Advances in Agronomy, Volume 15

Harcout, Corporate ADVANCES IN AGRONOMY VOLUME 15 This Page Intentionally Left Blank ADVANCES IN AGRONOMY Prepa...

Author: A. G. Norman (Editor)

17 downloads 908 Views 22MB Size Report

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

Report copyright / DMCA form

DOWNLOAD PDF

Harcout,

Corporate

ADVANCES IN

AGRONOMY VOLUME 15

This Page Intentionally Left Blank

ADVANCES IN

AGRONOMY Prepared under the Auspices

of the

AMERICAN SOCIETY OF AGRONOMY

VOLUME 15 Edited by A.

G. NORMAN

The University of Michigan, Ann Arbor, Michigan

ADVISORY BOARD

H. D. MORRIS F. L. PATTERSON G. M. VOLK

W. H. ALLAWAY C. 0. GAFDNER E. G. HEYNE

1963

ACADEMIC PRESS

New York and London

COPYRIGHT @ 1963, BY ACADEMIC PRESS INC. A L L RIGHTS RESERVED. N O PART O F THIS BOOK MAY B E REPRODUCED IN ANY FORM, BY PHOTOSTAT, MICROFILM, OR ANY OTHER MEANS, WITHOUT

WRITTEN PERMISSION FROM T H E PUBLISHERS.

ACADEMIC PRESS INC. 111 Fifth Avenue, New York 3, New York

United Kingdom Edition published by ACADEMIC PRESS INC. (LONDON) LTD. Berkeley Square House, London W.l

LIBRARY OF CONGRESS CATALOG CARDNUMBER:50-5598

PRINTED I N T H E UNITED STATES OF AMERICA

CONTRIBUTORS TO VOLUME 15 Numbers in parentheses indicate the pages on which the authors’ contributions begin,

HAROLD L. BARROWS (303), Research Soil Scientist, United States Soils Laboratory, Soil and Water Conservation Research Division, Agricultural Research Service, United States Department of Agriculture, Beltsuille, Maryland

W. S. CHEPU (211), Research Investigations Leader, Soil and Water Conservation Research Division, Agricultural Research Service, United States Department of Agriculture, Manhattan, Kansas

M. G. CLINE(339), Professor of Soil Science, Department of Agronomy, Cornell University, Ithaca, New York L. L. DANIELSON( 161 ), Leader, Weed Investigations-Horticultural Crops, Crops Protection Research Branch, Crops Research Division, Agricultural Research Service, United States Department of Agriculture, Beltmille, M ayland C. M . DONALD( l ) ,Professor of Agriculture and Head, Department of Agronomy, Waite Agricultural Research Institute, The University of Adelaide, Adelaide, South Australia

W. B. ENNIS,Jn. ( 181), Chief, Crops Protection Research Branch, Crops Research Division, Agricultural Research Service, United States Department of Agriculture, Beltsuille, Mayland D. H. HEINRICHS (317), Head, Forage Crops Section, Experimental Farm, Research Branch, Canada Department of Agriculture, Swift Current, Saskatchewan, Canada J. F. HODCSON (119), Soil Scientist, United States Plant, Soil and Nutrition Laboratoy , Soil and Water Conservation Research Division, Agricultural Research Seroice, United States Department of Agriculture, Ithaca, New York

VICXORJ. KILMER* (303), Soil Scientist, Soil and Water Conservation Research Division, Agricultural Research Service, United States Department of Agriculture, Beltsuille, M a yland

* Present address : Tennessee Valley Authority, Wilson Dam, Alabama v

vi

CONTRIBUTORS

D. L. KLINCMAN( 161), Leader, Weed Znvestigations-Grazing Lands, Crops Protection Research Branch, Crops Research Division, Agricultural Research Service, United States Department of Agriculture, Beltsville, Maryland J. A. MCKEAGUE(339), Research Oficer, Soil Research Institute, Canada Department of Agriculture, Ottawa, Canada

W. C. Smw ( 161), Leader, Weed Investigations-Agronomic Crops, Crops Protection Research Branch, Crops Research Division, Agricultural Research Service, United States Department of Agriculture, Beltsville, M a yland F. L. TIMMONS (161), Leader, Weed Investigatiom-Aquatic and Noncrop Arem, Crops Protection Research Branch, Crops Research Division, Agricultural Research Service, United States Department of Agriculture, Laramie, Wyoming N. P. WOODRUFF (211), Agricultural Engineer, Soil and Water Conservation Research Division, Agricultural Research Service, United States Department of Agriculture, Manhattan, Kansas

PREFACE The over-all objective of this serial publication remains unchanged. It is to survey and review progress in agronomic research and practice. In selecting topics the Editor does not fret about identifying the exact boundaries of agronomy. Indeed, he considers them frontiers where important battles are being staged and fought, the outcome of which extends the holdings of knowledge. There is emphasis on soil-plant relationships, where applicable. It is his belief that soil science and crop science so totally merge at the center that, realistically, there cannot be a separate “soils” treatment or “crops” treatment of the problems involved in the growth of a crop plant in the field. It is odd that this dichotomy should have developed in agronomic science in the United States; perhaps it is less sharp now than two decades ago. It is much less evident in other countries practicing scientific agriculture and will rarely be apparent in these pages. The lead article by Professor Donald is an elegant example of the modern analytical approach to what in the past would have been highly empirical “management” studies. Australian workers have pioneered in this field and have made outstanding contributions both to science and practice. The chapter by Chepil on wind erosion and its control provides yet another example of the advances that are possible when the basic physical principles of what may be considered a natural phenomenon become understood. Soil chemistry is strongly represented in this volume by no less than three articles. McKeague and Cline discuss the forms of silica in soils, an old problem with some new developments. Hodgson reviews the chemistry of the micronutrient elements in soils and their availability. On the more applied side, and an aspect of soil erosion that is often overlooked, Barrows and Kilmer bring together the available information on losses of plant nutrients from soils through erosional processes. The remaining two chapters deal primarily with plants. Heinrichs draws attention to the potential that exists in certain areas for more widespread use of alfalfas with a creeping or recumbent habit of growth, while Ennis and his colleagues discuss the impact of the adoption of chemical weed control on farm management practices and the resulting lowering of production costs. A. G . NORMAN Ann Arbor, Michigan August, 1963 vii

This Page Intentionally Left Blank

CONTENTS Page

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

v

PREFACE...........................................................

vii

CONTRI~UTORS TO VOLUME15

COMPETITION AMONG CROP AND PASTURE PLANTS

BY C. M. DONALD I. I1. I11. IV. V. VI . VII . VIII . IX . X. XI . XI1. XI11. XIV. XV . XVI . XVII. XVIII . XIX. XX.

Introduction ................................................. The Nature of Plant Competition .............................. The Influence of Density on the Community .................... The Influence of Density on the Plant .......................... The Associated Growth of Pairs of Species ..................... Equilibria in Mixtures ....................................... Competition between Crops and Associate Plants . . . . . . . . . . . . . . . . . Competition between Crass and Clover ......................... Plant Arrangement ........................................... Competition for Nutrients ..................................... Competition for Water . . . . . . . . . . .................... Competition for Light ......................................... The Leaf Canopy and Growth ................................ Fluctuations in Leaf Area ..................................... Leaf Arrangement and Competition for Light . . . . . . . . . . . . Height and Competition ............................... The Interaction of Competition for Two or More Factors .......... The Heritability of Competitive A .............. Competitive Ability and Yield . . . . . .............. Concluding Comments ....................................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 2 9

17 28 39 46 52 63 70 77 82 85 90 103 106 109 113 114

CHEMISTRY OF THE MICRONUTRIENT ELEMENTS IN SOILS BY J . F. HODCSON Introduction ................................................. Geochemistry of Micronutrients and Its Relation to Soils . . . . . . . . . . Forms of Micronutrients in Soils .............................. Distribution of Micronutrients in Soils ......................... V . Factors Affecting the Availability of Micronutrients . . . . . . . . . . . . . . . . VI . Needs for Future Research .................................... References ..................................................

I. I1. 111. I V.

ix

119 120 124 136 140 152 154

X

CONTENTS

IMPACT OF CHEMICAL WEED CONTROL O N FARM MANAGEMENT PRACTICES BY W . B. ENNIS.JR., W . C. SHAW.L . L . DANIELSON. D . L . KLINGMAN. AND F. L. TIMMONS

I. I1. I11. IV. V. VI .

Introduction ................................................. Production of Cultivated Crops ................................ Production of Deciduous Fruits and Tree Nuts . . . . . . . . . . . . . . . . . . . Production of Pastures and Rangelands .......................... Changes in Management of Farm Water Systems . . . . . . . . . . . . . . . . . Summary ................................................... References ..................................................

162 169 185 192 200 208 209

THE PHYSICS OF WIND EROSION AND ITS CONTROL BY W . S . CHEPILAND N . P. WOODRUFF I. I1. 111. IV. V. VI . VII . VIII . IX .

Notation ................................................ Introduction ................................................. The Surface Wind ............................................ Equilibrium Forces on Soil Grains ............................. The Cycle of Wind Erosion ................................... Soil Properties that Influence Wind Erosion ..................... Wind Erosion Control ........................................ The Wind Erosion Equation .................................. Needed Research ............................................ Conclusion .................................................. References ..................................................

211 214 218 222 229 249 270 291 296 298 299

PLANT NUTRIENT LOSSES FROM SOILS BY WATER EROSION BY HAHOLDL . BARROWS AND VICTOR J . KILMER I. I1. I11. IV. V. VI . VII . VIII . IX. X. XI .

Introduction ................................................. Methods and Conditions of Sampling Runoff . . . . . . . . . . . . . . . . . . . Organic Matter Losses ....................................... Nitrogen Losses ............................................. Phosphorus Losses ........................................... Potassium Losses ............................................ Calcium Losses .............................................. Magnesium Losses ............................................ Sulfur Losses ............................................... Interpretation of Runoff Data .................................. Summary ................................................... References ..................................................

303 305 306 307 309 311 312 313 313 313 315 315

CONTENTS

xi

CREEPING ALFALFAS BY D . H . HEINRICHS I. I1. 111 IV . V. VI VII . VIII

. . .

Introduction ................................................ Types of Root Systems in Alfalfa .............................. Physiological Considerations ................................... Breeding for the Creeping-Root Character ...................... Genetics of the Creeping-Root Character ........................ Association of the Creeping-Root Character with Other Plant Characters Agricultural Performance of Spreading Alfalfas .................. The Future of Spreading Alfalfas ............................. References ..................................................

317 319 325 327 331 332 333 335 336

SILICA I N SOILS BY J . A . MCKEAGUEAND M . G. CLINE

I. I1. I11. IV. V.

Introduction ................................................ Silica in Solid Forms .......................................... Silica in Solution ............................................ Deposition of Silica in Soils ................................... Silica in Relation to Kinds of Soils ............................. References ..................................................

339 340 353 377 384 389

AUTHORINDEX ......................................................

397

SUBJECT INDEX.....................................................

409

This Page Intentionally Left Blank

COMPETITION AMONG CROP AND PASTURE PLANTS

. .

C M Donald Waite Agricultural Research Institute. University of Adelaide. Adelaide. Aurtralia

I.

I1.

111.

IV . V. VI . VII . VIII . IX . X. XI . XI1 . XI11. XIV . XV . XVI . XVII . XVIII . XIX . XX .

Page Introduction ............................. . . . . . . . . . . . . . . . . . . . 1 The Nature of Plant Competition ............................... 2 The Influence of Density on the Community ..................... 9 The Influence of Density on the Plant ........................... 17 The Associated Growth of Pairs of Species ....................... 28 Equilibria in Mixtures ........................................ 39 Competition between Crops and Associate Plants . . . . . . . . . . . . . . . . . 46 Competition between Grass and Clover .......................... 52 Plant Arrangement ........................................... 63 Competition for Nutrients ..................................... 70 Competition for Water ........................................ 77 Competition for Light ........................................ 82 The Leaf Canopy and Growth ................................. 85 Fluctuations in Leaf Area ..................................... 90 Leaf Arrangement and Competition for Light .................... 95 Height and Competition ...................................... 99 The Interaction of Competition for Two or More Factors . . . . . . . . . . 103 The Heritability of Competitive Ability .......................... 106 Competitive Ability and Yield .................................. 109 Concluding Comments ........................................ 113 References .................................................. 114

1

.

Introduction

Man has long been familiar with competition among plants . The adverse effects of weeds on the growth of crops was surely apparent to the earliest cultivators. and several thousand years later. this everyday phenomenon provided a simple illustration in the parable of the sower and the seed: “And some [seeds] fell among thorns; and the thorns sprang up and choked them” (Matt. XIII. 7 ) . Here we have interspecific competition between a crop. presumably wheat or barley. and a weed. believed to be Centaurea calcitrapa . In the early days of agriculture. man must have learned too of the competition among the individual plants within a crop. of intraspeciiic competition. even though his knowledge was in purely empirical terms. 1

2

C. M. DONALD

He must have learned by experience that if the sowing rate were sparse, his harvest would be lean, and conversely that if the seed rate were increased beyond a certain value, the plants would be spindly and poorly grown. Thus he came to adopt an arbitrary seed rate, at least roughly suited to the crop and the environment, Another direct use of observation of intraspecific competition lay in the thinning of seedlings and the spacing of trees, practices which may not have lagged far behind the earliest use of root crops and cultivated trees. To complete this brief anthology on the early appreciation of plant competition, the reader may wish to speculate on how long ago man also recognized the effect of competition between parts of a single plant-of intraplant competition. When did he first pluck away some of the young fruits of a cluster so that the remainder might grow better-surely long ago? Whatever the times of these events may have been, we can safely claim that the effects of competition have long been known to man. The understanding of competition itself, as apart from its effects, has been left for study in more recent times. For many decades competition has been of interest to plant ecologists working on natural vegetation, though much of the work has been descriptive rather than experimental. When succession was recognized as a phenomenon of general importance in natural communities, competition was named as a basic component of the process. But there was great tardiness in extending thoughts on competition to man’s crops and pastures, partly because early ecologists regarded artificial communities as of little interest. Further, though competition and succession could be appreciated in the species mixtures found in pastures, competition within monospecific annual crops was long overlooked and only now receives the examination it warrants. In this paper our present knowledge of competition in crops and pastures is reviewed. Some data are drawn from natural communities, but more and more of the experimental studies on competition are now being done with artificially established communities of crop, pasture, and weed species. II. The Nature of Plant Competition

Milne (1961) points out that the original meaning of the Latin verb competere, which was “to ask or sue for the same thing that another does,” is fully preserved in the modern meaning of the word “competition”; this is given in the Oxford English Dictionary as “the action of endeavouring to gain what another endeavours to gain at the same time, the striving of two or more for the same object, rivalry.”

COMPETITION AMONG CROP AND PASTURE PLANTS

3

Unfortunately there has been no such continued clarity in the meaning of the word when applied to biological situations, though confusion has been far less serious in the studies of competition among plants than among animals, where predator-prey relationships and the “struggle for existence” have caused a chaotic range of usages of the term “competition” (Milne, 1961), Plant ecologists are fortunate that an effective analysis and definition of plant competition was given half a century ago as a result of the experimentation and the clear-mindedness of F. E. Clements. In 1907, as quoted by Clements et al. (1929), he wrote: “Competition is purely a physical process. With few exceptions, such as the crowding of tuberous plants when grown too closely, an actual struggle between competing plants never occurs. Competition arises from the reaction of one plant upon the physical factors about it and the effect of the modified factors upon its competitors. In the exact sense, two plants, no matter how close, do not compete with each other so long as the water content, the nutrient material, the light and the heat are in excess of the needs of both. When the immediate supply of a single necessary factor falls below the combined demands of the plants, competition begins.” This definition, with but slight qualification, is still acceptable as an account both of the nature and the modus operandi of competition among plants. It is of interest also that the great diversity of viewpoints regarding competition among animals seems to have converged to a definition of essentially the same meaning. Milne (1961), who has reviewed the definitions and the interpretation of competition among animals from Darwin’s time to the present day, regards the concept of competition held by Clements and Shelford (1939) as being the most acceptable and scientifically meaningful. He amended it slightly to read: “Competition among animals is the endeavour of two (or more) animals to gain the same particular thing, or to gain the measure each wants from the supply of a thing when that supply is not suficient for both (or all).” We see then that Clements’ definition of plant competition and Milne’s definition of animal competition are substantially the same, and there seems no reason why they should not be fused to cover all competition among living organisms-that among plants, that among animals, and, less commonly, that between plant and animal, e.g., between insectivorous plants and insect-eating birds, or between aquatic flora and fauna for dissolved oxygen. We can then write: Competition occurs when each of two or more organisms seeks the measure it wants of any particular factor or thing and when the immediate supply of the factor or thing is below the combined demand of the organisms. Harper (1961) prefers to avoid the word “competition” altogether,

4

C. M. DONALD

because of its varied shades of meaning in sport, games, and economics and because he considers that it lacks independent scientific meaning. He uses the word “interference” as “a blanket word to describe those hardships which are caused by the proximity of neighbours . . . . I hope that by using this less controversial word I may avoid judgement of what is or is not ‘competition’.” It is certainly true that the lay terms commonly applied in plant competition are drawn not from science but from man’s personal experience of physical competition-“the thorns Sprang up and choked them”; we speak of ‘‘stranglingby vines” and of “the smothering” and “crowding out” of unsuccessful species. But these inaccuracies are not serious if we accept them only as lay terms. “Interference” is not devoid of connotations similar to those of “competition,” and in any case it suffers the disability that it does not convey so vividly as does the word “competition” that there may be mutual effects of organisms one upon the other. Despite the confusion that has at times occurred in the use of the word “competition,” it seems that if it is used in its original meaning, and according to the biological concepts of Clements, it is well suited to a clearly delineated set of biological situations. There is no reason to discontinue the use of this simple and effective term. The factors for which competition may occur among plants are water, nutrients, light, oxygen, and carbon dioxide; in the reproductive phase, agents of pollination and dispersal must be added. There are other factors affecting growth, such as temperature and humidity, but these are not commodities in finite supply and hence are not the subject of competition. Water, nutrients, and light are the factors most commonly deficient, but in rapidly photosynthesizing crops, carbon dioxide may also be depleted by competing plants. Competition for soil oxygen, until recently disregarded, may be of significance in poorly structured soils. Clements et al. (1929) emphasized that competition for space is exceptional and occurs only among crowded tuberous plants, as when crowded carrots are of polygonal cross section or are forced from the ground. In nearly all crops there is plenty of space-for more plants, for more branches, for more leaves. This is vividly illustrated when a highyielding crop is harvested into sheaves or bales which then occupy only a small fraction of the apparent volume of the standing crop. Yet references to competition for space continue in the literature, often evading any attempt to recognize the real factors for which competition is occurring. We can perhaps recognize a few other instances of what may truly be competition for space, in addition to the crowding of tuberous plants to which Clements referred. If a seed falls on a rather bare surface, it

--

COMPETITION AMONG CROP AND PASTURE PLANTS

5

may lodge or be blown or washed into a crevice and thereby reach a potential site for germination. If, when the seed has fallen in the crevice, there is not room for another, then it may be said that there is competition for space, in the same sense that birds compete for nesting sites. Harper (1961) looked at this phenomenon by scattering the mixed seed of two annual grasses (Bromus rigidus and B. d r i t e n s i s ) on two soil surfaces, both prepared from a silty loam. One was a rough surface prepared with %-inchaggregates, the other a watered, dried, and cracked surface. As shown in Fig. 1 the establishment on the rough surface was linear 40

0 Y

30

P

E e a

5 20

:

Q

L

0

10

0

Hard, cracked

surface

80

I00

60 80 No of r e e d s d r o p p e d o n ground

I00

20

40

20

40

I

I

60 I

1

I

FIG.1. The relationship between the physical condition of the soil surface and the establishment of two Bromus species (Harper, 1961).

on the sowing rate, but on the caked, cracked surface, it reached a saturation value at relatively low seed densities. Harper explains his results in terms of the limited number of crack sites suitable for the germination of these Bromus seeds on the caked soil, in contrast to the more abundant germination sites on the rough soil surface. It is conceivable that roots may also compete for sites-that is to say for space-when they penetrate stony layers or hard clods through which there are relatively few channels. These are less common circumstances in crops and pastures. Yet inaccurate references to “competition for space” have continued, as the following examples will illustrate. “When grass and clover are growing together, they are competing for space, light, water and nutrients” (Robinson and Sprague, 1947) or “In the interaction of

6

C. M. DONALD

nurse crop and undersown herbage, competition for water, plant nutrients and space has to be considered. If the growth of the undersown species is rapid, they might also compete with the cereal for light” (Charles, 1958). Here are two circumstances in which competition for space has not been shown to occur. When de Wit (1960) speaks of “competition for the same space,” he uses the term “space” explicitly and by definition to mean “growing factors” or “requisites” like water, minerals, light and so on which are homogeneously distributed over and in the field where the plants grow. “Such a description” he adds, “is, however, not necessary, always inaccurate and therefore inadvisable.” While this use of the term “space” may be convenient shorthand for mass competition, it evades the need to pursue and recognize the real factors for which competition is occurring. The loose use of this term, whether wittingly or otherwise, could be discontinued to advantage. Most of the factors for which there is competition are found as a pool of material from which competitors draw their supplies. If the pool is of limited volume, or if it is subject to intermittent depletion by the competing plants, then the successful competitor is the plant which draws most rapidly from the pool or which can continue to withdraw from the pool when it is at low ebb or when its contents can no longer be tapped by other plants. If all the plants in the community are nearly equal in competitive ability, as in a wheat crop, they will tend to share equally in the supply until it is exhausted, and then, simultaneously, to suffer the effect of depletion of the pool. This simple picture applies broadly to water supply. If we turn to nutrients, the capacity to draw from the pool is in varying degree an expression of the differing ability of plants to make use of the nutrient in different chemical and physical forms. With carbon dioxide, for which competition perhaps occurs in dense crops in still weather, the relationship is again one of the use of C 0 2 from a pool of air within the crop, a pool subject to steady or intermittent depletion, and subsequent restoration from the external atmosphere. We must note, however, that the factor in short supply may not move freely to the points at which it is being tapped. Thus if a plant depletes the factor in its immediate neighborhood, the remaining supply of the factor still available to other plants may or may not flow toward it. There may be considerable diffusive flow of COZ, or mass flow of water, but movement toward depleted centers is, for example, much slower and more restricted in the case of phosphorus. But the effective competitor is not wholly dependent on the flow of factors along concentration or potential gradients. On the contrary the effectiveness of a competitor is

COMPETITION AMONG CROP AND PASTURE PLANTS

7

commonly an expression of its capacity to make rapid use of its immediate supplies and then, by growth of its roots or foliage, to extend its exploitation into a greater spatial part of the environment. The capacity to exploit the environment quickly may alone give success over competitors, We can again usefully quote Clements and his colleagues (1929). “It is evident that practically all the advantages or weapons of competing species are epitomized in two words-amount and rate. Greater storage in seed or rootstock, more rapid and complete germination, earlier start, more rapid growth of roots and shoots, taller and more branching stems, deeper and more spreading roots, more tillers, larger leaves and more numerous flowers are all of the essence of success.” This stands as a valid analysis. We would perhaps now wish to add the genetic and physiological attributes which stand behind these morphological charactersthe capacity to withdraw water or oxygen from dry or wet soils, to dispose leaves advantageously for light utilization, to take up nutrients of lesser availability, or to respond in development to particular climatic or seasonal conditions. It must not be assumed that competition necessarily occurs just because a factor is in short supply. All plants in a community may be short of the factor, but if the environment of each plant is independent of that of its neighbors, then there is no interference in the growth of one plant by another. For example, the germination and growth of wheat seedlings may be delayed and reduced by the poor oxygen supply in overwet, poorly structured soils. The factor governing oxygen supply to the seed is predominantly the gas exchange with the external atmosphere; each seed lies independently in its own micro-atmosphere, uninfluenced by its neighbor some inches away. There is no competition. The same is true for seeds lacking sufficient water for germination. But the lack of competition at the seed or seedling stage is the exceptional circumstance in crops and pastures, and competition is likely to begin soon thereafter. When we turn to competition for light, the concept of a pool or store of material is not valid. There is no store of light energy in the immediate environs of the plant. Instead light is available as a passing stream which must be intercepted by the leaves if it is not to be permanently lost to the plant. A dense canopy will intercept all light, but the young crop characteristically covers only a small portion of the soil surface and most of the energy is absorbed or reflected by the soil. It is perhaps because competition for light is so bound up with energy relations, photosynthesis, and growth that this phenomenon is at present better understood than competition for other factors. Mather ( 1981) discusses the interrelationships of competition, in which organisms are mutually harmful, and cooperation in which they

8

C. M. DONALD

are mutually beneficial. If the relationship is density independent, then it can only be neutral or cooperative, but if it is density dependent then there may be cooperation at low densities, followed by a neutral relationship and finally by active Competition as the density increases. In some instances, cooperation and competition may overlap, as when both beneficial and harmful effects occur simultaneously. The most obvious example of cooperation occurs when organisms of different sex mate to produce offspring, and this is of importance in dioecious crop plants. But clear instances of cooperation among plants are few. One of the commonest expressions of cooperation, and one which shows a continuity of meaning and effect with competition, is the effect of density on height. Height gives competitive advantage by enabling the shading of neighbors. Yet as density increases and competition for light is intensified, plant height may be substantially increased, a cooperative effect. For example wheat rows spaced at 12, 6, and 3 inches, with competition for no factor other than light (the roots of each row were in separate, standard soil cultures) had heights of 63, 78, and 85 cm at 16 weeks (Wassermann, 1963). If, in a monoculture of cereals, a plant is of less stature than its neighbors, it will in fact tend to elongate faster, the same effect of shading on height as that found in a crop of high density. In rows of maize, Hozumi et al. (1955) showed that there was a negative correlation between individual plant height and rate of elongation, bringing about a tendency toward equalization of height. A most notable example of cooperation occurs between germinating seeds of Trifolium subterraneum. The dormancy found in some varieties of this species for many weeks after harvesting may be broken by exposure to an atmosphere containing 0.5 per cent carbon dioxide (Ballard, 1958). If a small heap of seeds of marked dormancy is placed in a normal atmosphere under temperature and moisture conditions favorable to germination, and if any one seed in the heap germinates, it will produce enough carbon dioxide to trigger the germination of the other seeds, providing a classical instance of cooperation. Seeds per pile: Germination:

1 1.4%

5 50%

15 96%

Despite these examples, the relationship among plants is almost invariably competitive, and this is especially true of man’s crops, in which he seeks to ensure a maximum exploitation of the environment. This he can achieve, and unwittingly does partly achieve, only by creating such a relationship between the plants and their environment as will ensure maximum pressure on the factors needed for growth. It is a regime in which there is intense competition among his plants.

