taken into account when nitrogen declines in tops are studied. They appear to be of minor importance, even for a crop such as cotton, which can shed large amounts. In order to account for the nitrogen losses observed by Basinski et al. (1971) (Fig. 13), P. Jakobsen (personal communication) intercepted all shed material from this crop, but was not able to account for the nitrogen losses from the tops. Studies of nitrogen losses through the shedding of pollen are required. It seems unlikely, however, that this is the major cause of the observed average decline of 46 kg N ha-' (Table I).

2 . Leaves Leaves, or parts thereof, can be removed through mechanical agitation after senescence or through insect grazing. If we assume a wheat crop containing 120

284

R. WETSELAAR A N D G . D. FARQUHAR

kg ha-' of nitrogen in the tops with a NHI of 0.68, the leaves plus stems contain about 38 kg N ha-'. For a nitrogen loss of 28 and 39 kg N ha-', as obtained, respectively, by Gasser (1962) and Barley and Naidu (1964) (see Figs. 2 and 3), a major part of the leaves would have to be removed. This would certainly have resulted in a decline in dry matter content, despite any compensation by continued dry matter production. Gasser did not observe a decline in dry matter (dry matter yields over time were not given by Barley and Naidu). Richardson et al. ( 1931) observed nitrogen losses in tops of Lolium subulatum in spite of special precautions taken to prevent losses of leaves and seed. The large losses of nitrogen from rice obtained by Tanaka and Navasero (1964) (Fig. 5) could not, according to them, have been due to leaf drop. The decreases in dry matter and N contents of sorghum also exceeded the amounts that could be attributed to loss of leaves (Herron et al., 1963). It appears, therefore, that losses of nitrogen from tops occur over and above those associated with losses of leaves, and parts thereof. 3 . Insects

Cotton is particularly susceptible to insect damage. There is some evidence that numbers and fecundity of some insect species are increased when high nitrogen plants are infested (A. G. L. Wilson, personal communication). This would indeed mean that a greater nitrogen decline in the tops of cotton would be obtained with such plants. But to be of significance, at least 2 0 4 0 % of the leaves would have to be removed before seed setting, and 40-80% thereafter, assuming that the fruits were not affected by insects. Wilson et al. ( 1972) observed in the field a greater decline in the N content of tops of irrigated cotton with continuous chemical pest control than with a delayed control. In the latter case, some insects selectively destroyed small reproductive plant parts, causing the cotton plant to continue vegetative and reproductive development, which resulted ultimately in a higher nitrogen uptake. 4 . Birds

Birds can harvest or damage seeds or whole seed heads in gramineous crops. Since the seeds can contain 75 kg N ha-', a 20% seed removal would be equivalent to a nitrogen loss of 15 kg N ha-'. Such losses appear to be possible and would occur during the period when the N content in tops declines.

5 . Particulates Mechanical agitation due to wind may induce shedding of different plant parts; it may also cause losses of minerals that accumulate on leaf surfaces due to leaf

NITROGEN LOSSES FROM TOPS OF PLANTS

285

rubbing (Kingsley et al., 1957). Viets (1965) has observed dried salt, largely amides, on the tips of grasses within a few days after nitrogen fertilizer applications. In addition, plants give off airborne, submicron-sized particulates, even during the young stages of growth. This can occur either through fragmentation and loss of wax rodlets during rapid leaf expansion andor through the production of airborne salt crystals generated during rapid transpiration (Beauford et a l . , 1977). The N content of such particulates is not known.

6 . Microorganisms When plants are “diseased,” heterotrophs can oxidize amino acids, in addition to carbohydrates, as primary sources of carbon and energy according to RCH(NY)COOH

+ YO = RCOCOOH + NH, + 2H

(1)

(Fry, 1955). If the biochemical environment is such that the ammonia is volatilized, losses of plant nitrogen will occur. Nothing seems to be known about the extent of plant nitrogen losses through this mechanism. It is unlikely, however, that in all cases in Section I1 the test plants were substantially diseased. 7 . Leaching As nitrogen can be translocated from older to younger leaves and from leaves to grain, a part of the nitrogen in tops must be in a water-soluble form. Some of this can diffuse into water on leaf surfaces. If this water is removed, e.g., by rain, dew dripping, sprinkler irrigation, or spraying with pesticides, the plant will lose nitrogen. Leaching of minerals from a number of species has been established using radioisotopes and is generally small for young leaves, increases with age, and is largest when leaves approach senescence (Tukey and Tukey, 1959). In arid climates where rain needs to be supplemented by irrigation, the overhead sprinkler system induces more leaching losses than furrow or sheet irrigation. According to Tukey and Tukey (1959) leaching losses will be increased when the overhead irrigation water contains appreciable quantities of dissolved salts. Oil and various chemicals used for pest control also appear to increase leaching losses, in particular when surface-active agents are employed, since such agents tend to reduce the hydrophobicity of leaves (Tukey and Tukey, 1959). Kingsley et al. (1957) subjected mature wheat and flax (Linum usitatissimum) to simulated rain and concluded that leaching of plant nutrients is a continuous process in the field. However, it is known that such leaching losses due to rainfall or dew tend to decrease with consecutive leachings (Lausberg, 1935; Kingsley et al., 1957).

286

R. WETSELAAR AND G. D. FARQUHAR

Few studies have been made in natural environments of losses of nitrogen due to leaching. In forests, however, extensive measurements have been made of N contents of throughfall and stemflow. For instance, Likens et al. (1977) observed an annual flux of 9.3 kg N ha-' through these mechanisms for an undisturbed forested ecosystem at Hubbard Brook, New York. We are unaware of such measurements having been made on annual crops. Long er al. (1956) report 1 and 1 1 % nitrogen losses, respectively, from cereals and beans when their leaves were subjected to artificial rain. Recent laboratory leaching experiments by I. H. Hunter (personal communication) using winter wheat suggest a high leachability of leaf nitrate. The actual amount of nitrate-N removed from the leaves appears to be dependent on leaf age and amount of rainfall (Fig. 17). Up to 70% of the nitrogen contained in the leaf could be removed by subjecting a fully senescent leaf to artificial rain. Tanaka and Navasero (1964) investigated the effect of artificial rain on rice, grown under flooded conditions in the field. They sprayed a high-nitrogen plant with 300 ml of distilled water per day for 2 days at ear initiation stage. The N content (composition not specified) of the collected water was 46 ppm, which is equivalent to a removal of 0.56 kg N ha-' per mm rainfall. They observed in fact, from plants of the same nitrogen treatment, a decrease in nitrogen of 47 kg ha-' between August 24 (ear initiation stage) and October 22 (maturity) (Fig. 5). This would require a total rainfall of 84 mm if nitrogen were extracted at the same rate as given above for the whole period. Tanaka and Navasero (1964) also observed a nitrogen concentration of 102

70

c

50

a

n U

<

YI 4

-

30 25% s e n e s c e n t A 10% s e n e s c e n t

0

1

2

3

4

5

6

7

8

RAINFALL (mm)

FIG. 17. Dependence on the amount of artificial rain of the amount of nitrate-N leached from wheat leaves as affected by degree of senescence (I. H. Hunter, personal communication).

NITROGEN LOSSES FROM TOPS OF PLANTS

287

ppm in dew. Assuming that one such dew event removed the nitrogen it contained permanently from the leaves, that the leaf area index (LAI) was 7, and that the dew covered all leaves on one side with an average thickness of 1 mm,then 7 kg ha-' of nitrogen would be removed with each dew event. In some cases, nitrogen leaching losses can be inferred. Penning de Vries and Djiteye (1980) report that in the natural rangelands of the Sahel, W. Africa, the greatest nitrogen losses by the annual pastures occur in the last few weeks of the wet season. In addition they found on one occasion dew to contain 0.5 kg N ha-', mainly in the organic form. In the monsoonal area of northwestern Australia, Norman and Wetselaar (1960) obtained for a series of fodder crops nitrogen losses during the 1956-1957 wet season, but not in the following season (Fig. 18). The former season had a much higher rainfall, especially during the later stages of growth. In contrast, in the same area the N and dry matter content in the tops of native pastures increased until the end of the wet season, and only decreased afer the rains had ceased (Fig. 12). In other cases, the absence of leaching losses can be assumed, because the plants were grown under cover and the pots rather than the plants were watered (Basinski et al., 1971; Campbell and Davidson, 1979; Lapins and Watson, 1970; Norman, 1963; Power et al., 1970; Richardson et al., 1931; Terman and Allen, 1974; Watson and Petrie, 1940), or because the time of nitrogen decline occurred during a dry period. Lapins and Watson (1970) took special precautions to avoid leaching by rain, but still observed losses from a maturing legume and nonlegume. Although they suggested ammonia volatilization as a likely loss path-

E IY

Z

8

WEEKS FROM SOWING

FIG. 18. Time course of mean N content of tops of six fodder crops (grain sorghum, fodder sorghum, bullrush millet, sudan grass, cowpea, and Cyamopsis tetragonolobu) without N fertilizer (N,,) or with 105 kg ha& of ammonium sulfate-N (NILfor wet seasons 1956-1957 and 1957-1958, in northern Australia. The former season had high rainfalls in the latter stages of crop growth (Norman and Wetselaar, 1960). Redrawn by permission of CSIRO,Australia.

288

R. WETSELAAR AND G. D. FARQUHAR

way, losses via dew and wind could not be excluded. The figure obtained by extrapolation of the results of Tanaka and Navasero could be an overestimate. Their results do suggest, however, that substantial losses due to leaching can occur and may in fact explain some of the observations of nitrogen losses from maturing plants and crops.

8 . Guttation Guttation is another avenue for loss of nitrogen compounds from plant tops. Very little appears to be known about amount, frequency, chemical composition, or actual dislodging of the guttation droplet from the leaf. Goatley and Lewis ( 1966) determined the chemical composition of guttation fluid on 3-cm-tall seedlings of rye, wheat, and barley. The concentrations in ppm N for the three species were, respectively, nitrate: 0.2, 0.2, and 0.2; ammonium: 4.6, 4.1, and 7.3; asparagine: 0.5, 0.4, and 2.0; glutamine: 0.2, 0.1, and 0.2. Let us assume an average loss of guttation fluid of 0.5 ml per plant per day containing on average 5 ppm of nitrogen. Over 100 days only 40 g N ha-' will be lost for a crop such as flooded rice at a 25 x 25 cm spacing. This example suggests that guttation alone could not account for major losses from plant tops. C. GASEOUS LOSSESFROM PLANTS

Recent reports (Stutte and Weiland, 1978; Silva and Stutte, 1979a; Stutte et a l . , 1979; Weiland and Stutte, 1978a, 1980) suggest that plants can lose large amounts of nitrogenous compounds directly to the atmosphere. Stutte and Weiland (1978) extrapolated their results for soybean to a field crop for one season, and estimated losses via this pathway of 45 kg N ha-'. Gaseous losses may be important in contributing to the nitrogen declines reviewed in Section 11. There are a number of nitrogenous compounds that could volatilize from leaves, including ammonia and some other amines, dinitrogen, oxides of nitrogen, hydrogen cyanide, and some alkaloids (McKee, 1962). Of these we first examine ammonia and other amines. 1 . Ammonia and Other Amines

In 1928 Klein and Steiner examined 78 different dicotyledonous and 21 different monocotyledonous species, mainly nonagronomic ones, for their capacities to volatilize ammonia and other amines. Tests were made on the flowers of 70 species and on the leaves of 14 species. All gave off ammonia at an average rate of 3 p g per 100 g fresh weight in 24 hours. The rates were higher for certain species, for younger leaves (senescing leaves were not investigated), at

NITROGEN LOSSES FROM TOPS OF PLANTS

289

high light intensities and at high temperatures. With 12 species it was established that other amines were volatilized, particularly from flowers. Such losses suggested to them the presence of ammonia (or ammonium) and other volatile amines in leaves and flowers. Therefore, Steiner and Lijffler (193 1) made painstaking investigations on hundreds of plant species and found ammonia or ammonium in all species, although the amount varied greatly. The concentrations were low in leaf buds, increased to a maximum in young leaves, and then decreased. A second peak was obtained during “autumn yellowing.” Yemm (1937) also found that ammonium levels in leaves increased with senescence, when expressed on a fresh weight basis. Steiner and Loffler (1931) observed that some flowers and ripening seeds and fruits contained high ammonia (or ammonium) concentrations. There was also an indication that shaded leaves had a higher concentration than nonshaded ones. Other volatile amines were found in 48 species, especially in flowers, with I -amylamine, and 1-butylamine being conspicuous. In our view, the techniques used by Steiner and Loffler were such that the presence of the compounds was established, but there is doubt about the actual volatilization losses. To assess their rates we first note from Table I1 that losses in the field can be approximately 1.2 kg N ha-’ day-’. Assuming a LA1 of 5 , this would correspond to a continuous flux from the leaves of 20 nmol m-2 s-l. The average rate of volatilization of ammonia observed by Klein and Steiner (1928) can be estimated as 25 pmol m-* sec-I (1.5 g N ha-’ day-’ at LA1 5 ) which is negligible. It was not until 40 years later that further indications of ammonia losses from plants were obtained (Martin and Ross, 1968). Rhodes grass (Chforis gayana) was grown in a soil with a pH of 5.7 to which KI5NO3 or (I5NH4)$3O4 was applied. The plant and soil were contained in a gas lysimeter in which any volatilized ammonia was trapped. The authors thought, while not excluding its possibility, that gaseous losses of 15NH3 from their acidic soil were unlikely, especially when 15NO; was applied. They suggested ammonia may have evolved from the plant. Such a suggestion was strengthened by similar experiments (Crasswell and Martin, 1975) in which higher rates of ammonia volatilization were obtained during leaf senescence. In the past decade, interest increased in the gaseous exchange of ammonia between leaves and the environment. Several groups showed that leaves can take up (and metabolize) ammonia (Porter et af., 1972; Hutchinson et al., 1972; Meyer, 1973; Aneja, 1977; Faller, 1972; Denmead et al., 1974, 1976, 1978). Porter et af. (1972) reported that trace amounts of gaseous NH, or amide may leak from maize. Meyer (1973) passed ammonia-free air over several species. The plants released ammonia, but the loss per month was always less than or equal to 0.3% of the total plant N content. An important discovery was that of the massive release and refixation of

290

R. WETSELAAR AND G . D. FARQUHAR

ammonia (or ammonium) in the photorespiratory carbon cycle (Woo et al., 1978; Keys et a l . , 1978). Farquhar et al. (1979) inferred that there was a finite partial pressure of ammonia in the intercellular spaces of maize. They detected no fluxes of NH, into or out of maize leaves when realistically low partial pressures of ammonia were imposed (5 ? 3 nbar), but observed an evolution of 0.6 nmol mP2sec -'(36 g N ha-lday-' at LA1 5 ) from senescing leaves. Farquhar et al. (1980) showed that when the atmospheric partial pressure of gaseous ammonia is above a certain level, called the ammonia compensation point, there is a net uptake of ammonia. When the atmospheric partial pressure is below this compensation point, ammonia is evolved by the leaves. They found that the compensation point increased with temperature. Farquhar er al. (1980) suggested that the molar flux of NH, into a leaf, J, could be predicted from

where g is the stomatal conductance to the diffusion of NH,, n, is the ambient (or atmospheric) partial pressure of NH,, y is the compensation point, and P is the total atmospheric pressure. It is appropriate now to assess the likely magnitude of fluxes into and out of leaves in a crop. Assuming that Eq. (2) is valid, losses (negative J ) should occur when the compensation point is high, i.e., at high temperatures and, possibly, in senescing leaves. Further, the losses will be larger when stomatal conductance is greatest, as occurs with high light intensities, ample moisture, and high levels of nutrition, in particular of nitrogen (Wong et a l . , 1979). With a generous estimate for stomatal conductance of 0.4 mol m-2 sec-', the compensation point, y, would have to be 50 nbar higher than the partial pressure in the atmosphere in order to obtain a continuous efflux of 20 nmol rn-' sec -'(1.2 kg N ha-' day -'at LAI 5 ) . Considering that conductance is usually less, and that losses would occur mainly during the day, a more reasonable estimate would be 100 nbar. The actual values of y observed by Farquhar et al. (1980) were all less than 10 nbar, which makes it appear that ammonia volatilization may account for a small part, only, of losses reviewed in Section 11. However, these experiments were canied out at low light intensities and the effect of light on the compensation point is unknown. The observation (Meyer, 1973) that Pinus virginiana and Amaranthus retrofexus raised the outgoing partial pressures of ammonia in a chamber to 16 and 18 nbar, respectively, emphasizes the need for further research in this area, using both physiological (Farquhar er a l . , 1980) and micrometeorological techniques (e.g., Denmead et al., 1978).

2 . Dinirrogen Many early workers (Eggleton, 1935; Pearsall and Billimoria, 1937, 1939) attributed gaseous nitrogenous losses to the Van Slyke reaction (Van Slyke,

NITROGEN LOSSES FROM TOPS OF PLANTS

29 1

1914) in which nitrous acid reacts with amino acid to release dinitrogen: HNOZ

+ R'CH'NYCOOH + Nz + YO + R'CH'OH*COOH

(3)

Pearsall and Billimoria ( 1937) floated Narcissus pseudonarcissus leaves on media containing N h N 0 3 and found that, under certain conditions, the nitrogen contents of the tissues were lowered considerably. These losses were greatest in older tissues when proteolysis was in progress. They later (1939) observed that in darkness the nitrogen losses were approximately double the amount of NQ--N taken up, and they concluded that most of the NO; formed by reduction of NO; was lost in a stoichiometry consistent with the Van Slyke reaction. Similar losses occurred from senescing tissue in the light. In nonsenescing tissue in light, the losses were a smaller proportion of the assimilated nitrate, since protein was formed and nitrite did not accumulate. Subsequent work in other laboratories did not confirm these observations and the reaction has fallen from favor in a physiological context (McKee, 1962). However, Porter (1969) concluded from experiments with nitrite and anthranilic acid at 50°C, that aromatic amines may undergo the classic Van Slyke reaction. In general, this reaction is favored by high nitrite concentrations and a pH less than 5, as undissociated nitrous acid is the actual reactant. The disparity between the earlier sets of results may be caused by differing pH conditions and degrees of nitrite accumulation. Vanecko and Varner (1955) infiltrated wheat leaves with K15N02and observed that illumination led to evolution of 15N2.The majority of the gas evolved (always greater than 80%) was oxygen, and the authors suggested that the dominant overall reaction was 2HN0, + 2&0

+ 2NH3

+ 34

(4)

with the ammonia being assimilated. Their observation of 1.4 moles of gas evolved per mole of nitrite metabolized was close to the predicted value of 1.5. The reaction of nitrite with amino nitrogen was assumed to have caused the evolution of dinitrogen.

3 . Nitric Oxide and Nitrogen Dioxide Reaction (4) represents the complete reduction of nitrite, as catalyzed in vivo by nitrite reductase, production of the necessary reductant being associated with the photolysis of water and evolution of oxygen. Evans and McAuliffe (1956) observed the nonenzymatic reduction of nitrite by ascorbic acid and by NADPH (reduced pyridine nucleotide). About 80% of the nitrite-N appeared as nitric oxide (NO) and nitrous oxide (N,O), while dinitrogen was also evolved. The reaction was slow at pH 6 and the rate increased rapidly with increasing acidity. Porter (1969) found that when nitrite reacted with oximes (which are present in plants: McKee, 1962) or organic matter, the major gaseous product was NO.

292

R. WETSELAAR AND G. D. FARQUHAR

Application of certain photosynthetic inhibitors blocks nitrite reduction. Evolution of NO, and to a lesser extent NO,, then occurs from soybean leaves (Klepper, 1979a). Application of 2,4-D is thought to promote nitrate reduction in the dark, also causing nitrite assimilation and the emission of NO, (NO and N q ) from soybean (Klepper, 1979a). The greatest rate of evolution observed was 2.4 pg NO and 0.07 pg NO, (g fresh weight)-' min-'. Assuming a leaf fresh weight of 1 kg per square meter of leaf surface, this corresponds to a flux of 1360 nmol m-' sec-I (82 kg N ha-' day-' at LA1 5). Some aspects of this research have been verified in wheat (Churchill and Klepper, 1979)but with smaller amounts of NO, detected. Klepper (1979b) has also reported that hydration products of SO, are effective in causing NO, emission from soybean. The rates just quoted are enormous, and Klepper's conclusion that untreated leaves evolve no NO, needs confirmation; techniques capable of measuring fluxes of 1360 nmol m-, sec-' may not be sufficiently sensitive to detect those of interest here, i.e., fluxes of about 20 nmol m-, sec-' (1.2 kg N ha-' day-' at LA1 5 ) . Nitrogen dioxide reacts with water to yield nitric and nitrous acids. Plants take up and metabolize NO,, the reaction being apparently first order with respect to NO, concentration (Rogers et al., 1979). However, the partial pressure of NO, to which Rogers et al. exposed their plants was never less than 80 nbar, and extrapolation of their data suggests that a compensation point for NO, of 33 nbar may have existed in their leaves. Plants also assimilate NO (Anderson and Mansfield, 1979), although the extent to which the NO must first convert to NO, appears to be unknown. The data of Galbally and Roy (1978) suggest that a compensation point may also exist for NO. When they placed a chamber over grass, the soil plant system gave off NO, and the partial pressure in the chamber rose, but leveled at 15 nbar. 4 . Total Gaseous Fluxes

There are other volatile nitrogenous compounds that may be evolved by plants. For example many species possess the capacity to produce HCN and the cyanogenic glycoside, dhumn, can reach a concentration of 3-5% of the dry weight of sorghum seedlings (Beevers, 1976). Franzke and Hume (1945) detected emission of trace amounts of HCN from sorghum plants, but the amounts involved, less than 2.5 pmol m-2 sec-' (0.15 g N ha-' day-' at LA1 5 ) , had no impact on the plant nitrogen budget. In assessing the extent to which nitrogenous compounds are exchanged between leaves and the atmosphere, it is important that the atmospheric level of the compound be neither depleted nor enhanced. In few experiments is this ensured, and rates of volatilization may often be underestimated if the partial pressure in the chamber approaches a compensation point. This probably occurred, for example, in the experiments of Klein and Steiner (1928).

NITROGEN LOSSES FROM TOPS OF PLANTS

293

Pyrochemiluminescence is a new and sensitive technique for measurements of total nitrogen, apart from molecular nitrogen. Stutte and Weiland (1978) made the first attempt to apply this technique to the assessment of the total evolution of nitrogenous compounds from leaves, apart from dinitrogen. These workers enclosed leaves of various species in plastic bags. Air was passed over the leaves, through a cold finger in dry ice for condensation, and then recycled over the leaves. The condensate was then analyzed for nonelemental nitrogen using the pyrochemiluminescent technique. Nitrogen losses of, typically, 9 nmol m-2 sec - l (0.5 kg N ha-' day-' at LA1 5) were observed in a number of plant species. The losses increased with increasing temperature, the highest rate being 27 nmol sec -l(1.6 kg N ha-' day -' at LA1 5) from Amurunrhuspulmeri at 35°C. A disadvantage of their procedure is that the partial pressures of the unknown nitrogenous compounds are close to zero in the air recirculating over the leaf. Overestimation of the effluxes will then result unless the leaf is incapable of metabolizing the gases. The early results of Stutte and co-workers (Stutte and Weiland, 1978; Stutte et al., 1979; Silva and Stutte, 1979a,b; Weiland and Stutte, 1978a,b, 1979a,b) are consistent with the presence in leaves of a number of nitrogenous compounds, each exerting a finite partial pressure. For example, a leaf with compensation points for NH3, NOz, and NO of 3, 33, and 15 nbar, respectively, and with a stomatal conductance to the diffusion of water vapor of 0.4 mol m-' sec-', would evolve nitrogen at a rate of 9 nmol m P sec-' (0.5 kg N ha-' day-' at LA1 5) if the partial pressure of each compound outside the leaf were zero. Recent data from Stutte's laboratory (Stutte etul., 1979; Weiland and Stutte, 1980) suggest that the majority of the nitrogen evolved is in reduced forms. I t is reasonable that the unknown compounds should also have compensation points and that the compensation points should increase with temperature. The effluxes should then increase with temperature, especially if the partial pressure of each compound outside the leaf is zero. Nevertheless the atmospheric partial pressures are not zero and therefore these exciting findings may be extrapolated to the field situation only with caution.

IV. ASSOCIATED METHODOLOGY PROBLEMS

In order to assess the N content of tops, plant samples representing the true situation in the field or greenhouse must be obtained. These samples must then be processed in such a way that they remain representative. Nitrogen can be lost from samples during collection, transport and storage, oven drying, grinding, and chemical analysis. These will be discussed in the following sections.