COMPETITION AMONG CROP AND PASTURE PLANTS

9

The principal factor which might lead to a revision of the definition of competition given earlier is the exudation by a plant of toxins which depress the growth of a neighbor. There is an extensive literature on this subject, and it has been amply demonstrated that identifiable substances found in fallen leaves or exuded from roots are in fact harmful to the germination or growth of other species. But the significance of these substances under field conditions is greatly in question, and this is especially so in the relationship of crops, pastures, and weeds. They are excluded from the discussions in this paper except in one instance where they were evoked to explain competitive relationships among familiar pasture plants. 111. The Influence of Density on the Community

Competition in man’s sown crops is commonly between plants of like or similar genotype, all sown at the same time and each with closely similar environmental conditions. The crop of wheat or corn, barley or cotton is a typical monospecific community in a nearly uniform environment. We can first examine the relationship of density to the total yield of dry matter (the biological yield) of various crops. Where the immediate objective in planting studies has been to determine the optimum sowing rate, the data rarely include a sufficiently wide range of densities to permit the definition of the relationship of density to yield, but a few studies have varied density from low to very high values. Figure 2a shows the relationship of the yield of dry matter to density at 0 days (weight of embryos), 131 days, and 181 days in subterranean clover, grown under conditions of liberal water and nutrient supply. At the time of sowing there was a linear relationship of density and dry matter, expressed here as embryo weight. Competition (in this instance for light) first became operative at the highest densities and then progressively, with advancing stage of growth, at lower and lower densities. At the highest densities, competition became more and more intense until growth was completely arrested. Thus the yield at lower densities approached progressively that at the higher densities. The original linear relationship of density to yield of dry matter was replaced by a curve in which the yield rose sharply with increasing density to a maximum which was constant for all higher densities. Yield per plant ( w ) at these higher densities ( d ) was now inversely proportional to density ( wd = k). Similar results were secured with Lolium rigidurn, but here the factor governing the final yield was the supply of nitrogen. The constancy of the final yield of dry matter per unit area at moderate to high densities has also been shown for perennial ryegrass (Hol-

10

C. M. DONALD

liday, 1953), rape (Holliday, 1%0a), various crop plants (Shinozaki and Kira, 1956), and barley and white persicaria (Fig. 2b) (Aspinall and Milthorpe, 1959). Holliday (1960b) further points out (quoting Davies, 1954) that this relationship applies not only to vegetative parts such as 40 -

-

181 D A Y S

1330

-

Y

J

EMBRYOS

0 0

100

50

150

D E N S I T Y (PLANTS PER

a

so

250

xx)

LK)

I-

2

8 30

-

L

w n

ui

5

v

I-

20

-

I

P

$!

0 4 8

b

16

32

64 DENSITY (plants per pot)

128

FIG. 2. The relationship of density, yield, and time. ( a ) Subterranean clover (Donald, 1951). ( b ) Persicaria (Aspinall and Milthorpe, 1959).

tops, but also to other vegetative yields such as the total yield of potato tubers. Various mathematical expressions have been proposed for this relationship of density to production. These include Eqs. ( 1 3 ) . Equation (1)was proposed by Mitscherlich (1919) w = M (l-e-ks)

(1)

COMPETITION AMONG CROP AND PASTURE PLANTS

11

where w is weight per plant, M is the asymptote of w, s is the density, and k is a constant. This equation has failed, however, to give a statistically good fit to the data shown in Fig. 2 or to other similar data (Kira et al., 1953). Warne ( 1951), working with vegetables (truck crops) proposed the relationship

Y = Kda where Y is the yield per unit area, d is the density, and K and a are constants. Kira et al. (1953) presented this same relationship in the form w&=K (2) where w is the yield per plant. They refer to it as the C-D (CompetitionDensity) Effect.

\

119 days

Density ( ~ o q scale)

FIG.3. The linear relationship of log plant weight to log density, based on the “power equation” wda = K. Soybeans (Kira et al., 1953).

12

C. M. DONALD

Data conforming to this empirical equation will lie on a straight line B log d ) if log w is plotted against log d, as is shown in (log w = A Fig. 3 from Kira et al. (1953). Various workers regard this “power equation,” for which neither Kira nor they claim any biologically meaningful basis, as a useful working formula which fits many data as well or better than any other equation [Goodall (1960) with mangolds; Harper ( 1961) with Bromus species], However, the equation fails to accommodate the situation in which widely spaced plants have attained a ceiling value of W . To meet this point Kira et a2. proposed the introduction of an arbitrary value giving the ceiling yield per plant at the lower limit of density for the occurrence of competition. This is shown for example at 12 and 21 days in Fig. 3. This arbitrary limitation, together with the empiricism of the equation, detracts from its usefulness. Perhaps the most satisfactory expression governing the relationship of density and yield is the simple expression -1= A d + B (3) W

+

where w is the yield per plant, d is the density and A and B are constants. Shinozaki and Kira proposed this relationship of yield and density in 1956, based on four assumptions: ( i ) The growth (increase in dry weight) of a plant is represented by the simple logistic equation 1 dw --=k w

dt

( 1--

3

where w is the weight per plant, W is the asymptote of w, k the coefficient of growth (the initial relative growth rate), and t the time. (ii) The initial relative growth rate ( k )is independent of density ( d ) . (iii) Final yield is constant at high densities, namely, Wd = Y = constant. (iv) Time is measured from a common date of sowing, t = 0. Essentially, this logistic equation is based on the concept of a ceiling yield per unit area, and on a relative growth rate that is initially constant but decreases as the ceiling yield is approached. It can be simply tested by plotting l/w against density, when the relationship will be linear. De Wit (1959), Reestman and de Wit (1959) and Holliday (1960a) have independently derived this equation and, like Shinozaki and Kira (1956), have found it to fit data from density-yield studies with a great variety of crops ( Fig. 4 ) . If we now turn to the yield of reproductive parts, of seed or fruit, a differentrelationship pertains. Studies in many circumstances have shown

COMPETITION AMONG CROP AND PASTURE PLANTS

13

that the relationship of density to the yield of grain is depicted by a curve rising to a maximum and then showing a slow decline at higher densities. This has been found to apply to the seed yield of wheat, corn, soybeans, peas, and ryegrass (see review by Holliday, 1960b). The yield of seed increases with increasing seed rate and then declines; further, there is a considerable range of seed rates giving near-maximal yield.

POPULATION

Oo

20

40

i

IO'/ACRE

m PLANTS / M a

no

100

120

600

000

loo0

RANTS /ARE

FIG.4. The linearity of the reciprocal of yield per plant on density. Top left: rape (Holliday, 196Oa). Bottom left: soybeans (Shinozaki and Kira, 1956). Right: beets (redrawn from de Wit, 1960). (The data used by Shinozaki and Kira are identical with those on which Kira et uZ. based Fig. 3. )

Though density and dry matter, or density and grain yield, have been determined in many studies, the three parameters, density, yield of dry matter, and yield of grain have not often been measured together. Most agronomists and plant breeders have measured only their seed yield or only their forage yield, and our knowledge of dry matter/grain relations is most fragmentary. In Fig. 5, data from six studies are examined: three experiments with corn, one with wheat, one with an annual grass, and one with an annual clover. The curves for dry matter, which it must be emphasized relate only to the tops, are hand-fitted curves based on the concept of a ceiling

14

C . M. DONALD

yield at high densities. The relationship is generally satisfactory up to very high densities. The curves for the yield of seed show a peak value and then fall by 10 to 40 per cent at high densities. It is now of interest to consider the interdependence, ifany, of the two curves; it will be seen that in each instance the peak of the grain A

Grain

52 0 - -.

0 0

1 2 Cw(. seed

/ acre

3

4

link

D

C Total yield

Tola1 veld

0

8

P

I - 2

20-

0 0

20 40 60 b m / 6 4 ’ length O( row

0

40

Plants /are

F

E

0

20 10’

1 0 Plants

/ m?

20

0

100

800

Aants/sq link

FIG.5. Experiments showing the relationship of density, total yield of tops, and yield of seed of various crops. Derived from data by the following authors: ( A ) Holliday ( 1960b), wheat; (B) Donald ( 1954), Wimmera ryegrass; ( C ) Haynes and Sayre (1956),corn; ( D ) Morrow and Hunt (1891),corn; ( E ) Bunting and Willey (quoted by Harper, 196l),corn; (F) Donald ( 1954), subterranean clover.

curve occurs approximately at the density at which the yield of dry matter, the biological yield, is leveling off. Because of this, it is tentatively suggested that the minimum density giving the ceiling biological yield (where wd = K ) may also be the density giving the maximum grain yield. This is the density at which any gain in total yield per acre due to the addition of extra plants is offset by a loss due to the decrease in the weight per plant.

COMPETITION AMONG CROP AND PASTURE PLANTS

15

There seems little doubt that these relationships would be altered by environmental factors; they would presumably not hold, for example, under conditions in which water supply becomes exhausted before grain is formed. It is suggested that in the instances shown in Fig. 5, light or nutrients were the factors governing production. These relations of density, total yield, and grain yield are here stated in the crude terms of yield per unit area. Obviously these phenomena depend on the behavior of the individual plant, and in Section IV it is suggested that maximum grain yield may be given at densities which give low intensity of competition within the plant. We now turn to the relationship of density, biological yield, and time. By redrawing Fig. 2a, as is done in Fig. 6 (and adding values for an

0

62

131 DAYS

FROM SOWING

FIG. 6. The relationship of density, yield, and time-subterranean clover (Donald, 1951). The curves are for four swards of different density (plants per square link).

extremely high density), we can see that the density required to give the maximum yield per acre will depend on the date of harvest. If the crop is to proceed to maturity, then a wide range of densities (in this case 60 to 1300 per square link1) will give the maximum biological yield attainable by the genotype within that particular environment, and a relatively low population will suffice. On the other hand if a crop is to be used at an early date, then the greater the density, the greater the yield. In this instance the heaviest sowing gave 17 times the yield of the 1

One square link = 1/100,000 acre = 0.44 sq. ft. = 0.040 sq. m.

16

C . M. DONALD

62-rate at 62 days, and 2.3 times at 131 days, although the yields were the same at 181 days. The determination of optimum density for an early harvest is more difficult than at maturity. How much does seed cost? How much is this early production worth? It is certain that the economically optimum seed rate for an early harvest will be higher than for a harvest at maturity, And the basic reason is that the earlier the harvest the less intense will be the interplant competition and the less the depression of yield per plant. This general principle is well illustrated in the case of fodder crops such as oats. The optimum rate of seeding for early green fodder production will be higher than for grain production. Similarly Warne has pointed out (1951) that the optimum seeding rate for carrots to be harvested while young and immature is less than that for carrots harvested at a later stage of growth. In many crops the density-yield relationship may not stabilize for several years. In simple form, it is clear that the horticulturist will not plant his fruit trees at a spacing giving the greatest production per acre in the first bearing year, since this would lead to a serious restriction of tree growth in subsequent years. A less obvious example is provided in Holliday’s studies (1953) of perennial ryegrass, where the attainment of constant yield did not occur till the third year, and where the plant as the ecological unit was succeeded by the tiller. In the first year yields were almost linear on seed rates from 5 to 40 pounds per acre, with some further increase to 160 pounds per acre. By the third year both yield and tiller number were constant (500 tillers per square foot) in all treatments. A major factor influencing optimum seed rate for any particular crop is the genotype. Though there is little information of this kind, the Illinois studies on corn (Lang et al., 1956) are of particular interest. Until recently the standard crop density in the central corn belt of the United States was 12,000 plants per acre, but, as will be discussed later, the increasing use of fertilizer nitrogen has led to higher seeding rates. All the hybrids for use in this region had been bred to give single-stemmed plants with a single ear and high yield per acre at the standard 12,000 spacing, but when they were grown at higher densities they no longer showed common behavior. Some hybrids maintained one ear per plant, but others showed considerable proportions of barren plants in these denser communities. This led to a study of genotype-density interaction. The hybrids were grown at densities from 4,000 to 20,000 plants per acre. At low densities, some hybrids had multiple stems and multiple ears; others showed little branching and produced only a poor second ear. It was of great significance that the varieties which had the greatest tendency to branch and

COMPETITION AMONG CROP AND PASTURE PLANTS

17

to carry multiple ears at low densities, proved to be the varieties with the lowest percentage of barren stalks at high densities. The varieties with the least number of second ears at low densities showed the highest sterility in crowded stands. A particularly striking interaction of density and genotype is reported by Nelson and Ohlrogge (1957) in a comparison of the normal Hy 2 corn hybrid and a semidwarf mutant, ‘compact’. While the grain yield of ‘normal’ declined by 78 per cent as the population was raised from 26,000 to 52,000 plants per acre, that of ‘compact’ fell by only 2 per cent. In a furthw study (Sowell d al., 196l),it was shown that at high densities ‘normal’ continues to produce additional vegetative parts after tasseling, while in ‘compact’ all growth after tasseling is expressed as an increase in the weight of the ear shoot. In studies with soybeans at spacings from 2 inches to 32 inches in 38-inch rows, Hinson and Hanson (1962) similarly found an interaction in yield between genotype and density. They considered that at close spacing varieties display their different susceptibilities to competitive effects, whereas at wide spacing they show their different capacity to use a more extensive environment. A further example of the interaction of density and genotype is available from Western Australia, where Rossiter (1959) examined subterranean clover at wide spacing (1plant per 6 sq. ft.) and in swards (66 pounds seed per acre). He found the ranking of strains for production to be quite different at these two densities. With the spaced plants seed yield increased linearly with the lateness of flowering of the variety, but in swards the reverse was true. The main factor responsible for these production patterns was apparently the available soil moisture which was sufficient to sustain spaced plants but imposed a progressive limitation on later flowering swards. These studies draw attention to the risks inherent in the assessment of any genotype at noncommercial spacing. It cannot be assumed that the relationship of, say, yields of wheat varieties at 14-inch row spacing (though perhaps of great interest) will be reproduced at 7-inch row spacing. It is also clear from these studies that information on the genotype-density-fertility relation of many crop plants could be of great consequence in the development of more productive varieties under improved fertility conditions. IV. The influence of Density on the Plant

Plants show extreme plasticity, responding remarkably in size and form to environmental conditions. One of the most potent of these external forces is the presence of competing neighbors, which may reduce a

18

C. M. DONALD

plant to diminutive size. Table I shows the effect of the progressive crowding of wheat plants, with reduction of the yield of dry matter and the yield of grain to a few per cent of that of widely spaced plants. It is characteristic of crop plants that they show on the one hand a capacity to exploit a more favorable or extensive environment, and on the other a capacity to survive and reproduce in an acutely restricted habitat. TABLE I The Plasticity of the Wheat Plant in Response to Changes in Density Seed rate equivalent to the initial densities ( pounds/acre)

0.4

2.1

10.5

Density (plants/m.Z) At germination At 119 days At 182 days

1.4 1.4 1.4

7 7 7

35 35 35

At 119 days Dry matter per plant ( g. ) Height (cm. ) Area of leaves per plant (cm.2) Area per leaf (cm.2) Tillers/plant Weight/tiller (g.) At 182 days Grains per spikelet Grains per ear Ears per plant Weight per grain (mg.) Yield grain per plant (g.) Seeds per plant Yield grain per meter2 ( 8. )

184 190 154

326.4 1078 694 447

4.2 1.1 108 Lodged 199 50 25.7 21.0 3.0 1.2 1.39 0.96

2.03 2.20 1.89 1.60 21.5 32.9 37.8 29.9 29.4 18.6 2.2 7.2 32.7 34.2 35.0 33.2 1.52 33.2 24.7 7.05 215 46 970 705 247 46 173 234 - INSIGNIA Wheat, Waite Institute, Adelaide (Puckridge, 1962). Y

5

17.2 45.9 47.6 101 75 83 2550 2660 861 17.2 24.0 25.5 40.5 29.5 10.6 1.12 1.62 1.73

55.2

1.75 18.8 0.7 33.1 0.42 12 185

It is because of this plasticity that when the seed rate is cut by half, crop yield may be little af3ected. The fewer plants simply approach more closely their potential yield. It is nevertheless a surprising thought that man, in growing a successful, “healthy” field crop, creates such intense competition that the individual plants are, in quantitative terms, markedly subnormal. It is the community of suppressed plants which gives the greatest yield. This can be illustrated by comparing plant size at wide spacing with that at densities which give the maximum crop yield (Table 11). The contrast between the growth of the isolated plant and the plant growing under competitive stress is also shown by a comparison of relative growth rates. When plants are widely spaced, under favorable conditions of water, light, and nutrient supply, so that no competition occurs

19

COMPETITION AMONG CROP AND PASTURE PLANTS

between them, they may continue to grow at a nearly constant rate until they pass from the vegetative stage, or until they develop such a canopy of foliage as to suffer internal competition for light. In contrast, plants entering into early competition with their neighbors immediately show a reduction in growth rate which becomes progressively more marked as competition intensifies. A study by Black (1957) demonstrates these relationships. TABLE I1 Competition and the Reduction in Yield per Plant at Densities Giving the Maximum Yield per Unit Area

At wide spacing

At spacing giving maximum yield per unit area

Density

Yield/plant

Density

Yield/plant

Buckwheat, dry matterb (Iwaki, 1958)

25/m.2

19.6 g.

400/m.2

1.6 g.

Subterranean clover, dry niatterb (Donald, 1954)

6/m.2

1500/m.2

0.6 g.

7.1 g.

Crop

34 g.

Wheat, grain (Puckridge, 1962)

1.4/rn.2

33.2 g.

35/m.2

Corn, dry matterb (Ilaynes and Sayre, 1956)

1/64 in. of row

1.35lb.

16/64 in. of row

ll/m.z

29.7 g.

66/m.2

Broad beans, seed (Hodgson and Blackman, 1956)

0.42 lb.

9.3 g.

a This initial density may not have been maintained in some of these stands, and in such cases the yield per plant at the final harvest would be higher than that shown. b Dry matter refers to tops only.

The work was done with subterranean clover, which, when grown at wide spacing, is a trailing herb with little more than a monolayer of leaves and thus with limited internal shading. In dense swards the individual plants are of upright form and the community forms a heavy canopy of leaves. Black ( 1957) grew spaced plants (area per plant = 25 sq. links = approximately 1m.’) from seed weighing 3, 5, and 8 mg. The ratios of the weight of the plants at emergence and also of their cotyledonary areas were 3:5:8(Fig. 7). Harvests were made at intervaIs up to

Y E 5

0

YIL 1

o

Ma

I

a

d

1

4b

w

I

10

mn

FIW

I

100 Y)IIM

1

I20

uo

1 YQ

21

COMPETITION AMONG CROP AND PASTURE P U N T S

194 days; it will be seen that the relative growth rates were almost maintained, that the rates were the same for each group, and that, as a consequence, the relative weights at the final harvest showed close relationship to the weights of the seeds and seedlings; the final plant weights were in the ratio 3:4.5:6. Contrast the situation in swards, with 625 plants per square meter. Here plants from three seed sizes, in this case with a weight ratio of 3:6:12, grew at the same rate only until competition became operative, first among the largest plants and finally in each sward until the growth 600

500

a

5.

400

f

n Y v1 Y

\

2

\

300

I-

:

3

D 200

I00

I

I

0 LOG

2

3

DENSITY

FIG. 8. The influence of density on various plant characters in subterranean clover. (Drawn from Donald, 1954.)

rates and the ceiling yield, governed mainly by light, were the same. The ratio of the final weights of the plants was 3:3.1:3.1. No more striking illustration could be given of the effect of competition on plant growth than is provided by this study. Plants at normal field spacing are smaller, have fewer stems and leaves, fewer inflorescences, and fewer seeds than widely spaced plants. Yet there is not a downward trend in all plant characters as density is increased. Figure 8 shows that in subterranean clover, although the dry matter, and the seeds per plant (as well as the number of inflorescences FIG. 7. The relative growth rates of subterranean clover, initially of different seed weights, growing ( a ) under widely spaced conditions, and ( b ) in dense swards ( Black, 1857).

22

C. M. DONALD

per plant) all fell progressively with increasing density, the seeds per inflorescence and the weight per seed actually rose to a peak value as density was increased, and then fell. Table I11 summarizes the data known to the author on this point; it is rather meager both in quantity and quality, but the phenomenon has recurred in these four experiments and seems to be meaningful. TABLE 111 The Influence of Density on Seeds per Inflorescence and Weight per Seed Lolium rieidum Plants per sq. k. 0.5 5.2 Seeds per plant 2152 365 Seeds per ear-bearing tiller 26.0 33.5 2.02 Weight per seed (mg. ) 2.25 (Donald, 1954) Y

Trifolium subterramum Plants per sq. lk. Seeds per plant Seeds per inflorescence Weight per seed (mg.) (Donald, 1954)

3.2 0.5 184 448 2.19 1.20 6.60 7.26

Triticum vulgare Plants per m.2 Seeds per plant Seeds per ear Weight per grain (mg.) ( Puckridge, 1962 )

1.4 970 32.9 34.2

Zea muys Plants per 64-inch of row 1 Weighi per ear (lb. ) 0.42 (Haynes and Sayre, 1956) a

7 705 37.8 35.0

2 0.55

16 127

79 28

921 1.7

22.8 2.29

10.4 1.94

11 91 2.69 7.91

78 750 11 1.0 2.17 2.40 7.31 7.17

35 215 29.9 32.7

8 0.75

3.0 1.87

184 1078 45 12 21.5 18.8 33.2 33.1

16 0.55

64 0.19

N.S. 0.27@

0.8 0.9

6.6

N.S.

All differ

Values in this column = least significant difference ( 5 per cent).

It has been suggested (Donald, 1954) that the greater seed weight and number of seeds per inflorescence at intermediate densities are due to the varying time of incidenoe of interplant and intraplant competition. At the widest spacing, competition is absent during the early stages of growth. Flower primordia are laid down by each plant in large numbers. As growth proceeds, interplant competition becomes progressively operative, until, when flowering and seed setting occur, the load of inflorescences is so great as to lead to competition among the inflorescences themselves and thereby to reduce the efficiency of seed production in the individual inflorescence; this loss of efficiency at the widest spacing shows in reduced numbers of seeds per inflorescence and reduced seed

COMPETITION AMONG CROP AND PASTURE PLANTS

23

size compared with somewhat denser stands. Thus intraplant competition may be intense at low densities. It seems that in moderately dense stands, interplant competition becomes operative at the time of flower initiation or formation. The number of floral primordia laid down or developed by each plant is considerably reduced, and this reduced load lies more closely within the capacity of the plant as interplant competition intensifies; seeds per inflorescence and seeds per unit area achieve maximum values. In extremely dense stands the Competition at the time of laying down of primordia is presumably already intense; plants in such dense communities suffer both interplant and intraplant competition. Figure 8 suggests that the attainment of maximum seed yield per unit area is the result of the interaction of competition between plants on one hand, and competition within plants on the other. In fact there is some indication that competition within the plant, which is an integral of the physical environment and of the competition by neighbors, may be the governing factor, with maximum yields per unit area when the competition within the plant is at its least intensity. It is suggested that these relationships warrant much fuller examination. There have been few studies of the individual plants within crops and pasture communities. Most of our knowledge of competition is based on mean plant values, derived by dividing yield per unit area by density. For this reason the few studies of individual plant values are of particular interest. Koyama and Kira (1956) have examined the distribution of plant weight in monospecific communities. They have found, both in their own experiments and in data from other sources, that a population initially of nearly uniform plant weight, will move progressively toward a skew distribution. As growth continues, there will be an increasing proportion of small plants and a decreasing proportion of large plants (Fig. 9 ) . Koyama and Kira suggest that, theoretically at least, this phenomenon will occur independently of competition among the plants in the community. This view is based on Blackman’s definition of relative growth rate:

1 dw 1 r=--or T=(log wt-log w o ) w at t where u) is plant weight, t is time and T is the relative growth rate. The latter formula is an average relative growth rate over the interval (0, t ). If wo (the initial plant or seed weight) has a normal distribution and if the relative growth rate T has a normal distribution, then log wt will also have a normal distribution, and thus wt will have a skew distribution. A normal distribution of seed weight is regarded by Koyama and Kira as usual, and they also present a single example showing a normal dis-

24

C. M. DONALD

tribution of the relative growth rate in spaced (noncompeting) plants of rose mallow. Nevertheless the evidence that skewness develops among noncompeting plants must be regarded as rather inconclusive. It is certainly clear, as shown in Fig. 9, that crowding greatly accentuates and accelerates the progress toward skewness; Koyama and Kira suggest that this is due to the increase in the variability of the relative growth rate in crowded communities. They give the following data from an experiment with rose mallow to support this point. When rose mallow was grown at spacings of 8 cm., 4 cm., and 2 cm., the range of relative growth rates

Ih

22 6€qS

3Qd-

61 days

5.0

cm.

93 days

FIG. 9. The frequency ( f ) distribution of fresh weight per plant of soybeans at 5, 10, and 28 cm. spacing (Koyama and Kira, 1956).

among the plants was least at the wide spacing ( 8 cm.), and greatest in the densest community ( 2 cm.). Further, the correlation between plant weight and relative growth rate was highest in the 2-cm. spacing and least in the 8-cm. spacing. This would mean that the tendency for the heaviest plants to become relatively heavier and the light plants to become relatively lighter is increased as density rises. Koyama and Kira quote results with touch-me-not (Zrnpatiens balsurnina), radish, soybean, red pine (Pinus densiflora), Azuki bean (Phaseolus chrysanthos), turnip, and dent corn showing the progress with time from normal to skew distribution of plant weight and showing the change to be more rapid, and the skewness to be ultimately more extreme (L-shaped), when the density is high.

25

COMPETITION AMONG CROP AND PASTURE PLANTS

On the other hand Harper (1961) points out that in his experiments with Gajic on Agrostemma githago, the variability was greatest ( 1 to 24 capsules per plant) at low densities and least (almost all plants with a single capsule) at high densities, and Koyama and Kira also observe that exceptional cases exist in which density apparently has no influence on the progress toward skewness. Stern (1960) recorded the individual weights of plants in swards of three densities (4, 16, and 36 plants per square dm.) of subterranean

DAV5

fIOM

5OWlNG

I

4

16 P L A N T S PER OM?

PLANTS PER DM?

36

,--4

20

60

100

140

10

60

100

140

20

60

100

140

20

60

100

140

20

60

1 0 0

140

20

60

100

140

DAYS

FROM

SOWING

FIG. 10. Showing the range of weights of the individual plants in three clover swards of different density, and the coefficient of variation of the plant weights (Stem, 1960).

clover. His results, in terms of the range and variability of plant weights, are shown in Fig. 10; here, the coefficient of variation is nearly the same at all densities until day 90, when it increases sharply at the higher densities. In general terms, then, Stern’s results are in agreement with those of Koyama and Kira-that variation in growth rate is greatest at high densities.