294

R. WETSELAAR AND G.D . FARQUHAR

A. PLANT SAMPLING

To obtain all aboveground plant material, plant tops should be cut exactly at ground level. The older the plant gets, the thicker its stems become, and thus the more difficult i t i s to cut the stem at ground level. At these stages the cut is therefore often taken several centimeters above ground level, leaving ‘‘the stubble” in the field; this systematic error, although small, will contribute to an artifactual decline of N content in the tops. With increasing plant age the amount of shed material, such as leaves, also increases, and when this material is not included in the sample some nitrogen in the tops is “lost.” For the measurement of small changes in N content, it is preferable to intercept, and collect daily, shed material to avoid losses of this material due to wind, trampling, leaching, and decomposition. B. FRESH SAMPLE STORAGE

The length of time and environmental conditions between collection and drying of sample can affect the dry matter content through respiration and could also conceivably change the N content. Little systematic work seems to have been done on this aspect. Cochrane and Brown (1974) could not find any change in dry matter and crude protein content of freshly cut, immature Phafaris tuberosa when the plant material was stored, for 0.5 or 24 hours after cutting, in open trays in an air-conditioned laboratory at 21°C. Melvin and Simpson (1963) dried freshly cut Lolium sp. at 21°C and 65 or 85% relative humidity for 30 hours. The decreases in dry matter were 0.1, 8.4, and 7.3% when harvested at booting, head emergence, and full flower stage, respectively. The decreases in total nitrogen were only in the order of 0.1% of the total amount of N. In the more mature material, soluble nitrogen, particularly amide and amino compounds, increased after the 30-hour drying period. They presumed that this was due to decomposition of some of the nitrogenous constituents into carbon dioxide and ammonia, the latter then being converted to amide. C. OVEN-DRYING

Samples of plant material are dried to arrest any chemical changes of the material, to assess their dry matter contents, and to facilitate pulverization in order to obtain small, representative, samples for chemical analyses. In order to determine the N content of the material as it was at the time of sampling the drying process should be such that no changes in dry matter or nitrogen percentage occur. If the drying temperature is too low, below 50-60°C,

NITROGEN LOSSES FROM TOPS OF PLANTS

295

respiration losses can be induced (Mayland, 1968; Smith, 1969), whereas drying for a prolonged period at high temperatures (100°C) can cause thermochemical degradation (Mayland, 1968; Sharkey, 1970). Prolonged wilting, resulting in formation of volatile amides, can induce nitrogen losses when such material is dried at 100°C (Johnson, 1978). Yemm and Willis (1956) observed a loss of 90 and 42% of free ammonia and glutamine, respectively, when drying root material at 80°C as compared with direct extraction. In contrast, Novozamsky and Houba (1977) found no difference in “free ammonia” between drying at 45 and 70°C and only marginally higher values when they freeze-dried leaf and stem material. According to Johnson (1978) volatile nitrogen losses due to drying of 3 to 10% have been widely reported. However, the proportion of leaf nitrogen present as volatile compounds is usually small. Changes in nitrogen content due to different methods of drying appear to be small in comparison with changes in dry matter content (Cochrane and Brown, 1974; Mayland, 1968; Sharkey, 1970), although high temperatures appear to give greater nitrogen losses than lower temperatures (Roschach, 1957; Sharkey, 1970). The effects of temperature on dry matter changes have been examined by Mayland (1968) and Smith (1969). Both advocate drying at 100°C for 60 to 90 minutes to arrest respiration. Smith then recommends drying at 70°C. Alternative methods to oven-drying are freeze-drying and exposure to microwave radiation. The disadvantage of freeze-drying is that when the dried material is exposed to ambient conditions rehydration and subsequent reactivation of enzymatic activity can occur (Mayland, 1968). Grinding of freeze-dried material should therefore take place at below 0°C. According to Yemm and Willis (1956) a loss of 50% of the free ammonia is still possible even with freeze-drying. Microwave drying overcomes most of the dry matter and nitrogen loss problems just discussed (Johnson, 1978), presumably because plant material is more equally exposed to the drying effect. In this context it would seem that with oven-drying the desired result can only be obtained if all plant parts in the oven are immediately exposed to air temperatures as advocated by Smith (1969). Thus, very loose packing and a forced draft of air appear to be prerequisites for minimizing dry matter and nitrogen losses. It would appear that after drying at any temperature or with any method various amounts of water remain in the dried plant material. Mayland (1968) exposed plant material that was air-dried, dried at 60, 80, l W C , or freeze-dried, to evacuated P,O, and found a moisture percentage of between 2.4 and 3.9. It is possible, of course, that in some cases, this moisture balances respiration loss (Raymond and Harris, 1954). In general, it is better to minimize respiration losses and to measure the residual water content. Finally, ground plant material should be redried before storage, and should be

296

R. WETSELAAR AND G . D. FARQUHAR

stored in sealed bottles (Smith, 1969) to avoid uptake of ammonia from the atmosphere. D. GRINDING

Immediately after oven-drying, plant material is ground t6 facilitate subsampling and chemical analysis. For field bulk samples, either all material is ground in a big hammer mill, or a subsample is taken and pulverized in a small, highspeed hammer or cross-beater mill. In the former case, the ground material has to travel several meters through encasement which is not always airtight, while at the outlet escape of ground material is possible. In the smaller, laboratory size mills, the inlet is often continuously open for plant material feeding; the collector at the outlet often consists of porous, hessian-type material to withstand air pressures. Thus, in both cases, the powdered material can escape. As the plant material is extremely dry and finely powdered during and after grinding, fractional separation and loss through segregation may occur. The less fibrous material, which is likely to have a higher N content, can escape into the air or precipitate electrostatically onto internal parts of the grinder. The loss due to segregation will be higher with older plants and could therefore result in apparent nitrogen losses. It is difficult to estimate the magnitude of nitrogen losses thus obtained. Let us assume at maturity a dry matter yield without grain of 10,000 kg ha-' (the grain being harvested separately), with 30% having very little fiber, 30% some fiber, and 40% a high fiber content, and with these fractions having N contents of, respectively, 2, 1, and 0.5%. If 20% of the highest nitrogen fraction were ''lost'' during the grinding operation, then the N content of the tops would be underestimated by 12 kg ha-'. However, the degree of underestimation is unlikely to change substantially between anthesis and maturity. E. KJELDAHL NITROGEN

A part of the nitrogen in plant tops can be in the nitrate form. The actual concentration will depend on the balance between nitrate supply from the roots and the nitrate reductase activity (NRA) in the tops. Concentrations of nitrate-N as high as I .89% of the dry weight have been found in 7-week-old wheat plants (Freney and Lipsett, 1965) and in the petioles of young cotton plants (MacKenzie et al., 1963). It is therefore essential that nitrate-N be quantitatively reduced to the ammonium form during Kjeldahl digestion. In many cases, however, no special precautions are taken to ensure that this occurs, and an unpredictable proportion of the nitrate will be reduced and determined (Bremner, 1965). This leads to an underestimate of the actual N content in the tops.

NITROGEN LOSSES FROM TOPS OF PLANTS

297

As plants get older, NRA will in general become greater than the rate of nitrate supply from the root system, unless split applications of nitrogen fertilizers are given. This will result in a more or less asymptotic decrease in nitrate content with time (MacKenzie et al., 1963). Consequently, the underestimate of total nitrogen, due to the absence of a full reduction of nitrate during digestion, will decrease with age. This could mask, partly or fully, declines in N content of tops as plants get older. Details of the Kjeldahl method are not always provided in publications. In the papers reviewed, Lapins and Watson (1970) and Basinski et al. (1971) indicated that nitrate reduction was included in their method; both observed marked declines in N content of tops (Figs. 9 and l l ) . F. ROOT NITROGEN

Whether or not any nitrogen is translocated to roots, or via the roots to the soil, is confounded by the problems of complete root nitrogen recovery. First, when roots are recovered from the soil under realistic (field) conditions, it seems inevitable that at least a portion of the finer roots remains a part of the “soil.” This could lead to a false conclusion that nitrogen had been translocated from the roots to the soil. The second problem is that soil adheres to the roots. It is common practice to wash off this soil, but this may leach nitrogen compounds from the roots. An alternative approach is to remove carefully the roots and shake off excess soil. After oven-drying, the mass of the adhering soil in one subsample is then determined by dry combustion, after correction for soil organic matter. With another subsample the total N content is determined. Correction is made for the N content of the adhering soil, based on that of the bulk soil. However, this correction is problematic as the adhering soil will most likely have come from, or include, a small zone of 1 to 2 mm around the roots. It is to this zone that any exudation from the roots will have gone (Rovira, 1969a). As far as we know a technique has yet to be established to overcome the problems just outlined. G . DISCUSSION

It is important in the present context to determine the time course of the N content in tops of plants with minimal error. To achieve this, each step in the process that assesses this content needs to be as accurate as possible. Within the sample area all aboveground parts of the plant material should be collected, and shed material should be intercepted and collected daily. Thereaf-

298

R. WETSELAAR AND G. D . FARQUHAR

ter, the samples should be oven-dried as soon as possible to avoid respiration losses. Substantial dry matter and nitrogen losses can occur during oven-drying. These can be minimized by loose packing of the material in the oven, and by submitting the material to a force draft of air, at 100°C for the first hour, followed by 70°C until the moisture content does not change. A final assessment of this moisture content is required. Apparent nitrogen losses can be induced during grinding. These can be overcome by having an air-tight system during the grinding operation and by including all ground material that precipitates on the internal parts. When the total N content of the ground material is determined, precautions should be taken to reduce NO,-N quantitatively to NHi-N. For a nitrogen decrease in time to be induced artifactually by errors in the assessment of N content, such errors, individually or collectively, must be systematic and must increase with time in high nitrogen plants. We recognize that this might occur to a small extent at the sample collection stage. Nevertheless, we believe that despite this and despite the array of other problems of methodology discussed previously, declines in N content of tops of annuals (Section 11) are real phenomena. In the context of explaining nitrogen decreases in the tops of plants each species must be examined to determine whether or not substantial amounts of nitrogen are translocated to the roots. This can be achieved by applying compounds labeled with lSN to the leaves and subsequently determining the amount of 15N in the soil root compartment. If little is found in this compartment, the question is answered. If substantial amounts are found, the difficult task of resolving root and soil N contents remains.

+

V. CONCLUSIONS We draw the following conclusions: Significant losses of nitrogen can occur from the tops of annual and perennial species, and these are not due to systematic errors in methodology. For perennial species such losses are at least partially accounted for by accumulation in below-ground storage organs. For annuals these losses are greater from plants with high concentrations of nitrogen. Such losses either occur only after flowering, or occur continuously. When continuous, the losses become apparent as declines in N content when they exceed the rate of uptake through the roots. Careful attention must be paid to the fate of nitrogen in the tops of annuals between anthesis and maturity.

NITROGEN LOSSES FROM TOPS OF PLANTS

299

The relation between plant nitrogen content and the capacity of the inflorescences to absorb nitrogen should be investigated. Net losses from tops may result from the cumulative effects of a large number of small losses, some of which increase with N content or with senescence. Leaching of mobile nitrogenous compounds from tops may be an important mechanism of loss. Loss in gaseous form directly from the tops may be another important mechanism. In the latter case, nitrogen is lost from the soil-plant system and balance sheets are affected accordingly. Failure to recognize this may have led workers in the past to overestimate losses from the soil by denitrification, leaching, and ammonia volatilization. ACKNOWLEDGMENTS We thank Mr. B. Weir and Mrs. J. Hardy for assistance in preparation of the manuscript.

REFERENCES Aldous, A. E. 1930. J. A m . SOC.Agron. 22, 385-392. Allison, F. E. 1955. Adv. Agron. 7, 213-250. Allison, F. E. 1966. Adv. Agron. 18, 219-258. Aneja, V. P. 1977. Ph.D. Thesis. North Carolina State University, Raleigh. Anderson, L. S., and Mansfield. T. A. 1979. Env. Pollution 20, 113-121. Anderson, T. 1866. Trans. Highland Agric. Soc. Scotl. 4 1, 167. Cited by McKee, H. S. 1962. “Nitrogen Metabolism in Plants.” Oxford Univ. Press (Clarendon), London and New York. Austin, R. B., Ford, M. A., Edrich, J. A., and Blackwell, R. D. 1977. J. Agric. Sci. Camb. 88, 159- 167.

Barley, K. P., and Naidu, N. A. 1964. Ausr. J. Exp. Agric. Anim. Husb. 4, 39-48. Bartholomew, W. B. 1972. Inr. Soil Ferr. Eval. Improv. Progr. Tech. Bull. No. 6. Basinski, J. J., Beech, D. F., Evenson, J. P., and Wetselaar, R. 1971. Corron Grow. Rev. 48, 175-193.

Beauford, W., Barber, J., and Baninger, A. R. 1977. Science 195, 571-573. Beevers, L. 1976. “Nitrogen Metabolism in Plants.” Arnold, London. Bird, J. N. 1943. J. Am. SOL-.Agron. 35, 845-861. Boatwright, G . O., and Haas, H. J. 1961. Agron. J . 53, 33-35. Bremner, J. M. 1965. In “Methods of Soil Analysis, Part 2: Chemical and Microbiological Properties” (C. A. Black, ed.). Agronomy No. 9, 1149-1 178. Am. SOC. Agron., Madison, Wisconsin. Burd, J. S. 1919. J . Agric. Res. 18, 51-72. Campbell, C. A,, and Davidson, H. R. 1979. Can. J. Plant Sci. 59, 963-974. Churchill, K., and Klepper, L. A. 1979. Pesric. Biochem. Physiol. 12, 156-162. Cochrane, M. J., and Brown, D. C. 1974. J. Ausr. Insr. Agric. Sci. 40, 67-69. Crasswell, E. T., and Martin, A. E. 1975. Aust. J. Soil. Res. 13, 43-52. Daigger, L. A , , Sander, D. H., and Peterson, G. A. 1976. Agron. J. 68, 815-818. Denmead, 0. T., Simpson, J . R., and Freney, J. R. 1974. Science 185, 609-610.

300

R. WETSELAAR AND G. D. FARQUHAR

Denmead, 0. T., Freney, J. R., and Simpson, J. R. 1976. Soil Biol. Biochem. 8, 161-164. Denmead, 0. T.. Nulsen, R., and Thurtell, G. W. 1978. Soil Sci. SOC. Am. J . 42, 840-842. Dilz, K., and Woldendorp, J. W. 1960. Proc. 8rh Int. Grassl. Congr. pp. 150-152. Doneen, L. D. 1934. Bull. State Coll. Washington Agric. Exp. Sm. 2%. Eggleton, W. G. E. 1935. Biochem. J . 29, 1389-1397. Evans, H. J., and McAuliffe, C. 1956. In “Inorganic Nitrogen Metabolism” (W. D. McElroy and B. Glass, eds.), p. 189. John Hopkins Ress, Baltimore, Maryland. Faller, V. N. 1972. 2. Pflantenernaehr. Bodenkd. 131, 120-130. Farquhar, G. D., Wetselaar, R., and Firth, P. M. 1979. Science 203, 1257-1258. Farquhar, G. D., Firth, P. M., Wetselaar, R., and Weir, B. 1980. Plant Physiol. (in press). Firth, P., Hansa Thitipoca, Sumalee Suthipradit, Wetselaar, R., and Beech, D. F. 1973. Soil Biol. Biochem. 5, 41-46. Frank, H. 1954. Planru 44, 319-340. Franzke, C. J., and Hume, A. N. 1945. J . Am. Soc. Agron. 37, 848-851. Freney, J. R., and Lipsett, J. 1965. Nature (London) 205,616-617. Fry, B. A. 1955. “The Nitrogen Metabolism of Micro-organism.” Methuen, London. Galbally, I. E.,and Roy, L. R. 1978. Nature (London) 275, 734-735. Gasser, J . K. K. 1961. J. Sci. Food Agric. 12, 375-380. Gasser, J. K. K. 1962. J . Sci. Food Agric. 13, 367-375. Goatley, J. L., and Lewis, R. W. 1966. Plani Physiol. 41, 373-375. Grandfield, C. 0. 1930. J . Am. Soc. Agron. 22, 709-713. Hale, M. G.. Foy, C. L., and Shay, F. J . 1971. Adv. Agron. 23, 89-109. Herron, G. M., Grimes, D. W., and Musick, J. T. 1963. Agron. J. 55, 393-396. Huntjens, J. L. M. 1971. Plant Soil 35, 77-94. Hutchinson, G. L., Millington, R. J., and Peters, D. B. 1972. Science 175, 771-772. Johnson, A. D. 1978. I n “Measurement of Grassland Vegetation and Animal Production” (L.’t Mannetje, ed.) Commonw. Bur. Past. Field Crops Bull. 52, 96-102. Keys, A. J., Bird, 1. F., Cornelius, M. J., Lea, P. J., Wallsgrove, R. M., and Miflin, B. J. 1978. Nature (London) 275, 74 1-743. Kingsley, A. F., Clogett, C. O., Klosterman, H. J., and Stoa, T. E. 1957. Agron. J. 49, 37-39. Klein, G., and Steiner, M. 1928. Jahrb. Wiss. Bot. 68, 607-710. Klepper, L. A. 1979a. Amos. Environ. 13, 537-742. Klepper, L. A. 1979b. Amos. Environ. 13, 1475. Lapins, P., and Watson, E. K. 1970. Aust. J . Exp. Agric. Anim. Husb. 10, 599-603. Lausberg, T. 1935. Juhrb. Wiss. Bor. 81, 769-806. Likens, G. E., Bormann, F. H.. Pierce, R. S., Eaton, J. S . , and Johnson, N. M. 1977. “Biogeochemistry of a Forested Ecosystem. ’* Springer-Verlag. Berlin and New York. Long, W. G., Sweet, D. V., and Tukey, N. B. 1956. Quart. Bull. Michigan Agric. Exp. Sm. 38, 528-532.

McKee, H. S. 1962. “Nitrogen Metabolism in Plants.” Oxford Univ. Press (Clarendon), London and New York. MacKenzie, A. J., Spencer, W. F., Stockinger, K. R.. and Krantz, B. A. 1963. Agron. J. 55, 55-59.

Martin, A. E., and Ross, P. J. 1968. PIanr Soil 28, 182-186. Mayland, H. F. 1968. Agron. J . 60, 658-659. Melvin, J. F.,and Simpson, B. 1963. J . Sci. Food Agric. 14, 228-234. Meyer, M. W. 1973. Ph.D. Thesis. University of Maryland, College Park. Miller, E. C. 1939. Tech. Bull. Kansas State CON.Agric. Appl. Sci. 47. Morgan, M. A., Volk, R. J., and Jackson, W. A. 1973. Plant Physiol. 51, 267-272. Mothes, K.,and Engelbrecht, L. 1952. Flora 139, 1-27.

NITROGEN LOSSES FROM TOPS OF PLANTS

30 1

Nikitishev. V. 1. 1974. Agrochimiya No. 12, 7-15. Nikitishev, V. I . , and Krusser, 1. F. 1974. Agrochimiya No. 3, 49-58. Norman, M. J . T. 1963. Aust. J . Exp. Agric. Anim. Husb. 3, 119-124. Norman, M. J . T., and Wetselaar, R. 1960. CSIRO Aust. Div. Land Res. Reg. Surv. Tech. Paper No. 9. Novozamsky, I . , and Houba, V. J. G. 1977. Neth. J . Agric. Sci. 25, 26-31. Oades, J . M. 1978. J . Soil Sci. 29, 1-16. Pearsall, W. H.. and Billirnoria. M. C. 1937. Biochem. J . 31, 1743-1750. Pearsall, W. H., and Billirnoria, M. C. 1939. Ann. Bot. 3, 601-618. Penning De Vries, F. W. T., and Djiteye, M. A. 1980. Production Prirnaire des Paturages Saheliens. Agric. Res. Rep., DOC Wageningen, Netherlands (in press). Penston, N. L. 1935. Nature (London) 136, 268-269. Porter, L. K . 1969. Soil Sci. Soc. Am. Proc. 33, 696-701, Porter, L. K . , Viets, F. G.,Jr., and Hutchinson, G. L. 1972. Science 175, 759-761. Power, J . F., Gnmes, D. L., Reichman, G . A., and Willis. W. 0. 1970. Agron. J . 62, 567-571. Raymond, W. F., and Harris, C. E. 1954. J. Er. Grassl. Soc. 9, I 19-130. Richardson, A. E. V., Trumble, H. C., and Shapter. R. E. 1931. CSIRO Aust. Bull. No. 49. Richardson, A. E. V., Trumble, H.C., and Shapter, R. E. 1932. CSIRO Aust. Bull. No. 66. Rogers, H. H., Campbell, J. C., and Volk, R. J. 1979. Science 206, 333-335. Roschach, H. 1957. Z. Pflanzenernaehr. Dueng. Bodenkd. 77, 37-52. Rovira, A. D. 1965. In “Ecology of Soil-Borne Plant Pathogens-Prelude to Biological Control” (K. F. Baker and W. C. Snyder, eds.), pp. 170-186. Univ. of California Press, Berkeley. Rovira, A. D. 1969a. Bot. Rev. 35, 35-57. Rovira, A. D. 1969b. Aust. J . Biol. Sci. 22, 1285-1290. Rumberg, C. B., and Sneva, F. A. 1970. Agron. J . 62, 31 1-313. Sharkey, M. J . 1970. J . Br. Grassl. Soc. 9, 1 19-130. Silva, P. R. F., da, and Stutte, C. A. 1979a. Arkansas Farm Res. 28, 4 , 3. Silva, P. R. F., da, and Stutte, C. A. 1979b. Proc. Plant Growth Regulator Working Group 6, 35-39. Slyke, D. D., Van. 1914. J . Biol. Chem. 16, 121-124. Smith, D. 1969. Res. Rep. CON.Agric. Life Sci. Univ. Wisconsin 41. Sneva, F . A. 1967. Annu. Rep. Range Res. Squaw Butte Exp. Stn. Burns, Oregon. Pp. 40-41. Sneva, F. A., and Hyder, D. N. 1963. Oregon State Univ. Agric. Exp. Stn. Bull. No. 592. Sneva, F. A , , Hyder, D. N., and Cooper, C. S. 1958. Agron. J . 50, 40-44. Steiner, M., and bffler, H. 1931. Jahrb. Wiss. Bot. 71, 463-532. Stonier, R. R. 1962. Aust. J . Exp. Agric. Anim. Husb. 2, 185-192. Stutte, C. A., and Weiland. R. T. 1978. Crop. Sci. 18, 887-889. Stutte, C. A,, Weiland, R. T., and Blem, A. R. 1979. Agron. J . 71, 95-97. Tanaka, A., and Navasero, S. A. 1964. Soil Sci. Plant Nutr. 10, 36-39. Terman, G. L . , and Allen, S. E. 1974. Proc. Soil Sci. Soc. Am. 38, 322-326. Tukey, H. B., and Tukey, H. B., Jr. 1959. Am. Soc. Hort. Sci. 74, 671-676. Vanecko, S., and Vamer, J. E. 1955. Plant Physiol. 30, 388-390. Viets, F. G . , Jr. 1965. In “Soil Nitrogen” (W. V. Bartholomew and F. E. Clark, eds.). Agronomy No. 10, 503-549. Am. SOC. Agron., Madison, Wisconsin. Watson, R., and Petrie, A . H. K. 1940. Aust. J . Exp. Biol. Med. Sci. 18, 313-340. Weiland, R. T., and Stutte, C. A. 1978a. Arkansas Farm Res. 27 (2), 16. Weiland, R. T., and Stutte, C. A. 1978b. Proc. Plant Growth Regulator Working Group 5, 78-85. Weiland, R. T., and Stutte, C. A. 1979a. Proc. Plant Growth Regulator Working Group 6, 17-23. Weiland, R. T . , and Stutte, C. A. 1979b. Crop Sci. 19, 545-547. Weiland, R. T.. and Stutte, C. A. 1980. Plant Physiol. 65, 403-406.

302

R. WETSELAAR AND G. D.FARQUHAR

Weiland, R. T., Stutte. C. A., and Talbert, R. E. 1979. Weed Sci. 27, 545-548. Weinmann, H. 1940. Plant Physiol. 15, 467-484. Wetselaar, R. 1972.Aust. Rice Res. Conf. 1972, Leeton N.S.W. Australia, pp. 2(c)l-2(c)9. Wetselaar, R. 1974. Proc. Conf. Thai-Aust. Chao Phya Res. Proj. Chainat: 1966-1975 pp. 91-1 15, Australian Development Assistance Agency, Canberra. Wilfarth, H..Romer, H., and Wimmer, G. 1905. “On the Assimilation of the Elements of Nutrition by Plants during Different Periods of Their Growth.” Vinton, London. Williams, R. F. 1955.Annu. Rev. Plant Physiol. pp. 25-42. Wilson, A. G. L., Basinski, J. J., and Thomson, N. J. 1972. Cotton Grow. Rev. 49, 308-340. Woldendorp. J. W. 1962.Plant Soil 17, 267-270. Wong, S. C., Cowan, I. R., and Farquhar, G. D. 1979. Nature (London) 282, 424-426. Woo, K. C.,Berry, J. A,, and Turner, G. L. 1978. Carnegie Inst. Yearb. 77, 240-245. Yemm, E. W. 1937. Proc. R. Soc. B 123, 243-273. Yemm, E. W., and Willis, A. J. 1956.New Phyrol. 55, 229-252.