26

C . M.DONALD

What the significance of these phenomena may be in crop production we do not yet know. Many of the data relate to total plant weight, and we need more information about the relation, if any, of these distribution patterns to the yield of fruit or grain. It may be noted for example that when plant height, as distinct from plant weight, was examined in corn growing in single rows (long boxes) the frequency distribution showed a peak moving progressively to the right; i.e., there was an ever decreasing proportion of taller plants (Hozumi et al., 1955). There was a tendency, as noted earlier, for plants of lesser height relative to their neighbors to have the greatest rate of shoot elongation, and this led toward equalization at the upper values, On the other hand no such tendency was apparent in respect to plant weight, the lighter plants presumably having a lesser leaf area and a less advantageous leaf display. A further feature of individual plant performance emerged in this study on corn. It was found that the weight of any plant in the row tended to be inversely related to the weight of its immediate neighbors and directly related to the weight of the “second neighbor” in each direction (Fig. 11). In other words there was a trend toward an alterna-

-I

L

-

c’l AUC II

AUG

15

AUC 27

W AUG 27

FIG. 11. The correlation between the weight of the shoot (estimated, c21, or actual, W ) of a corn plant and its first to fifth neighbors, respectively. The corn was growing in a single row in a long box (Hozumi et al., 1955).

tion of heavier and lighter plants along the row. One may note, however, that the effect is less evident in the measured weights ( w ) at the final harvest on August 27 than in the estimated weights ( c21, where c is the stem circumference and I the stem length), presumably because of a greater variation in height than in weight. Despite the great plasticity of plants, the competition at high densities may be so severe that considerable numbers will die. Table I shows the

27

COMPETITION AMONG CROP AM) PASTUlW PLANTS

heavy loss of wheat plants at extremely high rates of sowing, and the death of perennial and annual pasture plants is shown in Table IV. Inspection of the wheat data shows that the “self-thinning” did not reduce the 184-density or the 1078-density to the density giving the maximum yield of grain per unit area (35/m.2); in fact, grain yield at the TABLE 1V Mortality among Perennial and Annual Pasture Plants at High Densities A. Perennial grasses competing together in low prairie (Clements et al., 1929) Plants per 100 square inches Species

Elymus canadensis Panicurn virgatum

June 27, 1924

Aug. 7, 1924

Sept. 8, 1924

May 28, 1925

Aug. 17, 1925

242 367

191 345

90 99

58 18

Not distinguishable 6

B. An annual clover, sown at various densities in Adelaide on June 11 ( Davidson and Donald, 1958 ) Plants per square link August 18 November 10

1.1 1.05

4.1 3.9

14.1 11.7

50.4 25.9

very high density was less by 25 per cent. Thus plant survival had “precedence” over total seed production per unit area; if this holds in natural communities of annuals, it means a preservation of the gene pool at the expense of the seedling population in the ensuing year. The clover data (Table IV) show clearly that mortality is related to density, or perhaps more precisely, in the terms used by Goodall (1960) and Harper (196l),to the distance of the nearest neighboring plants. There is some evidence (Sagar, quoted by Harper) that the proximity of plants of the same species may even have an influence in a complex community with many other species present. It was found in an old pasture that the survival of seedlings of plantain (Plantago lanceolatu) became greater as did the distance of the seedling from the nearest “established” plantain. There is good evidence (see later sections) that in subterranean clover swards, mortality may be due especially to competition for light, but this explanation seems unlikely for Sagar’s results in a mixed community. Studies of the performance of individual plants within crops and pastures are all too few and our thinking is too directly based on yield per unit area because this is the obvious criterion of success in agriculture. Not only do we need to know what is happening to the individual plants, but furthermore, as Goodall (1980) writes, “The relationship between yield and plant population has perhaps been obscured rather than

28

C . M. DONALD

clarified by the common practice of expressing yield in terms of unit area. This has the effect of introducing the independent variable again in the form of expression of the dependent variable, which can be avoided if yields are expressed per plant.” V. The Associated Growth of Pairs of Species

In natural communities monospecific stands are less common than stands of several species; indeed the “pure culture” is rarely found except in extreme environments such as swamps or saline areas. Yet in most of his arable cropping, man attempts to grow cultures of a single species and to regard all other plants as weeds. There are, however, some arable cropping programs involving mixtures of two species, and this is especially so in regions where hand implements are used at sowing and harvest. The degree of intimacy of mixing of the species varies greatlythe use of pastures or bananas beneath coconut palms in Ceylon, the planting of wheat among olives or cork oaks in the Mediterranean, the sowing of alternate rows of corn and cotton under irrigation in Greece, or the close intermingling of oats and vetch as a hay mixture, the “foin blanc” of the south of France. Some of these are very old practices and one may ask whether long experience has shown that mixed cropping is advantageous. A more recent development, especially of the past half century, is the sowing together of two or more species in artificially established pastures. Here a particular circumstance pertains, namely that the foliage of a great variety of herbage species can, with little or no complication, be harvested simultaneously by the grazing animal. In the case of all crops other than fodder crops, the mechanized harvesting of mixed species usually presents technical problems. It is the pasture worker in particular who asks whether he will gain or lose by sowing a mixture of species. He observes that native pasture communities commonly show considerable complexity, with several layers each of several species. May not this structure enable the community to exploit the environment to a maximum degree? Species tolerant of shade exploit the light ben’eath the upper canopy of foliage. May not the roots, with their varying patterns of penetration and concentration, better exploit the water and nutrient supply? If this is so, then a pasture of two species may well give a greater production of dry matter per acre than a pure culture. Some pastures are sown with a mixture of species or strains with the further expectation that the constituent genotypes will give production at different seasons, so that the peak of seasonal growth will be ex-

COMPETITION AMONG CROP AND PASTURE PLANTS

29

tended and total production increased. In many regions it is also common practice to sow a grass and a legume, the latter primarily to provide cheap nitrogen, the grass to exploit the nitrogen and to provide a better forage for animals than is given by pure legume swards. It is first useful to consider the simplest of mixed cultures and by the simplest of criteria. These are the instances in which the mixture is limited to two species, in which the donation of nitrogen from one species to the other is nil or negligible (grass with grass, or grass with clover for one season only or at high nitrogen levels), and in which production is measured simply as the final yield of dry matter, not of any individual component such as grain or fiber. The question then becomes-can two species fix more carbon than either species growing alone? In considering this interspecific competition it is obviously important to take into account the density of the species in the pure cultures and in the mixed sward. For example, the following alternative interpretations of “mixing two species” will give different comparisons.

Pure cultures ( i ) x plants of A per unit area (ii) y plants of B per unit area

Mixtures ( i ) x plants of A plus y plants of B per unit area (ii) x/2 plants of A plus y/2 plants of B per unit area

The first mixture, a simple addition of the pure cultures, will reduce the yield per plant of both A and B, since all plants will be exposed to an intensified level of competition compared with the corresponding pure culture. In the second situation, x / 2 plants of A with y/2 plants of B, no part of the outcome is predictable. The total yield and that of either species may be either more or less than that of the pure cultures. Several postulates can be submitted as to what may happen in this circumstance:

It is reasonable to suggest, as indicated above, that two species of contrasting habit, with respect to branching, leaf distribution, height, root distribution, mineral uptake, or other morphological or physiological character, will together be able to exploit the total environment more effectively than a monoculture, and will thereby give increased overall yield. When two species grow together one can envisage in terms of the general concept of the nature of competition by Clements, that one species may be more successful than the other in securing an undue “share” of the light, the water, or the nutrients, and that as a consequence its yield per plant will be increased while the yield per

30

C. M. DONALD

plant of the other species will be decreased. Such is the general meaning of dominance and suppression. 3. In direct contrast to this viewpoint, Ahlgren and Aamodt (1939) suggested that when various common mesophytic pasture plants are associated in pairs, the yield per plant of both species in the mixture may be less than the yield per plant in each of the corresponding pure cultures. For example, their experiments included the following findings: Yield per plant (g.) Pure cultures: 36 plants of 36 plants of Mixture: 18 plants of with 18 plants of

redtop per sq. ft. timothy per sq. ft. redtop per sq. ft. timothy per sq. ft.

0.369 0.577 0.294 0.360 They postulated that this result was due to “harmful root interactions,” presumably toxic excretions. 4. It is reasonable to suggest that any or all of the foregoing phenomena, if in fact real, may be intensified, weakened, eliminated, or reversed by changes in the total density or in the level of supply of any of the factors needed for growth. The experimental data on these points are all too few, for though there are innumerable qualitative accounts of two or more species growing together, there have not been many properly controlled experiments. It is first convenient to examine, when two species are mixed in equal proportions, how the total yield compares with that of the lower- and higher-yielding cultures, respectively. Table V shows a summary of these relationships, calculated from various studies of forage species in which dry matter production has been recorded. It can be seen that the relationship between the pure cultures and the mixtures of type (ii) can be summarized as follows: ( a ) The yield of the mixture will usually be less than that of the higheryielding pure culture. ( b ) The yield of the mixture will usually be greater than that of the lower-yielding pure culture. ( c ) The yield of the mixture may be greater or less than the mean yield of the two pure cultures. ( d ) There is no substantial evidence from these experiments that two pasture species can exploit the environment better than one. These relationships are summarized in Fig. 12.

31

COMPETITION AMONG CROP AND PASTURE PLANTS

We may next ask-what is the fate of each of the component species in such an equal mixture of two species? Since the yield of the mixture is normally less than the heavier-yielding pure culture, it follows that both species cannot have a higher yield per plant than the corresponding TABLE V The Relationship of the Yield of Dry Matter per Unit Area from a Mixture of TWO Forage Species to That of the Constituent Species Grown in Pure Culture Number of mixtures testeda

Relationship

Roberts Aberg a" ( 1943) and Olsen (1942) ( A ) (B)

Erdmann and Harrison Donald (1946) (1947) ( A ) ( B ) ( C )

( a ) The relationship of yield of the mixture to yield of higher-yielding pure culture Above Below

1 11

0 6

1 14

2 8

1 5

( b ) The relationship of yield of the mixture to yield of lower-yielding pure culture Above Below

11 1

5 1

14 1

9 1

6 0

5 1

( c ) The relationship of the yield of the mixture to the mean yield of the two pure cultures Above Below

8 4

3 3

9

7

6

3

5 1

5 1

0 6

6 9

15 0

1 5

Total

1 9

65 5

13 50 2 2 0

a Calculated from various authors. Roberts and Olsen (1942): 2 grasses, 6 legumes alone and in grass-legume pairs in the field. Aberg et al. ( 1943) : ( A ) 4 grasses alone and in pairs in the field; ( B ) 3 grasses, 3 legumes alone and in all pairs in pots. Erdmann and Harrison (1947) : 4 grasses, alone and in all pairs; numbers of plants not equal in the mixtures. Donald (1946): ( A ) 4 grasses alone and in all pairs, glasshouse bench; ( B ) 4 grasses alone and in all pairs in pots; ( C ) 3 grasses, 3 legumes alone and in all pairs, in field.

pure cultures. But are yields of both reduced, or is one above and one below that of the corresponding pure culture? Table VI summarizes the data from the same studies; it will be seen that the only significant departures (other than one) of the yield in the mixture from the yield per

32

C. M. DONALD

plant in pure culture are those in which the yield of one species is above and, of the other below, that of the corresponding pure culture. It seems that when two fodder species are grown together they give no advantage in terms of the yield of dry matter over the higher-yielding

Pure culture of speciesA

Equal mixture of A+B

Pure culture of species B

FIG. 12. A diagrammatic account of the yield relationship of two species growing alone and in a mixture,

pure culture. This presents a simple picture of the competitive relationships. One of the species is the aggressor, able to exploit more than its “share” of the factors of the environment, while the other is suppressed because it is able to secure only a lesser part of the light, water, or nutrients. This relationship is shown diagrammatically in Fig. 13. There has been no further evidence to support the earlier study by Ahlgren and Aamodt (1939) suggesting a mutual depression in yield. Because it is not usual to mix varieties when sowing a crop, there 10

10

10

10

mvIvIiiii SPECIES A

Ynld per planl 10 Ydd pv unit orto 40

SPECIES

A

+ B

Yield per planl A 14 8 3 Yield ptr untl area

34

SPECIES B

Yoeld p t r planl YKM

5

mil area 2 0

FIG. 13. A diagrammatic representation of the yield relationships commonly found when two species are grown separately and in association at “normal” secd rates.

8

5

TABLE VI The Relationship of the Yield per Unit Area of Each Species in a Two-Species Mixture to Its Yield in Pure Culture

$

Numbers of instancesa Relationship of yield of the two species in the mixture to their respective pure cultures One species greater, the other depressed Both species greater

Roberts and Olsen ( 1942)

Aberg et al. (1943) (A) (B)

Erdmann and Hamson (1947)c

Donald ( 1946 ) (A)

(B)

(C)

5 (1)

5 (-)

13 ( 6 )

9

4 (3) 4 ( 2 ) 11 (5)

5 (1)

I(-)

2(-)

1

I(-)

-

-

-

-

2(-)

3(-)

Totalb

51 (17) 15 (1)

Both species depressed 2 (-1 - 1 (-) 3 (-1 a Calculated from various authors; see Table V for a brief account of each experiment. b The numbers in parentheses show the number of instances in which the yields of both species in the mixture differ sign&cantly from those of the corresponding pure cultures. 0 No statistical analysis presented.

8

1

'd

r

P

3

v,

34

C . M. DONALD

are even fewer data on associated crop plants than on associated fodder plants. Stringfield (1959) found that when he grew mixtures of pairs of corn hybrids the yield of grain per acre from the mixture showed remarkable equality to the mean yield of the two hybrids growing separately. Any absolute increase in yield by one of the hybrids in the mixture was compensated by an equal decrease by the other. Similarly Hanson et al. (196l), and Hinson and Hanson (1962) found that any advantage gained by one of the competing pair will be lost by the other, and they considered that competition between genotypes is an additive system. On the other hand Pendleton and Seif (1962) found no such compensating loss and gain when they mixed two corn genotypes differing markedly in stature, They planted alternate rows of US-13 and of a brachytic 2 dwarf version of US-13, the former 106 inches in height, the latter 72 inches. They secured the following equivalent yields in bushels per acre: Normal US-13 Brachytic US-13

Alone

Alternate rows

Gain or loss

91.1 63.5

96.6 44.1

5.1 -19.4

+

Thus the mixture yielded only 70.4 bushels compared with a mean for the pure cultures of 77.3. It would seem that the explanation for the contrast between these results and those of Stringfield, who showed no such loss due to mixing, lay in the nature and degree of the difference between the genotypes. Stringfield’s hybrids though differing considerably in yield were presumably of closely similar stature and habit, so that there was little true intergenotypic competition, each genotype producing to its own capacity in the mixture, or with modest compensatory effects. In Pendleton and Seifs study, one of the genotypes, normal, stood over 30 inches above the associate dwarf genotype. As these authors point out, there was presumably considerable shading of the dwarf by the taller plant, but very little shading, except of the basal leaves, in the other direction. Thus the depression of the dwarf was far greater than the advantage accruing to the normal. Yet all these crop results, whether the yield was equal to the means of the pure stands, or nearer the lesser-yielding variety fall within the same general relationships as those cited for fodder plants. Despite the simplicity of these relationships, it is easy to become bemused in making comparisons of pure cultures and mixtures. Papadakis (1941) made extensive comparisons of cereal-legume mixtures for grain production, and concluded “as a general average of all mixtures and all stations, the cereal grain produced by the mixture was 61% more than the grain produced

COMPETITION AMONG CROP AND PASTURE PLANTS

35

by % hectare of the cereal grain alone. On the other hand, the grain of the leguminous plant produced by 1 hectare of the mixture was 9% less than that produced by 5 hectare of the legume grown alone. The total yield was 21% higher than the average of the yields of the two plants grown alone.” The use of mixtures was therefore advocated. His results can alternatively be stated as shown in the tabulation. Kg./ha.

Alone Cereal Legume

966 1268

Mean

1117

Kg./ha.

Mixture Cereal component Legume component

Total

777 579

1356

The mixture did not differ significantly from the higher-yielding pure culture; it was higher in 11 instances and lower in 23 instances. These data fall therefore within the general pattern of the instances already cited, where the yield of the mixture lies within the span of the yields of the two pure cultures. The advantage in practice of sowing such a cereal-legume mixture is that the grower would not need to predict whether the cereal or legume would give the higher yield. He could sow the mixture and know that he would have reasonable success. Although these relationships show high consistency in the studies here reported, they must be examined more fully in relation to density. The comparisons have been made at a single density. A mixture of x/2 plants of A with y/2 plants of B, has been compared with pure stands of x plants of A or y of B. But in all experiments, the seed rates or plant numbers have been quite arbitrarily chosen by the workers concerned, Perhaps it can be said of the rates that they approximated to “normal” rates based on local experience. And if this is so, then we can accept the findings recorded in these studies as in fact indicating the situation likely to be met in practice. But if we wish to determine with some precision whether two species have a mutually beneficial or harmful effect, then a more sophisticated approach is necessary. The problem can be illustrated as follows. Suppose two species A and B have yields in pure culture of 100 and 80, respectively, at densities of, say, 24 plants per unit area in each case. It may at first be propounded that if, when 12 plants of A are mixed with 12 plants of B the yield is above 90,there has been a beneficial interaction, whereas if the yield is below 90, there has been a harmful interaction. But suppose, first, that A will give its maximum attainable yield with only 12 plants per unit area and, second, that it is able strongly to suppress B. It would now seem likely that 12 A and 12 B will yield 100though there is no positive benefit in the sense that the environment is

36

C. M. DONALD

more effectively exploited than by A alone. On the other hand, if 500 A or 500 B are needed in order to give the full yield of either pure culture, then 12 A 12 B will presumably give a yield precisely intermediate between that of 24 A and that of 24 B, since in all three cultures each plant will grow to its full, unrestricted size. Clearly then the assessment of the effect of association must take into account the effect of density. A technique for such an analysis has been devised by G. A. MacIntyre (private communication) who has formulated a “competition index.” Suppose first that in a mixed culture of 10 plants of A and 10 plants of B, it is found ( a ) that the mean yield per plant for the A species is equal to that obtained in a pure culture over the same area of 15 A plants, and ( b ) , that the mean yield per plant of the B species is equal to that obtained in a pure culture of 30 B plants. Then the competitive effect on the 10 A plants of the 10 B plants in the mixed culture is the same as the competitive effect of 5 A plants in the pure A culture, and, likewise, the effect on the 10 B plants of the 10 A plants in the mixed culture is the same as the effect of 20 B plants in the pure B culture. In symbolic terms:

+

Competitive effect on 10 A plants: Competitive effect on 10 B plants:

10 B = 5 A; A/B = 2 10 A = 20 B; A/B = 2

In both equations, 1 A plant is equivalent to 2 B plants and the two species may be described as being competitively equivalent under the conditions which give these yields per plant. To construct a “competition index” independent of the equivalence factor which is 2 in this case, one might either determine the difference between the ratios 10/5, 20/10 (obtaining 0 ) or take their ratio (obtaining 1 ) . In either case a comparison of the products 10 x 10 ( actual mixed culture numbers) and 5 X 20 (pure culture equivalents) is involved. A convenient index is therefore:

5 x 20 (pure culture equivalents)

= 1 in this case. 10 x 10 (actual numbers) Suppose now that in the above example the numbers of plants in the pure cultures giving the same yields per plant as in the mixed culture were 12 A and 25 B, respectively, instead of 15 A and 30 B. In symbolic terms: Competitive effect on 10 A plants: Competitive effect on 10 B plants:

10 B = 2 A; A/B = 5 10 A = 15 B; A/B = 1.5

The equivalence factor for competitive effect between the species is now different for the two comparisons. Species A will tolerate without effect

COMPETITION AMONG CROP AND PASTURE PLANTS

37

the substitution into a pure culture of A, of 5 B plants for one of its own; whereas for species B the competitive effect of 1 A plant is equivalent to that of only 1.5 of its own plants. (The argument is similar if the inverse ratios B/A are considered.) There remains the question of whether the association of the two species is beneficial or not as a whole. Maclntyre’s index provides a measure of this. In the present case the index is:

2 X 15 (pure culture equivalents)

= 0.33 10 x 10 (actual numbers) and thus the mixed culture “tolerates” competition among greater numbers (in the uniquely appropriate mean sense indicated by the product of numbers) of the competing species than would be predicted on the basis of intraspecific competition. In this sense the association of the two species is beneficial. Conversely if the index were greater than unity, a harmful association would have been indicated. In generalized terms, when

N A

plants of species A compete with N B plants of species

B on a unit area, and the yield per plant of A in the mixture equals the yield per plant of N ’ A plants in a pure sward on a unit area, and the yield per plant of B in the mixture equals the yield per plant of N’B plants in pure sward on a unit area, then if

(N’A

-N A )

( N’B - N B )

< 1

N AN B

there has been a positive benefit. If the index is greater than unity then there has been a harmful association. In order to determine the “competition index” it is necessary to grow each species in pure culture at a sufficient range of densities to enable a density-yield curve to be constructed. Figure 14 shows the results of a field experiment in which Wimmera ryegrass and subterranean clover were grown in broadcast swards alone and in association at four densities and in three proportions (Donald, unpublished). It will be seen that the yield from the mixture exceeded the “proportional yield of the two species in almost all instances. Yet when the “competition index” was calculated for each plot it showed a positive effect in 28 instances and a negative effect in 32 instances (4 densities and 3 ratios times 5 reps). Of the aggregate indices for the four densities, two were above unity and two below. Thus there was no evidence of any beneficial association in this experiment.

38

C.

M. DONALD

A similar calculation was applied to Mann and Barnes’ (1953) data with ryegrass and clover growing in pots ( McIntyre, private communication). Although in this instance the data are of uncertain value because of the side lighting which occurs in pots, and which may interact with density, the analysis showed that subject to any biases introduced by boundary effects, there was a positive benefit from association of the two species. 40

30

r ._ -C

2

20

0

.-Y

>-

lo

0

I

Pure Clover

I

I

25Gr. 75CI.

5 0 Gr. 50CI.

I

75Gr. 25 C I

I

Pure Grass

Proportion of grass and clover

FIG.14. The aggregate yield from mixtures of varying proportions of Wimmern ryegrass (Gr.) and clover ( C l ) (as shown on the abscissa) and of varying density ( d is the number of plants per square link).

The present situation, then, is that there are very few sets of data lending themselves to full analysis for determination of the beneficial or harmful effects of associating species in pairs. Meanwhile the mixing of species at empirical rates approximating commercial practice shows that the yield of the mixture will usually be less than that of the higheryielding pure culture, greater than that of the lower-yielding pure culture, and about equally likely to be above or below the mean yield of the two pure cultures. In terms simply of yield there seems to be no benefit from mixing forage species. But whether this is true when pasture is grown beneath coconut palms, or when other such dissimilar species are associated, is not known.

COMPETITION AMONG CROP AND PASTURE PLANTS

39

VI. Equilibria in Mixtures

In any constant environment, whether the constancy is absolute or subject to regular diurnal or seasonal change, the vegetation will move toward an equilibrium, again either absolute or subject to regular seasonal change. This equilibrium will be expressed in terms of both the yield and the botanical composition of the community. In annual crops sown in pure culture, this phenomenon is of little significance, since the farmer, with his accumulated experience, sows at a rate at which mortality will be low or nil. He is not concerned except in freak circumstances, with the crop produced by natural reseeding, nor is he interested, except in more primitive agricultures, with harvesting and resowing the grain of mixed crops. In effect an equilibrium for the harvest time is sought at seeding time. In pastures, however, the situation is very different. Several species may be sown, the ratios of the seeds in the mixture may vary widely, the plants will persist for a number of years either by perennation or by natural reseeding, and the management will vary greatly. Reference has already been made to Holliday’s studies (1953) with perennial ryegrass, where, irrespective of plant density, the equilibrium community was one of 500 tillers per square foot in the third year; the mortality which occurs in dense stands both of perennials and annuals is similarly a movement toward an equilibrium. The less favorable the environment, the fewer plants needed to exploit it. This relationship is expressed in the density of the equilibrium community and is reflected in seed rates. For example, in Australia lucerne is sown at about 12 pounds per acre on well-watered, fertile flats, but at only 1 to 2 pounds at its low-rainfall limit. In these lessfavored areas, dense stands move rapidly toward an equilibrium of widely spaced plants. Thus in some pastures there is a choice of practice. In general, if the farmer seeks full production in the first year of growth of his perennial pasture, then he must be prepared to sow at a rate which will lead to heavy mortality of his plants in the second or subsequent years. Alternatively, if he sows the minimum seed needed to give a full stand by the second or third years, then his first year production will be low. The former philosophy tends to be adopted in regions of intensive production, the latter in land development. The same relationships apply to production early in the season and to mortality late in the season by annual pasture plants. If we turn to mixed communities the picture is more complex.

40

C. M. DONALD

Table IV has shown the move toward an equilibrium population in one of the many associations studied by Clements et al. (1929), and this undoubtedly occurs extensively in natural communities. In this instance, and in Holliday's study, the move to dominance and equilibrium was time dependent. On the other hand, in studies by Mann and Barnes (1953) an equilibrium between two species was attained within a few weeks of sowing where the density was high, i.e., attainment of the equilibrium was density dependent. Figure 15 is derived from their data.

ISO

' Rytgrarr

alone

.

loot

s

c

5 0 1

P

:+

v

o 0 No. of Ryrgrarr Plants

2

4

8

6

No of Clover Plonts

Nunb.r of Clortr Planlr wllh 2 R ) q a r s M

r

FIG.15. The yield of dry matter of ryegrass and clover, alone and in mixtures, showing the constant yield and proportions in mixtures above a threshold density. (Derived from data by Mann and Barnes, 1953.)

Ryegrass alone gave a ceiling yield of c. 100 g. with one or more plants per pot; clover gave a ceiling yield of c. 150 g. with 6 or more plants per pot. When ryegrass was added to clover or clover to ryegrass, the ceiling yield of the mixture in each instance was 150 g., i.e., it was equal to the ceiling yield of the heavier-yielding pure culture. A mixture of 6 clover with increasing ryegrass showed constant composition when the number of ryegrass plants was 4 or more; and a mixture of 2 ryegrass

COMPETITION AMONG CROP AND PASTURE PLANTS

41

with increasing numbers of clover plants showed constant composition when the number of clover plants was 6 or more. It can be inferred from these results that any mixture containing more than two ryegrass plants and more than six clover plants would, under the conditions of this experiment, have had a yield of 15Og. and a ratio of grass to clover, 1:l. Similar equilibria can be derived from data by Black (1960b) with varying densities of clover and lucerne. An example is shown in Fig. 16. 400r

LKERNE t 2 5 0

0

.

CLOVER

_ -

I

50

250

1250

6250

INITIN. N U W R OF LUCERNE PLANTSIM?

FIG.16. The trend toward constancy of yield and composition at high densities in lucerne-clover mixtures. (Derived from data of Black, 1960b.)

When increasing numbers of lucerne plants were added to a population of 250 clover plants per square meter, the mixture was stable in total yield and composition with any lucerne component above about 800 plants. This experiment was run only for 66 days and the populations of young plants would not have reached their final equilibria, which would certainly be only a small proportion of these very high densities. It seems evident that if the density of the species in a mixture is raised above certain threshold values, then the mixture will be constant in yield and the contribution of each species will also be constant. This principle is of importance in compounding seed mixtures. It indicates that the composition of the forage can be influenced only within certain limits by manipulation of the seed mixture. Botanical composition beyond these limits can be attained only through modification of the environment. For example in Mann and Barnes’ study, both the total yield and the proportion of grass to clover could have been altered far outside the equilibrium values recorded in the experiment by a change in the nitrogen status, as will be shown later in Table X. But, within

42

C. M. DONALD

any particular environment, it will be wasteful of seed to sow any constituent of a mixture at a rate exceeding that at which it will make its maximum useful contribution. We can now turn to the consideration of equilibria determined by the relative reproductive rates of associated plants. Work in this field has been the especial interest of de Wit and his colleagues ( d e Wit, 1959, 1960, 1961; de Wit and Ennik, 1958) and can best be reported by a direct account of their work. Two species (say oats and barley) are grown together in a mixture.

PURE Q4TS

PROPORTION OF SEEDS PLANTED

PURE BARLEY

FIG.17. The relationship of seeds planted to seeds harvested in mixtures of two cereals, when there is no competition, but only a different rate of seed production per plant, by each species. De Wit refers to this theoretical association as “peaceful co-existence” (de Wit, 1960).

The first situation envisaged by de Wit is one in which each plant uses its share of the environment (area or environment/total number of plants) without impinging on the environment of its neighbors; i.e., there is no competition. Nevertheless, one species, say the barley, produces more seeds per plant (or, expressed differently, more seeds per unit of environment) than does the other species. In the absence of competition environment available to barley environment available to oats

-

number of barley seeds sown (S,) number of oats seeds sown ( S o )

COMPETITION AMONG CROP AND PASTURE PLANTS

43

In this situation the relationship of seed sown to seed harvested, for a mixture of two species, will be as shown in Fig. 17. The harvested seed will have a greater proportion of barley than the seed mixture, and if the mixture were sown and harvested for a number of generations, then it would finally be composed wholly of barley. In these circumstances, without competition, the reproductive rate of barley (ratio of number of barley seeds harvested to number of barley seeds sown) is greater than that of oats, and we can express the relative reproductive rate of barley to oats as the ratio of the reproduc-

PURE QaTS

PURE BAR LEY

PROPORTION OF SEEDS PLANTED

FIG.18. The relationship of seeds planted to seeds harvested in mixtures of two cereals, when one is a more successful competitor than the other (de Wit, 1960).

tive rate of one to that of the other. In Fig. 17 the barley has a higher reproductive rate than the oats, and in this situation the relative reproductive rates would be 2 for barley and 0.5 for oats, one rate being the reciprocal of the other. However, when any two crop species are grown together in a mixture at normal rates they do in fact compete, and the plants of one species will gain more than their share (environment/total number of plants) of the environment while the plants of the other will gain less. With barley and oats, the relationship shown in Fig. 18 was commonly recorded by de Wit; the barley is the aggressor species, and the oats fails to “hold its share of the environment. When barley gains a greater

44

C. M. DONALD

proportion of the environment than the proportion of barley sown, the allocation of the environment will be: Eb - kl s b Ell kZS“

where k , and k, are coefficients showing the extent to which barley gains and oats loses its share of the environment. The ratio k,/k, is termed by de Wit “the relative crowding coefficient” of the first species in relation to the second; an alternative term would be “the relative competitive ability.” The relative reproductive rates (R.R.R. or a ) can be expressed, say for barley and oats, as ratio of proportion of barley in harvest mixture to the proportion of barley in the seed mixture a = ratio of proportion of oats in harvest mixture to the proportion of oats in seed mixture

log( H H O, ) = l o g a + l o g (

2)

If the community is entirely stable, the relative reproductive rate is unity (log a = 0) and the relationship of the harvest mixture to the seed mixture is linear at 45 degrees through the origin (Fig. 19a). If one species has a higher reproductive rate than the other, the relationship of the harvest mixture to the seed mixture will be expressed by lines lying either wholly above or wholly below the line for the stable community. One species will become less in each succeeding generation and will finally disappear because it is the less effective competitor. Two unstable situations are shown in Fig. 19b in which effects other than competition for factors of growth may be involved; here the two species have a harmful interaction, as might happen if one produces substances toxic to the other, or a beneficial interaction as when one species produces substances beneficial to the other, or when some component of the environment is distinct for the two species, such as depth of rooting, source of nitrogen, or season of growth. De Wit was not able to find a harmful interaction in his studies of flax and Camelina (which reputedly produces substances toxic to flax) though he found the relative crowding coefficient of Camelina greatly to exceed that of flax. He found the second relationship, a movement toward an intermediate

COMPETITION AMONG CROP AND PASTURE PLANTS

45

equilibrium, to operate in grass-clover mixtures. Clover-dominant swards or grass-dominant swards, subjected to an alternation of artificial winters and summers in growth cabinets each moved toward an equilibrium composition of intermediate value. This was attributed to the different source of nitrogen for each species (soil or air) whereby a greater nitro-

a

LOG

*

IN SEED S O W N

/

b

LOG

8

inlrrmcdialc

I N SEED S O W N

FIG.19. The four relationships that may exist between two species; see text for discussion. The four dotted lines may not necessarily be straight. (Drawn from various figures by de Wit, 1959, 1960, 1961.)

gen supply was possible with both species present than with a pure culture of either. Alternatively, this same phenomenon can be described by saying that when clover becomes dominant, the nitrogen accretion encourages grass growth in the next season; when grass becomes dominant and is harvested, the decrease in the nitrogen status of the soil encourages the growth of the clover.

46

C. M. DONALD

VII. Competition between Crops and Associate Plants

The two situations in which a crop is subject to competition are when there are weeds in the crop or when a pasture is being established beneath it. There is no biological difference between these situations, except that in one instance man hopes that the crop will thrive and the weeds be suppressed, and in the other that crop and pasture, growing in competition, will each do well. For the purpose of this discussion it is proposed to refer to the crop and its associate, this latter term to include either weed or forage species. The usual terms applied to a crop growing with undersown forage plants are “cover crop,” “nurse crop,” or “companion crop,” the last two terms presumably based on the odd concept that smaller plants enjoy having their light cut off or their water supply depleted. Yet the most obvious point regarding any community of crop and associate species is that each will be reduced in yield compared with a pure culture of like density of that species. (This is the simple addition of two communities, not the admixture of two “half communities”-see Section V.) It is scarcely necessary to document the reduction in crop yields which occurs when there is weed infestation: equally it is axiomatic that cereals will be reduced in yield if, without change of crop density or spacing, they are interplanted with a forage species. The reduction, especially in the yield of grain, may be very slight, but this will especially be the case only ifthe fodder species is heavily suppressed. Pendleton et al. (1957) recorded depressions in the yield of corn of about 15 bushels per acre (about 15 per cent) due to interplanting with alfalfa, and Staniforth and Weber (1956) showed a depression of 10 per cent in the yield of soybeans due to infestation with weeds. Australian data in Table VII similarly show a 25 per cent depression in the grain yield of wheat when forage species were drilled in the same 7-inch rows, or of 16 per cent in the case of 14-inch rows. The depression of growth of the associate species due to competition will generally be far greater than that of the crop, though unfortunately, few workers have determined the production of the associate species when grown alone. The study reported in Table VII shows a depression in the growth of the pasture plants of over 70 per cent when they compete with wheat in 7- or 14-inch rows. When forage plants are sown with crops, it is implicit in the adoption of the practice that only a small depression of crop yield is expected or is acceptable, but that a depressed growth of the forage species will still provide an adequate base for the following year’s pasture production. Accordingly the growth form and

47

COMPETITION AMONG CROP AND PASTURE PLANTS

the rate of seeding of the selected forage species are always such that the forage species will be the suppressed component of the association, In particular, they are always of less stature than the crop, This means that while they may be strong potential competitors for water and nutrients they suffer such heavy shading that their growth, and in turn their realized capacity to compete for water and nutrients, is greatly reduced. TABLE VII The Competitive Relations between Wheat and Undersown Pasture Species (Lolium rigidum and Trifolium subterruneum) at Adelaide, South Australian Total dry matter production (g./m.2)

~

Treatment 1 2

Associate pasture species

Wheat crop Spacing 7-in. rows

-

Yield 896

All

Grain yieldby wheat (g./m.2)

Spacing

Yield

species Yield

-

-

896

482

~

-

7-in. rows

854

854

-

-

3

7-in. rows

815

7-in. rows (in 215 same rows as wheat )

1030

363

4

14-in. rows (alternate 7-in. rows)

630

14-in. rows (alternate 7-in. rows)

355

985

252

5

14-in. rows

870

-

-

870

333

6

-

-

14-in. rows

652

652

-

7

14-in. rows

759

14-in. rows

185

944

281

(in same

rows as wheat) a

Santhirasegaram (1982).

In contrast to the depression of crop yields quoted above, there was no effect of the forage species on the yield of the barley in the six-year studies by Jarvis et al. (1958)of the effect of undersown grasses and legumes on the yield of barley in England, and this is by no means a unique finding. It seems that only one interpretation of these results is possible. Under the conditions of their study, there was presumably no measurable competition for water or nutrients, for then the undersown species would have reduced the supplies to the barley; there was competition only for light. In competing for this factor the barley had a m plete “priority” because of its greater stature. Nevertheless, some light

48

C. M. DONALD

is unexploited by cereal crops, and if water and nutrients are nonlimiting, it can permit a weak growth of undersown species without affecting the crop. It seems to be well established in the relationship of crop and associate that if the distance between the crop rows is increased or the density of crop within the rows is decreased, then there will be a stronger development of the associate species. When the crop is thus reduced in density, the decrease in the yield of grain due to wider spacing of the rows or plants will, however, not be proportional to the wider spacing, despite the stronger growth of the associate species. These two points are illustrated in the study by Pendleton and Dungan (1953) on oats and red clover. Here a doubling of the space between the cereal rows (halving the number of rows) still permitted 82% of the yield of grain with 8-inch rows. The more widely spaced rows of oats were less competitive with each other than at normal spacing-or, stated otherwise, they were each able to command a greater quota of light, water, or nutrients than the more closely spaced rows. In these studies the growth of the clover was not obviously affected (yield was not measured) in normal seasons by the spacing of the oats, but was greatly enhanced by wide spacing of the oats in a dry season (e.g., 1950, Table VIII) indicating that competition for water was a dominant factor in that season. A striking effect of the seed rate of the cereal on the growth of the associate species (mainly weeds but also including some undersown TABLE VIII The Influence of the Spacing of Oats on the Growth of the Cereal and of an Associate Species (Red Clover) ( 1950 Season)" Spacing - of oats rows, constant rate in the row ( inches )

Yield of oats

Manner of spacing 8-in. oats rows All rows sown

8

Per cent of rows sown

Bu. per acre

Relative yield per acre

Relative yield per row

Growth estimate of red clover

100

78

100

100

19

67

68

87

128

25

16, 8, 8, 16 16

2 rows in 3 sown Each second row sown

50

64

82

164

48

24

Each third row sown

33

51

65

195

71

Clover alone

No oats sown

-

-

-

100

Pendleton and Dungan (1953).

0

49

COMPETITION AMONG CROP AND PASTURE PLANTS

species) was shown in the study by Bula et al. ( 1954). The total yield of dry matter was constant at seed rates from 0.5 to 3 bushels, but the associate species were reduced from nearly 50 per cent of the total yield to less than 10 per cent (Fig. 20). Similarly Weber and Staniforth (1957) showed a marked reduction in the yield of weeds (foxtail and

r

Total

yield

40 e 0

\

s3 0 L

2

e ?i

0

6 2o

-U 0

.-

>. 10

0

0

0.5

I .o

1.5

2.0

2.5

3.0

Seed rate of oats (bush)

FIG. 20. The relationship of the seed rate of oats to the yield of associate species. (Drawn from data by Bula et al., 1954.)

smartweed) as the seeding rate of soybeans was increased. Figure 21 shows the effect of the seeding rate of wheat at Adelaide on the growth of associate clover. In all these cases, it is reasonable to assume that the higher population of crop plants was able to gain greater dominance than could a lesser population because the photosynthetic surface of the crop per unit of land surface (its leaf area index) at the time of germination was far greater. Crop growth rate would initially be linear on density and thus give greater opportunity for the crop to gain an earlier advantage in the competition for water, nutrients, and light. In further studies in Iowa (Jorge and Staniforth, 1961), the interaction between fertility status ( level of nitrogen application) density of the corn population, yield of corn, and growth of the associate species (two weedy Setaria species) was examined (Table IX). The marked competition for nitrogen was relieved as nitrogen levels were raised, so

50

C. M. DONALD

that the yield both of corn and foxtail rose, and the depression in corn yields due to foxtail became less. The competition for light was in favor of the corn because of its greater height; the foxtail was heavily reduced by the shading in dense populations, and its percentage effect on corn yields became less as the corn population increased. Thus the com0-0-0

HARVEST

1

m-m-¤

HARVEST

2

600 -

-

-

500

N.

2

-

400J

Y

-

n W

;I 300-

-

z

0

3 200

0

100

A

OL 0'

I

I

I

15

30

45

RATE

OF

SOWING

I 60

(WHEAT, LBS./AC.)

FIG.21. The influence of the rate of sowing of wheat on the yield of the undersown clover. (Dry matter of the clover at two dates and production of clover seed. ) ( Santhirasegaram, 1962.)

bination of high nitrogen (removing the effect on the corn of competition for that nutrient) and high density (placing the foxtail at maximum disadvantage in the competition for light) were the effective means of minimizing the depression of corn yield due to these weeds. Detailed study of the environmental conditions affecting associate species within crops have been made by Larson and Willis (1957) and Santhirasegaram (1962), the former in a corn crop, the latter in a wheat crop. Larson and Willis found a considerable range of environmental conditions between E-W rows of corn 80 inches apart under Northern

51

COMPETITION AMONG CROP AND PASTURF, PLANTS

Hemisphere conditions. The zone lying 20 inches to the south of any row had greater radiation, higher soil temperature, and lower soil moisture than did the zone 20 inches to the north of the row. The establishment of alfalfa seedlings was far higher to the north of the rows, where moisture conditions were more favorable. TABLE IX The Influence of Nitrogen Level and Crop Density on the Growth of Corn and Foxtaila

Parameter

Nitrogen level, No9 N70, N1.40

Corn density, 8000 to 20,000 plants per acre

Yield of corn (grain) in The highest yields were at high nitrogen and high absence of foxtail density. The greatest response to nitrogen occurred at the highest densities. Yield of foxtail (dry mat- There was an increase ter) in the mixed comfrom No, to N,,, but munities (pounds/acre) not from N,, to NI4,: At No : 1580 At N,o: 2190

The yield of halved at density: At 8,000: At 20,000:

Influence of foxtail on The depression of corn corn yield in the mixed yield was least at communities high nitrogen At No : 29% At N,o: 15% At N,,,: 11%

The depression of corn yield was least at high densities At 8,000: 19% At 12,000: 19% At 18,000: 17% At 20,000: 14%

a

foxtail was high corn 2880 1410

From data of Jorge and Staniforth ( 196l), Table 2.

Santhirasegaram (1962) grew subterranean clover as an undersown species in wheat, placing it in rows at distances of 0 (along wheat row) 3%, 7, and 10% inches from the 14-inch wheat rows in both N-S and E-W plantings. His results, obtained in the Southern Hemisphere, are shown in Fig. 22. It will be seen that the growth of the clover showed direct relationship to the light density available to it. In the N-S rows, the best-lit clover was in the 7-inch position, halfway between the 14inch rows. In the E-W rows, the clover immediately south of the rows was the most heavily shaded and made the poorest growth. These results are the converse of those of Larson and Willis, who found that where radiation was highest, the establishment and growth were poorest because of adverse water relations. Santhirasegaram points out that, in his own studies, the crop was grown during the winter months of a Mediterranean environment, when water is commonly nonlimiting for

52

C. M. DONALD

several months (it normally becomes the basis of intense and critical competition in the late spring) and that radiation values are not high during these months. In contrast, in the United States studies, summer radiation values were involved and water relationships were more

.-*-.

HARVEST 1 N-5

E-W

ROWS

LIGHT ABOVE CLWER A1

N-S

ROWS

# z

k 9 ctr

Po

2 E-W

ROWS

HARVEST

HARVEST 3

n

T.S.N.

HARVEST

k9

ROWS

A

pi&

FIG. 22. The relationship between the yield of clover lying in various positions between 14-inch rows of wheat, and the light intensity in each position (Santhirasegaram, 1962 ).

critical. Thus the influence of high radiation values, with their opposing effects on the light status and moisture status of the undersown species, will depend on local and seasonal circumstances. VIII. Competition between Grass and Clover

The balance between grass and clover in mixed swards is highly susceptible to environmental change. Many studies have shown that a sward can become grass-dominant or clover-dominant according to the nutrient status or grazing management imposed on it, as the following few examples will show. Jones (1933) applied a number of

COMPETITION AMONG CROP AND PAS-

PLANTS

53

grazing treatments to a sward principally of perennial ryegrass and white clover. The original sward had 33 per cent white clover. When heavy spring grazing was imposed, this proportion was raised to 66 per cent, whereas with light spring grazing the clover was reduced to 13 per cent. This effect on botanical composition was due in part at least to the defoliation by the grazing animals, because many experiments have shown the influence of the height of clipping on the proportion of clover in mixed swards. For example, Robinson and Sprague (1947) found that there was 57 per cent clover in a grass-clover sward when cut closely (4inches to % inch), but only 36 per cent when the sward was allowed to grow taller (cut at 5 to 2 inches). Similarly Mott ( 1943) found that a mixed sward dipped when 4 to 6 inches in height throughout the season yielded 75 per cent clover, but when cut only twice during the season, it gave only 44 per cent clover; Nelson and Robins (1956) showed a reduction from 38 per cent clover to 20 per cent clover when an orchardgrass-ladino clover sward was cut less frequently. Nevertheless if cutting is very close or very frequent, the percentage of clover may actually fall, because all its laminas are removed while the grass, especially if of prostrate habit, retains some photosynthetic surface. For example Brown (1939) found that when swards were clipped regularly to % inch from 2 inches, a very severe treatment, the percentage of clover was less than when clipped to 5 inch from 5 inches. Turning now to the second well-known factor affecting the proportion of clover-there is abundant evidence that a low nitrogen level favors clover whereas high nitrogen favors grass. This is equally striking in pot culture or in the field, as shown by the examples in Table X. We may now consider the modus operandi of these two factors, the height of cutting and the application of nitrogen. It was considered at first (Blackman, 1934) that the effect of sulfate of ammonia on mixed swards might be a direct toxic effect on the clover. Blackman and Templeman ( 1938b) then studied this clover depression by examining the behavior of pure stands of grass and clover, respectively, and extrapolating their findings to mixed swards. They found that nitrogen applied as sulfate of ammonia or calcium nitrate had no adverse effect on pure stands of clover, but a reduced light supply greatly curtailed yields. As a consequence they concluded that the effect of nitrogen on a mixed sward was to stimulate grass growth, SO that increased shading depressed the yield of clover, without any direct effect of nitrogen on the clover. This phenomenon has recently been examined directly in mixed

TABLE X Examples of the Influence of Nitrogen Status on the Competitive Relations of Grass and Clover Author

Nitrogen rates and yields of herbage

Species

Trumble and Shapter (1937)

Lolium rigidum and Trifolium subterraneum in pot cultures

N ( g./pot) Grass ( g./pot) Clover ( g./pot )

0 10.4 54.5

0.5 65.9 26.5

1.0 126.6 8.3

2.0 192.8 2.2

Stem and Donald (1962a)

L. rigidum and T . subterraneum swards in sand cultures

N ( g./mz) Grass (gJdm.2) Clover ( g./dm.z)

0

1.2 12.8

2.5 2.8 11.6

7.5 5.3 8.2

22.5 12.9 0.8

Petersen and Bendixen ( 1961)

Dactylis glomerata and Trifolium repens, field plots

N (lb./ac.) Grass ( 100 lb./ac. ) Clover ( 100 Ib./ac.)

0 31 42

80 42 30

120 52 25

160 50 21

Nelson and Robins ( 19%)

D . glomerata and T . repens, field plots

N ( lb./ac. ) Grass ( 100 lb./ac. ) Clover ( 100 Ib./ac. )

0

44 38

50 60 36

100 65 30

200 79 20

.n !

2 6

COMPETITION AMONG CROP AND PAS=

55

PLANTS

14. IX . 5 7 .

67 days

I . X . 57.

c1or.r

tro..

ClO..,

Grel,

c1.r.r

tl...

04 d a y s

? b

16.X.57. 99days

-I

W

-I

n

z

a 0 a

r

SO.

0 W

>

b0

.

0 m

30.X.57. 113 d a y s

c1or.r

GIO..

I-

I

2 W

I

m.

SO

c1n.r

Gml.

c1or.r 60

. I G

I /

I

Gm..

ClO".,

I /

l9.XI.57. I33 d a y s 40

10

0

20

40

60

SO

KK)

LIGHT NO

0

0

DENSITY N2.5

(PER

20

4 0

LO

80

100

CENT DAYLIGHT) N7.5

N22.5

FIG. 23. The dcveloprnent of the leaf area, and its vertical distribution, in mixed swards of grass and clover. The swards vary from a clover dominant community at low nitrogen to a grass dominant community at high nitrogen. The curve shows the profile of light intensity (Stem and Donald, 1962a).

56

C . M. DONALD

swards by the study of leaf area and light profiles (Stem and Donald, 1962a). Figure 23 shows the influence of nitrogen status on the leaf area index at various nitrogen levels. Low nitrogen gave almost pure clover, whereas at a high nitrogen level, grass was markedly dominant (see Table X for yields). Figures 23 and 24 show the sequence of effects following the application of nitrogen to have been: I

O

-

I

N7.5

N225

~-

10 V l f L D OF CRASS

5.

/

I

1.L

or

GlliSS

A B O V E CLOVER

2.

O A V I flOU SOWING

FIG. 24. The sequence of effects following the application of nitrogen to a grass-clover sward, leading to the reduction of the clover (Stem and Donald, 1962a). See also Fig. 23.

/ 4

1'1

I icr. 40

m

Ix)

0

40

FIG.25. The level of radiation at the surface of nant on the left to grass dominant on the right. The part of the columns the radiation at the surface of of clover present at each nitrogen level. Southern

120

three grass-clover swards, varying from clover domicolumns show the level of radiation, and the unhatched the clover. The superimposed graphs show the weight Hemisphere. (From Stem and Donald, 196213.)

58

C. M. DONALD

( a ) Increasing rates of nitrogen application gave increased yields of grass (positive). ( b ) Increased yields of grass gave higher leaf areas of grass disposed above the clover leaf canopy (positive). ( c ) Higher leaf areas of grass above the clover reduced the light intensity at the clover leaf canopy (negative). ( d ) Reduced light intensity at the clover leaf canopy caused a reduced growth of the clover (positive). The decrease in the weight of clover in the N2Z,,-swards was an expression of two effects of low light intensity. There was a decline from

PURE CLOVER

CLOVER IN GRASS

120 D\Y5

MAY

I

JUNE

I

JULY

I

AUGUST I SEPT

80

40

0

CAY5

FROM SOWING

MAY

I

JUNE

I

I20

FROU SOWING

JULY

1

AUGUST

I SEPT

FIG.26. The level of radiation received by a pure clover sward (left) and a mixed sward (right). The unhatched part of the columns in the mixed sward shows the level of radiation at the surface of the clover. The superimposed graphs show the weight of clover present. Southem Hemisphere. (From Stem and Donald, 1962b.)

25 to 6 plants per square decimeter and, in addition, a decrease in the weight of the surviving plants. Figure 25 shows the trend in light intensity at the surface of the clover in these mixed swards. At low nitrogen little light was intercepted by the grass, but at high nitrogen the light intensity was reduced to a very low IeveI beneath the heavy grass canopy and the clover receded markedly in weight. Figure 26 shows a comparison in which the grass and the clover were competing only for light. Pure cultures of grass and clover, respectively,

COMPETITION AMONG CROP AND PASTURE PLANTS

59

were grown in slotlike containers that were held closely together with carpenter’s clamps (see Fig. 27). In this way the plants in neighboring slots were quite independent for nutrients and water supply, and competed only for light. It will be seen from Fig. 26 that whereas the clover in the pure culture grew vigorously, the clover in competition with the

FIG.27. Grass and clover growing under conditions permitting competition only for light. The plants are in narrow, slotlike, removable containers, held together by clamps. Figures 26 and 28 illustrate data derived with this technique (Stern and Donald, 1962b).

grass for light, even though it had ample water and nutrients, suffered such severe shading as to be eliminated from the sward. Figure 28 shows the relationship of the amount of clover present to the radiation at the surface of the clover in the same mixed sward. It will be seen that when the radiation fell below about 80 cal./cm.2/day due to competition for light by the grass, the photosynthesis of the clover could not keep pace with the respiration and the plants sharply lost weight. It is evident that the effects of clipping, grazing, or nitrogen application on the proportion of clover in grass-clover swards are largely accountable as a single phenomenon, namely the direct and indirect effects on the shading of the clover by the taller grass. Defoliation reduces the competition for light suffered by the clover, and nitrogen increases it. There is, however, another component of grazing which is the converse

60

C . M. DONALD

of defoliation in its effects on the proportion of clover in the sward. Much of the nitrogen voided by grazing animals, principally as urine, is immediately available for plant growth. In New Zealand it is estimated that of the total N voided, 70 per cent is in the urine. After taking account of the loss by volatilization (12 per cent) and the slow availability of some fractions (10 per cent), about 50 per cent of the

SLP I

u

-

0

MEAN

1

.

1

.

t

RADIATION

.

1

.

1

.

80

40 AT

SURFACE

1

.

1

I20 OF

CLOVER

.

1

.

1

.