ADVANCES IN AGRONOMY. VOL. 33

AGROTECHNOLOGY TRANSFER IN THE TROPICS BASED ON SOIL TAXONOMY F. H. Beinroth,* G. UeharaT J. A. Silva? R. W. Arnold,$ and F. B. Cadys ‘Department of Agronomy and Soils, College of Agricultural Sciences, University of Puerto Rico, Mayaguez, Puerto Rico tDepartment of Agronomy and Soil Science, College of Tropical Agriculture and Human Resources, University of Hawaii, Honolulu, Hawaii $Department of Agronomy and BBiometrics Unit, Department of Plant Breeding and Biometry, New York State College of Agriculture and Life Sciences, Cornell University, Ithaca, New York

I. Introduction A. General . . . . . . . . B. Role of Soil Surve .................................... C. Kind and Availability of Transferable Information .......................... 111. Soil Classification in Perspective . . . . . . . . . . . . . . . . . ...................... A. Some Needs for Soil Classification . . . . B. Development of Criteria for Soil Behavior C. Soil Taxonomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Taxonomy and Agrotechnology Transfer IV. Agrotechnology Transference Research .... A. The Benchmark Soils Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Developing Transferable Agrotechnology ................................. V. Quantitative Verification of Transferability within a Soil Family . . . . . . . . . . . . . . . . . B. The Transfer Model.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Application of Transfer Verification Methodology

306 307 309

312

316 322 323 324

A. Internationalization of Soil Taxonomy C. Land Evaluation . . . . . . . . . . . 334 D. Assembling Crop Performance Data.. .................................... 335 E. Information Systems and Data Banks

. . . . . . . . 338

L

J 303 Copyright @ 1980 by Academic R s s . Inc. All rights of reproduction in any form reserved. ISBN 012-000733-9

304

F. H.BEINROTH ET AL.

1. INTRODUCTION Agricultural production capacity in the less developed countries (LDCs) of the tropics must be expanded by increases both in the yields per hectare and in the area of cultivated land. This growth in resource productivity, however, should be guided by wise land-use decisions and based on sound methods of soil and crop management. Since most LDCs lack the trained manpower, the capital, and the institutional capacity to conduct all the research required to fill their needs in the short time available, it stands to reason that they should capitalize on experience gained elsewhere under conditions of similar ecology and factor endowment. The groundwork required to transform agriculture in the tropics from a natural resource-based enterprise to an industry founded on science and technology is now gradually being generated through research and development in many agencies and areas. Although there has been a lag in the large-scale dissemination and systematic application of this knowledge, transfers of applicable agrotechnology are of such critical importance to the agricultural and economic development of agrarian LDCs that they must now be facilitated. In the context of this perspective, this paper outlines the rationale, prospects, and limitations of an approach to the transfer of agroproduction technology that is based on soil classification. Although the scope of the paper is broad, reference is made to a research program of the Universities of Hawaii and Puerto Rico, the Benchmark Soils Project, to exemplify and substantiate some of the general statements with factual data.

II. THE TRANSFER OF AGROTECHNOLOGY A. GENERAL

International transfer of agricultural technology is not new. Diffusion of husbandry practices and information about crop varieties and livestock breeds was a major source of productivity growth even in prehistoric times (Sauer, 1969). It is also well known that, after the discovery of America, the introduction of new cultivars from the Americas to Europe had a profound impact on European agriculture. This “natural diffusion” in preindustrial time was largely a byproduct of nonagricultural human endeavors, characterized by long periods of gestation and delivery. The modem institutionalization of agricultural research and extension significantly expedited the process of dissemination, as evidenced by the recent promulgation of high-yielding varieties of rice and wheat in many tropical and subtropical countries. This apparent success, however, cannot obviate the extreme complexity of the transfer of agrotechnology. Only in recent years has this complexity become reasonably well understood (Wortman, 1976).

AGROTECHNOLOGY TRANSFER IN THE TROPICS

305

There are fundamental differences in the conditions that must rule the transfer of technology to the industrial and the agricultural sectors in developing economies. In the industrial sector, new industries can be established with minimal disturbance of the institutional framework of a country; in contrast, technology transfer in the agricultural sector must confront the attitudes and institutions that have so long held these rural societies stagnant (Myrdal, 1974). Also, whereas the transfer of industrial technology from developed to less-developed countries can be quite direct, the transfer of agricultural technology, particularly the biological component, is generally not possible. There is now a growing consensus that much agricultural technology is location specific and a function of the physical, biological, and socioeconomic conditions of the environment where it has been developed. The corollary of this statement is that, in order to be productive, “much of the agricultural research must be conducted and the results analyzed, tested, interpreted and applied within a relatively decentralized system” (Hayami and Ruttan, 1971). Thus, the technology for tropical agriculture must be developed primarily in the tropics. Inadequate appreciation of the location specificity of agricultural technology was a major reason for the lack of effectiveness of much of the technical assistance effort by national and international agencies in the 1950s and 1960s (Hayami and Ruttan, 1971). The “diffusion model” that, explicitly or implicitly, was the underlying rationale for technical assistance after World War I1 led to an “extension bias,” which, as Moseman (1970) pointed out, “met with only limited success because of the paucity of applicable indigenous technology and the general unsuitability of U.S. temperate zone materials and practices to tropical agriculture conditions. ” In a similar vein, Kamarck (1 972) observed that the tropics are littered with ruins of agricultural projects that refused to recognize the special conditions prevailing there. The Groundnut Scheme in East Africa (Wood, 1950) is probably the best-known case history among these failures. With only few exceptions, recent writings about economic development problems have largely ignored the possible influence of the environment on economic development, and, reflecting the economists’ temptation to find universal laws, the purely mathematical growth models make no provision for environmental variables (Kamarck, 1972). Boulding (1970) touched on this point, noting that the principal failure of economics has been in the field of economic development and wondering “whether culture-boundness may not have something to do with this relative failure. In this decade, a new perspective has emerged, tying the relative contribution of agricultural and industrial development to national economic growth. There has been a marked shift away from an earlier “industrial fundamentalism” to an emphasis on the significance of growth in the agricultural sector for the total development process (Hayami and Ruttan, 1971). It is now held that agricultural development in the LDCs should serve the dual purpose of securing increases in ”

306

F. H. BEINROTH ET AL.

both yields and employment. This should be achieved through agricultural technology that engages the latest results of scientific research adjusted to the highest possible utilization of the labor force (Myrdal, 1974). Agriculture is thus viewed not only as a necessary condition for meeting food requirements but also as a productive source of economic growth, and it is considered the best route to economic advancement for the agrarian LDCs (Hayami and Ruttan, 1971; Wortman, 1976). Rapid improvement of agricultural productivity in LDCs, consistent with required growth rates in the range of 3 to 6%, implies a transition from a natural resource-based agriculture to a science-based agriculture. To bring about this transition within a short period of time requires massive inputs of capital and research personnel. Many of the LDCs are small in size and population, however, and in view of their limited experiment-stationcapacity, an inelastic supply of scientific and technical manpower, and a general lack of capital, such nations cannot expect to generate by themselves the full range of scientific knowledge and expertise needed to develop and sustain a viable agriculture. They rely on assistance from external sources. The transfer of agroproduction technology, therefore, is critically important to the agricultural, and hence the overall, economic development in agrarian LDCs. B. ROLEOF SOILSURVEY A N D CLASSIFICATION

The site specificity of agricultural research and technology, mentioned in the preceding paragraph, results in large measure from differences in two environmental variables, soil and climate. In the tropics, as in the temperate zone, there is wide variability in both these factors; however, systematic groupings can be made to stratify the population of soils and climates. The seemingly infinite diversity can thus be reduced to a finite number of discrete entities with a limited range in characteristics that conform more nearly to agricultural management units and define the applicability of agronomic research. Because soils and climate are interrelated, both can be combined in one system of classification, as has been done in Soil Taxonomy, the U.S.system of soil classification described in Section 111. Thus, Soil Taxonomy, as the basis for soil interpretation, has the best potential for identifying agricultural land and consequently for the transfer of agrotechnology. Soil classification provides pragmatic groupings of soils for precise predictions about soil behavior, and its most important application is in soil survey. Historically, soil surveys have been geared to the improvement of agriculture, and most LDC governments presently support soil survey activities because they believe soil surveys supply reliable and accurate information for agricultural development and soil-resource management. Soil surveys, however, are only useful if

AGROTECHNOLOGY TRANSFER IN THE TROPICS

307

they follow reasonable scientific standards and if they are interpreted for practical purposes. Such interpretationsare predictions of soil behavior under stated conditions, which require the careful synthesis of many data in relation to soil qualities that are the result of the interaction between soil characteristics, crop requirements, and management practices (Kellogg, 1961). In fact, soil survey interpretation makes possible the most intensive use of soil science by integrating knowledge from many other disciplines (Smith, 1965). It is an implicit rationale of soil survey interpretation that soils classified in the same taxa have a common response to management practices. We make the basic assumption that experience with a particular kind of soil in one place can be applied to that particular kind of soil wherever it exists if consideration is taken of any climatic differences. The soil survey acts as a bridge that lets us transfer the knowledge gained by research or by the experience of cultivators from one place to all other places where it is applicable (Smith, I 965).

There are inherent constraints to this approach to knowledge transfers, however, relating to the nature of taxonomic soil classification, certain inadequacies in the classification of tropical and subtropical soils, and the restricted availability of soil-specific agronomic information. Moreover, transfers of agrotechnology based on soil classification cannot completely circumvent the need for on-site experimentation. Such transfers, however, can help avoid duplication of effort, maximize the utilization of knowledge generated elsewhere, and give direction to the kind of adaptive research required to adjust soil and agronomic experience to particular local conditions. Within these qualifications, soil survey and classification provide an effective basis for the scientific transfer of agrotechnology. C. KINDA N D AVAILABILITY OF TRANSFERABLE INFORMATION

Transfers of both the mechanical and biological components of agricultural technology can occur through (1) material transfers, (2) design and knowledge transfers, and (3) capacity transfers (Hayami and Ruttan, 1971). This paper is primarily concerned with knowledge transfers, which should be conducive to the production of locally adaptable technology after a prototype technology has been imported from elsewhere. More specifically, knowledge transference based on soil classification principally involves soil management and crop production practices, such as fertilization, tillage, irrigation, crop species selection, and so on. Although it must be noted that, in most cases, only principles and methodologies can be transferred, it should be emphasized that reasonably accurate estimates of the levels of management inputs and crop performance can be made. Examples of management principles include stratagems for phosphorus (P) fertilization in soils with a

308

F. H.BEINROTH ET AL.

high P-fixation capacity, methods of irrigation as they relate to crop phenology , and correction of subsoil acidity. This kind of transferable information is derived from empirical experimentation and induction from the knowledge of the interactions involved among soil characteristics, climatic conditions, crop requirements, and management practices (Kellogg, 1961). In this context, soil classification mainly facilitates “horizontal transfers”; that is, the transfer of experience from experiment station to experiment station at a scientific level. Since this information is not necessarily directly applicable to farm situations at new locations, it must be adapted to conform to the economic decision environment of the local farmer. The process of adaptation and extension to the farm level may be termed “vertical transfer,” and it is a critical element in the international transfer of agrotechnology (Hayami and Ruttan, 1971). An important question of consequence to agrotechnology transfer is whether there exists sufficient soil and agronomic knowledge in the tropics that can be transferred. At first glance, the answer to this question would seem affirmative. Many countries in the tropics have an impressive history of agricultural research, and a sizable body of relevant knowledge has accumulated over the years. The extent of this research is well documented in various publications, and special reference is made here to “A Review of Soils Research in Tropical Latin America” (Sanchez, 1972), ‘Soils of the Humid Tropics” (National Research Council, 1972), “Soil Management in Tropical America” (Bornemisza and Alvarado, 1975), a three-volume “Bibliography of Soils of the Tropics” (Orvedal, 1975, 1977, I978), and the annual reports of the International Agricultural Research Centers. These and other publications indicate that the aggregate of experience with tropical soils is significant. An analysis of the literature, however, shows that the knowledge is distributed unequally among tropical countries, often lacking where it is most needed, and that the experience is also unequally distributed by kind of soil, especially since many experiment stations are located on atypical soils. Consequently, the as yet underutilized land resources remote from population centers in the tropics have been largely neglected. Moreover, in many instances, the soils on which the agronomic research was conducted are not adequately identified and classified, which severely impedes the extrapolation of research results to other areas. Thus, it must be stressed emphatically that the classification of the soils in an internationally recognized system, in conjunction with soil survey, is a conditio sine qua non for the effective transference of the results of agronomic research. Knowledge gaps are now being partially filled through the work of the 11 International Agricultural Research Centers in Africa, Asia, and Latin America, six of which were established since 1970. In addition, bilateral agricultural assistance programs are underway in many LDCs, and several national gov-

AGROTECHNOLOGY TRANSFER IN THE TROPICS

309

ernments are greatly intensifying their own research efforts. Further, there now exists an operational network of financial institutions, including the World Bank, the Inter-American Development Bank, the Asian Development Bank, the African Development Bank, and various Common Market Banks. In recent years, these institutions have markedly increased their emphasis on agriculture and rural development (Wortman, 1976). For the first time, then, there now exists an institutional and technical infrastructure to fund and conduct the research needed to generate the scientific knowledge base for a productive tropical agriculture. But the results of this research must also be disseminated; experiment station and extension service capacity in many LDCs, however, remains critically limiting. The establishment of such institutions in these countries is of crucial importance to the successful transfer, adaptation, and application of the knowledge and technology developed elsewhere.

Ill. SOIL CLASSIFICATION IN PERSPECTIVE A. SOME NEEDSFOR SOILCLASSIFICATION

The soil mantle forms a continuum over the earth and is very difficult to study as a whole. Through the systematic subdivision of this whole, however, it is possible to describe many soils and predict their potential and response to management inputs. In a classification scheme, soil bodies are recognized by their properties and identified by names of classes; these soil names, then, provide identity to otherwise unidentified land areas and facilitate communication about features of soils and predictions about their behavior when managed or used in particular ways. Thus, soil classification is necessary for proper use of experience and to extend the results of research. Each soil is a unique combination of external and internal characteristics, with a defined range of expression that can be observed in the field and studied in the laboratory. These characteristics both describe the history of a soil and predict its potentials. The Soil Survey Manual (Soil Survey Staff, 1951) clearly points out that ‘‘the influences on soil behavior of any one characteristic, or of a variation in any one, depends upon the others in the combination.” Soil classification draws on the experience and results from all areas of fundamental and applied soil science, and the products of these areas can only be synthesized for accurate application through soil classification. Experimental field plot techniques deal with soils at small places. Through comparison of sets of data from different places, principles can be developed that fit the facts. Useful results from geographically separated experimental plots occur only when the plots themselves are representative samples of defined kinds of soil. For

3 10

F. H. BEINROTH ET AL.

interpretation, whether for an understanding of soil or for predictions about its behavior, results must be related to defined soil units. This is a major function of soil classification (Soil Survey Staff, 1951). A taxonomic classification of soils is a tool to help us organize our knowledge about soils and to remember that knowledge. Classification also enables us to see relationships among and between soils and their environment and to formulate principles of prediction value (Soil Survey Staff, 1951). The Soil Survey Manual of 1951 summarized many concepts and operational aspects of the Cooperative Soil Survey of the United States. In addition to describing procedures for conducting and publishing soil surveys, the book contains discussions of many concepts of applied soil science. The writers emphasized the importance of accurately predicting soil productivity for making economically valid agricultural and nonagricultural recommendations. Soil surveys require soil classification, of course, but other disciplines in soil science also require soil classification for appropriate application of results. In addition to classifying soils and plotting their boundaries on a map, a soil survey is able to correlate and predict the suitability of soils for various plants and the expected behavior and productivity of soils under different management systems. Most of these types of interpretations over the years have been based on reasonable estimates of similarity (correlation) rather than on specific experimental research on every defined soil. Because each experimental plot is a sample of a landscape, it should be an accurate and representative sample of a defined kind of soil worthy of being sampled and studied. The key to success is having defined kinds of soils-thus the urgent need for a classification that is relevant. In any soil classification system, a soil taxon provides a mental image of a collection of soils that are considered similar enough to be treated as one thing for a particular purpose (Cline, 1977). Consequently, the most important function of a taxon is to provide identity and class membership as determined by defined sets of physical, chemical, and biological characteristics. Every system attempts to provide the largest feasible number and most precise statements for its objective. A classification can best serve only one major objective. Conflicts can easily arise, therefore, in assessing the usefulness of a classification. An apparent conflict exists between an objective for agricultural land use and management on the one hand, and similarity of soils in terms of morphology and genesis on the other. The most widely accepted concepts and models of soils are based on the development of soil properties. Their relationships to observable landscape features form the basis for soil mapping, and, because of this reliance, systems of classification emphasizing soil morphology have been used by most soil survey organizations.

AGROTECHNOLOGY TRANSFER IN THE TROPICS

31 1

B. DEVELOPMENT OF CRITERIA FOR SOILBEHAVIOR

By 1950, the difficulties in consistently defining and applying the “Great Group/Series” classification system in the United States were considered to be insurmountable, and it was decided to develop a new, more refined, and more practical system of soil classification. Seven approximations over 25 years were produced by the Soil Survey Staff of the Soil Conservation Service of the U.S. Department of Agriculture, in cooperation with pedologists from the United States and abroad. The culmination was the publication of Soil Taxonomy (Soil Survey Staff, 1975), which is already generally regarded as a means to one worldwide system of soil classification. An important aspect of the development of this system was the need for one or more categories that would group soils by properties relevant for use and management. Categories called “high families” and “low families” were proposed for testing. High families dealt with central concepts and intergrades, and low families used properties thought to be significant for use and management for agriculture. The criteria suggested by Kellogg in 1951 (Cline, 1979) to group series into low families were drainage class, texture and coarse fragments, degree of development relative to modal profile, degree of weathering, and size (thickness) of profile relative to modal profile. The testing of high and low families by placing existing soil series into these categories revealed the lack of precise definition of criteria and also of the series themselves (Cline, 1979). Guy Smith, the principal author of Soil Taxonomy, insisted on classifying cultivated soils to assist in interchanging results of experience throughout the world, in contradiction to the then current emphasis on classifying “virgin” or relatively undisturbed soils. He also noted that if the Soil Survey Staff were to succeed in making each low family a group about which fairly specific management recommendations for given phases could be made, then soil temperature and soil moisture (or substitutes) must be in the system (Cline, 1979). In the third approximation of the new system, the criteria tested for grouping series into low families were texture below the A horizon, permeability below the A horizon, soil drainage class, moist and dry consistency below the A horizon, and soil depth. In the fifth approximation, it was hoped that classes of low families would be as nearly uniform as possible in those properties that affect development of roots and the movement and retention of soil moisture. At that time, there were nine classes of mineral soils in the highest category. For seven of those groups, the soil family criteria included available water capacity in the 6- to 60-inch depth, and also in the 60- to 240-inch depth if relevant; permeability to at least 60 inches; consistency, both moist and dry; and presence of toxic substances. In the other two groups, clay mineralogy and base saturation were also used.

312

F. H.BEINROTH ET AL.

Definitions of the criteria and their limits were tested again and again by placing series into the proposed classes. The sixth approximation established the criteria for the category called “family,” and the seventh approximation set the trend for family definitions by specifically defining classes in a systematic manner (Cline, 1979). The evolution of interpretive groupings was not haphazard; rather, it was the product of repeated testing. Data from the United States and other parts of the world (1) provided an expansion of genetic theory and a means of testing the model against the real world, (2) governed the choice of class limits, and (3) served to give an applied bias to the lowest two categories (family and series) by evaluating existing interpretations for those subdivisions. C. SOILTAXONOMY

I . Categories Soil Taxonomy (Soil Survey Staff, 1975) was developed on the premise that a classification which presents a model of genetic relationships among soils is the most effective system for understanding soils as natural bodies. The concept of polypedons was created to insure that the soil bodies we classify are the same bodies we attempt to map. The definitions of the higher categories are more abstract than those of the lower categories (family and series), yet the features used to satisfy the definitions of all categories are soil properties. The bases for differentiating categories are related to models of genesis (Table I). The recognition of classes within a category vary according to the combination of soil properties thought to best reflect, and be consistent with, the definitions of that category. The order category consists of 10 classes of soils whose features differ according to the degree and kind of dominant sets of soil-forming processes that have existed. Within the order classes, suborders are distinguished by soil properties that reflect the major control of current processes; climate, parent material, and biological activity are examples of such control. Within each suborder, great groups are defined by soil properties that provide additional influence on current processes not identified in the higher categories. Great groups are divided into subgroups whose properties represent departures from a central concept of the great group. These departures are usually due to intergradation of processes, but some are extragrades, with properties not related to other genetic pedons. Within each subgroup, the family classes contain soils having similar physical and chemical properties that affect their responses to management and manipulation for use. The soils in families can be further classed into series whose restricted ranges of properties provide further homogeneity of morphology and composition.

313

AGROTECHNOLOGY TRANSFER IN THE TROPICS

Table I Example of Relationshipsamong Category Subdivisions in Soil Taxonomy Category name Order Suborder Great group

Subgroup

Family

Series

Basis for differentiation

Example of class name

Dominant soil process that developed soil Major control of c m n t process Additional control of current process

Ultisol

Blending of processes (intergrades) or extragrades Internal features that influence soil-waterair relations Nature of materials that affect homogeneity of composition and morphology

Aquic Tropudult

Udult Tropudult

Fine loamy, mixed isothemic Aquic Tropudult Cerrada

Main features of the class Clay accumulation; depletion of bases Soil moist most of time; humid (udic) climate Fairly constant soil temperature all year; tropical environment Temporary wetness in rooting zone Texture and mineralogy in a control section, and soil temperature Soil forming in weathering diabase

2 . Significance of the Soil Family Category As is true with all multicategorical systems, the properties associated with classes accumulate from the higher and more abstract categories down to the lower categories. Thus, many more statements can be made about soils in a family than about the soils in a suborder. In addition, in contrast to the many genesis-related properties used at higher levels, the soil family category in Soil Taxonomy is based on properties without regard to their significance as marks of soil processes or the lack of them (Soil Survey Staff, 1975). The responses to management of comparable phases of all soils in a family are thought to be nearly enough the same to meet most needs for practical interpretations of such responses. For this reason, families are defined primarily to provide groupings with restricted ranges in particle-size distribution in horizons of major biological activity below plow depth, mineralogy of these same horizons, temperature regime, thickness of soil penetrable by roots, and a few other properties to provide additional homogeneity in some families, as shown in Table I1 for one specific subgroup. Eleven particle-size classes are commonly used; however, 40 additional contrasting particle-size combinations are recognized to identify changes in pore-size distributions that seriously affect movement and retention of water. Seventeen classes of mineralogy are recognized to assist in evaluating the presence of toxic

314

F. H.BEINROTH ET AL. Table I1 Currently Recognized FamiIies in the United States of a Selected Subgroup, Typic Paleudults" Family characteristics Particle-size classb

Mineralogyb

Soil temperaturec

Loamy-skeletal Loamy-skeletal Loamy-skeletal over clayey Clayey-skeletal Coarse-loamy Coarse-loamy Coarse-loamy Fine-loamy Fine-loamy Fine-silty Clayey Clayey Clayey Clayey Clayey Clayey

Siliceous Siliceous Siliceous Mixed Si1iceous Siliceous Siliceous Siliceous Siliceous Siliceous Kaolinitic Kaolinitic Mixed Mixed Mixed Oxidic

Mesic Thermic Mesic Mesic Hyperthermic Mesic Thermic Mesic Thermic Thermic Mesic Thermic Isohyperthermic Mesic Thermic Isohyperthermic

" Paleudults have thick zones of clay accumulation and only modest reserves of weatherable minerals. Typic subgroups are well-drained soils. bThe control section for both particle size and mineralogy is the upper 50 cm of the zone of clay accumulation (argillic horizon). Soil temperature is measured at a depth of 50 cm from the soil surface. materials and degrees of weathering that influence chemical or physical behavior. Eight classes of soil temperature are recognized in the family category, providing such information if not stated or implied at higher categorical levels for a particular soil. Use of family taxa rather than series taxa for identifying polypedons can affect the precision of interpretations, but if the nature of the soil or the state of knowledge is such that precise interpretation for a wide range of uses is impossible or not needed, then little is lost by naming soils as phases of families rather than as phases of series.