160

IN

PRECEDING

1

200 14

DAYS

C A L . / C M ~DAY

FIG. 28. The level of radiation at the “surface” of the clover canopy in a grassclover sward, showing the deterioration of light density and the eventual death of the clover. Southem hemisphere. (Stem and Donald, 1962b.)

total voided N is available for immediate re-use by plants. This component of grazing, by permitting a rapid recycling and use of the nitrogen, encourages grass growth and thereby clover suppression. Because of the progressive suppression of clover through shading when nitrogen is applied to mixed swards, it follows that fertilizer application and rhizobial activity cannot be used as additive sources of nitrogen. In general, and with some variation of pattern, applied nitrogen will continue to displace rhizobium nitrogen until displacement is complete. If fertilizer nitrogen is further increased it will give a net

61

COMPETITION AMONG CROP AND PASTURE PLANTS

improvement in nitrogen supply. This is well illustrated (Table XI) in the results of Linehan and Lowe (1960) in Northern Ireland. If legumes cannot be grown successfully in pastures, there is clearly a prima facie case for the use of fertilizer nitrogen, as on bermudagrass in the Gulf States, or on ryegrass on the heavy river clays of Holland, where white clover does not thrive. But where legumes can be grown TABLE XI The Influence of Fertilizer Nitrogen on a Grass-Clover Sward (Linehan and Lowe, 1960) N (Ib./ac./year) Per cent clover Yield of herbage (cwt. dry matter/ ac./year ) Yield of protein (cwt./ac./year)

0

35

70

140

36

28

21

7

59

60

62

68

10.1

9.9

\

9.3

-

-10

1.3

10.1 /

V

-21

350

0.3

78

93

11.3

16.6

Net increase in supply of nitrogen

Replacement of clover nitrogen

Net recovery of applied nitrogen ( % )

210

0

+ 10

+28

satisfactorily the application of nitrogen should be regarded as giving no gain in production until the clover has been lost by competition through the increased growth of grass, and even then it may be a small or negligible gain (Table XI1 ) . Nevertheless there are theoretical reasons why pure grass may exceed grass-clover in its production, though these are imperfectly understood. First, a substantial weight of substrate materials in the clover must be consumed as energy for nitrogen fixation. This may be of the order of 20 units dry matter per unit of nitrogen (Bond, 1941) so that if 100 pounds nitrogen is fixed, as much as a ton of dry weight may be lost. [In turn this point may be reduced in significance because the relationships of time, leaf area, and growth of the sward are by no means linear (Donald and Black, 1958).] Secondly, grass may have a greater potential yield in any given light environment than clover, because grass foliage tends to erectness, whereas that of clovers tends to be horizontal (Brougham, 1958a; Wilson, 1960). Again, this point is not adequately resolved. Thirdly, experience shows that when a mixed sward is given nitrogen, the grass always responds in growth; it follows as a

62

C. M. DONALD

reasonable deduction that in any grass-clover sward the grass is suffering some degree of nitrogen deficiency. Or stated in other terms, clover grows only under conditions of such partial nitrogen deficiency as will restrict grass growth and admit light to the clover. But again, this proposition cannot be regarded as proved. In the field there is no consistent picture of greater production by fertilized grass than by grassclover. TABLE XI1 A Comparison of Yields from Unfertilized Grass, Fertilized Grass, and Grass-Clovera Country and treatment New Zealandb ( + N = 620 lh. N/ac./year) United KingdomC ( + N = 210 lh. N/ac./year) United Kingdom6 ( + N = 70 lb. N/ac./year) a 0 c

6

Pounds dry matter per acre Grass

Grass

+N

Grass-Clover

2,200

10,500

11,500

1,300

7,870

6,230

2,263

4,430

6,861

Values are 3-year means. Sears (1960). Green and Cowling ( 1960). Herriott and Wells (1960).

The finding that the effect of low nitrogen dressings is simply to replace clover nitrogen has led some United Kingdom workers (e.g., Watkin, 1954) to the viewpoint that an all-or-nothing policy should be followed in the use of nitrogen-that it should not be applied at a light rate over the whole farm, but rather that a majority of the fields should carry grass-clover, while special-purpose areas ( e.g., for an “early bite” in the spring) should be sown to pure grass, such as Italian ryegrass, and heavily fertilized with nitrogen. Kennedy (1958) based his recommendations for the use of nitrogen in New York State on the amount and vigor of the clover in the sward; he advocated the use of 60 pounds nitrogen on an all-grass sod, only 25 pounds nitrogen where there were moderate amounts of legumes, and none on vigorous grass-clover mixtures with over 20 per cent clover. This discussion of grass-clover competition has been written wholly in terms of nitrogen status and light relationships, and there is no doubt that these govern the competitive pattern in most of the grass-clover pastures of the earth. There are, however, other circumstances in which the status of nutrients such as potassium, sulfur, phosphorus, or molybdenum or of water, may in large measure govern the grass-clover relationships. These are discussed in other sections.

COMPETITION AMONG CROP AND PASTURE PLANTS

63

IX. Plant Arrangement

Conventional spacing practices for crops have developed through experience of yield and of convenience. Wheat is grown in rows 7 or 8 inches apart, corn in rows about 42 inches apart, and pastures often as continuous broadcast swards. We may ask whether these patterns are optimal or whether departures from them make any difference at all. Assuming that the pattern is simple, we can recognize three components of arrangement that might influence yield, namely: ( 1) A square grid, on the one hand, or a progressively elongated rectangle, on the other; e.g., 4 X 4 inches or 2 x 8 inches or 1 x 16 inches. Elongated rectangles give a row distribution; e.g., 1 X 16 inches gives 16-inch rows with 1” between plants ( 2 ) Regular spacing as contrasted with irregular spacing in the row; e.g., 3-3-3-3 inches or 4 3 - 2 3 inches between plants. Alternatively, irregular spacing may arise in hills through variation in the number of plants per hill ( 3 ) The direction of the rows (N-S, E-W)

1 . Square Grid or Rectangle Reference has been made previously to Goodall’s (1960) emphasis on the value of looking at yield per plant rather than yield per acre. Goodall further adds: “Considering each plant as an individual, whose yield is affected by the plants around it, has the additional advantage of making it easier to treat the effects of different plant arrangements concurrently with those of different population densities.” He then postulates, on slender theoretical grounds, “that the reduction in photosynthesis through competitive reduction in light intensity will over a considerable range increase roughly in proportion to the inverse cube of the distance (of the neighbouring plants) , . . and that . . . the effects of below-ground competition on yield should increase with proximity rather more than in inverse proportion to the distance. . . . The (competitive) effect should increase inversely as a power of the distance (of the neighbours) ranging between the first and third power, according to the relative importance of below-ground and above-ground competition.” This is a most interesting approach both to planting pattern and density, providing a quantitative basis for thought on the d e c t of density on growth. We now turn to the cumulative effect of neighbors, as examined in

64

C.

M. DONALD

the work of Sakai (1955), in which he surrounded a single rice plant by six other plants arranged in a hexagon; these six plants included respectively 6, 5, 4, 3, 2, and 1 plants of “red rice” and 1, 2, 3, 4, 5, and 6 plants of an upland rice variety of weak competitive ability. As shown in Fig. 29, there was a linear increase in the yield of the central plant

2!

w 3

-

4-

I

1

6/0

1

I

I

NUMBER

OF STRONGLY AND

1 016

313

WEAKLY

COMPETITIVE NEIGHBORS

FIG. 29. The linear relationship between the number of surrounding competitors and the yield of a central plant (Sakai, 1955).

as the proportion of the strongly competitive red rice plants in the surrounding hexagon was decreased. (Sakai suggests that this gives maximum advantage to a mutant of greater competitive ability, because, initially, it will be surrounded by weak competitors. ) Goodall (1960) points out that the decrease in the logarithm of the weight of a plant is a satisfactory expression of the competitive effect of its neighbors. He then tested Sakai’s data and showed that the linearity of additive effect of the competitors ( u p to 6 in number) based on the log yield of the center plant is as good as the linearity based on the yields themselves. If these postulates are accepted-that the effect on the log yield of any plant in a crop depends on the distance and the additive effect of the neighbors, then, Goodall proposes, the yield of a plant in a row crop should be represented by log w = u

+ b, log + bz log XI

XZ

COMPETITION AMONG CROP AND PASTURE PLANTS

65

where w is the weight per plant and x1 and x2 are the spacing within the row and between the rows, respectively. He tested this formula on various data [especially those of Wiggans (1939), who varied the planting arrangement of soybeans] and found that it seemed to have general validity, although with some anomalies remaining to be examined. A possibly unsatisfactory feature of this equation is that it implies that if, say, b, > b2, then the optimum spacing at any given density would be that in which spacing in one direction is as wide as possible and in the other is as close as possible. This paper of Goodall is reported at some length because it provides a refreshing change from the scores of empirical studies of density and spacing, many of which establish only that in “County Cornstalk in 1957, 36-inch spacing was better than 42inches-or vice versa or more or less. Any difference in the value of bl and b2 in Goodall’s equation, such as he found in Wiggans’ data, is presumably due to the inequalities created by differences in cultural treatment, in particular to unidirectional soil cultivation, but possibly also, if any single experiment is examined, to the influence of row direction in relation to radiation. If however, there is no inequality of effect in the two directions, and if the direct distance of neighbors or any power of it (this is in contrast to the row or plant spacing) governs competitive effect, then square planting will give the minimum competitive effect of neighbors and this will presumably lead to maximal plant yield. The evidence that square planting will give the highest yield, although not wholly consistent, seems reasonably firm,as illustrated by the comparisons given in Table XIII. There is advantage to be gained by reducing to a minimum the distance from the center of the nearest plant TABLE XI11 The Effect on Yield per Plant (or per Unit Area) with Approach to Square Planting at Constant Seed Rate Corn (Hoff and Mederski, 1960)

Soybeans (Wiggans, 1939)

Irrigated grain sorghum (Porter ei al., 1960) Broad beans (Hodgson and Blackman, 1956)

In four experiments, 5-10 bu./acre greater yield from equidistant spacing than from 42-inch rows. Margin more or less constant from 10,000 to 20,000 plants per acre 8 x 2 inches gave 39.8 bu./acre 16 x 1 inches gave 36.5 bu./acre inches gave 32.0 bu./acre 32 x Yield (mean of 3 years, 3 rates, and 3 levels of N ) was 5880 lb. in 40-inch rows and 6609 lb. in 12-inch rows, i.e., at constant seed rate No effect on yield at constant seed rate at 18, 36, and 54 cm. row spacing.

66

C. M. DONALD

to any point in the soil, or in increasing to a maximum the distance of the nearest neighbor casting shade. For any given density this means hexagonal planting; square planting is a near practical approximation. In the United States there are three main spacing arrangements recognized for corn: ( a ) In drill rows, e.g., in rows 42 inches apart with plants 10% inches apart in the rows (14,224 plants per acre) ( b ) In hills (also known as check-row planting, checked planting, hill planting, and hill grouping) in which plants are grouped to give a square pattern of groups of plants, that is, to give crosswise rows, e.g., 4 plants per hill, in hills 42 X 42 inches (14,224 plants per acre) ( c ) As uniformly distributed plants, that is to say, in strictly square planting, e.g., 1 plant per hill in 21 x 21-inch spacing (14,224 plants per acre) Dungan eta,?. ( 1958) have reviewed the comparisons of these methods. It would seem that there has been no regular advantage in yield from drill rows or hills, respectively, but these studies have often been of very limited value as comparisons of competitive relations because of differences in weed infestation or amount of cultivation. The few experiments comparing hill grouping and uniformly distributed plants, at the same population per acre, are also reviewed by Dungan et al. (1958). They show that under favorable conditions, uniform spacing gives a greater yield than hill grouping, sometimes substantially so; this superiority might be expected on theoretical grounds. However, when conditions are adverse, e.g., dry weather, then the advantage is narrowed or may disappear. This suggests that the advantage of uniform distribution may lie in the reduction of competition for light, but that when the water supply is restricted it supersedes light as the prime limiting factor and reduces each planting system to a common yield. The influence of row width on the yield of small grains is reviewed by Holliday (1963). At constant seed rate a decrease in row width below the standard 7-8 inches (i.e. a change towards square planting) has generally given a small increase in yield, while increasing the row width above the standard has usually given some decrease in yield. It should be noted, however, that where it is customary to localize fertilizer along the row, the degree of localization becomes less as row spacing is decreased. Thus while the increase in yield achieved under good conditions by

COMPETITION AMONG CROP AND PASTURE PLANTS

67

square planting is sufficiently substantiated, there are so many other cultural factors involved in its application that, in many circumstances, it may not be worth while.

2. Regulurity of Distribution If wheat plants are distributed at 2-inch spacing in drill rows 7 inches apart, then irregularities in the 2-inch spacing will give alternations toward or away from a square planting pattern. Instead of the plants each occupying areas 7 X 2 inches, they will occupy areas 7 x 4, 7 x 1, 7 x 2, and 7 x 1 inches. The influence of these irregularities, such as normally occur with a wheat drill, has been examined by several workers. Engledow (1928) and his colleagues examined the relationship between the number of plants and the yield per foot length of row in wheat fields in England and found positive correlations ranging from 0.58 to 0.68. He concluded that “the yield per foot length is directly and causally connected with the number of plants . . . that the population density is an important limitation to yield in typical field crops in this country appears beyond doubt . , . even in good average fields of corn (wheat) in this country 2040% of the potential maximum yield was lost by reason of small gaps or low densities in population.” Sprague and Ferris (1931) and Smith (1937) examined Engledow’s proposition in some detail. They confirmed the correlations established by Engledow but postulated a different explanation of the phenomenon. Smith put forward two premises, namely, that in a crop with a normal distribution of density per foot length, such as he had found in the field: ( a ) The average densities of nearby foot lengths must be equal to the mean plant density of the field or vicinity and therefore a foot length with few plants will, on the average, be surrounded by more densely populated foot-lengths, and vice-versa, and ( b ) if the yield of a foot length is affected by surrounding densities in proportion to the ratio of their plant densities, then this alone may induce a correlation of plant density and yield. Smith tested this interpretation by several methods and established that his hypotheses did in fact account for the correlations recorded between the number of plants per foot length and the yield. He showed that there was a strong negative correlation between the yield per foot length and the number of plants in the adjacent rows, and that if the yield in any foot length were corrected for the number of plants in the adjacent foot lengths, then the correlation of yield and density disappeared.

68

C . M. DONALD

Both groups of workers tested the effect on yield of variation around a mean density per foot length and thereby gained direct evidence that Engledow’s interpretation of his findings was incorrect (Table XIV). Sprague and Ferris (1931) not only found that the yield did not differ significantly between regular and variable spacings, but further that the TABLE XIV The Influence of Density and Regularity on the Yield of Wheat

I.

From Sprague and Ferris (1931) Seeds per foot

Treatment (1) (2)

Range of values Constant 13-29

Mean value 21 21

Yield ( g./ft. 1

7.2 7.6

11. From Smith (1937) Seeds per foot Range of values

Treatment (a)

(b) (C)

(d) Q

Correlation

Constant (evenly spaced) Constant (variable spacing) 1 to 13 1 to 13 (I)

Mean value

Y.P

Yield per plot

12

-

356

12

-

7

0.7W 0.71a

38 1 393 387

7

I

of yield ( y ) and population ( p ) per foot length of row.

grain yield per foot length was just as variable in the constantly spaced foot lengths as in those of variable density. Turning to Smith‘s results, also reported in Table XIV, the typical correlations recorded in reference to his treatments ( c ) and ( d ) , would, if Engledow’s views were correct, lead to the attainment of higher yield when the spacings reported under treatments ( a ) or ( b ) were adopted. This did not occur. It is clear then that the positive correlation of yield and density in individual foot lengths in a wheat crop is the outcome of competitive relationships between rows and that because of plant plasticity uneven sowing will not affect wheat yield so long as plants are still “within reach” of distant soil and light. The degree of unevenness of sowing done by a wheat drill does not affect yield. The most explicit work of this kind on corn (Zeu mys) is that by Kiesselbach and Weihing (1933). They showed in most striking fashion that irregular planting of corn in hills had no influence on the mean

COMPETITION AMONG CROP AND PASTURE PLANTS

69

yield of corn per acre (Table XV). Here, as in wheat, the plasticity of the corn plant and the influence of competition by neighboring plants eliminate the effects of irregularity on yield. On the other hand, if plants are missing, with no compensatory effect of greater numbers of plants in other hills, then the increased producTABLE XV The Effect of Stand Irreeularitv in Corna Mean plants per hill in all treatments (mean of 14 years) = 3 Seeds per hill 3 2 or 4 1, 3, or 5 Yield (bu./acre) 49.9 50.6 49.3

1, 2, 3, 4, or 5 50.0

Mean plants per hill in all treatments (1931 only, a dry season) = 2.9 Mean deviation from 2.9 per hill 0.18 0.73 0.85 Yield (bu./acre) 25.0 25.5 25.3 0

1.00 25.1

From Kiesselbach and Weihing, 1933.

tion per plant from within the hill or from neighboring hills only partially compensates for the missing plants. Thus Dungan and Nelson (quoted by Dungan et nl., 1958) found that if one plant was missing from a 3-plant hill, 89 per cent of the loss was compensated within the hill or in contiguous hills; with two plants missing there was 68 per cent compensation, and with all three plants missing only 43 per cent compensation. There seem to have been few studies concerned with the planting arrangement of two species with constant numbers of each species, other than of cover crop and fodder species (see Section VII). Harper (1961) has conducted such a study with two Mediterranean annual grasses, Bromus rigidus and Bromus madritensk. He grew them together in pots, with both systematic and random variation of the arrangement of 50 plants of each species. The total yield per pot was unaffected by changes in the planting pattern, but the proportional contribution to yield by each species showed marked variation. No explanation was apparent; there is a possibility that the differences were an artifact brought about by the proportion of plants of each species growing in the outer circle of plants ( a considerable proportion of the total) virtually without competition for light. The results are nevertheless provocative and warrant study in more extensive communities.

3. Direction of the Rows Few experiments have been done on this aspect of plant arrangement; perhaps rather surprisingly, they all give the same result despite differences in latitude, season, and cmp height. All show that yield is greater

70

C. M. DONALD

from a crop with north-south rows than in a crop with east-west rows. The most comprehensive study is that by Pendleton and Dungan (1958) in which the two directions were compared at three row spacings of oats for seven years. There was an increase due to N-S rows of 7.7 per cent in grain yield at 2.4-inch spacing, 6.7 per cent at 16-inch, but of only 2.1 per cent at the normal 8-inch, spacing. The authors suggested that this was due to differences in the light regime, with superior lighting in N-S rows, as compared with the poor lighting on the N side of E-W rows. Others who have examined the effect of row direction are Perekaljskii (1951), who showed an increase of 1 3 centners per hectare in grain yields from N-S drills over E-W drills; Dungan et al. (1946), who showed a significant increase in both the fodder and grain yield of corn in India; and Santhirasegaram (1962), who showed an increase of 6.5 per cent ( P > 0.05) in the dry matter of wheat in one experiment and an increase of 11 per cent in grain yield ( P < 0.05) in another. The consistent pattern of these results indicates that this effect of row direction is real. Pendleton and Dungan (1958) showed that light regime (reflected light from a board placed at or near the ground surface) was higher from 10 A.M. to 2 P.M. in a N-S crop than in an E-W crop, but lower in the earlier and later parts of the day. Soil moisture was higher in the E-W rows, but whether because of soil shading or less water use by the crop was not determined. Similar relations were recorded by Santhirasegaram ( 1962). The study of row direction is well worth fuller examination, both for its possible application in practice and for its theoretical significance in crop illumination. There are many cereal areas where a difference of a few per cent in production might make N-S planting well worth while. X. Competition for Nutrients

Competition presumably may occur for any nutrient needed for plant growth, but our knowledge of this subject is limited. We can envisage a situation in which there is a finite supply of readily available nutrient, such as nitrate nitrogen, and the competitive success of any plant is governed by the nuniber of individuals drawing on the supply, and by the relative rates at which they do so. An alternative or additive situation is that in which a nutrient is present in a range of chemical and physical forms; the competitive ability of different species may be determined by their capacity to make use of each of these forms. To these relationships in competition for nutrients, we must add a secondary effect-namely that success in gaining a greater share of the nutrient

71

COMPETITION AMONG CROP AND PASTURE PLANTS

supply may cause such increased growth, that dominance over a less successful species may be due as much to competition for light as to the competition for the nutrient. A simple example of competition for a nutrient, apparently with little or no secondary effect, is provided in the study by Drapala and Johnson (1961) of the border effects between fertilized and unfertilized plots. Alternate plots of sudangrass received 0 and 100 pounds of nitrogen, respectively; rows were 6 inches apart both within and between plots. Plants in the two border rows of contiguous plots showed the effects of competition for nitrogen, those in the fertilized plot being depressed by the competition (loss to their unfertilized neighbors of part of the 0

N3 0

0

1

01 3

1

I

1

I

6

12

25

50

I

DENSITY (PLANTS PER POT)

FIG.30. The ceiling imposed by nitrogen status on the yield of grass growing in pots, almost independent of density (Donald, 1951 ).

nitrogen applied to them) and those in the unfertilized plot showing advantage. Competition for nitrogen can also be readily seen in pot cultures, particularly if other nutrients, water, and light (amply spaced pots) are nonlimiting. Bromus catharticus was grown in river sand in pots, at densities of 1, 3, 6, 12, and 50 plants per pot, with three levels of nitrogen supply: 0, 150, and 700mg. N per pot (Donald, 1951). At the low and medium levels of nitrogen the yield was not affected by the number of plants (Fig. 30). If there was one plant in the pot, it secured all the nitrogen; if there were 50, then they all shared the nitrogen. Only at the high level of nitrogen was a single plant unable to make full use of the nutrient supply. Here three or more plants were needed.

72

C. M. DONALD

This experiment provides an illustration of competition in its simplest form. The plants were of the same genotype, they competed for a single factor, and they gained an equal share of that factor. With differing plant densities and a limited nitrogen supply under field conditions, the picture is less straightforward. Figure 31 shows the relationship of the yield of Loliurn rigidurn, the percentage nitrogen in the herbage, and the nitrogen uptake per unit area for a very wide range of densities. The percentage nitrogen fell from 1.8 at low densities to 1 in extremely crowded communities. But this was more than a simple

LOG

DENSITY

FIG.31. The difference in the form of the curve for the yield of dry matter and the yield of nitrogen in Lollurn rigidurn, implying the partial independence of competition for light and for nitrogen as density increases (Donald, 1951).

sharing of supply. At high densities the total nitrogen taken up by the sward was less than at lower densities. This secondary effect may have been due to a shallower depth of rooting or less efficient soil exploration by the extremely poorly grown plants at high densities. Reference is made in Section XIX to the study by Lee (1960) where more efficient exploration of the soil for water and nutrients by the root system appeared to be a factor in the competitive success of one barley variety over another. A particularly elegant study of the relationship between density of population and the competition for nutrient supply was made by Lang et al. (1956) on corn. They showed that at a low nitrogen level the maximum yield of corn (75 bushels per acre) was given by a population of 12,000 plants per acre; at a moderate nitrogen level, the maximum yield (92 bushels) was given by 16,000 plants; whereas with high nitrogen the

COMPETITION AMONG CROP AND PASTURE PLANTS

73

peak yield (118 bushels) was not achieved till the population was raised to 20,000 plants (Fig. 32). These results must be considered alongside the fact that competition for light was undoubtedly occurring at each nitrogen level and was presumably intensified with each increment of nitrogen or density. At the low level of N, wider spacing would have given a greater yield per plant owing to less acute competition both for

401 4

I

I

1

I

I

8

12

I6

20

24

THOUSANDS OF PLANTS

'

PER ACRE

FIG. 32. The relationship of density, fertility, and yield in corn (Lang et d., 1956).

nitrogen and for light. With a better nitrogen supply and without change in light regime, a greater population of plants is needed to attain maximum exploitation of both nitrogen and light. These results were confirmed with sweet corn in Nova Scotia; Chipman and Mackay (1960) found that with nil or 500 pounds of mixed fertilizers the highest yield was given with 14,500 plants per acre; with 1000 pounds of fertilizer, 19,400 plants gave the peak yield, while with heavier fertilizer dressings 29,000 plants gave a greater yield than any lesser population. This study

74

C. M. DONALD

was in terms of “marketable ears”; had it been in terms of grain, then even higher densities may have been advantageous at high fertility levels. There is sufficient evidence to derive the general principle that as fertility status is improved, so the density required to give maximum yield by annual crops will increase. Conversely, as density is increased, so the response to an added nutrient will continue to a higher level of application. All the foregoing examples relate to competition for nutrients between annuals in a monoculture; competition of rather different kind may occur between species in a mixed community. Competition between grass and clover for potassium is described in a study on pastures in New York by Blaser and Brady (1950). In their first trial, it was found that potassium alone gave an increase in total yield, due almost entirely to an increased yield of clover. When nitrogen as well as potassium was applied, there was a marked drop in clover yield, yet with very little increase in grass yield. The authors postulated that there was competition for potassium between the grass and clover and that this competition was intensified by the application of nitrogen. The clover on the N K plots actually had a lower K content than on the control plots. This hypothesis was tested in the following year, using three levels of nitrogen and three of potassium. As the level of nitrogen increased, there was more grass, more removal of potassium by the grass, and less growth of clover. As the potassium was increased, the effect of nitrogen fertilizer in depressing the production of clover became less. Blaser and Brady (1950) suggested that the grasses, able to grow at lower temperatures than the clovers in the spring, and thus to grow earlier, depleted the potassium, so that competition for potassium became acute as temperatures rose. The application of nitrogen, by stimulating grass growth accentuated this competition. They recommended that grasses modes’t in stature and unadapted to low temperatures should be chosen, together with strains of legumes effective in nutrient uptake. Walker and Adams (1958) found competition of similar kind between grass and clover for sulfur in field experiments in the Canterbury foothills of New Zealand. At low sulfur levels, almost all the sulfur was taken up by the grass, and the clover was greatly depressed. When nitrogen was added the position was aggravated, but the clover suppression could then be considerably relieved by heavy sulfur application. Similarly competition between grass and clover has been recorded for phosphorus (Trumble and Shapter, 1937). Gray and his colleagues (1953) examined the proposition that the competition between grass and clover for potassium might relate directly

+

75

COMPETITION AMONG CROP AND PASTURE PLANTS

to the cation exchange capacity of the root systems of the grass and the clover. First they found that three grasses, when grown in pure culture, showed K uptake inversely related to the cation exchange capacity of the roots. Then, when these grasses were grown in association with ladino clover, the K uptake and the yield of the clover were inverse to the uptake of K by the grass and directly related to the cation exchange capacity of the grass roots (Table XVI). While all this is no TABLE XVI The Root Cation Exchange Capacity of Three Grasses in Relation to Their Potassium Uptake and Their Competitive Ability against Clover@ Cation exchange Relative capacity of the roots K uptake by (m.e./ grass in 100 9.) pure culture

Relative

K uptake in mixtures

Y

Grass Bentgrass Kentucky bluegrass Smooth bromegrass a

16 21 24

100 72 27

Grass

Clover

100 40 26

14 27 38

Yield of clover in mixture

(g.) 2.68 4.67 4.85

From Gray et ul. ( 1953).

more than circumstantial evidence (the grasses may lie in the table only in order of their vigor-their order of relative growth rate), it is nevertheless of great interest as a possible mechanism of competition for nutrients. Mouat and Walker (1959) concluded that the basis of competition for phosphorus was also a function of root cation exchange capacity, and that differences in competitive ability for both potassium and phosphorus, due to a wide difference in cation exchange capacity, is the cause of the severe competition between brown top and white clover. Competition for nitrogen is the most widespread of the forms of nutrient competition, affecting most of the crops and pastures of the earth in some degree. Its role in crop production has already been discussed, and it perhaps suffices to refer to the influence of nitrogen on yield depression by weeds. Blackman and Templeman in their early studies on cereals in England (1938a) showed that in years of normal rainfall, competition for nitrogen was the most important factor in the depression of cereal yields by cruciferous annuals, though light was important when the weeds were tall and dense. In many fields nitrogen application alone raised the yield of weedy crops to levels as high as those of clean crops. Under very different conditions in the wheat belt of New South Wales, it was found that the depression of wheat yield

76

C. M. DONALD

due to skeleton weed (Chondrillu juncea) in normal seasons was due to early competition for nitrogen and that by spraying the weeds two months before the crop was sown, yield could be greatly enhanced. Similar yields could alternately be attained by the application of liberal nitrogen when the crop was sown (Myers and Lipsett, 1958). The competition for nitrogen between grass and clover presents a unique though common circumstance because of the interaction not only of the species, but of their source of nitrogen. Walker et al. (1956) using labeled nitrogen ( N16) in pot cultures, found that when growing alone, grass took up considerably more soil N than did clover; and further, that with increasing levels of combined N, the clover took up more nitrogen and fixed less. When the two species were grown together the grass took up practically all the nitrogen (over 90 per cent) and indeed took up almost as much as it did when grown alone. Under these conditions the clover became almost fully dependent on symbiotic nitrogen (100 per cent at low N, 85 per cent at high N ) . But this is by no means necessarily a constant picture. It will certainly depend for example on temperature relations. Willoughby (1954) in Australia found that in the Canberra environment, where “winter growing” Mediterranean pasture plants are used, an early application of nitrogen (i.e., in February while temperatures were still high) gave clover dominance because of the more rapid growth of the clover seedlings than of grass seedlings at high temperature. In these circumstances there was rapid uptake of nitrogen by the clover, and the subsequent growth of the grass was greatly reduced by nitrogen deficiency. Conversely if nitrogen was applied late, when temperatures had fallen, the growth of grass was extremely vigorous, while the clover was much reduced. Thus it is known that grass and clover may compete actively for phosphorus, potassium, sulfur, and nitrogen, and no doubt for other elements also. A striking presumed instance of interaction between competition for phosphorus and for nitrogen is provided in Table XVII. On a soil acutely deficient in these two elements, gramineous and leguminous cover crops were undersown with grass and legume pasture species. The grasses were seriously depressed by the gramineous cereal crops, but not by the peas; and conversely the clovers were depressed by the peas, but unaffected by the barley or oats. It would seem that the main factor of competition between the gramineous species was nitrogen, whereas between the leguminous species, which were “independent” for their nitrogen supply, the competition was mainly for phosphorus.

COMPETITION AMONG CROP AND PASTURE PLANTS

77

The competition between two grasses for two nutrients, phosphorus and potassium, has been examined by van den Bergh and Elberse (1962). One of these grasses, Lolium perenne, is regarded as an indicator of high soil fertility, the other, Anthoxanthum odoratum, as an indicator of low fertility. During the first 2 months (in the third month disease interTABLE XVII The Yield of Undersown Pasture Species under Gramineous and Leguminous Cover Cropso, IJ

Undersown species Grassesc Legumesd a

b c

d

Cover crop None

Barley

Oats

Peas

11.9 10.2

0.6 9.5

1.2 9.0

10.7 4.6

Donald and Neal-Smith (1938). Values: hundredweight dry matter per acre. Lolium rigidum (mainly) and Phalaris tuberosa. Trifolium subterraneum and Medicago sativa.

fered) the series at the high P-K level showed an increase of LoZium and a decrease of Anthoxanthum, while at low P-K the reverse took place. Though this particular experiment was somewhat inconclusive, it illustrates the need for an understanding of the vexed question of the differential capacity of genotypes to tolerate a low nutrient status or to exploit a high one. There are of course lots of field examples of this phenomenon. Almost any soil, no matter what the deficiency or how acute it may be, has some plants growing reasonably well on it. The intensely copper-deficient calcareous sand dunes of coastal South Australia support the vigorous growth of two grasses, Bromus mudritensis and Lagurus ouatus, without fertilizer. Only with the use of copper fertilizer will such grasses as Lolium spp., or legumes such as Medicago saliua, compete successfully on these soils ( Riceman et al., 1940). Even within a species the differing capacity of genotypes to withdraw nutrients from less available sources has been widely recorded and offers a potential means of overcoming or relieving expensive fertilizer problems on some soils. The significance of such differences in competitive relations between genotypes deserves attention. XI. Competition for Water

Water deficiency restricts the growth of a considerable proportion of the crops of the earth and is the commonest factor determining the geographic limits of crop production. Competition for water usually

78

C. M. DONALD

occurs together with other forms of competition, especially for nitrogen and for light, but it is by no means of parallel intensity with these other forms. Indeed when the competition for water or nitrogen is intense, growth may be so restricted that competition for light is of reduced importance, whereas if water and nutrients are nonlimiting, shading will be a major factor. The 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, and this capacity rests in turn on several attributes of the genotype within the particular environment. These are the relative growth rate and the corresponding earliness of water demand, and the rate of root extension, both in the sense of “reaching lateral unexploited soil before other plants and in gaining access to deeper supplies through a deep-rooted habit of growth. The capacity of the plant to remove water to a high pF, or to higher pF than its neighbors, is usually of little importance in competitive relations, as water content does not change much in most soils as the limiting pF is approached. In a mixed association such as a pasture, the species may differ markedly in their capacity to compete for water, but in most agricultural crops the plants are of like or similar genotype so that they experience a common intensity of competition for water from germination to maturity. It has been long established in agricultural practice, perhaps for centuries, that the optimum density of any annual crop will be less in a dry environment than in a wet one, again illustrating the point that the more favorable the environment for any reason whatever, the higher will be the optimum population. Wheat, grown over a very wide range of rainfall and fertility conditions, is sown at rates as low at 40 pounds per acre at its arid limits, and as high as 200 pounds in favored areas. The studies by Karper (1929) of grain sorghum production at Lubbock, Texas, during the ten years 191&1925, illustrate this relationship of rainfall to optimum density, but also emphasize the difference in rainfall-density relationships between genotypes. Two varieties, KAFIR and MILO, were grown at spacings from 3 to 36 inches in the row, with rows at 36 inches. KAFIR showed a very marked reaction to season, the optimum density in wet seasons being ten times that in dry seasons, whereas MILO gave is maximum yield at wide spacings in either wet or dry seasons. This was directly associated with the degree to which each of the varieties tillered and was able to exploit more favorable conditions. (Table XVIII.) A similar relationship was shown in an irrigationdensity study on sorghum in Kansas by Grimes and Musick (1960). When they varied the number of irrigations in a dry season (3.2 inches

COMPETITION AMONG CROP AND PASTURE PLANTS

79 in 4 months), the optimum density for production became higher as the number of irrigations was increased. In some crops the density-water relationship is complicated by the need to consider quality as well as aggregate yield. This especially applies to horticultural crops and is well illustrated in Salter’s studies TABLE XVIII The Difference between Varieties of Grain Sorghum in the Density-Yield Relationship in Wet and Dry Seasons@ Variety: Seasons: Optimum spacing in rows (in.) Yield at optimum spacing ( bu. ) Characteristics of the variety

* 0

MILO

KAFIR

2 driest yearsb

36

2 wettest yearsC

2 driest yearsb

2 wettest year9

3

24 and 27

21 and 27

14.9 32.8 does not tiller freely. Population of stems largely dependent on density of plants. Optimum density rises as conditions more favorable KAFIR

16.4 31.2 tillers freely. Is able to “adjust” to conditions by degree of tillering. Yields increase as spacing increases from 3 in. to 18 in. and then are rather uniform to 36 inches MILO

From data by Karper (1929). 1917 with 8.73 inches and 1924 with 9.45 inches. 1919 with 31.61 inches and 1923 with 26.17 inches.

on cauliflowers (1961).He found, as in the foregoing studies on cereals, that irrigation gave its greatest benefit to production at high densities, but further than this, that it maintained the yield of marketable heads of cauliflower at high densities, even though they were small (Table XIX). In contrast the nonirrigated plots suffered such intense competition at high densities that most of the heads were defective for market. All too little is known of this interaction between density and competition for water, beyond the broad generalization that as the water status is improved, so can more plants share the supply without suffering stress. Milthorpe (1961) states the general principle that the greater the amount of leaf growth made before plants come into contact with each other, the more extensive will be the root system and the less likely is the plant to suffer from drought. The higher the density, then the smaller the plant at any time during ontogeny, and the higher the water content at which shortage of water is experienced. In mixed associations,

80

C . M. DONALD

Milthorpe continues, the individuals of one or more species (the suppressed species) are likely to suffer long before the other species. In a pure culture all individuals will tend to be affected together; those survive the longest that-because of earlier emergence, larger embryos, or higher growth rates-have grown the most extensive root system. TABLE XIX The Interaction of Density and Irrigation on the Production of Cauliflowersa, b Plant spacing (inches) Relative densities Weight per plant ( 9 ) . marketable or not Without irrigation With irrigation Weight ( 9 ) per unit area (34 x 34 inches) Without irrigation With irrigation Per cent increase Yield of marketable heads (tons per acre) Without irrigation With irrigation

34 x 34 1

24 x 24 2.01

17 x 17 4.00

12 x 12 8.03

1352 1260

805 1194

488 765

287 488

1352 1260 -7

1618 2400 48

1952 3060 57

2304 3919 70

7.2 6.7

7.6 12.2

2.8 14.8

1 .o 14.6

From data of Salter (1961). The values are the means for three nitrogen levels, which were not significantly different in their effects. b

Striking differences of patterns of root systems at various spacings has been examined by Haynes and Sayre (1956) for corn in Ohio. The density varied from l-inch spacing to 64-inch spacing of plants in single rows presumably free from lateral competition ( 8 feet 6 inches spacing between rows). When plants were widely spaced, they had a more or less circular distribution of roots (i.e., a circle of about 30 inches at 36-inch spacing), but when closely crowded there was a greatly increased penetration of roots laterally from the row (at l-inch spacing the roots penetrated 16 inches along the row and 41 inches laterally). The authors suggest that these patterns could in part be predicted if it were accepted that any se,aent of a root system in moist soil will develop more rapidly than another segment of the same root system in dry soil. The aspect of their study of greatest relevance to population relationships is that at the high density (l-inch spacing), where each plant had a spread of roots 16 inches along the row in each direction, there were presumably no less than 32 overlapping root systems at any

COMPETITION AMONG CROP AND PASlVRJ3 PLANTS

a1

point beneath or near the row! In contrast at 32-inch spacing only two root systems overlapped. Here lies clear evidence of the reality of competition for water or nutrients among crowded plants. Differences in the advent of competition for water between crops in a common environment may be due to these patterns of horizontal or vertical root distribution or to the internal physiology of the plant or both. Slatyer (1955), studying the water relations of three crops in an unfavorable environment in the Northern Territory of Australia, found that cotton, as compared with peanuts and grain sorghum, suffered greater water stress partly because of the lack of regulation of internal water loss with a consequent loss of turgor, but more particularly because cotton had a less extensive root system, poorly developed below 2 feet, and apparently even failing to exploit some areas of soil above that depth. Though these crops were not grown in association, the contrasting capacity to compete for water was clearly defined. The evidence on the influence of density or water supply on the competition for water and the efficiency of water use, is not wholly consistent. Russell and Danielson (1956), growing corn under three contrasting water regimes, found no difference in the yield per inch of water loss, whether competition for water was almost absent or was very intense. Grimes and Musick ( l W ) ,using three populations and four row widths in grain sorghum in Kansas, also reported no effect on either the seasonal use of water or the efficiency of water use. In contrast, Porter et al. (1980) in a study with irrigated sorghum in Texas, found a difference in yield at spacings of 12, 20, 30, and 40 inches, though the amount of water used was the same in all treatments, The greater yield at the higher densities was therefore reflected in a greater efficiency (approximately 15 per cent) in water use. It should be emphasized, however, that in this study all treatments were irrigated, SO that water was virtually nonlimiting. The equal use of water by the. full canopy of foliage on all plots follows the principles enunciated by Penman (e.g., 1948), whereas yield advantage at the narrow spacings may have been due only to influence of the row spacing on the harvest index (ratio of yield of grain to total yield of dry matter). An aspect of great interest in pasture management in drier areas or seasons is the influence of defoliation on the competitive ability for water. Army and Kozlowski (1951) showed the much weaker capacity of plants without tops to absorb water held at high osmotic pressure than of those with tops intact. Jantti and Kramer (1956) looked at this phenomenon in the field. These authors grew swards on soils previously irrigated to three different water states. After defoliation to 1,4, and 12 cm.,

82

C. M. DONALD

the regrowth was much stronger, and the water uptake faster, where the longer stubble was left. This could, of course, have been explained purely in terms of the residual leaf area, but the differences due to stubble heights were far greater on the dry soil than on the intermediate or moist soil. The results therefore conform to Army and Kozlowski’s studies and indicate the check to competitive ability for water caused by heavy defoliation. Finally, the interrelationship of competition for water and other factors is reemphasized. Kurtz et al. (1952a, b ) tested their viewpoint that the competition in corn, or between corn crop and intercrop, in the Corn Belt of the United States is mainly for the mobile nutrients, water and nitrogen. They found that when water and nitrogen were applied to corn grown in narrow tilled bands in competing pasture swards, the yield could be raised from about 25 per cent (minus N, minus water) of the yield under a conventional system to about 85 to 90 per cent (plus N, plus water). The residual deficit in yield was unexplained, but the experiment provides a good example of the analysis of competition into its main components. The competition by the intercrop for nitrogen was also shown by the reduced nitrogen uptake by the corn, while the competition for water was reflected in the lower soil moisture where an intercrop was present. XII. Competition for Lighi

Competition for light may occur whenever one plant casts a shadow on another, or, within a plant, when one leaf shades another leaf. It occurs in almost all crops and pastures, and is perhaps absent only among newly emerged crop seedlings, where there may be no competition of any kind, or in arid regions, where the density of the community is limited by water supply. Yet man has been slow to recognize this factor to the full in his crops and pastures. Though he has been aware of the deep shade within forest communities, and has recognized the role of competition for light in the succession from colonization to a tall climax, he has until recently had little appreciation of the total darkness beneath his crops and pastures. Monsi and Saeki (1953) noted the high light density on the floor of several forest communities (Pinus 28% of daylight, Castanea 13%, and Cryptomeria 5 % ) compared with the values for three communities of herbs (Helianthus 4.5%, Phragmites 1.2%, and Miscanthus 0.7%). Light intensities of this latter order are characteristic of crops and pastures. Solar radiation is the factor governing the ultimate yield of any particular genotype or community. If water and nutrients are available

COMPETITION AMONG CROP AND PASTURE PLANTS

83

in adequate supply, so that competition for these factors ceases, then light becomes the sole limiting factor to production (Donald, 1951; Blackman and Black, 1959). This situation is becoming more and more common in world agriculture, because of the extension of irrigation and the increased use of fertilizers. There are already many situations in which production is controlled, if not for the full season, at least for a part'of it, dominantly or wholly by light supply and by the capacity of the crop to intercept and use the light in photosynthesis. Heavily fertilized cereal crops or clover-rich pastures in areas of favorable rainfall, or irrigated and fertilized grass pastures or crops, are examples of communities in which the dominant factor governing production is the utilization of light. Perhaps the tea estate provides the classic example of this situation. Here the tea bush, growing with liberal fertilizer and in a wet climate, is cultivated to present a tablelike surface of foliage for the interception of light and the production of a weekly crop of young shoots. On the other hand, even where there is a shortage of water or nutrients, as occurs for most of the world's crops, competition for light remains a factor of major importance. For example, Donald (1958) showed that when two grasses were growing at a level of nitrogen supply which severely restricted yield, competition for light was an important factor in the total competition. Except in the poorest of crops with acutely limiting supplies of nutrients or water, light is significant in the competition both within crops and between crops and weeds. A crop or pasture may develop sufficient foliage to intercept all measurable light within a few weeks of planting. A subterranean clover sward sown at the rate of 350 plants per square foot, equivalent to a strong natural germination, intercepted a11 light within 40 days of emergence (Black, 1958); buckwheat sown at 100 plants per square meter reduced the light at ground level to 5 per cent of daylight only 37 days from sowing, while with 400 plants/m.2 the light was reduced to 2 per cent ( Iwaki, 1958). In a study of wheat at Adelaide (Puckridge, 1962), the light intensity at ground level was reduced in a stand sown at normal seed rate (184 plants/m.2 = 55 pounds/acre) to 4% daylight after 14 weeks. At a high sowing rate (1078 plants/m.2 = 326 pounds/acre) the light was reduced to this value within 10 weeks. Clearly in these circumstances the lower leaves of the canopy will suffer competition for light from leaves lying above. Indeed, as emphasized by Clements, competition for light may begin in very dense communities almost immediately the seedlings emerge; Bennett ( 1960),growing subterranean clover under conditions of adequate water and nutrient

84

C. M. DONALD

supply, compared densities of 12.5 plants/dm.2 and 142 plants/dm? The individual weight of the more crowded seedlings at 10 days was already 10% less, and at 17 days was 25% less, than that of the sparse seedlings -a direct effect of early mutual shading. It has been noted earlier that competition for light differs from that for nutrients, water, or carbon dioxide in that there is no common pool from which plants continue to draw their supplies until it is depleted or exhausted. Light energy is instantaneously available, and it must be instantaneously intercepted, or it will be lost as a source of energy for photosynthesis. It follows that the successful plant is not necessarily the plant with more foliage, but the plant which has its foliage in an advantageous position relative to the foliage of its competitors for light interception. A cereal crop may suppress a weed or undersown fodder plant not because it is potentially or actually more leafy but because it is taller and has foliage so displayed as to intercept the light and leave its more dwarf competitor in relative darkness. There are many examples of competitive success of one species over another based on competition for light. Pickworth Farrow (1917) describes how the fronds of bracken reduced the light intensity available to heather to as little as one-seventieth of daylight, and thereby suppressed the heather. The role of light in the competition between grass and clover has been discussed earlier. Yet the process of competition for light is not immediately one of competition between species, nor even of competition between plants. It is competition between leaves. If one leaf lies above another, then the depression of the photosynthetic rate of the lower leaf will be the same, whether the superior leaf is of the same plant or another. This competition between leaves is especially evident in a dense crop where the leaves of neighboring plants form a continuous canopy in which the foliage of each plant will be intermingled with that of several neighbors. Competition between leaves for light is also expressed within a single isolated plant. Here, although the plant may be solitary and suffering no competition of any kind from its neighbors, the lower leaves may be so heavily shaded by those above as to die. This is intraplant competition of the most intense kind. Though competition may occur within a plant for water or nutrients, usually with preferential protection of meristematic or reproductive parts, it is rarely of such intensity as will occur within a single plant for light. Basically this is because light is not redistributed. A clover plant may be well nourished throughout with nitrogen from a single root nodule; or a wheat plant with a local concentration of fertilizer to a few of its roots will do better than a plant with a lesser

COMPETITION AMONG CROP AND PASTURE PLANTS

85

concentration available to its whole root system, as when fertilizer is drilled. But this redistribution does not apply to light received by leaves. If a leaf is heavily shaded, it will be unable to maintain itself above compensation point and will die, Finally, in this general discussion of light as a factor in competition, it is of interest to consider the influence of the climatic environment. Here again there is a notable contrast from water and nutrients. In a region or season of adequate rainfall, competition for water will be reduced or absent. And the same is broadly true of nitrogen or phosphorus-that if these nutrients are available in liberal supply, competition will be absent. Not so for light. The depth of canopy will commonly be such that the basal leaves will be heavily shaded and much reduced in photosynthetic activity; they will perhaps be at or near compensation point. Yet if the light is more intense, the same relationship is likely to prevail. Again the basal leaves will be suffering acute competition for light; the difference in the two situations will reside not in the intensity of competition, but rather in the depth of the canopy. In these terms and if other factors are not limiting, competition for light will be as acute in brightly lit regions as in those of overcast skies. XIII. The leaf Canopy and Growth

Many of the earlier studies of the efficiency of leaves were done with isolated plants growing in pot cultures or growth chambers, or on individual leaves. This led to emphasis on the efficiency of the foliage -how much carbon is fixed per square centimeter of leaf per day? The expression developed for this purpose was “net assimilation rate” (N.A.R.) expressed as _1. - dw

L where

dt

L = total leaf area of the plant w = weight of the plant and t = time.

The limitation of this concept as applied to agriculture is that plants do not grow as isolated individuals, but as communities of individuals under strong competitive stress. Furthermore, as a crop or pasture develops its canopy of leaves, there is a very wide range of foliar environments, from that of a well lit leaf at the surface of the canopy to that of the poorly lit leaf below. Thus, while net assimilation rate can be a useful criterion of the efficiency of an individual leaf or plant, we should recognize that for the crop as a whole it is no more than an average

86

C . M. DONALD

value for all leaves, which may have a range in their contribution from strongly positive to negative values. The translation of thought from single plants to communities requires modification also in regard to light intensity. Many workers have shown that the photosynthesis of an individual leaf reaches its maximum rate at about 1500 foot-candles, compared with values for bright sunlight of up to 10,000 foot-candles. It was at first assumed that this value of 1500 foot-candles might represent the upper limit of usefulness of daylight, but in fact dense communities of leaves, because they are competing for light, will benefit from increased light intensity up to extremely high values. As light intensity increases, so will leaves lying deeper and deeper within the canopy achieve a positive balance of photosynthesis over respiration. Thus, Heinecke and Childers (1937) found a notable contrast between the apple leaf and the apple tree. Whereas the individual leaf attained its maximum photosynthetic rate at one-quarter to one-third full sunlight, the rate of photosynthesis by the whole tree showed almost linear dependence on light intensity up to values of 8000 foot-candles (9:30 A.M. to 2:30 P . M . ) . A large part of the foliage of the tree was positively functional only on the brighter days. A similar comparison is available for pine trees; whereas individual needles were light-saturated at 3500 foot-candles, young pine trees showed an increased photosynthetic rate to 10,000 foot-candles ( Kramer and Clarke, 1947). A third example is available with Cynoclon ductylon (Alexander and McCloud, 1962); isolated leaves were saturated at 2500 to 3000 footcandles, but the light required to saturate a sward, depending on the height and foliage pattern, ranged u p to 7000 foot-candles, the highest intensity used in these studies. The most notable exception to these general relationships is the work by Hesketh and Musgrave (1962) in which they record a progressive increase in the photosynthetic rate of single leaves of corn up to values of 10,OOO foot-candles. As yet, it is difficult to assess this finding. The approach to light relationships and competition for light was greatly aided by the concept of the leaf area index (LA1 or L ) . Watson (1947) suggested that “the measure of leaf area which is relevant to agricultural yields, that is, the weights of different crops produced per unit area of land, is the leaf area per unit of land, which it is proposed to call the Leaf Area Index.” With the development of this concept has come also a growing understanding that it is the area of foliage in an acre of crop, and the light relationships of that foliage, rather than any variation of inherent efficiency of the leaf, which will have by far the greater influence on photosynthesis and crop growth.

COMPETITION AMONG CROP AND PASTURE PLANTS

87

Competition for light within a crop is governed by its manner of interception. Several workers (Monsi and Saeki, 1953; Kasanaga and Monsi, 1954; Davidson and Phillip, 1958) point out that light interception by a canopy of leaves is exponential (Beer‘s law), whereby

I =l 0 e c k L where I = light intensity beneath a leaf area index of L I,, = light intensity above the crop L = leaf area index above the point of measurement k = coefficient of extinction. This means that light intensity will fall sharply as it penetrates into a sward. As the leaf area index increases, a stage is reached at which the lowest leaves are only sufficiently lit to be at or just above compensation point. Such a leaf area index has been termed “the optimum leaf area index” (Kasanaga and Monsi, 1954; Davidson and Donald, 1958) or “the marginal compensation area” ( Davidson and Phillip, 1958). This is the situation in which all leaves are making a positive or neutral contribution (Figs. 33 and 34). If there is a further increase in leaf area index, the lowest leaves will be so poorly lit that their respiration will exceed their photosynthesis. They are in negative balance, their weight declines, and the rate of dry matter increment of the whole sward is less than it was at the optimum leaf area index. The existence of an optimum leaf area index has been demonstrated in a few crops. Watson (1958) studied the relationship of crop growth rate to the values of leaf area index in kale and sugarbeet; different leaf areas were achieved by thinning, so that at the beginning of the observation period all plants were of the same chronological and physiological age, He found that thinning kale to 21, 16, 11, and 6 plants/m.2 gave mean values for the leaf area index in the ensuing period of 5.3, 4.2, 2.9, and 1.6. These in turn gave crop growth rates of 78, 130, 101, and 88 g./m.:/week. In a series of experiments, kale gave values of 3.0 to 5.4 for its optimiim leaf area, but in sugarbeet there was no fall in crop growth rate up to a leaf area index of 6. Several experiments at Adelaide (Davidson and Donald, 1958; Stern and Donald, 1962b; Black, 1963 ) have demonstrated an optimum leaf area index in swards of subterranean clover. For example, in one study the crop growth rate fell by about 30 per cent as the leaf area index rose from 4.5 to about 8.7. As might be expected, the optimum leaf area is not a static value, but will change as the competitive relations between leaves are influenced by the light intensity. As the light intensity

88

C. M. DONALD

increases, so will the optimum leaf area be greater (Stern and Donald, 1961; Black, 1963); or conversely, as the leaf area increases so is more light needed to attain the maximum crop growth rate. If the leaf area continues to increase beyond the optimum, a further definable stage will be attained. The canopy deepens, the light intensity

n

rLw dt

<

L.- A.. 1

T IME-

FIG.33. A schematic account of the relationship of the rate of production of dry matter to the leaf area index ( L A I ) . The vertical rules show three states of the plant community, the optimum leaf area, the ceiling leaf area index, and the ceiling yield. The LA1 scale is equivalent to a time scale if the plant community is undefoliatecl.

at the base of the crop or pasture becomes progressively less, there is a more and more rapid death of bottom leaves, until leaves are dying as rapidly at the base of the canopy as they are produced at the top. The leaf area index is now static and maximal; it is conveniently known as the ceiling leaf area index (Davidson and Donald, 1958). In clover swards this ceiling leaf area may far exceed the optimum (for example in one experiment it was 8.5, compared with an optimum of 4.5), but whether this is the case for most crop plants is not known. The following data for subterranean clover illustrate the approach toward a ceiling leaf area index in swards of varying density. Density (plants/m.2) Leaf area index Oct. 13 (mid-spring) Leaf area index Nov. 10 (late sprinz)

25 2.8

7.4

100 5.3 7.5

250

7.0 8.0

1250 8.7 8.7

These data provide an illustration of competition for light, and of a maximum canopy of foliage determined by the light regime. Brougham (1958b) has provided a clear-cut account of the dynamic equilibrium of a white clover sward at the ceiling leaf area index. It showed a total value of 5.5, made up of 0.5 units of young, folded leaves, 3.5 units of apparently actively photosynthesing leaves and 1.5 of senesc-

89

COMPETITION AMONG CROP AND PASTURE PLANTS

ing leaves. The sward was static simply because the rate of senescence was equal to the rate of appearance of young new leaves, a typical situation in a community at its ceiling leaf area. If now a crop a t the ceiling leaf area continues growth, no more Optimum Leaf Area Index T h e net assimilation by the whole folii e has attained a maximum value. eaves are making a ositive contribution to dry we& increase, though t h e contribution by the lowest leaves may be very small.