3 . The Future of Soil Taxonomy It seems obvious that when a system employs available information and cumulative experience, a classification can be devised that organizes and arranges the data in a desired manner. Because Soil Taxonomy was based on and evolved from knowledge and experience with soils in the United States, the placement of those soils and predictions of their behavior have been very accept-

AGROTECHNOLOGY TRANSFER IN THE TROPICS

315

able. Class limits were tested and modified by examination of the data base available in the United States at the time, and the physical and chemical properties and estimated crop yields under several levels of management were repeatedly considered. A global soil-data base and worldwide experiences were not extensively tested for Soil Taxonomy, however, partly because the main emphasis was the need to classify soils in the United States and partly because of the lack of familiarity with the existing world data base. In the United States, there are only a few areas on which to develop criteria for tropical soils. As scientists in the tropics and subtropics began to evaluate Soil Taxonomy by attempting to fit their soil information into the system, it became apparent that definitions and class limits similar to those used for soils of the United States did not provide satisfactory groupings for the broad range of soils and conditions in the tropics. Soil-moisture regimes and some ranges of soil temperature, for instance, may need refinement and modification to provide useful separations or groupings. The importance of translocated clay in understanding the slightly weathered soils of the temperate regions tends to be far less significant in the highly weathered soils of the tropics. In addition, different methods for determining available water, cationexchange capacity, and extractable ions often reflect past experience and training and vary in standards across the world. Moreover, lack of correlation among laboratory values makes consensus about criteria difficult, even when there is a desire for standardization and uniformity. Thus, what may appear to some observers to be a deficient classification system is merely a reflection of the stage of its development. As the data base and experiences throughout the world, particularly in the tropics, become available and can easily be shared, compared, and tested, the same kinds of changes associated with the early development of Soil Taxonomy will occur frequently until there is available a more complete and precise data base on soils of the lower latitudes. D. TAXONOMY A N D AGROTECHNOLOGY TRANSFER

The concept of agrotechnology transfer involves at least two aspects if the transfer is to be successful. The first is research data about soil behavior or response when used for a particular crop or farming system with a specified set of management practices. The work must be done on a representative sample of a defined soil, otherwise the places where transfer could possibly occur are unknown. These predictions depend upon conscientiously designed experiments conducted by competent researchers. The second deals with the economic, social, and political conditions against which to evaluate the experimental results in order to provide recommendations. Predictions of soil-crop responses are transferable to the extent that site- or

316

F. H . BEJNROTH ET AL.

location-specific influences are quantified and understood. Soils record a long and complex history, and it is entirely possible that two soils may have followed two different pathways but now may have the same set of properties used to identify taxa. Soils may also exhibit features of agronomic significance that are not used to identify soil taxa. For instance, general predictions can be made about a soil’s behavior related to its internal composition, but crop repsonse may vary in any year because of the marked influence of locational factors, such as solar radiation, precipitation, and pests that respond to the uniqueness of a particular season at a particular site. These latter items usually are not transferable and must be measured and evaluated throughout a decentralized system if effective agrotechnology transfer is to take place. Thus, the shape of response curves at different sites may be reasonably predicted (see Section V), but the magnitude of the response may depend on conditions that occur at a given location during a given growing season. Tillage problems, erosion hazards, commonly adapted crops, potential hazards from naturally occurring toxic compounds, and general nutrient supply can often be estimated for soils classified in higher level taxa. If crop responses to management inputs on particular kinds of soils can be predicted for lower level taxa, then it is clear that many other types of information can also be transferred among soil taxa. Agronomic experiments are usually replicated by treatment but seldom by kind of soil because of an assumed similarity of predicted response for similar soils. Results from experiments conducted on the same phases of the same kinds of soils can be evaluated for their correlation with common soil properties. If these experiments are conducted on the same kinds of soils but at different locations, then site-specific conditions can also be evaluated and predictions of crop responses can be improved. Agronomic interpretations vary in their specificity, just as taxa within Soil Taxonomy differ in their specificity. As additional precision and accuracy are desired, similarity of all inputs at each site must be increased, otherwise the predictions cannot be improved. They will be limited by the degrees of similarity of inputs. A crucial input is proper identification and classification of the soil in which the cropping experiments are conducted.

IV. AGROTECHNOLOGY TRANSFERENCE RESEARCH A. THEBENCHMARK SOILSPROJECT

I . Rationale As pointed out in Section 11, it has long been assumed that an adequate system of soil classification provides a suitable vehicle for the extrapolation of soil-

AGROTECHNOLOGY TRANSFER IN THE TROPICS

317

related experience. Nonetheless, this basic assumption has never been subjected to the scrutiny of field experimentation, neither in temperate nor tropical areas. In an attempt to determine scientifically the transferability of agroproduction technology in the tropics on the basis of soil taxonomic units, the Universities of Hawaii and Puerto Rico have established companion projects under contracts with the U.S.Agency for International Development. The parallel projects were implemented in 1974 and 1975, respectively, and are commonly referred to as the Benchmark Soils Project. The term “benchmark soils” was drawn from and conforms to the concept originally advanced by Kellogg (1961). The project is briefly described here to present an example of agrotechnology transference research. The primary objective of the Benchmark Soils Project is to demonstrate that and how soil management and crop production knowledge can be transferred among tropical and subtropical countries, using soil family taxa as a point of reference. The central tenet of the project is that the soil family, as defined in Soil Taxonomy (Soil Survey Staff, 1975) and discussed in Section 111, affords a sound basis for agrotechnology transfers. This hypothesis derives substance from the fact that soil families are narrowly defined taxa that have been contrived for the explicit purpose of providing groupings of soils that are reasonably homogeneous in characteristics important to the growth of plants. Soils belonging to the same family, therefore, should have essentially the same management requirements, analogous responses to soil manipulation, and similar potentials for crop production. The soil family thus stratifies the population of soils into pragmatic units that, through the differentiae which define them, have the inherent attribute to facilitate agrotechnology transfers. Knowledge transfers can be made at any categoric level of taxonomic system, with increasingly precise statements possible at lower levels. Since predictions and transfers of crop performance and soil management could be refined if they were based on information more specific than that contained in soil family taxa, an argument could be made in favor of using the lowest category of Soil Taxonomy, the soil series, as the basis for agrotechnology transfers. In addition, one could also use phases to indicate soil characteristics that are not considered in taxa definitions but are important to a specific land use. It is well to remember in this context that the differentiae used for series are mostly the same as those used for the classes in other categories to which the series belongs, but the range permitted is less than is permitted in the family or some other higher category. Taken collectively, however, the number of possible distinctions is too large to be comprehended readily and to be incorporated in a key (Soil Survey Staff, 1975). Phases, on the other hand, provide for a utilitarian classification that can be superimposed on the taxonomy at any categoric level to permit more precise interpretations and predictions. The purpose of phases, like that of soil series, is mainly practical, and no strict rules for their usage have been established.

318

F. H. BEMROTH ET AL.

Thus, although the most precise predictions can be made for phases of soil series, it would clearly be unrealistic to use this category in the process of international agrotechnology transfers. First, such detailed information rarely exists for LDCs; second, soil series and phases are not rigidly defined, and different rationales for establishing them are employed in different countries. As a consequence, the lowest categoric level of Soil Taxonomy that can be universally applied in a uniform and consistent manner is that of the soil family.

2 . Experimental Methodology In order to verify the transferability of agrotechnology, a test crop must be grown under similar management conditions on the same soil family in a network of sites. Because members of a soil family are not identical, but similar within a range of characteristics, the outcomes should be expected to vary within a similar narrow range. Three soil families from upland areas of the tropics that have the potential to be productive under appropriate management practices were selected by the Benchmark Soils Project to test the transfer hypothesis. The first family selected was the thixotropic, isothermic family of Hydric Dystrandepts, which is derived from volcanic ash, and sites have been established in Indonesia, Hawaii, and the Philippines. The clayey, kaolinitic, isohyperthermic family of Tropeptic Eutrustox, which links the Puerto Rico and Hawaii projects, was the second family selected, and sites have been established in Brazil, Hawaii, and Puerto Rico. The clayey, kaolinitic, isohyperthermic family of Typic Paleudults was included at the request of the participating cooperating countries because it is an important, underutilized soil in the tropics, and sites have been established in Cameroon, Indonesia, and the Philippines. The locations of the soil networks are shown in Fig. 1. The soils are described in Technical Report 1 and Technical Report 5 (Ikawa, 1979; Beinroth, 1979). The research design and methodology to test the transfer hypotheses were developed at the Workshop on Experimental Designs for Predicting Crop Productivity with Environmental and Economic Inputs (Silva and Beinroth, 1975). In the project design, distinction is made between two types of experimental sites, designated primary and secondary sites. Primary sites of about 8 ha are locations where three kinds of experiments, transfer, variety, and management experiments, are conducted, whereas secondary sites of about 2 ha are locations where only transfer experiments are conducted. Secondary sites may differ from primary sites in that they exhibit variations in soil properties important to plant growth, such as base saturation or particle size, within the range permitted by the definition of the soil family. A minimum of eight sites was established for each soil family network. The transfer experiments (N x P experiments) are designed to provide data to

FIG. 1. Location of the experiment sites on the three tropical soil families studied in the 3enchmark Soils Roject.

320

F. H. BEINROTH ET AL.

test the hypothesis of transferability. The variety experiments are intended to evaluate varieties of maize, the test crop, for both their suitability to the environment and their responsiveness to N and P. And the management experiments are designed to provide information to local government agencies on ways to increase resource productivity by appropriate land use at different levels of management. In the transfer experiments, the variables to test the transfer hypothesis were selected because they are (1) controllable, (2) important to the test plant, maize, and (3) key features of the soil families under study. Phosphorus (P)was selected because it is an important limiting factor to crop yields and is linked with the definition of the three soil families under study, since technical information contained in the soil family name indicates the presence of mineral constituents that determine the soil’s ability to “fix” P. Thus, the term “thixotropic” connotes mineral constituents with a high capacity to absorb and immobilize fertilizer P. Similarly, the term “clayey” indicates soils with finely divided particles that possess large surface areas. In addition, the affinity of such an extensive surface for P is implied by the suffixes “ox” in Eutrustox and “ult” in Paleudult. Nitrogen (N) was selected as the second variable because it is rapidly depleted in tropical soils when the land is cleared and continuously cropped, and it is often the most limiting nutrient to crop growth in the tropics. It also plays an important role in the utilization of fertilizer P by crops and influences the yield level attained. partial factorial modification by The treatment design selected was the Escobar described by Laird and Turrent ( 1979). It provides 5 levels each of the 2 variables combined, to give 13 of the 25 possible treatment combinations. It was found to be one of the outstanding designs on the basis of (1) appropriateness for use in graphic estimation of economic optima, (2) magnitude of bias error, (3) number of treatment combinations, and (4) magnitude of variance error. It also has good coverage of the factor space with its 13 treatments. A diagram of this design is shown in Fig. 2 and coded values for the treatments are given in Table 111. The same treatment combinations are used for all three families, but the rates used differ among families according to the different levels of these nutrients in the different soils. These treatments are replicated three or four times and installed in a randomized complete block design. The experiments have been designed so that all other controllable factors such as irrigation, other nutrients, plant protection, and plant density are maintained near optimum levels for clarity of treatment effects. Field operations, data collection, and data processing follow standardized procedures and guidelines to assure comparability of results throughout the network. Weather variables, disease incidence, and crop variety adaptions are not reflected in the soil family names, although predictions for crop response need to incorporate these data. Therefore, weather variables such as temperature, rain-

AGROTECHNOLOGY TRANSFER IN THE TROPICS

32 1

M.86

1

to.40

1 i0 b

i- -0.40 -0.85 -1

-1 -0.85

-0.40 0

+0.40 t0.85

t1

PhaphOrUB (mbd V d U d

FIG. 2. Treatment design for 52 partial factorial modification by Escobar used for the two-factor transfer experiments.

fall, relative humidity, solar radiation, and wind speed and direction are measured at each experimental site. Experimental grain yield results, with all related data on controlled and uncontrolled variables, are analyzed by project personnel, in conjunction with the project statistical consultants, with the ultimate objective of testing the transfer hypothesis and developing a model of agrotechnology Table 111 Coded Values of Treatments

Used in Transfer Experiments Treatment

Phosphorus

Nitrogen

A B C D E F G H

-0.85 -0.85 +0.85 +0.85 -0.40 -0.40 +0.40

-0.85 +0.85 -0.85 +0.85

J K

L M N 0 P

+0.40 0" -0.85 +0.85

0 0 Complete controlb Partial controlc

-0.40 +0.40 -0.40 +0.40 0 0 0 -0.85 +0.85

0 = optimum treatment, includes micronutrients. Complete control = no fertilizer applied. Partial control = no treatment variables applied but given blanket application of macro- and micronutrients.

322

F. H. BEINROTH ET AL.

transfer based on soil interpretation and soil classification. A discussion of the unique test of the transfer hypothesis is presented in Section V. The results of a total of 78 transfer experiments conducted by the project on three different soil families clearly support the general validity of the transfer hypothesis. The general shape of the response surface is similar among members of the same family (see Fig. 5 ) , but differences are apparent due to variations in soil characteristics within the limits of the soil family and soil N and P levels. The magnitude of the yield response likewise varies due to the individual site response-input relationship, which is affected by the specific environment of the site during the growing period. Nonetheless, there are general trends predictable across all sites of each soil family network. The common and specific behavior from all soils comprised in a soil family was further substantiated by the kind of response, or length thereof, to individual fertility variables. For example, no consistent response to lime on the Hydric Dystrandepts or to K on the Tropeptic Eutrustox was found over all sites. In addition, downy mildew disease occurred at very low levels (
The development of sound management practices that can be transferred is essential for increasing crop production and for wise land utilization. Experimentation to identify these practices can be costly and time consuming; however, the indication that agrotechnology can be transferred on the basis of the soil family offers an approach to reduce greatly the cost and time to acquire this knowledge. The key requirement is that the soils of research centers in the tropics, as well as those of important agricultural areas, be classified according to Soil Taxonomy. Management information developed at research centers throughout the tropics can then be transferred, economically and efficiently, to the appropriate soil families in agricultural areas of the tropics. It is likely, however, that not all management practices will be available for all soils or crops of interest in a particular country. In this case, the desired management practices can be developed on benchmark soils representing important agricultural areas of a country or a region. The basic objectives of the research to be conducted on carefully selected benchmark soils should be to ( 1) develop management practices that allow

AGROTECHNOLOGY TRANSFER IN THE TROPICS

323

sustained crop production conducive to maximum yields for different levels of economic inputs without causing soil degradation, and (2) quantify the soilhand qualities of benchmark soils in relation to the needs and performance of important crops. The latter necessitates knowledge of the soil requirements of specific crops and is aimed at providing a better basis for soil survey interpretation, land evaluation, and extrapolation of research results. The particular kind of research needed will vary with the nature of the problems associated with a particular benchmark soil, the crops to be grown, and the prevailing socioeconomic considerations. Soil-related constraints to crop production likely to be encountered and requiring investigation include the areas of soil fertility and plant nutrition, soil amendments, plant adaption, soil-water management, soil biology, soil erosion and conservation, and soil survey interpretation and land evaluation. Swindale (1979) has recently scrutinized the research needs on these and other aspects of tropical soils and reference is made to his paper. The Technical Advisory Committee (TAC) to the Consultative Group on International Agricultural Research (CGIAR) also has endorsed the use of benchmark soils sites and the need for soil, plant, and water studies upon them (TAC Secretariat, 1979). It is clearly impossible to conduct extensive research on all soil families. It may also be unnecessary. Different soil families differ in some, but not in all, respects; thus, we may assume that soils from different soil families will perform alike in some ways if they possess qualities common to, and crucial for, a particular type of land use. The sphere of transferability of knowledge and experience, therefore, is not restricted to members of the same soil family but extends to all families that possess qualities common to a particular use requirement. This commonality of soil qualities defines the sphere of transferability for a particular soil use. Since the sphere of transferability transcends soil family boundaries, the identification of the cause-and-effect relationships between soil characteristics or qualities and crop performance is essential to the wide, practical application of agrotechnology among soil families.

V. QUANTITATIVE VERIFICATION OF TRANSFERABILITY WITHIN A SOIL FAMILY A. GENERAL

Transferability of management practices may be accepted when the weight of evidence is sufficiently convincing. Based on data from a series of experiments conducted by the Benchmark Soils Project, where the management factors are the same for each experiment, the feasibility of transferring crop response was

324

F. H . BEINROTH ET AL.

CONSTANTS (soil. climalel CONTROLLED VARIABLES (applied leftilizsrs)

INPUTS

-b

DATA - !