I I 1

NET ASSIM. lOF FOLIAGE

*!,.f,

Ceiling Leaf Area Index Leaf area index has reached a maxi. mum. T h e rate of death of leaves a t the base nf the canopy d u e to low light intensity equals the rate of appearance of new young leaves. Net assimilation by the foliage is now below that at optimum L.

I I

I

NET ASSIM

106 FOLIAGE

I

I2

2

2 3

10

2

8

6

2

4

4

3

5

1

2 2 2

6 7

0

0 38 13 I

10

I

-1

-2 -I

I9 -

Ceiling Yield Non-photrisynthetic tissues have increased until the respiratory losses by the crop equal the photosynthetic gains. T h e dry weight of living material per unit area I S static.

TOTAL N E T ASSIMILATION

-

NIL

FIG.34. The basis of the concepts of optimum leaf area index, ceiling leaf area index, and ceiling yield. L denotes the successive layers of unit leaf area index; P, photosynthesis; R, respiration. From Donald ( 1961).

leaves will be added because accelerated leaf production means accelerated death of leaves by shading. Thus any additional dry matter will be built into nonfoliar parts, in particular storage organs of various kinds. A simple example is given by Went (1957), who refers to the build-up of nonphotosynthesing tissues in old strawberry plants until the plants

90

C. M. DONALD

were barely at compensation point at a light intensity of 1500 footcandles. In Fig. 5, a number of studies show that at all high densities (roughly speaking, all densities which are in excess of the density giving maximum grain yield) there was a constant yield of dry matter. While no experimental evidence was adduced on the factor or factors governing these ceiling values, it is probable that light was the factor in some instances. The curve for subterranean clover (Fig. 5 F ) is believed to have been governed by light alone, since water and nutrients were liberally available at all times. Light may well have been the factor also in some of the corn studies in this figure. Indeed we may return to the point made earlier, that if there is R nonlimiting supply of nutrients and water, and if there are enough plants, then the yield will be governed by the growth form of the plants in the community and their capacity to intercept and use the light. When this limit is reached, the only means of increasing yield is to turn to a genotype with greater capacity to intercept and use light. XIV. Fluctuations in Leaf Area

In the early stages of growth of a crop, the crop growth rate (rate of increase of dry matter per unit acre) will be almost linear on the area of the foliage. This is the period in which nutrients and water are usually nonlimiting and in which the competition between leaves for light is negligible at normal crop planting rates. If the seed rate is varied, then at the time of emergence the leaf area index (the cotyledonary area per unit ground area) will in turn be linear on seed rate. Thus the early crop growth rate will depend directly on seed rate and will be much faster at high densities than at low densities. This is illustrated in Fig. 36 where three densities of subterranean clover were compared. There was a tremendous divergence in leaf production up to 60 days. The dense stand was the first to reach its optimum leaf area index (maximum slope), but showed a sharp curtailment as it reached the ceiling value. The two lower rates, though slow to attain the optimum leaf area index, were eventually able to overtake the dense sward and to approach closely to the ceiling leaf area index. The relationship of density, leaf area, and light interception is shown in a study by Iwaki (1958) on buckwheat where the values for light at the ground surface provide an account of the effectiveness of light interception and of the increasing competition among the leaves for light (Fig. 35). It will be seen that in the dense stand the ceiling leaf

COMPETITION AMONG CROP AND PASTURE PLANTS

91

area index of 4 was achieved by June 25, but in the less dense stand not until the end of the season. The final yield (August 1) was greater from the dense stand (622 as compared with 544 g./m.?), and this was no doubt an expression of the greater integral of leaf area and time, Watson (1952) has used the term “Leaf Area Duration” as an expression of this integral of the leaf area index over the whole growth period. A rapid, early increase in leaf area index and crop growth rate within a constant environment can therefore be attained by increased density.

FIG.35. The pattern of leaf area increment and light interception in buckwheat at two densities. (Redrawn from Iwaki, 1958.)

It can equally be achieved by the use of large seeds which have a greater cotyledonary area at emergence than do small seeds. Black (1956) examined the rate of dry matter increment by seedlings of subterranean clover derived from both large and small seeds. He found that the growth rate depended not on the weight of the cotyledons at emergence, but on their area, which was unaffected by depth of sowing and was linear on seed size. Thus if seed of double the weight was used, the cotyledonary area was also double. Figure 36 shows that with a constant density (number of seeds), the effect of seed size is just the same as the effect of seed rate. The two effects are different expressions of the same factor-namely, the variation in the total photosynthetic area of the emerging seedlings. The leaf area of crops and pastures may be subject to sudden reduction when they are thinned (plant removal) or defoliated (leaf removal). Under conditions of nonlimiting water or nutrient supply, these treatments have very similar effects. They may lead to an increase or a

92

C. M. DONALD

decrease or to no change in the rate of leaf production and of crop growth, depending on the leaf area before and after reduction. If the leaf area is low, then any further reduction will depress the rate of growth; if the leaf area is reduced from a value somewhat above to somewhat below the optimum there may be no change in growth rate; if the leaf area is reduced from a ceiling value toward the optimum, there will be a great resurgence of leaf production because of the reduced competition among the remaining leaves for light. These effects can be predicted from Fig. 33. In an experiment on subterranean clover, thinning at 14 days or 28 days reduced leaf area disadvantageously

L

40

80

DAYS FROM SOWING

I20

0

20

40

60

80

100

120

DAYS FROM SOWING

FIG. 36. The influence of density at constant seed size, and of seed size at constant density, on the rate of increment of the leaf area index of subterranean clover. Left (Davidson and Donald, 1958); right (Black, 1957).

and checked the rate of increment of the leaf area index (Fig. 37). Yet each of these reduced swards attained the ceiling leaf area of the control sward by day 92. In the example of defoliation of a clover sward by clipping, the rate of increment of leaf area was almost unchanged by the clip at 96 days, but showed dramatic increase when the clip at 124 days reduced the leaf area from its static ceiling value. Total leaf production during the season was increased by 40 per cent over the control by defoliation at day 96, and by 80 per cent by defoliation at day 124. The relationship of height of sward and defoliation treatment was also studied by Alexander and McCloud (1962). The work was done with Cynodon dactylon, greenhouse grown; the swards were transferred to a closed system for the measurement of CO, uptake. Figure 38 shows data on the COZ uptake of swards of various heights, subsequently

COMPETITION AMONG CROP AND PASTURE PLANTS

93

defoliated to 8 inches. In the uncut swards net photosynthesis had slowed down in the 20- and 26-inch swards because of lodging, leaf crowding, greatly intensified competition for light, and death among the lower leaves. Defoliation to 8 inches reduced the leaf area index of the 14-inch sward to a less effective value. In the heavily overgrown 20- and 26-inch swards, defoliation to 8 inches removed all or almost all the functional leaf (i.e., to almost nil values of leaf area index) and C02 uptake feel1 to very low values. Thus the effect of defoliation must be considered in terms both of initial and of final leaf area index.

20

40

60

80

DAYS FROM EMERGENCE

100 DAYS FROM SOWING

FIG.37. The effect of thinning and of defoliation on the rate of increase in the leaf area index of subterranean clover. Left (Bennett, 1960); right (Davidson and Donald, 1958).

It is believed that considerations of the leaf area relations of pastures must be the basis of their proper management. If a sward is constantly and heavily defoliated, then at all times it has so little leaf that it must have a low crop growth rate. Ideally a pasture should be maintained at the optimum leaf area index, with leaf being removed at the rate it is being produced ( Davidson and Phillip, 1958; Donald and Black, 1958). Such a theoretical state cannot be maintained in practice, but the concept does scrve to emphasize the disability both of a low leaf area on the one hand or of a heavily overgrown sward on the other. These relations give no support to the concept of rotational grazing.

94

C. M. DONALD

Any system which leads to a gradual increase in leaf area and then to the rapid removal of a large part of the foliage must inevitably incur periods in which the crop growth rate is depressed because the leaf area index is above or, more commonly, below the optimum. These relationships also emphasize the importance at the beginning of the season of permitting pasture to develop sufficient leaf before grazing begins. Whether growth follows the autumn rain of the Mediterranean

3 $4

I I 0

1

2

3

4

5

6

7

ILLUMINATION ( I . c . x l 0 0 0 )

FIG.38. The rate of CO, uptake by swards of Cymdon dactylon. Upper graph: swards of three heights, 14, 20, and 26 inches. Lower graph: the same swards after defoliation to 8 inches. (Alexander and McCloud, 1962.)

region or the rising spring temperatures of the American midwest, it remains important to avoid the removal of foliage at a time when growth is almost linear on the modest leaf area. This may involve an extension of the period of hand feeding of livestock. There are complications in terms of the leaf area in the use of “autumn-saved pasture” in temperate areas such as the United Kingdom and New Zealand. Well-grown autumn pasture is carried into the winter

COMPETITION AMONG CROP AND PASTURE PLANTS

95

for use during the months of slow growth rate. But if the pasture is, say, at a ceiling leaf area in the autumn, so that leaf production and leaf death are in equilibrium, any fall in light intensity, such as normally occurs from autumn to winter, must lead to a greatly accelerated leaf senescence because the lesser light cannot maintain the same depth of canopy alive. Much feed will be wasted. As a principle it seems desirable to ensure that feed carried into a period of declining light intensity should have a suboptimal leaf area index. XV. Leaf Arrangement and Competition for Light

Leaves do not lie in continuous, horizontally disposed layers. They may approximate to a horizontal disposition in a clover sward, but there is no continuity of the “layers.” In a cereal crop, leaves may lie at all angles, influenced by genotype, stage of growth, density, and nutrient status. Thus many of the general considerations regarding competition for light are subject to modification according to the arrangement of the leaves and its influence on competition. Leaves have a low transmissibility of light. Kasanaga and Monsi (1954) measured transmission by the mature leaves of some 80 species and found that most values lay between 5 and 10 per cent, grasses and herbs showing a strong mode at 9 per cent. Corn leaves exposed to 10,000 foot-candles transmit only 6 per cent of the light (Hesketh and Musgrave, 1962). If a crop canopy were in fact made up of continuous layers of leaf, then the transmission would be about 10 per cent through a leaf area index of 1, and, following Beer‘s law, 1 per cent through a leaf area index of 2, and so on. But this is not the picture found in nature. For example, Brougham ( 1958a) found a “light intercepting capacity” (the relative reduction of light by each unit of leaf area index) of 0.26 in a sward of perennial ryegrass, equivalent to a mean transmissibility of 74 per cent per unit area of leaf index. This value means that 50 per cent light was transmitted with a leaf area index of 2.25. Clover had a transmissibility of 50 per cent per unit of leaf area index. These great differences in the transmissibility of light by a leaf and a sward reside in the leaf arrangement of the sward-the absence of continuous leaf layers, Light penetrates directly as sun flecks, it reflects from leaf to leaf and, its transmission through leaves is probably of little significance. In the characteristic situation, a great majority of the leaves lie in diffuse light of steadily falling intensity with increasing penetration into the sward. Monsi and his colleagues (Monsi and Saeki, 1953; Kasanaga and Monsi, 1954; Saeki, 1960) have emphasized the importance of the

96

C. M. DONALD

“density” of the leaf layers, i.e., the proportion of each leaf plane occupied by the leaves ( l / l O , 1/5, 10/10, etc.). To take an extreme case, a canopy of leaf index 1, made up of a continuous layer of leaves of zero transmissibility, would intercept all light. If, on the other hand, the same canopy of leaf area index 1, were made up of 2 layers each with an index of 0.5, then clearly, depending on the depth between the layers, a great deal of light would be transmitted. The theoretical analyses by Kasanaga and Monsi (1954) suggest that at low light the most efficient canopy is made up of continuous (10/10) layers, but that at high light intensities leaf layers of low density (1/10) may give a growth rate 40% greater than continuous layers with the same total leaf area index. Another aspect of leaf arrangement, the degree of leaf dispersion within any layer, has been examined theoretically by Wilson ( 1 W ) . If the leaves are distributed at random in a leaf layer, the variance will be 1, if uniformly dispersed (underdispersion) it will be less than 1, while if clustered with heavy overlap ( overdispersion) it will be greater than 1. Warren Wilson suggests that the more uniform the dispersion (i.e., the lower the variance) the greater will be the crop growth rate because less light will penetrate to the ground and the overlap of leaves will be minimized. Perhaps the aspect of leaf arrangement which has the most immediate promise of contributing to increased production is the leaf angle, though all too few experimental data are available. Monsi and Saeki (1953) calculated the relative light interception by horizontal and erect foliage to be 1:0.44. Further, they showed that in five species with a leaf angle from the horizontal ranging from 0 to 75 degrees, the value of the extinction coefficient ( k of Beer’s law) followed these theoretical expectations. Brougham’s figure of 0.5 for the light-intercepting capacity (which approximates to k when k is small) of horizontally disposed white clover leaves, and of 0.26 for upright ryegrass leaves, conforms to this relationship. Wilson (1960) has emphasized the theoretical value of the inclined leaf compared with the horizontal leaf. The light intensity, particularly at the surface of the foliage, may greatly exceed the value needed for maximum photosynthesis; an inclined leaf will receive adequate light over a much greater area. It is in fact surprising that so little account of this more obvious aspect of leaf arrangement has been taken by plant breeders and agronomists. Watson and Witts (1959) have gained most interesting circumstantial evidence of the influence of leaf angle on production by the sugarbeet. They compared wild beet with a modern cultivated form.

COMPETITION AMONG CROP AND PASTURE PLANTS

97

Early in the season there was no difference in the leaf efficiency (N.A.R. or E.) of the two varieties and this indicated that there was no difference in their basic leaf physiology (this broadly conforms with intervarietal comparisons in other species). As the season progressed, both varieties fell in leaf efficiency, but the fall was much more marked in the wild than in the cultivated beet. The fall was interpreted as being

FIG. 39. An extrcme difference in leaf angle between two wheat varieties: 86-A variety from Persia with horizontal leaves. 8~-DHEADNOUGHT from New Zealand with vertical leaves (Donald, 1962).

due to mutual leaf shading; this was far more pronounced in the wild beet with its rosette of horizontally disposed leaves than in the cultivated form with leaves displayed at steeper angles and lying less closely upon each other. In brief, the competition between leaves for light was less acute in the modern beet and a greatcr proportion of the foliage received favorable light. It may well be, both in this crop and others,

98

C. M. DONALD

that the selection of genotypes with higher yield has unwittingly been a selection for leaf angle or some other aspect of leaf arrangement. Figure 39 shows two wheat varieties with an extreme difference in leaf angle, picked from a world collection of wheats at the Waite Institute. What is the significance of horizontal or vertical leaf angle on yield? The two varieties have been crossed (they gave a simple 3:l segregation for horizontal leaf in Fz) and it is hoped to produce isogenic lines

-;

100

I

\

I-

*

00

n 6 W

u

a a

60

W

t I-

40

Z

W

I-

z

I-

I

2J

20

0 DEPTH FROM SURFACE OF SWARD IN UNITS OF LEAF AREA INDEX

FIG. 40. The contrast in the profile of light density in relation to leaf area index in a grassy sward and a clovery sward (Stem, 1960).

differing only in leaf angle; these could be very useful experimental material. Finally, among leaf arrangements, there is the vertical concentration of the foliage, a factor not unrelated to Monsi’s density of the leaf layers. When elongation occurs in grasses, including cereals, there is immediately a very marked increase in the light intensity at the ground surface, even though the leaf area may be unchanged or increased (Stern and Donald, 1962b). This means that the greater vertical spacing of leaves permits increased downward penetration of light. On the other hand, circumstantial evidence suggests that there is no advantage in extremely wide vertical spacing of leaves as in very tall corn, that any

COMPETITION AMONG CROP AND PASTURE PLANTS

99

improvement in light relationships reaches an asymptote at some particular node length for any particular light environment. The great difference in light profiles of pasture communities can be seen in Fig. 40, where grass- and clover-dominant swards are compared. Grass had its leaf area (and its dry matter) rather uniformly distributed over its total height whereas clover had a heavy concentration of dry matter and leaf within a much narrower vertical span. In addition, the clover leaf was nearly horizontal whereas the grass leaf was at a steeper angle. In accord with this profile of foliage, the light density fell gradually as light penetrated into the grassy sward ( N z a . 5 ) , but abruptly as it entered the concentrated clover canopy ( N o ) . For example, at the final harvest, up to 33 per cent of daylight was intercepted in a vertical interval of 3.5cm. in a mature, clover-dominant sward, while the maximum percentage intercepted in the same vertical interval in a grass-dominant sward was 16 per cent of the daylight. All features of leaf arrangement which influence the competition for light between the leaves are of great potential importance in crop production. Leaf layer density, the dispersion of the leaves, the leaf angle, and the vertical distribution are all aspects of leaf arrangement lending themselves to worthwhile, though difficult, study. Undoubtedly we must also add such leaf features as the reflectivity of the leaf surface, affecting both the back reflection to the sky and the complex reflection patterns within the crop. Would any genotype of perennial ryegrass give more or less production if the leaf surfaces were dull instead of being glossy and highly reflective? How important are glabrous or tomentose glumes in the wheat ear? Perhaps these are less pressing aspects than leaf arrangement. The whole field is wide open for profitable study. XVI. Height and Competition

The influence of height on the competition for light is familiar in agriculture-the suppression of crop plants by taller weeds, of weeds by taller crops or of clover by grass (see Section VIII). Plant height is not only one of the most potent influences governing competitive relations, but it is also an influence subject to considerable useful manipulation by man. In 1929, Clements emphasized the great importance of height as a factor when he wrote: “The plants may be so nearly the same height that the difference is only a millimetre, yet this may be decisive since one leaf overlaps the other.” This great influence of a slight difference in height has been demonstrated in studies of the competition between pairs of clover varieties. Black (1960a) grew three clover varieties in pure culture and

100

C. M . DONALD

recorded petiole length; then he combined them in pairs. In each instance the variety with the longer petioles dominated the sward and yielded 70 to 95 per cent of the dry matter. Figure 41 shows the results of studies of YARLOOP and BACCHUS MARSH, two varieties of subterranean clover which gave equal yields in pure culture and had respective modal petiole lengths of 18 cm. and 14 cm. When the two varieties were grown in mixture, light measurements showed that at 20 days the YARLOOP leaves were intercepting 55 per cent of the light, and the BACCHUS MARSH

r

2:

z * I . g Y

I -

L Y

BM 2 0 DAYS

Yl0M

Y

0M 3 4 DAYS

YIBM

Y

0M

Y IBM

6 2 DAYS

FIG.41. Plant height and the profile of leaf area in swards of YARLOOP subterranean clover ( Y ), BACCHUS MARSH subterranean clover ( BM ), and mixed swards of these two varieties (Black, 1960a).

leaves only 30 per cent. By day 62 nearly all the YARLOOP foliage was carried above the leaves of the BACCHUS MARSH, so that the BACCHUS MARSH foliage received only 5 per cent of the light, even though it had 21 per cent of the leaf area. Thus YARLOOP attained strong dominance (80 per cent of the final yield), apparently for no reason other than its slightly greater petiole length. In another study, Black (1958) examined the influence of seed size on height, leaf area, and competition. As shown in Fig. 42, small seed (4mg.) and large seed (10mg.) of subterranean clover each sown at 150 seeds per square link gave exactly comparable swards in respect to height and leaf area; their yields were also identical (0.27 g./plant). However, it will be observed that at the time of the first harvest in June, the sward from the large seed, because it had larger cotyledons at germination, was slightly taller and had a slightly greater leaf area than did that from the small seed. This relatively small early difference governed the outcome of competition; in the mixed community the

101

COMPETITION AMONG CROP AND PASTURE PLANTS

plants from the small seed were progressively shaded and finally received only 2 per cent of the light and gave only 8 per cent of the yield. The length of petiole not only governs the competition between clover genotypes, but also that among the leaves themselves in a pure stand of clover. Brougham (1958b) showed that in a young stand of white clover the petioles were relatively short ( approximately 100 mm. ) but that as the young leaves grew up through the sward to form the top of the canopy, petiole length increased until the young leaves had

-

2

! L t ' ! I

SMAU

2

LA%€

-

-

-

-

-

.-

- --

-

-

- -

-

-

-

petioles of 300mm., whereas older leaves of less stature senesced below. Figure 43 shows the results of similar studies by Stern (1960); the height at which the laminae were held became greater as young leaves progressively overtopped older leaves. Pendleton and Seifs study (1962) of the effect of the difference in stature of two associated corn hybrids has already been recounted (Section V ) . Because of the easy access into a corn crop, this material could be of unique value in advancing our understanding of competition for light. A typical example of a height relationship in weed control is that described by Stahler (1948) in the control of Conuoluulus sp. in Minnesota. Observation might have suggested that the suppression of convolvulus by various crops was due in considerable degree to competition for water. But experimentation showed that the water content of the soil

102 I4

12

I

01,

C . M. DONALD

4 R A N T S PER M?

, , ,

20

01,

40

,

20

, 40

,

, , , 60 80

,

,

60

,

,

,

100

,

,

110

,

,

, , , , , , , 10 la, 120 140

, 160

140

,

. 160

/

O L 20

,

,

.

,

b0

40

DLI5 I"*(

I

lUL.

,

,

,

,

80 FROM

I

,

."C"II

,

I20

100

,

,

,

, IbO

140

SOWING

I

5'97

I

oc.

FIG. 43. The progressive increase in the height of the leaf laminae in clover swards of three densities (Stem, 1960). The successive leaves of a plant in the sward are numbered L1, L2, etc.

COMPETITION AMONG CROP AND PASTURE PLANTS

103

was reduced to much lower values in a pure convolvulus plot than in a plot in which convolvulus was making poor growth in competition with such crops as soybeans or sorghum. The authors measured the light under these crops and concluded that the suppression of convolvulus was primarily due to the greater height of the crop and to its successful competition for light. The development of the “stratified clip technique” by Monsi and Saeki (1953) has been a considerable aid to the fuller understanding of the influence of height. This technique, whereby the plant or crop is harvested in successive shallow layers from the surface downward, permits the construction of a profile of dry matter or leaf area. By the use of a light probe, this can be accompanied by a simultaneous record of the light profile (Fig. 23). An alternative technique involves clipping at ground level, followed by measurement of the individual petiole lengths and leaf areas; it has proved useful for the reconstruction of leaf profiles in clover swards (Black, 1958). XVII. The Interaction of Competition for Two or More Factors

Competition may occur for a single factor of the environment, as when water and nutrients are nonlimiting and there is competition only for light; it may occur also in an arid environment among widely spaced plants competing only for water. But in many crops and pastures competition involves more than one factor. This is because, as already noted, the secondary effect of competition for a nutrient or for water is differential growth, differential stature, and hence competition for light. Even in a monoculture in which all plants are limited in their growth by competition for water or nitrogen, the plants are likely to be of sufficient stature to compete for light. It has long been postulated that competition for each of two factors will involve interaction, so that the aggressor species will gain competitive advantage exceeding the sum of the effects which occur when each factor operates alone (Clements et al., 1929). These authors sought to examine this interaction under field conditions, but the methods followed are open to objection (relief from competition for light, but not for minerals, by neighbors was sought by bending the neighbors away from the test plants; this would affect growth and mineral uptake by the neighbors themselves). Chippindale (1932) sought in somewhat similar fashion to demonstrate in seed boxes the interaction between competition for light and competition for minerals when Lolium italicum and Festuca pmtensis were growing together on an infertile medium. After 14 months the

104

C. M. DONALD

fescue was so heavily dominated by the ryegrass that it had failed to develop beyond small, hairlike “seedling plants. When the ryegrass was clipped to admit light to the fescue or when nutrients were applied without clipping, the fescue remained depauperate. But when both clipping of the ryegrass and application of nutrients were practiced, the fescue grew vigorously. This technique, while dramatic in its effects, was imperfect because the plants considered to be competing for nutrients (after the clipping) had no tops, and because in the untreated culture competition may have been solely for light, the available nitrogen having been taken entirely by the successful competitor for light, the ryegrass. In fact it is extremely difficult (or perhaps impossible) to study the interaction of competition for two factors under normal field cropping conditions. Experiments can be designed to show the interaction of nitrogen and phosphorus on growth, or of water and nitrogen on growth. But such a study takes no account of the influence of these factors on the competition for light which may be of prime importance in governing the outcome of changing water or nitrogen status. It seems inescapable that if two factors of competition are to be studied under rigorous conditions, one of the factors selected for examination must be light, and considerable departure from field conditions must be made. A study has been made of the interaction of competition for light and for nitrogen in pot cultures. Lolium perenne and Phalaris tuberosa were grown in specially designed cultures in which plantings could be so arranged as to permit no competition, or competition only (senszc stricto) for nitrogen, or only for light, or for both factors (Donald, 1958). Aspinall (1960) has also reported a design of similar principle for pot cultures intended to segregate the competition of roots from that of tops. The basic design is shown in Fig. 44; in brief, the tops of plants competing only for nitrogen were separated by a panel and could not SOIL

PARTITION

1

AERIAL PART-

NO

COMPETITION

C O M P E T I T I O N FOR

C O M P E T I T I O N FOR

LIGHT

NUTRIENTS

C O M P E T I T I O N FOR LIGHT AND

NUTRIENTS

FIG. 44. The structure of pot cultures used to segregate the competition for light and for nutrients. A and B denote the positions of two competing plants (Donald, 1958).