Experimentation

OUTPUTS

~~~~~l~~ (predictions) lunctions

UNCONTROLLED VARIABLES (weather))

FIG. 3. The schematic approach to testing the transfer hypothesis; sequence of steps required to implement the experimental design of the Benchmark Soils Project.

evaluated. Developed in this section are the criteria and the data analysis methodology needed for the verification of transferability within a soil family. Quantitative verification starts with formulation of a general transfer model for the relationship between response data and various inputs, as is schematically shown in Fig. 3. Two determinants of yield, soil and the covariant of long-term climate, are important inputs and assumed to be constant within a soil family. Designed as variable inputs are management factors, for example, applied fertility, which are intentionally controlled at several levels. Other controllable factors not part of the treatment design are experimentally maintained at a constant level so that the response to the treatment design factors is the only information to be transferred. Actual soil levels of the treatment design factors are not constant across the experimental sites within a soil family, however, due to natural and past management variabilities. These and other variable inputs, including weather factors such as temperature and solar radiation, cannot be controlled at a constant level, but they can be measured at each site.

B. THETRANSFER MODEL For developing a measure of the weight of evidence in order to evaluate transferability, the relationship between the response data and the known levels of controlled and uncontrolled variables is characterized by parameters of a transfer function equation. Neglecting the uncontrolled variables for now, we can express the observed plot data (Y), which depend on the controlled variables, for example, applied phosphorus (P) and applied nitrogen (N), as Y =f(P, N)

+E

(1)

where the error component E, the difference between the response data Y and the transfer function f(P, N), is attributable to unknown sources and assumed to be random variation. Different mathematical forms for f(P, N), including polynomials and exponentials, have been used by soil fertility specialists. Historically,

AGROTECHNOLOGY TRANSFER IN THE TROPICS

325

the parameters of the response-input relationship, f(P, N), are estimated from the data for each site and the estimated function, f(P, N), is constructed by substituting the calculated statistics for the parameters. The resulting prediction equation for each site is expressed as P = f(P, N) where the Y s are the predicted data for each plot. The differences between the observed plot data and the predicted data, Y - Y , are called the ordinary residuals. Agrotechnology transfer is the extrapolation of a response-input relationship, estimated from a series of experiments, to new sites. Practically, we would like to evaluate transferability with only one series of experiments; consequently, the evaluation needs to simulate the transfer to nonexperimental sites. As developed by Wood and Cady (1980), one approach is to predict yields, denoted as f(-{), for one of k experimental sites using a transfer function estimated from the other ( k - 1) sites. The subscript i is an index for sites, i = 1, 2 , . . . , k . This is then repeated for each of the k sites; that is, we predict yields for each site based on a transfer function estimated from the other ( k - 1) sites. If the transfer residuals, Y i - Y(-{),are approximately the same magnitude as the ordinary residuals, Y i - P i , calculated by fitting a response function individually to each of the k sites, we have evidence for agrotechnology transfer. A specific criterion for the evaluation is the prediction statistic P, defined as the ratio of the pooled sum of squared transfer residuals to the pooled sum of squared within-site ordinary residuals. Thus, k

P = Z(Y{ f=I

f

i=1

Two transfer function models will be considered.

I . Transfer Model 1. Assuming the data can be adequately fitted by a quadratic polynomial in the design variables, a simple transfer model (transfer model 1) is a second-order polynomial response surface that is common to all sites but that allows a different intercept for each site. For this model, the test statistic P - 1, multiplied by a known constant, follows the F distribution with 5 ( k - 1) and the pooled residual degrees of freedom ( d ! . The prediction statistic P is evaluating the adequacy of a model with design variables only for use as a transfer model. The prediction equation is based on the shape of the response surface estimated from the other sites coupled with the site mean. Algebraically,

Y

= boi

+ b,P + bzN + b,P + b,N2 + b5PN (Model 1)

(3)

If the P and N variables are coded around zero, then the site intercepts are the predicted yields in the middle of the design. The i subscript on boi indicates that the intercepts can vary from site to site, while the b , , b2,. . . , b, terms determine the common shape of the response surfaces. For determining the economically optimal combination of P and N,only the shape, not the height, of the

326

F. H. BEMROTH ET AL.

response surface is important. Consequently, differences in the average heights of the individual site response surfaces are allowed in the transfer model by centering the observed yields about the mean for each site. 2. Transfer Model 2. A second transfer model (transfer model 2) is model 1 augmented by additional variables for the uncontrolled but measured site variables. These additional variables account for differences in the shape of individual site response surfaces due to interactions between the response surface variables and the site variables, for example, between the P and N linear terms and site variables. For one site variable, denoted by p . the estimated transfer function is written as

An alternative expression is

P

= boi

+ ( b , + b@)P + (b2 + b7p)N + b3P + b,W + b5PN

(4b)

This last equation emphasizes that the interaction variables allow different shapes of the response surface for each site since the estimated coefficients for P and N now depend on p. For transfer model 2, the test statistic, P - 1, is no longer proportional to an F statistic, but the distribution and subsequent significance level can be evaluated by procedures given in Wood and Cady (1980). C. APPLICATION OF TRANSFER VERIFICATION METHODOLOGY

Data from eight maize experiments on the thixotropic, isothermic family of Hydric Dystrandepts were used as a numerical example for testing the transfer of yield response to applied P and N. Included are four sites (IOLE-E, KUK-A, KUK-C, and KUK-D) in Hawaii, two (PUC-K and BUR-B) in the Philippines, and two (PLP-G and LPH-E) in Indonesia. A general description of the experimental and treatment designs is given in Section IV. A quadratic polynomial in the two treatment variables

is the assumedf(P, N) and is calculated for each site. The i subscript is an index for site identification, i = l(I0LE-E), 2(KUK-A), . . . , 8(LPH-E); f., are the predicted yields for the ith site, and the b 's are the estimated quadratic polynomial response surface parameters. The six-parameter quadratic polynomial function is fitted well by the data (3 replications of 13 treatment combinations of P and N). Lack-of-fit terms for each site with seven (13 - 6) dfare not important and have been pooled with the experimental error sums of squares. As shown in the previous section, the adequacy of transfer model 1 can be tested by the prediction statistic, which follows the F distribution with 35 and

AGROTECHNOLOGY TRANSFER IN THE TROPICS

321

264 df. The calculated F is 2.99, and the probability of a value of this magnitude, on chance alone, is less than 0.01. Based on the weight of evidence of transfer model 1, a model with only design variables will not be sufficient. Stated differently, interactions between sites and the quadratic polynomial variables exist. Additional data analysis shows that the P, N , and Pa terms of transfer model 1 interact with sites-that is, the eight b, coefficients for P, the eight b2 coefficients for N , and the eight b3 coefficients for Pa have a systematic trend over the sites rather than the random pattern expected with no interaction. Consequently, the uncontrolled but measured site variables are introduced to describe the sites quantitatively. Insight on interactions between treatment design variables and site variables can be gained from plotting the estimated coefficients for P, N , and Pz against selected site variables as shown in Fig. 4.Block A shows b , vs. soil phosphorus as measured by the modified Truog method, Truog P; block B shows b2 vs soil nitrogen extracted by 2 N KCl, extr. N; block C shows b, vs the average daily minimum temperature during an 8-week period around 50% tasseling, min. temp.; and block D shows b3 vs. min. temp. The systematic trends in the four plots of Fig. 4 reflect the presence of interactions. Estimated regression coefficients calculated by fitting a quadratic polynomial to the data in Fig. 4A are identical to the coefficients obtained by adding two interaction variables to the transfer model Y = boi

+ b , P + b,N + b , P + b4N2 + b,NP + bGpP+ b7pzP

(6a)

or, alternatively,

f

=

boi

+ (b, + bGp + b7p2)P + b2N + b3PZ + b4W + b5NP

(6b)

The latter equation shows the effect of Truog phosphorus (p) on the shape of the response surface and, in particular, the effect of a site’s level of soil phosphorus on the linear response to applied phosphorus (P) for that site. Table IV compares b6p b7p2 of the augmented the P response coefficients, estimated by b, transfer model (first column), with the P response coefficients, b I i , estimated from the individual site analyses (second column) and from transfer model 1 (third column). The closeness of the first two columns, compared with the third column, indicates the need for site variable data to explain the differences in the P response data. Based on Fig. 4 and on other analyses, four additional interaction variables were added to transfer model 2, namely, extr. N by N , min. temp. by N , (min. temp.)2 by N , and min. temp. by Pa. Inclusion of these interactions allows the N and Pa coefficients of transfer model 1 to vary from site to site. With the six interactions incorporated in the transfer model, the transfer sums of squares were calculated and are shown in Table V along with the ordinary residual sums of

+

+

B

;1

500

yo0

D

Ei sm

C

6f -

0

8600

t

H

-Mo

(D

rB u m

0

L% -m

0 0

200

0

0 2y)o

0 0

'

.

S

R W ~ YIW MCEMNIIE PCI

~

Z

I

Z

Z

0

/ .6

~

~ I6

R

I9 IS 20 M I N I M TDpERlllR IT)

21

ZZ

P

FIG. 4. Relationship between regression coefficients for N and P and individual site variables from the eight experiments. (A) P X Truog P; (B) N x extractable N; (C) N x minimum temperature; (D) Pz x minimum temperature.

329

AGROTECHNOLOGY TRANSFER IN THE TROPICS

Table IV comparison of Regression Coefficients for P Response, with and without Site Variables

P response, site variables added ( h + bsP + b , p Z ) , Transfer model 2

Site Hawaii IOLE-E KUK-A KUK-C KUK-D Philippines PUC-K BUR-B Indonesia PLP-G LPH-E

447

P response, no site variables (b,) Individual site model

Transfer model 1

546 1110

443 779 535 890

822 780 814 762

I068 1336

I107 1430

732 685

418 69 1

462 554

824 81 1

634

squares and the site variable data. For example, when transfer model 2 is estimated from the last seven sites and used to predict the yields for IOLE-E, the - f(-i#is equal to 10,650,460, a 25% increase transfer sum of squares, over the IOLE-E residual sum of squares. Summed over the eight sites, the

z(Yi

Table V Comparison of Residual and Transfer Sums of Squares aner Adding Site Variables (Transfer Model 2) Site variables

Site Hawaii IOLE-E KUK-A KUK-C KUK-D Philippines PUC-K BUR-B Indonesia PLP-G LPH-E Total

Residual sums of squares

Transfer sums of squares

Modified Truog P (PPm)

Extr. N (PPm)

Min. temp. (“C)

8,550,008 25.7 14,266 13,602,424 25.599,726

10,650,460 34,309,230 15,412,100 30,825,610

42 54 49 62

17 13 46 29

18.9 20.5 18.9 17.9

5,869,074 25,055,225

9,035,600 32,2 19,560

11

5

79 29

23.0 21.5

29,893.4 12 17,880,236

40,316,440 30,916,190

36 22

35 119

15.9 16.8

152,164,299

203,685,190

330

F. H. BEINROTH ET AL.

prediction statistic is

P

=

203,685,1901152,164,299 = 1.34

(7)

This value of P is associated with a significance level of 0.32, giving evidence that the response surface for applied P and applied N can be transferred with an estimated transfer model including different intercepts and interactions between the site variables and the treatment design variables. The P statistic, a ratio of sums of squares, is a summary statistic for comparing the transfer ( Yi - Y ( - { J and ordinary ( Y i - Pi>residuals. The actual magnitudes of the differences between the ordinary Pi, based on the individual site data, and the transfer Y ( - * ) , based on data from the other sites, are given in Table VI. The for five treatment combinatabular values are absolute differences, Pi tions with increasing levels of both P and N, and are averaged over the three replications for each site. The differences display variability but are sufficiently small, especially at the middle levels, so that the transfer predictions, Y ( - { ) , could be used for practical purposes to predict response to P and N application for sites where an experiment had not been carried out. The predicted yields for each site, plotted three-dimensionally with P and N as the horizontal axes, form an estimated response surface showing the predicted yield response for any combination of P and N within the experimental ranges of the factors. The response surface plots in Fig. 5 summarize the results of the transfer analysis. Specifically, both transfer models can be graphically compared with the individual site predictions. In the middle row, the predicted yields (the Table VI Absolute Differences (kdha) between PI and Y ( + l )Using Transfer Model 2 for Five Treatment Combinations of Applied Phosphorus (P) and Nitrogen (N) Coded P and N treatment

Site Hawaii IOLE-E KUK-A KUK-C KUK-D Philippines PUC-K

-0.85 PI -0.85 N

-0.40 PI -0.40 N

O/ 0

+0.40 PI +0.40 N

+0.85 P/ +0.85 N

I64 684 246 47 1

78 193 183 I39

59 538 267 350

94 469 71 249

195 99 48 1 238

160

BUR-B

125

279 225

120 359

78 327

344 93

Indonesia PLP-G LPH-E

56 1 703

340 582

92 277

276 213

748 985

IOLE-E

KUk-A

KUK-c

KUK-D

PUC-I(

rn-8

FIP-G

LPH-E

KUK-C

KUK-D

PUC-K

BUI-8

PLP-0

LPH-E

KU-C

KU-D

PUC-K

Dun-.

PLP-0

LPH-E

Using Individual Sit. Mod.1

IOLE-E

KUK-PI

P,-u Using Transfor Modoll

IOLE-E

KU-A

FIG.5. Comparison of response surfaces of predicted value of the eight experiments for three different models: individual site model (center), transfer model 1 (bottom), and transfer model 2 (top).

332

F. H.BEINROTH ET AL.

vertical axis) from fitting a quadratic polynomial in applied P (the right horizontal axis) and applied N (the left horizontal axis) to each site individually are plotted. The best-fitting response surface is formed from the predicted yields and is represented by the 9 x 9 grid for each site. To simulate the transfer of technology, consider for a moment that an experiment was not done at the PUC-K site and one wanted to predict the nature of the response surface from the other seven sites. Using transfer model 1, the predicted response surface would be the PUC-Kplot in the bottom row, while the plot in the top row results from estimating transfer model 2 from the other seven sites. The top response surface is a closer approximation to the middle response surface than the bottom one for PUC-K and is generally true for all the sites. The Y , - p(-t)differences in Table VI are the differences between the transfer model 2 response surface in the top row and the response surface below in the middle row for five P and N combinations. The P and N combination of -0.85 and -0.85 is the front comer of each plot, the combination of +0.85 and +0.85 is in the back comer, and the other three points are on the diagonal line between the two comers. If one views across the eight response surfaces in the bottom row, the similarity of P and N response can be noticed, but the resulting transfer equations do not predict as well as the transfer equations from transfer model 2 as represented by the top row. Some of the commonality of the P and N response is retained, but the introduction of site-variable information into transfer model 2 allows each response surface in the top row to have a uniqueness that is associated with the particular site to be predicted. For the P response, the closeness of transfer model 2 to the actual response is also shown in Table IV. For both the P and N responses, (1) the resemblance between the response surfaces for the top and middle rows of Fig. 5 and (2) the P statistic value of 1.34 with the associated significance level of 0.32 for transfer model 2 are weights of evidence that the P and N responses can be transferred.

VI. PREREQUISITES FOR WORLDWIDE AGROTECHNOLOGY TRANSFER A. INTERNATIONALIZATION OF SOILTAXONOMY

The key to international agrotechnology transfer is the use of a single soil classification system by all participating countries. Thus, it is a matter of high priority to develop a new soil classification, or to choose one from several existing systems, for use as the standard for international reference and communication. While such action is pending, Soil Taxonomy (Soil Survey Staff, 1975) has emerged as the defacro international soil classification system. This

AGROTECHNOLOGY TRANSFER IN THE TROPICS

333

has occurred in spite of the fact that Soil Taxonomy has not been fully tested, especially for soils of the lower latitudes, and is undergoing continuing revision. In 1979, the Soil Conservation Service of the U.S. Department of Agriculture established an Office of International Soil Classification and Correlation. Functions of the Office are to consider recommendations for revision of Soil Taxonomy from outside the United States, to consult on questions relative to the application of Soil Taxonomy and international soil correlation, to organize workshops for the study of soil classification and soil correlation, and to help organize training in soil survey methods using Soil Taxonomy (Johnson, 1978). This initiative by the Soil Conservation Service is particularly timely as the internationalization of one soil classification system, even on an ad hoc basis, is the first essential step for worldwide agrotechnology transfer. B. SOILRESOURCE INVENTORIES

It is obvious that transfers based on soil classification require knowledge of the kind and distribution of the soils in the recipient area. Historically, soil surveys have been made to acquire and compile this information. Since the subject of soil resource inventories is adequately covered in the proceedings of a workshop held at Cornell University (1977), the discussion in this paper is confined to some general comments. Land variability may be either systematic or random, but in either case, spatial and temporal variability in land characteristics is a major cause of uncertainty in agricultural productivity and land performance. The objective of soil resource inventories is to stratify both spatial and temporal systematic variabilities so that the corresponding variabilities in productivity and performance can also be stratified. Systematic changes in land, such as climatic fluctuations due to seasons or soil variation related to soil-forming processes, are easily predicted; random variabilities are less predictable, of course, and they include such transient events as weather, a termite mound, an uprooted tree, or an ancient village site. A soil survey should nonetheless alert the users of a site to important random variations in the land that are not delineated on a soil survey map. It follows from the need for a common international soil classification system that uniform procedures for making and interpreting soil surveys are also needed. Fortunately, relevant guidelines are now becoming available that reflect the experience gained during more than 80 years of soil survey operations in the United States. Much of this knowledge is contained in the Soil Survey Manual (Soil Survey Staff, 195 l), which is currently undergoing a complete revision and updating. Laboratory methods and procedures for collecting soil samples have also been standardized (Soil Conservation Service, 1972). Procedural aspects of soil survey including planning, design, staffing, quality control, field operations,

334

F. H.BEINROTH ET AL.

cartography, documentation, management, and publication are recorded in the National Soils Handbook, which is now being developed by the Soil Conservation Service (Johnson, 1978). Soil surveys and classification are not ends in themselves; they are the means to plan for the proper use of land resources. The critical link between a generalpurpose soil resource inventory and land-use planning is soil interpretation. Interpretation involves the discovery and recording of the cause-and-effect relationships between observed or measured land characteristics and land performance. Properly developed, soil interpretation can become the means to derive land evaluation for a specific purpose from a general-purpose soil survey. Unfortunately, soil interpretation on an international scale today remains an art. It needs to be made a science. C. LANDEVALUATION

In the transfer of agrotechnology,particularly vertical transfers, there are many aspects of soils that are not reflected in their taxonomic names. Slope, for example, is not used as a differentia of classification, but is very important to land use. It is therefore essential to consider soils within the wider concept of land. Land, as defined by FAO, is . . . an area of the earth’s surface, the characteristics of which embrace all reasonably stable. or predictably cyclic, attributes of the biosphere. . . and the results of past and present human activity to the extent that these attributes exert a significant influence on present and future uses of the land by man (FAO, 1976).

Land evaluation is the process of assessment of land performance when used for specific purposes. A framework for a comprehensive land-evaluation system designed to appraise the suitability of land for crop production has been developed by F A 0 (1976). This system attempts to relate physical qualities or attributes of land to the environmental requirements of crops for specified levels of management. Examples of land qualities cited in the F A 0 publication are energy supply (solar radiation, day length, temperature), water supply (precipitation, rainfall distribution, soil-water storage capacity), nutrient supply (soil fertility), oxygen supply in the root zone (drainage), trafficability , erodibility, salinity, foothold, and other qualities. The kinds of crop requirements that must be matched by the land are photoperiodicity, shade tolerance, soil and air temperature, humidity, wind speed, water needs, mineral nutrition, foothold, soil aeration, tolerance to aluminum and manganese toxicities and trace-element deficiencies, and susceptibility to insects, weeds, and diseases. The Benchmark Soils Project (1979), for example, has found that the soil family of Soil Taxonomy, among other things, stratifies land according to incidence of insects, weeds, and plant diseases. This feature of Soil Taxonomy, which enables it to stratify land into agroecological zones,

AGROTECHNOLOGY TRANSFER IN THE TROPICS

335

provides the link between the scientific and technical grouping of soils. The technical information accessory to the differentiating criteria must be extracted from Soil Taxonomy and reconstructed in a land-evaluation scheme. Matching land characteristics to crop requirements in land evaluation requires that crops are also stratified in a similar scheme. The latter is attempted by Duke (1978) who is categorizing plants according to their environmental requirements. Such efforts significantly contribute toward achieving the goal of land classification, which is to eliminate trial and error by stratifying the soil-plant-atmosphere continuum into tight crop-performance classes. Land evaluation is a practical product of soil resource inventories. Its primary use is in land-use planning to guide decisions on land use in such a way that the resources of the environment are put to the most beneficial use for man while at the same time conserving those resources for the future (FAO, 1976). D. ASSEMBLING CROPPERFORMANCE DATA

Up to this point, we have assumed that the relationships between land characteristics, crop requirements, and crop performance are known. In fact, we have simply assumed that crop performance depends on the goodness of fit between crop requirement and land characteristics. Since perfect fit is rarely attained, the effort expended to rectify the mismatch between land characteristics and crop requirement constitutes the management level. The ultimate goal of land inventories is to provide planners with reasonably accurate estimates of crop performance under prescribed levels of inputs. It turns out, however, that the information required to proceed with the matching process to predict crop performance is not available, or has not been collected and assembled with the matching process in mind. This is largely due to the fact that performance data do not depend only on average, long-term systematic variations in land characteristics but on random weather conditions during critical periods in the crop’s life cycle as well. Good performance data are those for which the cause-and-effect relationships between environmental conditions and crop performance are quantitatively known. The time has come to design field experiments to generate such data. There are two potential sources of performance data which meet the requirements for use in the matching process. These are the International and National Agricultural Research Centers. The value of research data generated in the National and International Centers would be increased if crop performance could be related to land characteristics, crop requirements, management inputs, weather, and any other factor significantly affecting yield. To do this, some international coordination will be required, and recommendations supporting this concept were made in a workshop on “Operational Implications of Agrotechnology TransferenceResearch sponsored by the InternationalCrops Research Center for ”

336

F. H.BEINROTH ET AL.

the Semi-Arid Tropics and the Benchmark Soils Project of the Universities of Hawaii and Puerto Rico, held in October 1978 in Hyderabad, India. E. INFORMATIONSYSTEMS AND DATABANKS

If international coordination of soil research is achieved so that participating countries make soil resource inventories based on a common soil classification system and also conduct agronomic experiments on well-characterized and well-monitored sites, it will become profitable to assemble soil and crop performance data in a central place or in a network of standardized data banks. The purpose of assembling such data in readily retrievable form is twofold. First, the crop performance data and the inputs used to achieve that level of performance can be transferred to other similar sites located elsewhere in the world. Second, the assembled information can serve as the basis for soil interpretation and matching crop requirements to land characteristics. For example, the absence or low incidence of downy mildew on maize grown on the thixotropic, isothermic family of Hydric Dystrandepts is not unique to this soil family but to all soil families with isothermic or cooler temperature regimes. Similarly, the high P-fixing capacity noted in this family extends to all thixotropic soil families. Thus, it is not necessary to do research on every site or even on every soil family. Much of the cause-and-effect relations between crop performance and land characteristics can be derived from research on a limited number of carefully selected benchmark soils. Orderly manipulation of the large number of soil and crop performance data along with weather and management data can only be achieved by means of computerized data banks. The key to the success of such data banks is storage of high-quality soil and crop performance data, collected in a standard way and designed for economy of effort, ease of interpretation, and transferability of information. It follows that standard soil, climate, crop performance, and management data assembled in a network of data banks should also be coded for storage and retrieval in a standard way. Some uniformity in choice of computer language is also desirable.

VII. CONCLUSION: IMPLICATION FOR AGRICULTURAL DEVELOPMENT In the next 20 to 25 years, the world will have two billion more people to feed. There is little doubt that the world has the land and the technology to produce

AGROTECHNOLOGY TRANSFER IN THE TROPICS

337

enough food to meet the anticipated population increase. There is, however, considerable doubt as to whether the food will be produced where it is most needed or distributed in an equitable manner. For humanitarian, as well as for the more pressing reason to secure world order, self-reliance on national food production will continue to be a major international goal. But even the most optimistic forecasts indicate that the emerging nations of the tropics will not be able to avert food shortages. To prevent this, the developing nations and the international development agencies will need to increase research efficiency significantly in the area of food production. The basic structure for achieving research efficiency already exists in the form of a network of National and International Agricultural Research Centers (NARCs and IARCs). These two types of centers play somewhat different roles. The IARCs conduct research that is more amenable to horizontal technology transfer. For some years now, the IARCs have been developing transferable germplasm and farming systems for the ,major agroecological zones. It is almost certain that their efforts have lessened the severity of food shortages in the resource-poor regions of the world. Research conducted by the NARCs has a narrower horizontal scope and is heavily oriented toward vertical technology transfer-that is, rendering scientifically and technically sound technology appropriate for assimilation in local farming systems. Greater research efficiency in the research centers can be achieved by refining procedures to match and tailor agrotechnology for specific agroenvironments and socioeconomic situations. These procedures, if based on sound taxonomic principles, are the means to organize knowledge so that the behavior and performance of the object being classified and studied may be transferred to other locations where similar conditions exist. Agrotechnology transfer as a means to meet the world’s food requirement and demand will require international coordination. The role of the coordinating body created to perform this task should be to enable countries with probable food shortages to exploit fully the agricultural capacity in a timely manner. To achieve this goal, this body must ensure that (1) a common soil classification system is used by all participating countries, (2) a uniform procedure for making and interpreting soil surveys based on a common soil classification system is employed, (3) standard procedures for matching crop requirements to land characteristics are developed and used, (4)international soil correlation and quality control are maintained in soil surveys and land evaluation, and ( 5 ) a tightly knit network of data banks for storing, analyzing, and displaying soil and crop performance data is established. The result will be greater research efficiency and greater economy of action and will contribute to the expeditious agricultural development in regions of the world where it is most needed.

338

F. H. BEINROTH ET AL. ACKNOWLEDGMENTS

We gratefully acknowledge the editorial assistance of Cynthia L. Garver in the preparation of this manuscript and the data analysis assistance of Clement P. Y. Chan.

REFERENCES Beinroth, F. H. 1979. Dep. Agron. Soils, CON. Agr. Sci., Univ. Puerto Rico and Dep. Agron. Soil Sci., Coll. Trop. Agr. Human Resour., Univ. Hawaii Tech. Rep. No. 5. Benchmark Soils Project. 1978. Dep. Agron. Soil Sci., Coll. Trop. Agr., Univ. Hawaii and Dep. Agron. Soils, Coll. Agr. Sci., Univ. Pueno Rico Progr. Rep. No. 1 (through Dec. 1977). Benchmark Soils Project. 1979. Dep. Agron. Soil Sci., CON. Trop. Agr. Human Resour., Univ. Hawaii and Dep. Agron. Soils, Coll. Agr. Sci., Univ. Puerto Rico. Progr. Rep. No. 2 (Jan. 1978-June 1979). Bornemisza, E., and Alvarado, A. (eds.). 1975. “Soil Management in Tropical America.” Soil Sci. Dep., North Carolina State Univ., Raleigh. Boulding, K. E. 1970. Am. Econ. Rev. 60, (2). 406-411. Cline, M. G. 1977. I n “Soil Survey Quality: Conference Proceedings, Bergamo East, N.Y., December 5-7, 1977,” pp. 5-19. New York Cooperative Soil Survey. Cline, M. G. 1979. Agron. Mimeo 79-12. Cornell Univ., Ithaca, New York. Cornell University. 1977. Agron. Mimeo 77-23. Cornell Univ., Ithaca, New York. Duke, J. A. 1978. In “Crop Tolerance to Suboptimal Land Conditions’’ (G. A. Jung, ed.), Special Pub. 32, Am. SOC. Agron., Madison, Wisconsin. FAO. 1976. Soils Bull. 32, FAO, Rome. Hayami, Y.,and Ruttan. V. W. 1971. “Agricultural Development: An International Perspective.” Johns Hopkins Press, Baltimore, Maryland. Ikawa, H. 1979. Misc. Pub. 165 (BSP Tech. Rep. I ) , Hawaii Agr. Exp. Sin., Honolulu, Hawaii. Johnson, William M. 1978. Paper presented at Workshop on Operational Implications of Agrotechnology Transference Research (October 23-26), ICRISAT, Hyderabad, India. Kamarck, A. M. 1972. Seminar Paper 2 , Econ. Dev. Inst.. Int. BankReconstr. Dev.. Washington, D.C. Kellogg. C. E. 1961. “Soil Interpretation in the Soil Survey.” U.S. Dep. Agr., Soil Cons. Serv., Washington, D .C . Laird, R. J., and Turrent, A. 1979. I n “Roc. Workshop on Expt’I Designs for Predicting Crop Productivity with Environmental and Economic Inputs” (Silva, J. A., ed.), pp. 143-179, Dep. Paper 49. Hawaii Agr. Exp. Sin., Honolulu, Hawaii. Moseman, A. H. 1970. “Building Agricultural Research Systems in the Developing Nations.” Agr. Dev. Council, New York. Myrdal, G. 1974. Sci. Am. 231 (3). 173-182. National Research Council. 1972. ‘‘Soils of the Humid Tropics.” Natl. Acad. Sci., Washington, D.C. Orvedal, A. C. 1975. Vol. I , Tech. Ser. Bull. 17. Ofice of Agr.. Agency for Intern. Dev., Washington, D.C. Orvedal, A. C. 1977. Vol. 2 , Tech. Ser. Bull. 17, Ofice of Agr., Agency for Intern. Dev., Washington, D.C. Orvedal, A. C. 1978. Vol. 3 , Tech. Ser. Bull. 17. Ofice of Agr., Agency for Intern. Dev., Washington, D.C. Sauer, C. 0. 1969. “Agricultural Origins and Dispersals: The Domestication of Animals and Foodstuffs” (2nd ed.). MIT Press, Cambridge, Massachusetts.

AGROTECHNOLOGY TRANSFER IN THE TROPICS

339

Sanchez, P. A. (ed.). 1972. “A Review of Soils Research in Tropical Latin America.” North Carolina Agr. Exp. Stn., Raleigh, North Carolina. Silva, J. A., and Beinroth, F. H. 1975. Hawaii Agr. Exp. Srn. Dep. Paper No. 26. Honolulu, Hawaii. Smith, G. D. 1965. Pidologie (Spec. No. 4). 5-134. Soil Conservation Service. 1972. Soil Surv. Inv. Rep. I , US. Dep. Agr., Soil Cons. Serv., Washington, D.C.? Soil Survey Staff. 1951. U.S. Dep. Agr. Handb. 18, U.S. Govt. Printing Office, Washington, D.C. Soil Survey Staff. 1975. U . S . Dep. Agr. Handb. 436, U.S. Govt. Printing Office, Washington, D.C. Swindale, L. D. 1979. “Roc. Conf. on Priorities for Alleviating Soil-Related Constraints to Food Production in the Tropics.” IRRI, Los B h o s , Philippines, June 4-8. TAC Secretariat. 1979. TAC Document AGDITAC:IARI~~II Re I . Wood, A. 1950. “The Groundnut Affair.” Bodley Head, London. Wood, C. L., andCady, F. B. 1980. Biomerrics (in press). Wortman, S. 1976. Sci. Am. 235 (3). 31-39.

This Page Intentionally Left Blank

ADVANCES IN AGRONOMY. VOL. 33

THE PRODUCTION CHARACTERISTICS OF Bromus inermis LEYSS AND THEIR INHERITANCE P. D. Walton Department of Plant Science, The University of Alberta, Edmonton, Alberta, Canada

. . . . . . . . 341

1. A. Plant Characters

11.

........ ................................................ Seed Production and Establishment ...................... ........ A. Floral Initiation . . . . . .................................. B. Stomata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

111.

345 346 347 347

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

C. Germination and Seedling Growth . . . . . . . IV.

B. Digestibility . . . ........................................ C. Protein Content ................................. . . . . . . . . . . . . 352 V. Forage Yield . . . . . . . . . . .

Vl. A. Seed Yield

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

B. Dry Matter Yield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . .

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

I. INTRODUCTION A. PLANTCHARACTERS

The genus Bromus comprises about 60 species, both introduced and native to North America. The name Bromus, which is the Greek for oat, is indicative of 341

Copyright @ 1980 by Academic Ress. Inc. All rights of reproduction in my form reserved. ISBN 012-000733-9

342

P. D. WALTON

the nature of the panicle and the area of adaptation. Bromus inermis Leyss, which is known as bromegrass, Austrian brome, Hungarian brome, Russian brome, and most commonly, smooth bromegrass, belongs to the subgenus Zerna of the tribe Festuceae. It is normally an octoploid (2n = 8 X = 56), and two distinct types are recognized within the species. The southern type was introduced from France and Hungary into the United States and was later brought to Canada, where it is grown in the eastern part of the country and in southern British Columbia. The northern type was introduced to Canada directly from Germany in 1888 and is now extensively grown in the prairie provinces. Smooth bromegrass is a native of Europe and Asia and is adapted to most temperate climates. It is a leafy, sod-forming perennial which spreads vegetatively by underground rhizomes. It is resistant to drought and to extremes of temperature, being capable of withstanding both hot, dry summers and long, cold winters. The species is grown both alone and in mixtures with other grasses and legumes and is used for pasture, hay, and erosion control. The forage quality of smooth bromegrass ranks well among the cool-season grasses. The crude protein content is high, ranging from 12 to over 20%, during the time of rapid growth at the beginning of the season. The inflorescence consists of many spikelets, each of which contains several hermaphroditic florets. The pollen is disseminated by wind. Each panicle usually produces abundant seed and is highly crossfertilizing, naturally cross-pollinated, and rather self-sterile.

B. AGRICULTURAL USE Following the introduction of smooth bromegrass into the North American continent and the initial recognition of the value of this species, interest declined. It was not until farmers and scientists sought to combat the dust-bowl conditions generated in the early 1930s by a combination of extensive ploughing, overgrazing, and drought conditions that it was realized that smooth bromegrass was, among the introduced grasses, one of the principal and most widespread survivors. It was as a result of selection work and studies carried out in those years that the two types of bromegrass (northern and southern) were recognized and described. The species is now accepted as one of the most successful grasses used for erosion control on roadside shoulders and steep road cuts. It is frequently grown in mixtures with other grasses for the vegetation of waterways, irrigation canal banks, and terraces, as well as in areas where the soil has been extensively disturbed. The aggressive and extensive root system of this grass species brings about rapid improvement in soil structure. Where a legume has been incorporated, the decay of the legume root maintains a balance of available nitrogen and aids in the decomposition of the grass roots. In areas where subsoil has been exposed or where soil has been eroded on slopes, the establishment of the species is much enhanced by a dressing of a nitrogenous fertilizer.

CHARACTERISTICS OF Eromus inermis LEYSS

343

On the Canadian prairies, smooth bromegrass is one of the principal grass components of a grass-alfalfa seed mixture which is widely used on nonirrigated pastures. Frequently, such fields are first harvested for hay, whereas the regrowth is used for pasture. The most favorable time for the establishment of smooth bromegrass is in the early spring. Planting should be carried out as early as the establishment of arable crops will permit. Under Canadian farming conditions, both a companion crop or excessive weed growth will retard the establishment of such a spring sowing. Competition for nutrients or moisture may be reduced by early and repeated mowings at such a height that the bromegrass plants will not be clipped. Material cut in that way should be removed so as not to cover the seedlings. Where pastures are used for grazing only, rotational grazing will double the production of beef per hectare over that obtained from continuous grazing (Walton, 1979). Where continuous grazing is necessary smooth bromegrass stands should be understocked during the early part of the season to allow growth to accumulate for later use. Such practices do, however, present dangers, since forage quality decreases as the plants mature. Like many other tall grasses, the yield of smooth bromegrass decreases with frequent cutting (Wright er al., 1967). On the Canadian prairies, four defoliations, under grazing conditions, and two cuts, where the material is intended for hay, will give optimum productivity and a satisfactory balance between production and forage quality. For both hay and pasture, smooth bromegrass behaves satisfactorily, if not ideally, in a mixture with alfalfa. Bromegrass makes a substantial contribution to early harvests. Where haying is delayed so that the first harvest is taken during or after the time of bromegrass seed formation, alfalfa will increase and dominate the mixture (Walton, 1979). Where grazing is frequent or stocking rates are high, alfalfa will tend to be grazed out of the mixture. Maintaining a balance between the grass and the legume in the pasture is important, since the danger of bloat increases as the proportion of alfalfa becomes higher. However, the legume will improve palatability and intake of the feed, resulting in better animal performance.

II. THE NATURE OF THE SPECIES

A. PLOIDYLEVELSA N D CHROMOSOME NUMBERS

Smooth bromegrass is most often encountered as an octoploid (2n = 8x = 56), although a range of chromosome numbers (2n = 28, 42, 56, and 70 and some intermediate values), suggesting a polyploid series, has been reported. It appears that an active state of evolution still exists in what may well be an old polyploid complex. The perennial nature of the species, together with its persis-

344

P. D.WALTON

tent and aggressive growth habit, has perpetuated a number of chromosome aberrations of the type reported by La Flew and Jalal(l972) and Tan and Dunn ( 1977a). Initial attempts to determine the genomic constitution of the octoploid led to disagreement, and three main theories emerged. Hill and Carnahan (1957) considered it to be an autoalloploid, with the genome formula AAAA BBBB, in which A and B represent two distinct genomes. The high bivalent and low quadrivalent frequencies in the plants examined were attributed to a genetic mechanism suppressing quadrivalent formation between homologous chromosomes. Ghosh and Knowles (1964), Schulz-Schaeffer (1960), and Wilton (1965) observed that the chromosome complement contained three pairs of satellite chromosomes, whereas either two or four pairs might have been expected from the above theory. Two pairs of chromosomes, of similar appearance, bore large satellites, whereas the third pair bore minute satellites. On this basis the genome formula of AAAABBCC was proposed. Elliott and Love (1948), Elliott (1949), and Nielsen (1951) suggested an alloploid genome formula (A, A,A,A,B, B, B, B 2) in which homologous genomes had been differentiated by gross structural morphology. The low frequency of quadrivalent formation indicated a remnant homology of parental genomes. By means of interspecific hybrids between B . erectus (2n = 28) and B . inermis, Armstrong (1977) demonstrated that the genome of B . erectus, arbitrarily designated as the A genome, was a component in the tetrasomic condition of B. inermis. This genome consists of five median chromosomes, one subterminal chromosome, and one chromosome with a large satellite. In interspecific hybrids between B. arvensis (2n = 14) and B . inermis, the haploid genome of B . inermis could be studied, since the chromosomes of B. arvensis are much larger than those of B . inermis. The karyotype of the B genome, constructed in this manner, consisted of two pairs of subterminal and five pairs of median chromosomes. From the results of this study, Armstrong concluded that there was no evidence to indicate the presence of more than two genomes. In explaining the opposing conclusions reached by previous workers, Armstrong points out that polymorphic chromosome morphology is not unknown in cross-pollinated species; also, suppression of at least one pair of satellites, the possible differences in contraction of chromosomes, and mechanical stresses during preparation of the slides may influence the interpretation of the genome. However, he concedes that the conventional staining procedures which he used would not be expected to detect homologous genome differentiation. The Giemsa staining technique could, through the banding patterns it produces, detect karyotype differences between homologous genomes. Within the genus, recent studies have shown that it is possible to cross Bromus inermis with B . compelianus (Hanna, 1961), B . erectus (Armstrong, 1973), and B . arvensis (Armstrong, 1977). In all cases the hybrids were infertile.

CHARACTERISTICS OF Bromus inermis LEYSS

345

Tan and Dunn (1977b) found a high frequency of anaphase irregularities in tetraploid and hexaploid B. inermis individuals. The commercial octoploid was more stable than plants at the other two ploidy levels. The authors believed that chromosomal instability in somatic cells could account for interplant variations of some morphological characters. B . STOMATA

I . Relationships with Ploidy Level

Although the size of the pollen grains, stomata, and cell, and the plant height and leaf width and length all increase at the higher ploidy levels, not all of these characters are equally reliable in detecting differences in ploidy level. Tan and Dunn (1973) studied pollen grain size and stomata size frequency. They found that higher correlation coefficient values and smaller standard deviations were obtained at the three ploidy levels ( 4 x , 6 x , and 8 X ) for stomata length than for any of the other characters. Both mature plants grown in the field and seedlings grown in the greenhouse conformed to this pattern. By studying the length of the stomata on the cotyledons, these authors were able to screen populations of Bromus inermis for plants with different ploidy levels.

2 . Relationships with Plant Physiology Walton (1974b) first drew attention to the association between high yield, low stomatal frequency, and large stomatal size. Three important plant processes are effected by the nature and activity of the stomata. First, the photosynthetic efficiency of the leaf and leaf sheath depends on carbon dioxide uptake and oxygen liberation. Second, plant respiration releases, together with carbon dioxide, energy stored in the form of carbohydrates. Third, the translocation and distribution of both assimilates and breakdown products is influenced by the movement of water vapor. When light is not limiting, as is normally the case on the prairies, these fundamental plant processes determine herbage production. The synthesis and breakdown of storage products and the translocation of these subtances for both storage and utilization is vital to plant growth rate, development, and, in perennial plants, to winter survival. Extensive consideration (Fuehring et al., 1966; Waggoner, 1969) has been given to the function of the stomata in relation to crop water use. Tan and Dunn (1975) confirmed conclusions reached by Walton (1974b) and showed that, though stomatal size was relatively constant over most of the plant’s surface, stomatal frequency tended to vary at different locations. The same point on equivalent leaves should, consequently, be used to record stomatal frequency.

346

P. D. WALTON

Tan and Dunn (1975) were also able to demonstrate that large stomata, long, wide leaves, and high tiller weight were associated with stomatal frequency. In 1976, the same authors found that these characters were under genetic control. In all cases, specific combining ability was much smaller than general combining ability. The narrow-sense heritabilities for both stomatal length (0.69) and stomatal frequency (0.85) indicated that these characters were highly heritable. The authors interpreted their data as indicating that selection for low stomatal number and larger stomatal size would produce progeny with a high tiller weight and long, wide leaves. Tan et al. (1976a) studied the genetics of leaf vein characters and the association of these characters with stomata traits. Their evidence indicated that forage yield increases might be obtained at both the first and second harvest by selecting for increased numbers of vascular bundles per unit area of leaf or sheath width (i.e., narrow interveinal distances). Where interveinal distances are small, the proximity of mesophyll cells to vascular tissue should lead to more efficient translocation, enhanced photosynthetic capacity, and increased crop productivity. Lea et al. (1977a) also studied stomatal diffusion resistance at three ploidy levels. These workers were able to show that stomatal resistance was negatively correlated with total stomatal aperture. Since porometric measurements of stomatal resistance were more readily obtained than microscopic measurements of stornatal size, the authors advocated the use of porometric measurements as a means of examining the large populations essential to a plant breeding program. They pointed out that such measurements would provide an insight into stomatal activity as well as stomatal aperture size. Lea et al. (1977b) used a “stomatal index” (the percentage ratio of the stomatal frequency to the total number of epidermal cells and stomata on the plant’s surface) to study stomatal activity. The advantage of using a stornatal index was that it was unaffected by environmental conditions, whereas stomatal frequency was influenced by the environment through its effect on the growth of leaf blades. Lines that had larger stomatal size but similar stomatal indices would be expected to have higher diffusion rates. C. SELF-FERTILITY

In themain, plants of smooth bromegrass are cross-pollinated. This lack of selffertility is the outcome of two distince causes. First, structural differences in the chromosomes resulting from inversions and translocations have caused meiotic irregularities. Second, further irregularities arise from physiological imbalance which is frequently the outcome of hybridization. Studies of meiotic irregularity, pollen viability, and seed set have been carried out in self- and openpollinated progenies by Jalal and Nielsen (1965). These authors found low seed set to be associated with differences in the chromosome structure, which resulted

CHARACTERISTICS OF Bromus inermis LEYSS

347

in unorientated and lagging chromosomes, bridges, or micronuclei. Such structural differences in the chromosomes were not significantly correlated with pollen viability. Seed set was also uninfluenced by the proportion of stainable pollen. Metabolic or physiological causes appeared to explain anomalous behavior such as prophase pycnosis or stickiness at metaphase 11, since these occurrences were independent of structural differences among the chromosomes. Such factors are closely related to pollen grain abortion, as well as being associated with each other. In some cases, meiotic irregularity is associated with self-fertility. Drolsom and Nielsen (1969) believed that such irregularity also conferred the opportunity for gene interchange and would result in the production of new and possibly desirable combinations. They also showed that the progeny of self-pollinated plants were uniform. Only a small portion of the potential genetic combinations was expressed. Studies carried out by Nielsen and Drolsom (1972) showed that the progeny resulting from diallel crosses were also much more uniform than might be expected from crosses of the highly heterogeneous parents of an octoploid such as smooth bromegrass. Cytological studies of the early stages of meiosis of intergeneric and interspecific hybrids showed that many sporocytes aborted during the early stages. In other plants, nonfunctional pollen grains were shown to undergo successful meiotic divisions. The experimental synthetics produced from a number of bromegrass breeding programs have given disappointing forage yields after being passed through three or four generations of seed multiplications. This is surprising, since the Hardy-Weinberg law indicates that there should be zygotic equilibrium after a single generation of panmixis and that there should be no decline in vigor following the second synthetic generation. Evidently, one of the prerequisites for the Hardy-Weinberg law is not being fulfilled; either mating is not completely at random, or differential selections of zygotes exist. While the influence of natural selection cannot be entirely disregarded, evidence provided by Nielsen and Drolsom (19721, Mishra and Drolsom (1973b), and Pattanyak and Drolsom (1974) supports the hypothesis that nonrandom mating may be responsible for reductions in forage yield following seed multiplication and also for the uniform progeny produced from crossing and selfing.

Ill. SEED PRODUCTION AND ESTABLISHMENT A. FLORALINITIATION

There are few reported studies of floral initiations for smooth bromegrass. Much of the information presented in the literature has been derived by inference

348

P. D. WALTON

from work with other cool-season grasses. Canode et al. (1972) dissected tillers, at weekly intervals, from bromegrass plants grown under field conditions. They were able to show that in Pullman, Washington, on February 12, the first signs of floral development occurred. Clarke and Elliot (1974) obtained similar results in Northern Alberta, leading them to believe that both northern and southern types of bromegrass undergo floral initiation exclusively in the spring. However, Newell (1951), working in Nebraska, was able to show that smooth bromegrass plants did not head when moved from the field in August to long photoperiods in the greenhouse. Some heading took place when the plants were moved on September 15, but moving the plants on November 15 or December 15 produced the most heads. This indicates that fall induction had been extensive. The methods used by Clarke and Elliot (1 974) were such that induction could have taken place in the fall but the morphological manifestation of this was not evident before the following spring. However, Sass and Skogman (1951) showed that floral primordia initiated in the autumn did not, under field conditions, survive the winter, so that for all practical purposes, Clarke and Elliott’s conclusions were sound. In Northern Alberta, plants of smooth bromegrass are fully headed in midJune, pollinated by early July, and mature in mid-August. Under these circumstances, southern ecotypes of smooth bromegrass produce lower seed yields than the northern ecotypes. This contrasts with yield data from Nebraska and other parts of the United States, where the reverse is the case. Clarke and Elliot (1974) claimed that since there were no differences between the two ecotypes in the time or degree of floral initiation, the seed yield differences were due to environmental adaption. This view is supported by Russian workers (Romanova and Vasiliskov, 1974) who found that in extreme northern latitudes, seed set was frequently reduced and some cultivars produced no seed. It is evident that precise information concerning the environmental factors that bring about floral initiation in smooth bromegrass, and the time when they function, is lacking. As is common in most cool-season grasses, existing evidence indicates a need for short photoperiods and cool temperatures. Also, the environmental requirements for induction differ between the various cultivars and terms. B . SEEDYIELD

Working in Washington, Canode (1968) found that a wide row spacing (up to 60 cm apart) gave increased smooth bromegrass seed yields. Spacing between 60 and 90 cm produced neither an increase nor a decrease in seed yield. Fulkerson (1972), using close row spacings (35 and 71 cm apart), presented evidence to show that low density plant populations resulted in high seed yields. The aggressive nature and creeping habit of smooth bromegrass calls for regular interrow cultivation to maintain high seed production.

CHARACTERISTICS OF Bromus inermis LEYSS

349

In an attempt to determine whether seed multiplication in northern areas modified the performance of southern types of smooth bromegrass, Knowles and Christie (1972) compared, at two Canadian locations, the original breeder or foundation seed from Nebraska, New York, and Iowa, with seed produced in western Canada. The authors studied forage yield, height of growth, and seed volume weight, and found that for the Saskatoon site there were no significant differences; such significant differences as were observed in the forage yield of the different seed lots in the Guelph trial were very small. The use of multiplication areas in western Canadian provinces would in no way reduce production potential. Use of seed produced there could not be regarded as being detrimental to forage production in eastern Canadian provinces. C. GERMINATION A N D SEEDLING GROWTH

After three cycles of recurrent selections for high seed set, Trupp and Carlson (1971) achieved a substantial increase in that character, accompanied by an increase in seedling vigor, plant height, and disease resistance. The average realized heritabilities for seed weight were 42% in both the first and second selection cycles. The advantage in seedling vigor which large-seeded lines showed over small-seeded lines diminished as growth and development progressed, but large-seeded plants had a slightly higher forage yield in the first year of harvest. McElgunn (1974) showed that germination was initiated more rapidly where alternating, rather than constant, temperature (2°C for 12 hours, followed by 13°C for 12 hours) conditions prevailed. The resulting growth rate differences did not persist for longer than 6 days after germination started. The Russian workers Kirshin and Shitova (1972) compared seeds of Bromus inermis germinated in the dark with those germinated with exposure to incandescent light for periods varying from 2 to 8 hours. The incandescent light inhibited coleoptile growth. In general, the degree of inhibition was proportional to the length of exposure, but very short periods of illumination had no effect. The authors also showed that a close positive correlation exists between seeding depth and coleoptile size, as well as between the size of the first leaf and the coleoptile length. Variations in both of these characters could be obtained either by varying the planting depth or by excluding light. In general, studies of germination and germination rates with smooth bromegrass indicate a close positive correlation between seed size and seedling vigor, as well as emphasizing the importance of planting depth in forage establishment. Tan et al. (1978b), using artificial growth conditions, demonstrated a highly significant correlation between seedling characters such as tiller number, leaf area, leaf number, shoot-to-root ratio, and seedling dry weight and the relative growth rate of the plant. Since none of these seedling characters were signifi-

350

P. D. WALTON

cantly correlated with the net assimilation rate, which together with the leaf area ratio constitutes the relative growth rate, leaf area ratio and relative growth rate must be regarded as the important contributors to dry matter production. Further work (Tan et al., 1978c) demonstrated a significant and positive correlation between relative growth rate and net assimilation rate (0.72 to 0.98). Thus, the relative growth rate is determined by the net assimilation rate rather than the leaf area ratio, an important consideration in breeding for seedling vigor. The effect of a companion crop on the germination of undersown forage seedlings was studied by Genest and Steppler (1973). Where forages were established without a companion crop, soil moisture percentages were higher. Because, in this experiment, the light intercepted by the weed growth was greater than that intercepted by the companion crop, differences in forage yields observed by these authors reflected the greater advantage of improved soil moisture conditions rather than better light penetration.

IV. FORAGE QUALITY There are few circumstances under which pastures cannot provide a more economical feed for domestic livestock than any other cropping system. However, they do not always meet all the dietary needs of the grazing animals. Consequently, forage quality is a factor of great economic importance. The problem in determining forage quality is that the chemical constituents which are of importance to the grazing animal are difficult to analyze and measure accurately. The wide range of chemical methods that may be used to study forage quality usually measure some characteristic that is of indirect importance in animal nutrition. The methods themselves are frequently far from rapid, and since forage material is exceedingly variable by nature, accurate determinations call for a large number of samples. Consequently, if costs are to be reduced by feeding animals from pasture without losses due to animal ill health, there is a need to develop rapid methods for studying forage quality. A. ANIMAL FEEDING TRIALS

In the absence of satisfactory chemical methods of analysis, animal feeding trials, though expensive, are exceedingly important and informative. Using dairy heifers, Martin and Donker (1968) compared, over a 4-year period, animals’ preference for, and production from, smooth bromegrass and Reed canary grass swards. The heifers’ average daily gain (0.74 kg/ha) was identical after grazing from the two types of pastures. In the absence of choice, the animals’ preference

CHARACTERISTICS OF Bromus inermis LEYSS

35 1

for smooth bromegrass was of little practical significance. In fact, since Reed canary grass was higher yielding and more persistent at the trial location (Minnesota), it was, for those circumstances, to be recommended over smooth bromegrass. The chemical composition and nutritive value of smooth bromegrass was compared with that of timothy and orchard grass (Dactylisglomerutu), when all three were harvested at 50% inflorescence emergence, by Kureger et al. (1969), using dairy goats as test animals. Orchard grass, followed by smooth bromegrass, gave the highest values for digestibility of dry matter, crude protein, and acid detergent fiber. Smooth bromegrass gave the highest crude protein and in vitro dry matter digestibility values and the lowest percentage acid detergent fiber when the plant parts (leaf blade, leaf sheath, and stem) were compared. Calder (1 977) made silage from smooth bromegrass pastures harvested at a range of dates in July and August. He concluded that the largest economic return, but not the highest yield of dry matter, would be obtained when the material was cut at the vegetative stage. The performance of steers fed early-cut silage compared favorably with the results obtained with high energy concentrates. Many beef producers harvest crops intended for silage at the early head stage with a view to obtaining a higher yield; Calder’s evidence indicates that earlier cutting would provide better returns.

B. DIGESTIBILITY Smooth bromegrass was one of the species studied by Wurster et al. (1971) to determine in virro and in vivo digestibilities for a range of forage materials. The in vitro and in vivo dry matter digestibility values were highly correlated (R = 0.89). Comparisons of digestibility values for whole plants and plant parts collected over the growing season showed significant differences between the two smooth bromegrass cultivars Sac and Manchar. The authors believed that this association existed. Plant breeders should be able to select bromegrass strains with highly digestible stems and leaf sheaths. Such cultivars would be of value where it was intended to harvest bromegrass at a time when the dry matter yield was at a maximum (i.e., for hay). On the other hand, selections for high whole plant digestibility values for early growth stages would produce bromegrass strains suitable for pasture. Kunelius et ul. (1974) harvested pastures over a 3-year period, at eight developmental stages prior to the first harvest, and at two intervals during the regrowth. In vitro digestibility percentages declined as the season progressed and were higher when the growth period was shortest. This seasonal decline was most marked during the first year. Acid-pepsin dry matter disappearance was positively associated with leaf width, culm diameter, and dark green color. All were characters that showed a negative correlation with plant height (Sleper and Drolsom, 1974).

352

P. D.WALTON

In general, plant disease reduced forage quality. Gross et al. (1975) were able to demonstrate negative correlation (-0.46) between the percentage of the plant that was diseased and the percentage of in vitro digestible dry matter. Inoculation with Drechslera bromi (Died.) and Rhyncosporiurn secalis (Oud.) produced a significant decrease in in vitro digestible dry matter, but inoculation with bromegrass mosaic virus produced no change in forage quality. Bhat and Christie (1975) harvested a range of bromegrass genotypes at three stages of maturity and subjected the stems to acid detergent fiber, lignin, cellulose, silica, and total cell wall constituent analysis. Top and bottom portions of the stems were analyzed separately. The authors showed that in vitro digestibility values declined at the rate of about 1% per day between the time of head emergence and full head extension. In the period between the beginning of fully extended head stage and anthesis, in vitro digestibility values declined more slowly (0.1 % per day). The lower parts of the plants were 4 to 8% less digestible than the top parts. The bottom part of the plant gave higher percentage values for all cell wall components with the exception of silica. There was also a significant negative correlation between in vitro digestibility and the amounts of all cell wall components. When the authors calculated the phenotypic and genotypic correlations, high positive values were obtained between the percentage lignin and the percentage acid detergent fiber. It seems that the lignin content of the plant can be used as a criterion in a breeding program where selections are aimed at an increase in in vitro digestibility. Scanning electron microscope studies were used by Akin and Burdick (1975) to investigate relative rates of digestion for the various types of tissues found in the leaf blade of a range of tropical and temperate grasses, including Bromus inermis. In all cases, the mesophyll and the phloem degenerated first. The leaf laminas of the temperate-season grasses degenerated much more readily than those of the tropical (C-4) plants, since the vascular bundles occupy a higher percentage of the leaf area in the C-4 species, which explains why temperate grasses are more readily digestible, and hence more nutritious, than the tropical grasses. C. PROTEIN CONTENT

Quantitative and qualitative measurements of the main plant components of smooth bromegrass when grown on irrigated land in the Canadian Prairies were determined by Kilcher and Troelsen (1973). Sequential sampling during a 14week period showed that the highest yield of dry matter was obtained at flowering time and that the proportion of leaves (by weight) decreased to 40% when the plant was mature. The leaves contained 12% more crude protein than did the stem over most of the life of the plant. When the plant was mature, the cell wall

CHARACTERISTICS OF Bromus inermis LEYSS

353

lignin content of the stem was about 70%, whereas that of the leaves did not exceed 60%. At maturity, the digestibility of the leaves and stems declined, respectively, to 57 and 35%. Nutrient energy yields were highest from material harvested for hay during the 2-week period from heading to midbloom. These conclusions were supported by work carried out by Lawrence et al. (197 1) and Winch et a f . (1970). Both studies showed that early harvesting resulted in a substantially higher protein content. Differences in forage quality between a number of species, including smooth bromegrass, have been reported by Tingle and Elliott (1975). These authors did not detect cultivar differences within species. The use of an orange dye binding method to determine protein content was discussed by Smith and Lutwick (1975). Although it was possible to calculate a regression relationship between total nitrogen content as determined by conventional methods and the values obtained from the orange dye binding method, variations around the regression line were high when nitrogen contents were greater than 2.5%. The orange dye binding method is not satisfkctory for the determination of total nitrogen contents of the grasses.

V. FORAGE YIELD A. ENVIRONMENTAL INFLUENCES

I . Harvest Date Two factors are of importance in determining date of harvest. First are quality characteristics, discussed in the previous section, which decline as the season advances. Second is dry matter production, which increases toward the end of the growth period. If the harvest date is early, the productivity and persistence of the pasture may be decreased. In the case of smooth bromegrass, this decrease is not as marked as that found in many other forage species; some authors (Horrocks and Washko, 1968) have found that, whereas productivity was reduced by early harvesting, persistence was in no way affected. Kunelius et al. (1974) studied the effect of cutting management on smooth bromegrass in eastern Canada over a 3-year period. Harvesting prior to heading reduced the forage yield and crude protein production. These results confirmed earlier findings by Winch et al. (1970) and Rochat and Gervais (1975). In western Canada, McElgunn et al. (1972) used a smooth bromegrass and alfalfa sward to test the effect of six defoliation schedules in which the initial cutting date varied. Over a 5-year period, these authors were able to demonstrate that early defoliation was detrimental to yield. The conclusions drawn from

354

P. D. WALTON

these studies do not agree with the results of the animal feeding trials (Calder, 1977) presented in Section IV. This difference could well indicate that the protein analysis methods used by McElgunn et al. were unreliable. Further south, in Wisconsin, Smith et al. (1973, 1974) showed that frequent and severe cutting regimes were capable of substantially reducing the porportion of bromegrass in an alfalfa-bromegrass mixture. Where the pasture was harvested only once or twice during the year, the height of cut did not result in a significant difference in either yield or persistence. These findings in Wisconsin were the reverse of those in western Canada, where repeated cutting or overgrazing eliminates alfalfa from a bromegrass-alfalfa sward (Walton, 1979).

2 . Fertilizers While forage crops are now universally accepted as being vitally important for the maintenance of man’s livestock, and consequently for his welfare, the economics of hay and pasture production are such that they have frequently been relegated to poor or marginal lands. Hence, the possibilities for large yield increases by improving soil fertility and pH levels are enormous. The rates of fertilization commonly used for forage crops in the North American continent are grossly inadequate. Undoubtedly, one reason for inadequate fertilization is that livestock producers tend to underutilize the additional forage that fertilization could provide. Increases in the capital value of land in recent years make it essential that the maximum profit per hectare be obtained by combining the highest possible yields of good quality forage with efficient and full utilization (Morgan, 1971). It is doubtful if the low levels of productivity that are expected from the native rangelands of the Canadian prairie provinces and parts of British Columbia will continue to support economically viable farming systems. A great many workers have shown that nitrogen will consistently increase yields of smooth bromegrass haylands and pastures provided that soil moisture is adequate. This work has been summarized by Wedin (1974). Offutt and Hileman (1972) pointed out that the ranking of cultivars from both northern and southern bromegrass types remained unchanged by applications of nitrogenous fertilizer. Northern ecotypes in Canada showed a smaller response to fertilization than did the southern cultivars. As well as adequate soil moisture, high carbohydrate reserve levels during the fall and winter are essential for maximum responses to nitrogen fertilization in the following year. When fertility is high in the fall and both potassium and nitrogen are present, carbohydrate reserves increase, but high rates of potassium without nitrogen cause a reduction in reserves. In the spring and summer, nitrogenous fertilizer applications result in the utilization of photosynthates for the production of new top growth, and the development of roots and rhizomes for the accumulation of carbohydrate reserves does not take place

CHARACTERISTICS OF Bromus inermis LEYSS

355

(Paulsen and Smith, 1969). Thus, nitrogen fertilizers can reduce the “sodbound” condition of northern bromegrass and has been reported to be more effective than cultivation (Meyer et al., 1977). Meyer et al. (1977) also studied forage production, percentage crude protein, and nitrate nitrogen responses in smooth bromegrass over a 22-year period in North Dakota. They concluded that applications of nitrogen at rates of about 100 kg/ha were essential to obtain economical production of smooth bromegrass. Nitrogen applied at the rate of 66 kg/ha resulted in a 214% increase of yield of dry matter over that of nonfertilized bromegrass. An application of 133 kg/ha resulted in a yield increase 257% higher than the control. The higher rate of nitrogen application would be economical if the additional crude protein that the herbage contained was replacing an expensive protein supplement. Productivity was also increased by repeated annual applications of nitrogen, so that for the last seven years, forage yields and crude protein production were higher than at the beginning of the experiment. It is recommended that in areas where the previous year’s forage production was low and where rainfall was minimal, nitrogen application should be reduced, since it might be expected that there would be a nutrient carry-over. Where the reverse is the case, nitrogen rates should be increased accordingly. Thus, adequate nitrogen fertilization will maintain longterm productivity. Hanson et al. (1978) were able to show that split applications of nitrogen fertilizer (equal parts in the spring and after the first harvest) gave higher yields and led to a larger total nitrogen recovery than did a single application at the beginning of the season. In these trials, conducted under irrigated conditions, both the yield and the percentage recovery from smooth bromegrass were higher than that from the other species tested (Reed canary grass, creeping foxtail). Schou and Tesar (1977) compared applications of anhydrous ammonia with ammonium nitrate on a number of different cool-season grasses. Whereas both forms of nitrogen resulted in similar yield increases, the anhydrous ammonia acted more slowly. Lechtenberg et al. (1974) compared beef production from smooth bromegrass pastures fertilized with anhydrous ammonia with that from pastures fertilized with ammonium nitrate. They agreed that anhydrous ammonia and ammonium nitrate were effective in increasing animal production per hectare. No animal disorders were observed. a. Nitrate Poisoning. Classical symptoms of nitrate poisoning seldom occur until diets contain in excess of 0.35 to 0.45% nitrate-nitrogen, but the animal’s response to nitrate poisoning is influenced by other components of the ration, particularly the availability of carbohydrate. The plant produces nitrates because the first step in protein synthesis involves the use of that substance. Consequently, anything that influences protein synthesis may well result in the accumulation of nitrates in plant tissues. The most common causes are:

356

P. D. WALTON

1. High application of fertilizer or high soil fertility 2. Drought conditions 3. Damage to plant tissue by defoliation as a result of grazing or hail damage

The more frequent use of nitrogen fertilizers in recent years has made it important that the factors which govern nitrate accumulation be well defined. Vanderlip and Pesek (1970) have shown that while nitrate accumulation in the plant increased with rates of application of a nitrogenous fertilizer, the amounts accumulated varied among forage crops. Rates of up to 100 kg of nitrogen per hectare resulted in no serious nitrate accumulation in smooth bromegrass. Similar applications to orchard grass pastures would raise levels of nitrate-nitrogen to 0.74%. The amount of nitrate present in the forage varied considerably in relation to the time of harvesting. Potassium deficiency has also been shown to cause an accumulation of nitrates. The influence of this substance on nitrate accumulation was most variable and depended on the relative levels of both nitrogen and phosphorus. As Vanderlip and Pesek (1970) pointed out, all three major plant nutrients (N, P, and K) affect the nitrate content of forage material. In their experiments, MacLeod and MacLeod (1974) showed that under conditions prevailing in eastern Canada, high rates of a nitrogenous fertilizer (896 kg/ha) could increase the percentage of nitrate in smooth bromegrass to 0.43%. The rate of potash application had no effect on the percentage nitrate present in the herbage. Smith and Lutwick (1975) extended these studies to a range of forage species which were compared at three maturity stages and at four rates of nitrogen fertilizer (0 to 940 kg/ha). All six of the grasses tested could, in the stages prior to heading, accumulate dangerous levels of nitrate in the plant tissue if high dressings of fertilizer were used. Russian wild ryegrass showed the greatest increase in nitrates in response to fertilizer, whereas timothy accumulated the least. Smooth bromegrass was intermediate. b. Hypomugnesemiu. Grass tetany (hypomagnesemia), caused by low levels of magnesium in an animal’s blood serum, occurs when ruminants graze lush spring pasture. The disease appears to be associated with both poor absorption of magnesium in the animal’s intestinal tract and low levels of magnesium in the forage ingested. Thill and George (1975) showed that smooth bromegrass, Kentucky bluegrass, crested wheatgrass, tall wheatgrass, and meadow foxtail were less likely than other species tested to cause gas tetany in ruminants. Under all circumstances, grasses with K+ to (Ca+ Mg+) cation ratios exceeding 2.2 would place animals at a greater risk than forages, such as those listed above, that had lower cation ratios (Gross and Jung, 1978). Thill and George also presented evidence to indicate that the risk of hypomagnesemia in grazing ruminants was greater during periods of temperature fluctuation. A high level of long-chain fatty acids in the herbage was an important factor in increasing incidence of gas tetany

+

CHARACTERISTICS OF Bromus inermis LEYSS

357

(Barta, 1975), since they reduce magnesium availability. The mean percentage of long-chain fatty acids in bromegrass was greater (46%) than in orchard grass (25%). Follett et al. (1975) studied the chemical composition of bromegrass on the Northern Great Plains in relation to their gas tetany hazard. These authors concluded that, whereas the addition of nitrogenous fertilizers increased enormously the forage production potential and consequently the livestock carrying capacity, nitrogen fertilization might result in gas tetany in ruminants. Improved management practices, which might include the oral supplementation of magnesium intake for ruminant livestock, were needed if advantage was to be taken of increased forage production following nitrogen fertilization. 3 . Interaction between Harvest Date and Fertilizers

Smooth bromegrass, in common with other cool-season grasses, has a critical growth stage at which carbohydrate reserves are low and tillers are few (June et al., 1974). This coincides with the time when elongation of the apical meristem has just occurred. Intense defoliation at this point in the plant's life can easily lead to a reduction of the plant population. Paulsen and Smith (1968) showed that smooth bromegrass grown with alfalfa produced higher yields with frequent (five) cuts than with infrequent (three) cuts. The response to these management treatments was reversed when bromegrass was grown by itself, indicating the importance of maintaining fertility levels when pastures are cut frequently. Applications of a nitrogenous fertilizer increased both the rapidity of regrowth and the total yield. Tiller numbers and tiller growth rate both increased so that photosynthetic areas were quickly replaced. Thus, while applications of nitrogen will increase the vegetative yield of smooth bromegrass, as was evident from Section V,A,2, frequent cutting can be detrimental to carbohydrate storage and winter survival. 4 . Soil Temperature and Moisture

Compared with Reed canary grass, smooth bromegrass gives higher yields at the beginning of the season, but shows a poorer yield of regrowth toward the end of the season. Read and Ashford ( 1 968) studied the effect of soil temperature on these two species and found differences in response. Although the yield of both species was reduced at lower soil temperatures, the decrease was greater in Reed canary grass. These reductions took place under high soil fertility and good soil moisture conditions. In both cases the reduction in yield appeared to be the outcome of an inability on the part of the roots to take up nitrogen and phosphorus. The yield response of the two species to different levels of phosphate fertilizer was similar. Baker and Jung (1968) showed that for smooth bromegrass the optimum daytime temperature for top growth was between 18.3" and 24.9"C.

358

P. D. WALTON

Temperatures of 34.8"C, for all the species tested (timothy, orchard grass, and Kentucky bluegrass), gave decreased yields. This decrease was less for smooth bromegrass than for any other species. For all species an increase in night temperature from 18" to 18.3"C decreased the level of carbohydrate reserves. Morrow and Power (1979) conducted trials in which the air temperature was held constant and soil temperature was varied from 3.3" to 33.3"C. The optimum soil temperature for smooth bromegrass aboveground dry matter production was 18.3"C. At this temperature, most of the other grasses (crested wheatgrass, western wheatgrass, Alta wild rye, Russian wild rye, green needlegrass, side oats grama, and blue grama grasses) produced the greatest amount of root dry matter. Waddington (1973) used factor analysis in an attempt to determine the effect of meteorological variables on forage growth in the spring and on regrowth after cutting. The most important variable was precipitation. Snowmelt provided adequate moisture for early spring growth, so that soil nutrients and temperature were the major factors influencing growth at that time. In late spring and early summer, growth was mainly dependent on the rainfall, but temperatures were sometimes low. Temperatures from mid-June to mid-August were usually satisfactory or high for forage growth, so that during this period, production was entirely dependent on rainfall. Yield differences due to climate were much larger than differences between species. B. PLANTMORPHOLOGY

In the past, agronomists have attempted to study yield by dividing it into components. The weight of a given number of grains, the grains per head, and the heads per unit area have all been extensively studied in relation to a number of environment and management factors. Frequently, it was found that a factor that enhanced one yield component was detrimental to all or some of the others. The study of yield components appeared to be an unrewarding way of determining the underlying dependence structure for yield potential. As an alternative, agronomists and plant breeders have studied the associations between morphological characters and yield. As well as elucidating the dependence structure underlying yield, such studies might indicate genetic causes that arise from pleiotropic gene action or indicate changes brought about by natural selection or by selection in a breeding program. A number of statistical techniques, which include simple and multiple correlations (Walton, 1976), (Tan et al., 1976b), stepwise multiple regression analysis (Walton and Murchison, 1979a), path coefficient analysis (Mishra and Drolsom, 1973b; Tan et al., 1977), and factor analysis (Waddington, 1973; Walton, 1974b, 1976), have been used in these studies.

CHARACTERISTICS OF Eromus inermis LEYSS

359

I . Traits Associated with Yield and Their Genetics The single trait that has most frequently been reported to be closely correlated with the forage yield of smooth bromegrass is plant height (Mishra and Drolsom, 1973; Walton and Murchison, 1979a,b; Tan er al., 1976a,b). Mishra and Drolsom (1973) showed that there was a close positive phenotypic correlation between plant height and leaf weight, leaf length, leaf width, culm diameter, panicle length, and the number of spikelets per panicle. These same authors (1972b) also showed that there was a strong positive association between vegetative traits and certain reproductive characters. This is surprising in view of the well-known negative association between forage production and seed yield. Following their pathway analysis study of phenotypic correlations, Mishra and Drolsom (1973) drew attention to the decrease in the number of spikelets per panicle which accompanied an increase in the number of florets. They believed that the two characters underwent simultaneous, mutually exclusive development, competing for the utilization of biosynthetic products. The reverse relationship was evident in the development of leaf width and culm diameter; plants with wide leaves had thick culms. The authors believed that the photosynthates available at the time when these characters were developing would influence the size of both traits. In 1976, Walton, using factor analysis, was able to detect six factors that contributed to the total forage yield of smooth bromegrass in any one year. Ranked in descending order of importance, these were: 1. Tiller size and weight 2. Plant height and yield of the second cut 3. Leaf area 4. Winter survival and first harvest yield 5. Plant height early in the growing season 6. Leaf-stem ratio

This confirms evidence gathered in earlier studies (Walton, 1974a,b). Tan et al. (1976a,b) drew attention to the association between net assimilation rate, crop growth rate, leaf area index, and standard leaf weight and forage yield. They related these characters to vein number, stomata size, and stomata1 frequency. There was evidence to show that plants with a large leaf area, wide leaves, and a high specific leaf weight had higher vein numbers per unit width of leaf and larger, but fewer, stomata. This information is important, for although high forage yield depends on the buildup of labile assimilates, the utilization of these substances in turn depends on the speed of their mobilization and translocation to sink areas. Walton and Murchison (1979a) used tiller characters to predict smooth

360

P. D. WALTON

bromegrass forage yield and determined a plant ideotype for maximum forage yield from this species, when harvested twice a year. For the first and second harvest, “a dense population of nonelongated tillers with a high leaf weight and area, early in the season,” followed by a rapid increase in the number of longstemmed headed tillers would give the most promising yields. The same authors (Walton and Murchison, 1979b) also showed that for all growth stages, tiller weight has positively correlated with leaf area and standard leaf weight. Nonelongated tiller density increased when nonelongated tiller leaf numbers per tiller were high. The density of elongated and headed tillers increased when stems were long, whereas high standard leaf weight, stem weight, and leaf weight reduced tiller density. Tan ef al. (1976a) also showed that tiller density and tiller dry weight were two major components of forage yield. Many research workers, Ross et af. (1970), Mishra and Drolsom (1972a,b), Walton (1974a,b, 1976), Tan ef al. (1976a,b), Timothy et al. (1959), Robinson and Thomas (1963), and Dunn and Wright (1970) have found that most of the morphological traits associated with forage yield in smooth bromegrass give general combining ability mean square values larger than those for specific combining ability. Knowles (1950), Drolsom and Nielsen (1970), Mishra and Drolsom (1972b), Sleper and Drolsom ( 1974), and Tan et al. (1 977) were able to show that for some traits, specific combining ability was also important. The proportion of phenotypic variation that is heritable is measured by calculating the broad-sense heritability (H ,,), while narrow-sense heritability (H ”) measures the proportion of the total genetic variation due to additive genetic variance. The following values have been obtained for total annual yield: Hb

0.68 0.48

HIl 0.25 0.37

Reference Tan er al. (1977) Walton and Murchison (1979~)

Heritability values for a range of plant characters are shown in Table I. The narrow-sense and broad-sense heritability values (Table I) indicate that, while the values for yield itself are low, plant breeders should be able to make substantial progress in changing plant height, leaf length, leaf width, panicle length, stomatal length and frequency, tiller density, sheath vein number, sheath width, and leaf digestibility. Walton (1974a) has drawn attention to discrepancies in size of the additive genetic variance as determined by Griffin’s (1956) analysis and that shown by narrow-sense heritability estimates. Both methods would be influences (in different ways) by the meiotic irregularities reported for this octoploid species, while high ploidy levels themselves will also affect the calculation of these values. Heritability estimates depend on the type of material, experimental design, and the reproductive system involved. However, since these factors influence all characters equally, the relative ranking of the traits is of interest.

36 1

CHARACTERISTICS OF Bromus inermis LEYSS

Table I Heritability Character Leaf characters Leaf width Vein number Vein frequency Interveinal distance Stomatal length Stomatal frequency Leaf number Leaf length Leaf area Leaf angle Leaf rigidity Tiller density Canopy height Sheath characters Sheath width Vein number Vein frequency Interveinal distance Stomatal length Stomatal frequency Panicle characters Culm diameter x lo2 Panicle length x 10 Spikelets/panicle Florets/spikelet x lo2 Panicles/plant x 10 First harvest Forage yield Leaf-blade dry weightkiller Leaf-sheath dry weight tiller Stem dry weighthiller Acid detergent fiber Whole plant Leaf blade Leaf sheath Stem Crude protein Second harvest Forage yield Acid detergent fiber (whole plant) Crude protein Plant height

Hb

H"

0.80 0.84 0.49

0.62 0.68

0.46

;:"0

0.72 0.73 0.75 0.74 0.94 0.48 0.59 0.86 0.88

0.38 0.55 0.14

0.80 0.71 0.48 0.48 0.72 0.59

0.21 0.33 0.30

0.45 0.94 0.59 0.42 0.48

0.27 0.48

0.68 0.94 0.95 0.92

0.25 0.57 0.56 0.46

0.77 0.93 0.85 0.83 0.58

0.04

0.49 0.31 0.16 0.54

Reference

Tan et al. (1976a)

0.49

0.44

Tan er al. (1977)

Tan et al. (1976a)

Mishra and Drolsom (1972b)

0.46 0.19 0

\

Tan et al. (1978a)

0.03 0.22 0.15 0.08 0.44

Walton and Murchison (1979~)

362

P. D. WALTON

2 . Canopy Characteristics The canopy architecture of a forage crop influences efficient light utilization, photosynthetic activity and, consequently, crop productivity. Tan el a f . ( 1977) used pathway coefficient analysis to demonstrate that the two canopy characters that showed the closest positive association with the spring forage yield in smooth bromegrass were the leaf area per tiller and tiller density. There was a negative direct effect between yield and other characters such as leaf number, leaf length, leaf angle and rigidity, and canopy height. Leaf area per tiller and tiller density exerted opposite influences in most of the association pairs measured by path coefficient analysis. Those working with other forage crop species have also been confronted with this dilemma. Rhodes (1972), following a suggestion by Cooper and Edwards (1961), successfully used selection for high critical leaf area index as the means of overcoming the problem. The relationship between leaf area per tiller and tiller density reached its own equilibrium under such a selection program. Teare (1972), who used a formula that included leaf area index, leaf angle, and canopy height to measure the relationship between these characters and the attenuation of radiant energy in smooth bromegrass, found that the best single component to predict light attenuation was forage area per unit of soil surface (i.e., leaf area index). His data supported the hypothesis that a tall, erect growth habit (with leaves distributed evenly along toe culm) is conducive to uniform illumination of leaf area within the forage canopy and this, in turn, would lead to high yields. 3 . Genotype Interaction with the Environment

Perennial forage grasses are expected to yield well under climatic and management conditions that are much more diverse than those provided for any other crop. These broad expectations make the nature of genotype by environment interaction particularly important. Since this interaction is generally accepted as being large, it is surprising that few studies have been undertaken to quantify, for the grasses, the magnitude of their genotypic interaction with environment. An exception exists in the case of smooth bromegrass. Tan et al. studied the effect of genotype by environment interaction on forage yield (1979a) and on morphological characters associated with yield (1979b). These trials were carried out at four locations in the province of Alberta, Canada, using a seven-parent diallel without reciprocals. For most of the characters studied, specific and general combining ability gave highly significant interactions with environments. Evidently, there was a differential expression of gene action at the four locations. Tan et al. (1979~)further partitioned genotype by environment interaction into heterogeneity among regressions and the residual, thus enabling environments to

CHARACTERISTICS OF Bromus inermis LEYSS

363

be assessed in two ways: first, as a mean expression of all genotypes and second, as the average performance of several parental genotypes. The first method showed that, for all the characters studied, a substantial part of the interaction was derived from heterogeneity among regression lines. However, since the residual components were frequently significant, some unpredictable variation was present. The linear model would, however, be predictive of the behavior of characters such as tiller density, yield per area, and second harvest yield, since tests of heterogeneity for the regression lines were highly significant. Considering next the second approach, when environments were assessed using the parental means, it was found that the levels of significance differed considerably from those determined by the first method. However, the ranking of genotypes on the basis of their linear regression coefficients was similar, so that the way in which the environment was measured did not affect the conclusions that might be drawn from the regression data. For all characters, a significant part of the genotype by environment interaction was due to the heterogeneity of the regression lines. However, the predictability of the genotypes varied with the characters measured. It is evident that the various forms of regression analysis provide powerful tools which can elucidate complex genotype by environment interactions and transform them into a series of predictable linear responses. The most sophisticated of these methods, the Eberhardt and Russell (1966) modification of the Finlay and Wilkinson (1963) model, generates three parameters: (1) the regression coefficient, (2) the deviation from the mean square value, and (3) cultivar performance. The problem of determining the weighting to be attached to each of these statistics has not been solved.

IV. PLANT BREEDING Many of the studies and publications discussed in the previous sections were conducted and written to present basic information from which breeding programs could be developed. Consequently, the consideration of possible plant breeding programs for smooth bromegrass will summarize and draw conclusions from the material already presented. The forage breeder is interested in obtaining agronomic improvement in three areas: (1) seed yield, (2) dry matter yield, and (3) forage quality. These three topics will be considered in that order. A. SEEDYIELD

Knowles and Ghosh (1968) studied the isolation distances necessary to prevent interpollination in Bromus inermis by using a genetic marker; they showed that

364

P. D.WALTON

for isolation distances of 1, 61, and 183 meters, the average contamination of their plots was 9.6, 1.0, and 0.2%, respectively. Thus, the isolation distances required for seed multiplication under the Canada Seed Act are marginal to maintain cultivar purity. However, the extent of contamination could be reduced considerably if borders were removed and discarded prior to harvest. The possibility of using plant breeding techniques to improve both seed yield and seed quality in smooth bromegrass was considered by Knowles et al. (1970). Under the environmental conditions prevailing in Saskatoon where they worked, Knowles and associates were able to show that for southern bromegrass strains, there was a marked response to selection for higher seed yields. In the selected progenies, aneuploid plants (2n = 55) had been eliminated, accounting for much of the seed yield improvement in the selection lines over the base population. Progenies of crosses between northern and southern bromegrass types also yielded promising material when selected for improved seed set. Almost invariably, seed set was higher in the progeny of southern bromegrass types that had been outcrossed with northern cultivars. In no case did a decline in forage yield accompany improved seed set. Heritability for seed yield and seed quality were high, and the authors believed that continued mass selection would prove successful in improving these characters. The evidence discussed in Section II1,C indicates that selection for seedling vigor in smooth bromegrass could well be achieved by screening breeding material for high seed weight. Considered from the viewpoint of long-term research, plant breeding methods provide the most promising approach to increased seed production. For the seed producer, however, an adequate level of soil fertility, wide row spacings, and protection from insects and diseases should prove most satisfactory in achieving high seed production. B . DRYMATTERYIELD

Mass or recurrent selection for dry matter yield is a simple approach that has been used successfully in many grass breeding programs. When practiced in Edmonton, Alberta, it was found that the repeatability of yield results from year to year was poor; clones that gave high yields in one year ranked poorly in the next. Evidently, a more sophisticated approach is required. The poor repeatability may be explained partially by the low narrow-sense heritability values (about 0.25) obtained for forage yield, when expressed as dry matter (Tan et al., 1977, 1978a,b,c; Walton, 1974a; Walton and Murchison, 1979c; see Table I). Also, both additive and nonadditive genetic variances have been shown to be of importance for first, second, and total forage harvest yield (Walton, 1974a; Tan et al., 1977, 1978a,b,c; Walton and Murchison, 1979~).Further, in some cases (Tan et al., 1979a), these two types of genetic variance have been shown to be of equal importance. In addition, Tan et al. (1979a) showed that the interaction between

CHARACTERISTICS OF Bromus

inermis LEYSS

365

the environment and both general and specific combining ability was highly significant. All these influences would impede direct selection for yield. Such difficulties could be overcome by using a recurrent selection program combined with testing at a number of locations. Another way of obtaining consistently high-yielding strains by direct selection would be to divide the area for which the new cultivars were intended into ecological zones and develop a plant breeding program for each region, but this would be costly. Instead of making direct selections for forage yield, the use of yield-related characters or yield components shows promise. This view is supported by the many publications (listed in Section V,B) that show general combining ability is high for morphological traits and by the results of the pathway analysis conducted by Tan et al. (1979b). The selection for leaf area, tiller weight, and tiller density may well give more substantial yield increases than direct selection for yield. However, selection for large leaf area, high dry weight, values per tiller, and high tiller density may pose some difficulties, since tiller density was negatively associated with leaf area (-0.46) and tiller dry weight (-0.61). Large populations should be studied to determine if these opposing characters are separable. If a choice had to be made, using existing information, the emphasis should be placed on tiller density (Tan et al., 1976a; Walton, 1976; Walton and Murchison, 1979a,b,c). Yield increases might also be achieved by simultaneous selection for canopy characters, which have been shown by Tan et al. (1977) to be closely correlated with yield. Such characters may also be manipulated by management practices and other environmental influences. Also, in some cases, the narrow-sense heritability values are low for canopy characters (leaf angle = 0.06). The character most closely associated with forage yield is plant height (Walton, 1974a,b; Tan et al., 1976a,b); however, this character is also much influenced by environment (Tan et al., 1979b) and, hence, unsuitable for selection purposes. Of the plant characters studied, those that provide the most satisfactory alternative for selection purposes to direct forage yield, are the stomata and vein characters discussed in Section II,B,2. For both the stomata (Walton, 1974b) and vein (Tan et al., 1976a) characters studied, the nonadditive genetic variances were substantial. Under these circumstances, hybrid cultivars (two-clone synthetics) should provide a satisfactory expression of the desired traits. Hybrid material of this type could be simply adapted to a range of ecological zones of the type suggested earlier. C. FORAGEQUALITY

It is only in very recent years that plant breeders have given attention to forage quality. A six-clone diallel cross was used to investigate the inheritance of in vitro digestibility by Ross et al. (1970). A diallel cross analysis of data collected

366

P. D.WALTON

in two seasons showed that additive genetic effects were present and that general combining ability was a significant source of variation. Sleper et al. (1973) used the acid-pepsin dry matter disappearance technique to study the heritability of forage digestibility. General combining ability was highly significant, again indicating that additive gene action was more important than nonadditive gene action in the inheritance of this trait. These authors agreed with Bhat and Christie (1975) in concluding that significant progress could be made by selecting for high digestibilities within bromegrass populations. None of these findings, however, was in agreement with the results obtained by Kamstra et al. (1973), who was able to show that two cloned synthetics, which had been selected for high and low in v i m and in vivo digestibilities, showed no significant differences for those characters. Also, Christie (1977) found that while the correlation between parents and progeny for in vitro digestibility values was high (0.55),one cycle of phenotypic selection was ineffective in improving digestibility. Most recently, Tan and associates (1978a) confirmed earlier findings which showed that general combining ability was a significant source of the genetic variation for digestibility in bromegrass. These authors also found that the acid detergent fiber and the crude protein contents of whole plants showed nonadditive genetic effects, while leaf blade acid detergent fiber was the only character that gave a high narrow-sense heritability. With that exception, narrow-sense heritability values for quality characteristics were all less than 10%.Under these circumstances, advances due to genetic selection would be slow. It is not surprising that while the genetic evidence indicates that selection for improved digestibility in smooth bromegrass should be possible, such attempts as have been reported in the literature have not been successful. A number of authors attempted to overcome this difficulty by using morphological traits, rather than chemical analysis, as criteria for selection (Bhat and Christie, 1975; Sleper and Drolsom, 1974; and Christie and Mowat, 1968). Digestibility, stem diameter, and leaf width have been reported to be positively correlated, while negative correlations have been found between plant height and digestibility. Tan er al. (1976a) believed that wide interveinal distances could increase digestibility. There are, however, no reports of the approach being used in a plant breeding program. Obviously, the improvement of nutritional value by plant breeding methods introduces fundamental problems that still remain to be resolved. Possibly the most serious of these is the lack of an accurate and rapid method of determining forage quality. Shenk (1977) has drawn attention to the value of using a computerized system of spectroscopy, pointing out that this method has the capability of analyzing large numbers of samples for multiple quality factors. Such a system would be capable of simultaneously determining crude protein, acid detergent fiber, lignin, cellulose, and nonstructural carbohydrates. While such techniques may be of considerable value, the plant breeder is still faced with the problem of relating the

CHARACTERISTICS OF Bromus inermis LEYSS

367

quality characteristics that he has studied to yield and yield-related traits, which also form a part of his breeding program. In the case of long-lived perennial plants which must be evaluated over a number of seasons and which are markedly influenced by environmental factors in both time and space, this can be a formidable task. It is, however, a problem that both plant and animal breeders have considered for many years. In view of the low heritabilities encountered in forage quality characteristics, it is advisable that this be taken into account in weighing the different traits for which selection is to be made. This is, in fact, the process of developing a selection index for a series of traits weighted for differences in their heritabilities. Computer models, which have been developed for this purpose (Shenk, 1975), could be used on data from progeny tests of large populations of parental clones evaluated for forage yield and quality over a range of locations, harvests, and generations.

REFERENCES Akin, D. E., and Burdick, D. 1975. Crop. Sci. 15, 661-668. Armstrong, K. C. 1973. Can. J . Genet. Cytol. 15, 427-436. Armstrong, K. C. 1977. Z. Pjlanzenzuecht. 79, 6-13. Baker, B. S., and Jung, G. A. 1968. Agron. J. 60, 155-158. Barta, A. L. 1975. Crop. Sci. 15, 169-171. Bhat, A. N., and Christie, B. R. 1975. Crop. Sci. 15, 676-679. Calder, F. W. 1977. Can. J . Plant Sci. 57, 441-449. Canode, C. L. 1968. Agron. J . 60, 263-267. Canode, C. L., Anwar Maun, M., and Teare, I. D. 1972. Crop Sci. 12, 19-22. Christie, B. R. 1977. Can. J . Plant Sci. 57, 57-68. Christie, B. R., and Mowat, D. N. 1968. Can. J . Plant Sci. 48, 67-73. Clarke, J. M., and Elliott, C. R. 1974. Can. J . Plant Sci. 54, 475-477. Cooper, J. P., and Edwards, K. J. R. 1961. Heredity 16, 63-82. Drolsorn, P. N., and Nielsen, E. L. 1969. Crop Sci. 9, 710-713. Drolsorn, P. N., and Neilsen, E. L. 1970. Crop Sci. 10, 17-18. Dunn, G . M., and Wright, J. A. 1970. CropSci. 10, 56-58. Eberhart, S. A,, and Russell, W. A. 1966. Crop Sci. 6, 36-40. Elliott, F. C. 1949. Agron. J . 41, 298-303. Elliott, F. C., and Love, R. M. 1948. Agron. J . 40, 335-341. Finlay, K. W., and Wilkinson, G. N. 1963. Aust. J. Agric. Res. 14, 742-754. Follett, R. F., Power, J. F., Grunes. D. L., Hewes, D. A., and Mayland, H. F. 1975. Agron. J . 67, 819-824.

Fuehring. H. D., Mazaheri, A,, Bybordi, M., and Khan, A. K. S. 1966. Agron. J . 58, 195-198. Fulkerson, R. S. 1972. Can. J . Plant Sci. 52, 613-618. Genest, J., and Steppler, H. 1973. Can. J . Plant Sci. 53, 285-290. Ghosh, A. N.. and Knowles, R. P. 1964. Can. J . Genet. Cyrol. 6, 221-231. Griffin, B. 1956. Aust. J. Biol. Sci. 9, 463-493. Gross, C. F., and Jung, G. A. 1978. Agron. J . 70, 397-403. Gross, D. F., Mankin, C. J., and Ross, J. G. 1975. Crop Sci. 15. 273-275. Hanna, R. M. 1961. Can. J. Botany 39, 757-773.

368

P. D. WALTON

Hanson, C. L., Power, I . F., and Erickson, C. I. 1978. Agron. J . 70, 373-375. Hill, H. D., and Carnahan, H. L. 1957. Agron. J . 49, 449-452. Horrocks, R. D., and Washko, J. B . 1968. Pa. Agric. Exp. Sm. Bull. 753, 22. Jalal, S. M., and Nielsen, E. L. 1965. Crop Sci. 5, 401-403. Jung, G. A,, Balasko, I . A., Alt, F. L., and Stevens, L. P. 1974. Agron. J . 66, 517-521. Kamsua, L. D., Ross, J. G., and Ronning, D. C. 1973. Crop Sci. 13, 575-576. Kilcher, M. R., and Troelsen, J. E. 1973. Can. J . Plant Sci. 53, 767-771. Kirshin, I. K., and Shitova, L. G. 1972. Sov. J . Ecol. 2, 110-1 15. Knowles, R. P. 1950. Sci. Agr. 30, 275-302. Knowles, R. P., and Christie, B. R. 1972. Agron. J. 64, 801-804. Knowles, R. P., and Ghosh, A. N. 1968. Agron. J. 60, 371-374. Knowles, R. P., Cooke, D. A,, and Buglass, E. 1970. Crop Sci. 10, 539-542. Krueger, C. R., Hamilton, R. I., Schroll, J. M., and Baumgardt, B. R. 1969. Agron. J . 61, 659-663. Kunelius, H. T., Macleod, L. B., and Calder, F. W. 1974. Can. J. Plant Sci. 54, 55-64. LaFleur, T. D., and Jalal, S. M. 1972. Cytologia 37, 747-757. Lawrence, T., Warder, F. G., and Ashford, R. 1971. Can. J . Plant Sci. 51, 41-48. Lea, H. Z., Dunn, G. M., and Koch, D. W. 1977a. Crop Sci. 17, 91-93. Lea, H. Z . , Dunn, G. M., and Koch, D. W. 1977b. Crop Sci. 17, 669-670. Lechtenberg, V. C., Rhykerd, C. L., Mott, G. 0.. and Huber, D. A. 1974. Agron. J. 66, 92-97. McElgunn, J. D. 1974. Can. J . Plant Sci. 54, 265-270. McElgunn, J. D., Heinrichs, D. H., and Ashford, R . 1972. Can. J . Plant Sci. 52, 801-804. MacLeod, L. B . , and MacLeod, J. A. 1974. Can. J . Plant Sci. 54, 331-341. Martin, G. C., and Donker, J. D. 1968. Agron. J . 60, 703-705. Meyer, D. W . , Carter, J. F., and Vigil, F. R. 1977. N.D. Agric. Exp. Sin. 34, 13-17. Mishra, S. N., and Drolsom. P. N. 1972a. Crop Sci. 12, 389-391. Mishra, S. N., and Drolsom, P. N. 1972b. Crop Sci. 12, 497-499. Mishra, S. N., and Drolsom, P. N. 1973. J..Agric. Sci. 81, 69-76. Morgan, N. D. 1971. Proc. Fert. Prod. Market. Conf. 69-72. Morrow, L. A., and Power, J. F. 1979. Agron. J . 71, 7-10. Newell, L. C. 1951. Agron. J . 43, 417-424. Nielsen. E. L. 1951. Bor. Gaz. 113, 23-54. Nielsen, E. L., and Drolsom, P. N. 1972. Euphytica 21,90-96. Nielsen, E. L., Drolsom, P. N., and Voigt, P. W. 1969. Crop. Sci. 9, 785-787. Offut, M. S., and Hileman, L. H. 1972. Agric. Exp. Sin. Univ. Arkansas Bull. 776. Pattanayak, C. M., and Drolsom, P. N. 1974. Euphyrica 23, 479-484. Paulsen, G. M., and Smith, D. 1968. Agron. J . 60, 375-379. Paulsen, G. M., and Smith, D. 1969. Crop Sci. 9, 529-534. Read, D. W. L., and Ashford, R. 1968. Agron. J . 60, 680-682. Rhodes, 1. 1972. J . Agric. Sci. 78, 509-51 I . Robinson, L. R., andThomas, H. L. 1963. Crop. Sci. 3, 358-359. Rochat, E., and Gervais, P. 1975. Naturaliste Can. 102, 89-97. Romanova, L. V., and Vasiliskov, V. F. 1974. Soviet Plant Physiol. 21, 285-290. Ross, J. G.,Bullis, S . S., and Lin, K. C. 1970. Crop. Sci. 10, 672-673. Sass, J. E., and Skogman, J . 1951. Iowa State J . Sci. 25, 513-519. Schou, J. B . , and Tesar, M. B. 1977. Agron. J . 69, 440-446. Schulz-Schaeffer, J. 1960. J . Hered. 51, 269-277. Shenk, J. S. 1975. Agron. J . 67, 237-240. Shenk, J. S. 1977. J . Dairy Sci. 60, 300-326. Sleper, D. A., and Drolsom, P. N. 1974. Crop Sci. 14, 34-36.

CHARACTERISTICS OF Bromus inermis LEYSS

369

Sleper, D. A., Drolsom, P. N., and Jorgensen, N. A. 1973. Crop Sci. 13, 556-558. Smith, A. D., and Lutwick, L. E. 1975. Can. J. Plant Sci. 55, 573-577. Smith, D., Jacques, A. V. A,, and Balasko, J . A. 1973. Crop Sci. 13, 553-556. Smith, D., Rohweder. D. A,, and Jorgensen, N. A. 1974. Agron. J. 66, 817-819. Smith, J. D., and Knowles, R. P. 1973. Can. J. Plant Sci. 13, 332-334. Tan, G. Y . , and Dunn, G. M. 1973. Crop Sci. 13, 332-334. Tan, G. Y . , and Dunn, G. M. 1975. Crop Sci. 15, 283-286. Tan, G. Y . , and Dunn, G. M. 1976. Crop Sci. 16, 550-553. Tan, G. Y . , and Dunn, G . M. 1977a. Can. J . Genet. Cytol. 19, 531-536. Tan, G. Y . , and Dunn, G. M. 1977b. Cyrologia 42, 547-551. Tan, G. Y.,Tan, W. K., and Walton, P. D. 1976a. Crop Sci. 16, 722-724. Tan, G. Y . , Tan, W. K., and Walton, P.. D. 1976b. Pflanzenzuecht. 77, 339-346. Tan, W. K., Tan, G . Y . , and Walton, P. D. 1977. Crop Sci. 17, 7-10. Tan, W. K., Tan, G. Y . , and Walton, P. D. 1978a. Crop Sci. 18, 119-121. Tan, G. Y . , Tan, W. K., and Walton, P. D. 1978b. CropSci. 18, 133-136. Tan, G. Y . , Tan, W. K., and Walton, P. D. 1978~.Crop Sci. 18, 601-604. Tan, W. K., Tan, G. Y., and Walton, P. D. 1979a. Can. J . Genet. Cytol. 21, 57-63. Tan, W. K., Tan. G. Y . , and Walton, P. D. 1979b. Can. J. Genet. Cyrol. 21, 73-80. Tan, W. K., Tan, G. Y . , and Walton, P. D. 1979~.Crop Sci. 19, 393-396. Teare, 1. D. 1972. Phyton 29, 37-42. Thill, J. L., and George, J . R. 1975. Agron. J. 67, 64-68. Timothy, D. H.,Thomas, H. L., and Kernkamp, M. K. 1959. Agron. J . 51, 252-255. Tingle, J . N . , and Elliott, C. R. 1975. Can. J. Plant Sci. 55, 271-278. Trupp, C. R., and Carlson, I. T. 1971. Crop Sci. 11, 225-228. Vanderlip, R. L., and Pesek, J. 1970. Agron. J . 62, 491-496. Waddington, J. 1973. Can. J . Plant Sci. 53, 309-316. Waggoner, P. E. 1969. Crop Sci. 9, 315-321. Walton, P. D. 1974a. Can. J. Plant Sci. 54, 743-747. Walton, P. D. 1974b. Can. J . Plant Sci. 54, 749-954. Walton, P. D. 1976. 2. Pflanzenzuecht. 77, 43-55. Walton, P. D. 1979. Agric. Forest. Bull. 2(3), 20-22. Walton, P. D., and Murchison, C. 1979a. Euphytica 28, 801-806. Walton, P. D., and Murchison, C . 1979b. Genet. Agr. 33, 341-354. Walton, P. D., and Murchison, C. 1979~.A . Pflanzenzuecht. 3, 84, 35-41. Wedin, W. F. 1974. “Forage Fertilization.” pp. 95-1 18. American Society of Agronomy, Madison, Wisconsin. Wilton, A. C. 1965. Can. J. Bot. 43,723-730. Winch, J. E . , Sheard, R . W., and Mowat, D. N. 1970. Br. GrasslandSoc. J . 25,44-52. Wright, M. 1.. Jung, G. A , , Brown, C. S., Decker, A. M.,Vamey, K. E., and Wakefield, R. C. 1967. W . Va. Agric. Exp. Stn. Bull. 554T. Wurster, M. J . , Kamstra, L. D., and Ross, J . G . 1971. Agron. J . 63, 241-245.

This Page Intentionally Left Blank

Index A Acetylene reduction method, of nitrogen fixation1 measurement, 180- 183 Aerobic layer, of flooded soil, nitrogen fixation in, 160-161 Aggregation, of soil, tillage effects on, 44-45 Agrotechnology, transfer of development, 322-323 research on, 316-323 role in soil survey, 306-307 tropical, 303-339 verification, 323-332 worldwide, 332-336 Allelopathy, in grass-legume associations, 230-23 1 Ammonia, loss from plants, 288-290 Anaerobic layer, nitrogen fixation in, 161-163 Andepts, taxonomy of, 210 Annuals, nitrogen loss from tops of, 271-273 Azolla-Anabaena association, nitrogen fixation by, 175-176

6 Barley, tops, nitrogen loss from, 270-271 Biosystematics, of crop improvement, 113-1 18 Birds, role in plant-nitrogen loss, 284 Bridge crosses, involving wild species, 122-123 Bromus inermis Leyss agricultural use of, 342-343 animal feeding trials or, 350-351 digestibility of, 351-352 dry matter yield of, 364-365 environmental influences on, 353-354 fertilizers for, 354-357 harvest date and, 357 as forage, 350-363 breeding for, 365-367 yield, 353-363 plant breeding of, 363-367 plant character of, 34 1-342 plant morphology of, 358-363 ploidy levels of, 343-345 production characteristics of, 341 -369 371

protein content of, 352-353 seed production by, 347-350 breeding effects, 363-364 self-fertility of, 346-347 soil temperature and moisture effects on, 357 species characteristics of, 343-347 stomata of. 345-346

C Chromosomes, manipulation of, in wild species, 123-125 Coastal zones, wetlands of, nitrogen fixation in, 154-159 Crops earliness of, in wild species, 138 performance data for, 335-336 potassium requirement by, 59-1 10 quality improvement by wild species, 137138 sterility in, from wild species, 138-139 wild species improvement of, 111-147 yield improvement by wild species, 136-137 Conservation tillage systems, 1-58 chemical effects of, 48-49 crop yields and quality factors in, 13-16 definition of, 2 development in U S , 2-4 economics of, 49-51 environmental aspects of, 16-30 equipment and use of, 6-12 historical aspects of, 5-6 infiltration and water conservation by, 30-35 insect disease and, 38-39 microbial activity and, 48-49 plant diseases and, 39 residue yields and, 15-16 seedbed preparation for, 10-12 soil structure and, 43-48 soil temperature and, 39-43 survey on, 3-4 water erosion and, 21-28 weed control by, 35-38 wind erosion and. 17-21

312

INDEX

D Dinitrogen, losses from plants, 290-291

E Economics, of conservation tillage systems, 49-5 1 Energy, source of for nitrogen fixation, 174-175 Energy metabolism, potassium metabolism in, 83-85 Environment, conservation tillage systems and, 16-30 Erosion, tillage systems in control of, 16-30 Evaporation, of soil water, 33-35

F Flooded soil, nitrogen fixation in, 149-192 Flowers, nitrogen loss by, 283 Foliage, in grass-legume associations, 239-241 Forage, Bromus inermis as, 350-363 Freshwater ecosystems, nitrogen fixation in, 159- 160 Fruits, nitrogen loss by, 283

G Germplasm, collection and preservation of, 113-114 Gossypium spp., wild species hybridization of, 126- 127 Grains nitrogen loss from tops of, 264271 seedbed preparation for, 11 yields of, tillage factors in, 13 Grass-legume association, 227-261 allelopathy in, 230-231 competitive aspects of, 227-261 for light, 249-250 for nutrients, 250-253 for water, 253-254 foliage architecture in, 239-241 light requirements by, 237-238 morphological aspects of, 239-248 in pasture community, 229-23 1 periods and rates of growth of, 235-236 physiological aspects of, 231-239 rhizobium symbiosis and N transfer in, 232233

root-cation exchange capacity in, 238-239 root morphology of, 247-248 vesicular-arbuscular mycorrhizae in, 233-235 water use by, 238 Grasses evolution of, 229-230 seedbed preparation for, 12 Guttation, plant nitrogen loss by, 288

H Herbicides, weed control by, 37-38 Hybridization barriers to, 115-117 involving wild species barriers to, 115-117 utilization, 118-119

I Insects control of, tillage systems and, 38-39 role in plant-nitrogen loss, 284 wild species’ resistance to, 135-136

K Kjeldahl method for nitrogen, 296-297

L Land evaluation, description of, 334-335 Leaching, plant nitrogen loss by means of, 285-288 Leaves nitrogen fixation on, 170-174 nitrogen loss by, 283-284 Legume-grass association, competitive aspects of, 227-261 Legumes evolution of, 229-230 seedbed preparation for, 12 Light competition for, in grass-legume associations, 249-250 effect on nitrogen fixation, 177-178 requirement for. in grass-legume pastures, 231-238

373

INDEX

M Microorganisms plant nitrogen loss by means of, 285 tillage effects on, 48-49 Minerals, effect on nitrogen fixation, 177 Maize tops, nitrogen loss from, 271 Minerals, tillage system effects on, 14-15 Mycomhizae, vesicular-arbuscular, in grasslegume pastures, 233-235 N Nicofiunu spp., wild species hybridization of,

128- 130 Nitric oxide, loss from plants, 291-292 Nitrogen fixed, contribution to plant nitrogen requirement, 183-185 loss, from plant tops, 263-302 methodology problems, 293-298 pathways, 279-293 transfer, in grass-legume associations, 232233 Nitrogen-15, nitrogen fixation measurement by, 180-183 Nitrogen dioxide, loss from plants, 291-292 Nitrogen fixation in flooded soil, 149-192 aerobic layer, 160-161 anaerobic layer, I6 1- 163 environmental factors, 174-175 leaf and stem surface, 170-174 measurement, 180- 183 root zone, 163-170 inorganic nitrogen inhibition of, 175-177 Nutrients, competition for, in grass-legume pastures, 250-253

0 Oxygen, effect on nitrogen fixation, 178

P Particulates, plant nitrogen loss by means of, 284-285 Pasture species, nitrogen loss from tops of, 271-276 Pastures, grass-legume association in, 229-23 1

Perennials, nitrogen loss from tops of, 273 Pergelic soils, taxonomy of, 210-21 1 pH, effect on nitrogen fixation, 179 Plant(s) gaseous losses from, 288-293 protein, tillage effects on, 14-15 tops, nitrogen loss from, 263-302 Plant diseases tillage effects on, 39 wild species’ resistance to, 135-136 Pollen, nitrogen loss by, 283 Potassium application of, relation to crop growth, 91-103 yield components, 98-101 in cell turgor and water economy, 8 1-83 in crop production, 59-110 in energy metabolism, 83-85 in long-distance transport, 85-91 in plant physiology, 74-91 in soil, 60-74 transport, across biological membranes, 74-8 I

R Redox potential, effect on nitrogen fixation, 178- 179 Reduced tillage systems, description and use of, 8-10 Rhizobia, symbiosisof, in grass-legume association, 232-233 Rice nitrogen fixation in, 173 tops, nitrogen loss from, 268-269 Rice fields, nitrogen fixation in, 152-154, 164-165 Roots cation exchange capacity of, in grass-legume pastures, 238-239 growth of, tillage system and, 16 morphology of, in grass-legume associations, 247-248 role in plant nitrogen losses, 280-281, 291 Row crops, seedbed preparation for, 11-12 Runoff water infiltration and, 30-32 quality of, 29-30 Rye, tops, nitrogen loss from, 270

314

INDEX

S

T

Saccharum spp.. wild species hybridization of, Tillage systems, conservation t y p , 1-58 130-13 1 Tops, of plants, nitrogen losses from, 263-302 Salinity, effect on nitrogen fixation, 179-180 Triticale, development of, 140 Sea grass, nitrogen fixation by, 165, 166, 168 Triricum spp., wild species hybridization of, 133 Soil(s) Tropical soils, taxonomy of, 208-210 classificationof, needs for, 309-310 Tropics, agrotechnology transfer in, 303-339 flooded,nitrogen fixation in, 149-192 information and data banks for, 336 U potassium availability in, 60-74 assessment, 70-7 1 United States, soil taxonomy of, 193-226 factors affecting, 6 4 7 0 Universal soil loss equation (USLE), 22 potassium fractions in, 60-64 root soil interactions, 71-74 V resource inventories of, 333-334 structure, tillage effects on, 43-48 Vertisols, taxonomy of, 21 1 temperature, tillage effects on, 39-43 Soil Taxonomy w agrotechnology transfer and, 3 15-3 I6 Water application of criteria for, 203-205 competition for, in grass-legume pastures, categories of, 312-313 253-254 complexity of, 202-203 conservation of, tillage systems for, 30-35 diagnostic criteria for, 21 1-212 Water erosion future of, 314-315 control of, 22 international impact of, 213-222 tillage methods for, 22-28 objectives, 218-220 factors affecting, 21-22 soil classification, 213-215 Wheat tops, nitrogen loss from, 264-268 soil maps, 216-218, 220-222 Weeds soil surveys, 215-216 control by tillage systems, 35-37 international use of, 197-202, 332-333 herbicide control of, 37-38 infrequent use, 201-202 rotation control of, 38 as primary system, 198-199 Wild species as secondary system, 200-201 crop improvement by, 1 11-147 personnel training for use of, 205-207 approaches to, 119-126 problems for users of, 202-203 bridge crosses, 122-123 significance of family category in, 313-314 chromosome manipulations, 123- 125 taxonomic problems of, 207-213 direct, 120-122 of United States, 193-226 examples. 126-135 use in agrotechnology transfer, 303-339 physiological manipulations, 125- 127 quantitative varification, 323-332 uses, 135-140 Solanum spp., wild species hybridization of, genotype buffering in, 117-1 18 131-133 hybridization barriers of, 115- I i7 Sorghum, tops, nitrogen loss from, 270 Wind erosion Species, concepts of, 114-1 15 control of, 18-21 Stem, nitrogen fixation on, 170-174 tillage methods for, 19-20 Stomata, of Bromus inermis, 345-346 equation for, 18 Strawberries, cultivated, 140 Strip tillage, description and use of, 10 z Stubble mulch tillage, description and use of, 6-8 Zea spp., wild species hybridization of, 134-135