105

COMPETITION AMONG CROP AND PASTURE PLANTS

intermingle and compete for light; alternatively, the roots of plants competing only for light were separated by a panel in the soil and could not invade each other’s root zone and compete for nitrogen. Some of the results (high level of nitrogen supply at 105 days) are shown in Table XX. TABLE XX Competition for Light, for Nitrogen and for Both by Lolium perenne and Phalaris tuberosu (Donald, 1958)a Crass

Loliuni Phalaris a

No competition

Competition for light

Competition for nitrogen

Competition for both

4.71 4.67

4.19 3.19

4.31 1.17

4.72 0.32

Values stated as yield of dry matter in grams,

The Phalaris was depressed by 22 per cent by competition for light in the absence of competition for nutrients, and by 73 per cent by competition for light in the presence of competition for nutrients. Thus the interaction of the two forms of competition was strong and positive. As described by Clements et al. (1929) in relation to light and water: “A larger, deeper or more active root system enables one plant to secure a larger amount of the chresard [available water], and the immediate reaction is to reduce the amount obtainable by the other. The stems and leaves of the former grow in size and number, and thus require more water; the roots respond . . , to supply the demand, and automatically reduce the water still further, . . . , At the same time, the correlated growth of stems and leaves is producing a reaction on light by absorption, leaving less energy for the leaves of the competitor beneath it, . . . .” The successful species acquires a continuously increasing share of each of the factors in short supply. A principle which emerged from this study with divided pot cultures was that competition for light, or for any other single factor, is not a simple effect but has at least two components which themselves interact. A plant heavily shaded by its neighbor suffers reduced photosynthetic activity. This leads to lesser growth, a smaller root system, a reduced exploration of the soil, and thus a redriced capacity to take up water or minerals. This effect on water or mineral uptake is quite independent of competition by a neighbor for water or minerals. Conversely a plant with reduced nitrogen supply because of competition has less foliage and a reduced capacity to intercept light, even though it is suffering no competition for this factor. In the instance cited above, the Phalaris suffering competition only for light had 20 per cent less uptake

106

C . M. DONALD

of nitrogen, while the Phalaris suffering competition only for nutrients had 73 per cent less foliage (weight of tops) for the interception of light. These relationships are schematically presented in Table XXI. TABLE XXI The Components and the Interaction of Competition for Light and Competition for Nutrients by Two Species, Namely an Aggressor Species ( A ) and a Suppressed Species ( B ) (Donald, 1958) Competition for Competition for Competition for both light Effects light only nutrients only and nutrients Direct

( a ) Intrusion of A into light environment of B Reduced light supply for B

( c ) Intrusion of A into nutrient supply of B Reduced nutrient supply for B

B suffers: ( a ) Reduced light supply ( c ) Reduced nutrient supply

Indirect

( b ) As a result of reduced light supply B has reduced capacity to exploit its own nutrient supply

( d ) As a result of reduced nutrient supply B has reduced capacity to exploit its own light supply

( b ) Reduced capacity to exploit the nutrient supply ( d ) Reduced capacity to exploit the light supply

Interaction of ( a ) and ( b )

Interaction of ( c ) and ( d )

Interactions ab, ac, ad, bc, bd, and cd plus any higher order interactions

Interactions

When a plant suffers competition for two factors, it not only suffers the interaction outlined by Clements and his colleagues-the interaction of ( a ) and ( c ) in Table XXI-but in fact is subject to the multiple interacting effects of four factors. This gives background to the observation that a suppressed species may suffer such a cumulative effect of direct and interacting factors as to be eliminated. XVIII. The Heritability of Competitive Ability

Most of the studies on the heritability of competitive ability in plants have been done by Sakai and his colleagues at Misima, Japan. He has made extensive use of the technique of comparing the performance in pure culture and in mixtures of a number of genotypes of single cereal species. These studies show, first, that competitive ability has a genetic basis. When nine wheat varieties were grown alone and in all possible pairs (Sakai, 196l), the data could be analyzed in the manner of a diallelic study to give a value for “general competitive ability” for each

COMPETITION AMONG CROP AND PASTURE PLANTS

107

variety (based on mean increment or decrement in yield when grown in all eight mixtures) and for “specific competitive ability” (based on the performance of a variety in a particular mixture). Both general and specific competitive ability gave highly significant values-that is to say some varieties were strong competitors and others were weak competitors in all mixtures or in a particular mixture. This finding is not surprising; it is reported elsewhere, as in Black‘s clover varieties (1960a). But whereas Black related success to the effect of a specific character on competitive ability (length of petiole on competition for light), Sakai and Gotoh (1955) found in studies of 12 barley varieties that competitive ability was not associated with any observed or measured botanical features (height, maturity, seed size, plant habit, heading habit or grain yield). Indeed they found that vigor of growth in pure cultures may even show an inverse relationship to competitive ability. Thus 10 F, barley hybrids superior in pure stands to their midparents in several major attributes (plant weight 136 per cent, culm length 116 per cent, culm number 113 per cent, and weight of ears 128 per cent) had lower competitive ability than did their parents when grown in mixtures with tester varieties. These research findings led Sakai to the view that “competitive ability should be accepted as it stands as a genetic character, simple or aggregate” (Sakai, 1955). He further adds (Sakai, 1961) his belief “that the competitive ability of plants is not necessarily associated with such characters as are supposed to be of advantage in competition for light, water or nutrients.” Although it is clear that competitive ability, in any particular set of circumstances, has a genetic basis, it is more difficult to accept the further proposition that such ability may be independent of those plant attributes giving competitive advantage for the factors needed for growth, Though a number of characters were measured by Sakai, there are very many other attributes of possible significance in competition among barley plants, Some of them may be of great complexity or subtlety such as the leaf arrangement, the capacity for nutrient uptake, or the biochemistry of the tissues. It is certainly unsafe to assume that because no plant characters governing competitive ability for factors needed for growth were identified, they did not exist, or to assume that competitive ability has some independent significance or meaning. When examined as a genetic character, competitive ability has shown very low heritability (Oka, 1960). A Japonica and an Indica type of rice were crossed; the parental lines, the F2 population, and the F3 lines were tested for competitive ability against a standard variety. When com-

108

C. M. DONALD

petitive ability was measured by panicle number, h2 = 0.1177; when measured by plant weight, h2 = 0.0254. This perhaps supports the view that if “competitive ability” is measured under conditions in which competition can occur for many factors-for light, water, and various nutrients-a great number of plant characters is likely to be involved. On the other hand, it may well be that where competitive advantage depends on a single morphological feature (e.g., plant height), competitive ability may show higher heritability. Because the range of plant characters which can affect competition is so wide and so diverse, it seems most improbable that any uniform heritability pattern will emerge for competitive ability. On the whole I believe that to regard competitive ability as a genetic character is a generalization of doubtful value, since it may tend to defer the identification and analysis of the specific characters that govern competitive success. The interest of the geneticist in competition lies especially in its influence on natural selection. In agriculture this phenomenon is subject to major interference because most crops are pure cultures and a “weak competitor” may be exposed to no interspecific composition. In selffertilized crops in particular, it is normal in agriculture to preserve a genotype without apparent change through many generations. Some wheat, oats, and barley varieties in use today were bred or selected half a century ago; their ability to compete in the field with other varieties is of no consequence in their commercial usefulness or survival. On the other hand, in cross-fertilized crop species, selection can operate with great rapidity, and this phenomenon is of considerable importance in respect to seed production. A bred strain of a crop or pasture plant with a considerable genotypic variability may display an equilibrium composition in a particular environment. This means that the interaction of environment and genotype leads to a seed crop of substantially constant genotypic constitution from year to year. But if this complex of genotypes is taken to a different environment, the competitive ability of the constituent genotypes will change and the seed crop may differ greatly within a single generation. It is not possible to review this topic in detail in this paper, but a few examples may be cited. When the lucerne variety, RANGER, bred for the northern part of the United States, was grown in Arizona or Mexico, the resultant seed gave a community of plants differing appreciably after one generation from the northern material (Smith, 1955). When grown in Wisconsin, plants from southern seed were of greater height and showed muchincreased winter kill. Sylvh (1937) similarly showed that great changes

COMPETITION AMONG CROP AND PASTURE PLANTS

109

occurred in the genotypic constitution of a strain of red clover, bred at Svalof, when it was “grown on” in Ostergotland. At Svalof natural selection ensured the elimination of genotypes susceptible to stem rot and eelworm. In Ostergotland the severe winters are unfavorable to each of these pests and consequently the susceptible segregates were able not only to survive, but to compete successfully with the resistant segregates and to increase their successive contribution to the seed crop. When crops were grown together at Svalof from seed of each source, the yield of the Ostergotland material was seriously affected by parasitic attack and gave much reduced yields. Competition can thus be a direct factor in natural selection within man’s cultivars. On the other hand, it is important to keep in mind that natural selection can proceed without the operation of competition, though the phenomena may be superficially similar. When white clover from Germany was grown at Svalof, Sweden, its production increased by 37 per cent in two generations (Sylvh, 1937). This was due to the direct effect of the severe winters in killing the cold-susceptible segregatesnot to the competitive influence of the cold-resistant plants in the community. Similarly Middleton and Chapman (1941) found that smoothawned segregates disappeared from composite hybrid populations of barley. This was due not to the greater competitive ability of the roughawned types, but to a direct effect of the environment in reducing or preventing seed production by the smooth-awned types; these types showed a strong fall growth and consequently were susceptible to spring frosts. Day length can similarly effect gross changes in the genetic composition of crop and pasture cultivars without the operation of competition. XtX. Competitive Ability and Yield

The studies by Harlan and Martini (1938), Suneson (1949), and Sakai and Cotoh (1955) have clearly demonstrated that the adaptation of selffertilized cereal varieties to a particular environment, as judged by grain yield in a pure culture, is independent of competitive ability against other varieties. Harlan and Martini (1938) sowed a mixture of 11 barley varieties at 10 centers in the United States; after harvesting and resowing the mixed culture for periods of 4 to 12 years, a single variety became dominant. The leading variety was different at the various centers across the United States; furthermore, there were several instances in which the variety which was successful in farm use in a particular region was reduced to a very low proportion of the mixture in that region. Suneson’s results in California (Table XXII) likewise showed that varieties with

110

C . M. DONALD

a lower yield in pure culture may be highly successful competitors in a mixture. A comparison of the root systems of the most successful and least successful of Suneson’s varieties in the mixture has indicated a possible factor leading to the success of ATLAS over VAUGHN when they are TABLE XXII The Yield of Several Cereal Varieties in Pure Culture and Their Survival in a Mixture ( Suneson. 1949)

Yield in pure

ATLAS CLUB MARIOUT

HERO VAUGHN

culture (Mean 1929-1943)

Disease susceptibility

( = 100)

High Medium Medium Low

100 105 107

Composition of mixture in 1948 after 16 generations (equal proportions in 1933)

(k) 88.0 10.5 0.7

0.4

in competition (Lee, 1960). The competitive advantage of ATLAS begins just before jointing; a t this time numerous strong roots are produced at the crown. These branch freely, so that ATLAS develops a far more dense though somewhat shallower root system than VAUGHN. This may give ATLAS a greater capacity to exploit water and nutrients in the surface soil when the two varieties are growing in competition. These results seemed to suggest that the bulked hybrid population method of breeding, whereby hybrid populations are repeatedly resown for many generations in the expectation that the most productive genotypes will become dominant, may be of doubtful value. In fact, however, Suneson’s studies (1956) of this method have shown that when composite crosses are compared in successive generations with standard, conventionally bred varieties, they progressively approach and even pass the standard variety. This is illustrated in Table XXIII. At first sight these results, which show the emergence of productive genotypes under competition, seem to be in conflict with the studies of varietal mixtures. But two points must be made. In the varietal mixtures there were few genotypes and these were stable. In the hybrid populations each plant differed from its neighbors and, for many generations, from its parents. Thus the opportunity for the selection of “adapted genotypes (those producing many seeds) was high. The second point is that although bulked hybrids may, after a num-

111

COMPETITION AMONG CROP AND PASTURE PLANTS

her of generations, give yields as great or slightly greater than commercial varieties, they still fall far short of the potential maximum of the hybrid material. For example, Suneson (1956) was able to derive from an Fzr hybrid population three selections which gave yields over four years no less than 56 per cent above the standard commercial variety. TABLE XXIII Productivity of Successive Generations of a Bulked Hybrid Population, Compared with a Standard Local Variety@ Number of years tested

2 2 4 6 4 5 Q

Generations

F,-F, F7-F8

F,,-F,, F*,-F,o Fzi-F,, F,,-F!!,

Composite as % of standard variety

67 85 89 106 101 103

From Suneson ( 1956).

Thus it may be said that while bulked hybrid populations will show progressive adaptation to a point at which their yields may approximate or exceed those of conventionally bred varieties, there remains a sufficient degree of independence of yield and competitive ability to prevent genotypes of much greater potential yield in pure culture from gaining dominance. In the development of pasture plants, there may well have been a serious error of approach and procedure due to the tacit assumption that competitive ability and yield were identical attributes. This is illustrated by the work in the United Kingdom on perennial grasses and legumes. In the 1920s Stapledon (1928) conducted his classic studies on Dactylis glomerata ( orchardgrass ) ; he showed the great contrast between the ecotypes to be found in hedgerows and other ungrazed places and those found in old grazed pastures. The hedgerow types were tall, lax, and sparsely tillered, whereas those of old pastures were short, dense, and multitillered. It was reasoned from these studies that the most suitable and productive types for use in sown pastures were of the form and habit found in heavily grazed swards; as a consequence many bred strains of perennial grasses of the dense pasture habit were developed for use under grazing and have been extensively grown for some decades. Yet, surprisingly, these “pasture strains” have proved little or no more

112

C . M. DONALD

productive of meat or milk or even of herbage than the upright, freeseeding types which had been regarded as so inferior for pasture purposes. Erect, free-seeding strains such as Irish ryegrass or Danish cocksfoot, classed as poor material for pastures, have generally given much the same production as have the “pasture” strains. In retrospect we can identify the wrong step unconsciously taken in the reasoning of the plant breeders at that time, and generally followed by grassland workers in all countries. It was assumed that competitive ability and productivity were synonymous-that the ability to compete and persist under grazing conferred also an ability to produce under grazing-that the most productive grasses for pastures were the grasses to be found in permanent pastures. This now appears to have been a false premise. For example, Prendergast (1959) found no significant difference in the number of grazing days provided by “commercial” and pasture varieties of perennial ryegrass over a four-year period. (The term “commercial” is here used as in the United Kingdom to mean naturally occurring cultivars developed for seed production and carrying only commercial sponsorship.) The extreme pasture type ( S 23) showed better persistence and better resistance to weed invasion (the criteria by which it was originally selected), but it had much poorer production in spring and early summer, though better in the autumn and early winter. Similarly, replicated trials of liveweight gains on “commercial” and bred strains in Berkshire ( Grassland Research Institute, 1959) have shown a somewhat better performance by commercial strains than by pedigree strains of perennial grasses. There is in fact a good deal of circumstantial evidence that dense, compact tussocks may be less productive than more erect, lax types, simply because dense so-called pasture types suffer a greater degree of mutual shading of the leaves. There has been a notable trend away from the extreme pasture types, as illustrated by the popularity of the New Zealand short-rotation ( H I ) ryegrass (developed by hybridization of L. perenne and L . multiflmum), a grass of much more erect habit than the traditional pasture types of L. perenne. In the United Kingdom there has been a trend from pasture types of ryegrass toward timothy grass and meadow fescue, both grasses of more erect and less compact habit. It may well be that the superior production by coastal bermudagrass ( Cynodon ductylon ) over “common” bermudagrass is due almost wholly to the more erect habit, more widely spaced leaves and better utilization of light. It is, then, of some importance both in crop and pasture work that productivity and competitive ability should be clearly distinguished.

COMPETITION AMONG CROP AND PASTURE PLANTS

113

XX. Concluding Comments

The fuller understanding of competition among plants requires, almost by definition, a greater knowledge of the response of plants to their environment, and especially of the response to those environmental stresses created by neighbors. It seems at first surprising that we know so little of this basic relationship among plants within a crop, yet the reason is perhaps not too far to seek-studies of the behavior of interacting plants are so much more difficult than those of either the isolated plant or the community as a whole. Plant physiologists have studied the single plant, and agronomists have looked at the whole crop, but the plant within the community has scarcely been investigated. This is a field which promises both scientific depth and great potential reward in terms of crop production. It is a salutary thought that we do not know-nor have we even given the matter much consideration-what determines the density of population of cereal plants giving maximum yield. Yet until we know this, and especially until we understand the interaction of density with such factors as water and nitrogen, then the development of suitable varieties of plants must depend in the future-as in the past-on empirical plant breeding. We can claim great advances in genetics, and great advances in producing plants with drought escape or disease resistance, fatter pods or finer flowers. And the breeder can point, too, to varieties which, quite apart from these specific virtues, are able under the keen intraplant competition of a commercial crop, to yield more grain, more leaf, more dry matter. Why? The breeder has no idea. Indeed the answer to such a question will often be that it yields more because it has more ears, or more florets or more fertility or less abortion, which of course, is little more than a paraphrase of the statement that it yields more. Actually what happened was that the breeder selected it because it yielded more, not that it yielded more because it was consciously bred to do so. Why does a modern wheat variety, whether in Greece or New Zealand, yield more than a variety of like maturity and disease resistance of fifty years ago? Because it either ( a ) fixes more carbon or ( b ) has a greater proportion of the carbon in the grain. Why? No one knows. Perhaps it has a different root system, better leaf arrangement and light utilization, more glume surface, or one of many factors affecting growth and photosynthesis. And in particular, it has these desired characteristics when growing under the acute stress conditions of a commercial crop. And of these aspects affecting photosynthesis, a field of enquiry which offers especial promise is the study

114

C . M. DONALD

of leaf arrangement and its effect on competition for light within the crop and on the level of photosynthesis. It is suggested in this review that the maximum yield for any particular genotype and environment may be given at that density of plants at which competition within the plants is minimal. Certainly we can say that at much lower densities or at much higher densities intraplant competition is more intense. Alternatively we can say that maximum yield per acre is an integral of competition between plants and competition within plants. Surely it is reasonable to regard the understanding of this relationship which faces us constantly in the field, as of immediate importance in agricultural research. A notable deficiency in our understanding of competition is the substantial independence of competitive ability and yield. As recounted in this review, some instances are partly understood, but these are all too few. Obviously the interpretation of these relationships would add greatly to our understanding of crop growth and production, and would serve also to safeguard us from the assumption in selection programs that ability to compete gives ability to render maximum yields in pure culture. I should like to conclude on a recurring note, by emphasizing once more that the plant in a crop is a plant under acute competitive stress, and I make a plea that we should direct more research to this depauperate crop plant which is the source of all our agricultural production, ACKNOWLE~CMENTS Part of the material in this article was assembled or prepared during 1959 when I spent my sabbatical leave from the University of Adelaide in looking at overseas work on plant competition. The assistance of the Rockefeller Foundation, the Royal Society and Nuffield Foundation, and Come11 University in enabling me to visit Europe and the United States is gratefully acknowledged. I am indebted to Dr. W. R. Stem, Dr. K. Santhirasegaram and Mr. V. D. Wassermann, who willingly allowed the presentation of data from their theses in the University of Adelaide. I acknowledge with thanks the comments by several of my University colleagues on various parts of this article. Mr. G. N. Wilkinson provided comment on several statistical aspects and Mr. C . A. McIntyre of C.S.I.R.O. kindly agreed to the publication of his “competition index,” which he devised to examine the data in Fig. 14. In the section on competition for light I have drawn considerably on a paper prepared in 1960 for the Society for Experimental Biology. REFERENCES Aberg, E., Johnson, I. J., and Wilsie, C. P. 1943. J . Am. Soc. Agron. 35, 357-369. Ahlgren, H. L., and Aamodt, 0. S. 1939. J. Am. Soc. Agron. 31, 982-985. Alexander, C. W., and McCloud, D. E. 1962. Crop. Sci. 2, 132-135. Army, T. J., and Kozlowski, T. T. 1951. Hunt Physiol. 26, 353-362. Aspinall, D. 1960. Ann. Appl. Biol. 48, 637-654.

COMPETITION AMONG CROP AND PASTURE PLANTS

115

Aspinall, D., and Milthorpe, F. L. 1959. Ann. Appl. Biol. 47, 156-172. Ballard, L. A. T. 1958. Australian J. Biol. Sci. 11, 246-260. Bennett, C. F. 1960. Project Report, Agronomy Department, University of Adelaide, Adelaide, South Australia. Black, J. N. 1956. Australian J. Agr. Res. 7 , 98-109. Black, J. N. 1957. Australian J. Agr. Res. 8, 335-351. Black, J. N. 1958. Australian J. Agr. Res. 9, 299-318. Black, J. N. 1960a. Australian J. Agr. Res. 11, 277-291. Black, J. N. 1960b. Oikos 11, 26-42. Black, J. N. 1963. Australian J. Agr. Res. 14, 20-39. Blackman, G. E. 1934. Ann. Botany (London) 48, 975-1001. Blackman, G. E., and Black, J. N. 1959. Ann. Botany (London) [N.S.] 23, 131-145. Blackman, G . E., and Templeman, W. G. 1938a. 1. Agr. Sci. 28, 247-271. Blackman, G. E., and Templeman, W. G . 1938h. Ann. Botany (London) [N.S.] 2, 257-280. Blaser, R. E., and Brady, N. C. 1950. Agron. J. 42, 128-135. Bond, G. 1941. Ann. Botany (London) [N.S.] 5, 313-337. Brougham, R. W. 1958a. Australian J . Agr. Res. 9, 39-52. Brougham, R. W. 1958b. New Zealand J. Agr. Res. 1, 707-718. Brown, B. A. 1939. J. Am. SOC. Agron. 31, 322-332. Bula, R. I., Smith, D., and Miller, E. E. 1954. Botan. Gaz. 115, 271-278. Charles, A. H. 1958. Field Crop Abstr. 11, 233-239. Chipman, E. W., and Mackay, D. C. (1960). Proc. Am. SOC. Hort. S c i . 76, 442-447. Chippindale, H. G. 1932. Ann. Appl. Biol. 19, 221-242. Clements, F. E., and Shelford, V. E. 1939. “Bio-ecology.” Wiley, New York. Clements, F. E., Weaver, J. E., and Hanson, H. 1929. Carnegie lmt. Wash. Publ. 398. Davidson, J. L., and Donald, C. M. 1958. Atrstroliun J. Agr. Res. 9, 53-72. Davidson, J. L., and Phillip, J. R. 1958. In “Climatology and Micro-Climatology,” Proc. Canberra Symposium, pp. 181-187. UNESCO, Paris. Davies, G. J. 1954. Rept. No. 38, National Institute of Agricultural Engineering (Silsoe, England), 28 pp. [Quoted by Holliday ( 1960b ).I de Wit, C. T. 1959. Jaarb. lnst. Biol. Scheik. Onderz. Wngeningen for 1959. de Wit, C. T. 1960. Verslag. Landbouwk. Onderzoek. 66, 8. de Wit, C. T. 1961. Symp. Soc. Exptl. Biol. 15, 314-329. de Wit, C. T., and Ennik, G. C. 1958. Jaarb. lnst. Biol. Scheik. Onderz. Wageningen for 1958. Donald, C. M. 1946. 3. Council Sci. Znd. Res. 19, 32-37. Donald, C. h4. 1951. Australian J . Agr. Res. 2, 355-376. Donald, C. M. 1954. Australian 1. Agr. Res. 5, 585-597. Donald, C. M. 1958. Austialiun J. Agr. Res. 9, 421-435. Donald, C. M. 1961. Symp. Soc. Exptl. B i d . 15, 282-313. Donald, C. M. 1962. J. Australian lnst. Agr. Sci. 28, 171-178. Donald, C. M., and Black, J. N. 1958. Herbage Abstr. 28, 1-6. Donald, C. M., and Neal-Smith, C. A. 1938. Bull. Council Sci. lnd. Res. 122. Drapala, W. I., and Johnson, C. M. 1961. Agron. J. 53, 17-19. Dungan, G . H., Sisodia, U. S., and Singh, G. D. 1946. Allahabad Farmer ( l n d i u ) 29, 8-13 (quoted by Pendleton and Dungan, 1958). Dungan, C. H., Lang, A. L., and Pendleton, J. W. 1958. Aduan. Agron. 10, 435473.

116

C . M. DONALD

Engledow, F. L. 1928. J. Agr. Sci. 18, 1-40 Erdmann, M. H., and Harrison, C. M. 1947. J . Am. SOC. Agron. 39, 682-689. Goodall, D. W. 1960. Bull. Res. Council Israel 8D, 181-194. Grassland Research Institute, Hurley, U. K. 1959. Expts. in Prog., NO. 12, 29. Gray, B., Drake, M., and Coley, W. G. 1953. Soil Sci. SOC. Am. Proc. 17, 235-239. Green, J. O., and Cowling, D. W. 1960. Proc. 8th Intern. Grassland Congr. Reading, England, 1960 pp. 126-129. Grimes, D. W., and Musick, J. T. 1960. Agron. J. 52, 647-650. Hanson, W. D., Brim, C. A,, and Hinson, K. 1961. Crop. Sci. 1, 255-258. Harlan, H. V., and Martini, M. L. 1938. 1. Agr. Res. 57, 189-199. Harper, J. L. 1961. Symp. SOC. Exptl. Biol. 15, 1-39. Haynes, J. L., and Sayre, J. D. 1956. Agron. J. 48, 362-364 Heinecke, A. J., and Childers, N. F. 1937. Cornell Uniu. Agr. Expt. Sta. Mem. 201. Herriott, J. B. D., and Wells, D. J. 1960. J. Brit. Grassland SOC. 15, 63-69. Hesketh, J. D., and Musgrave, R. B. 1962. Crop Sci. 2, 311-315. Hinson, K., and Hanson, W. D. 1962. Crop Sci. 2, 117-123. Hodgson, G. L., and Blackman, G. E. 1956. J. Exptl. Botany 7, 147-165. Hoff, D. J., and Mederski, H. J. 1960. Agron. J. 52, 295-297. Holliday, R. 1953. Agr. Progr. 27, 108-123. Holliday, R. 1960a. Nature 186 (4718), 22-24. Holliday, R. 1960b. Field Crop Abstr. 13, 1-16. Holliday, R. 1963. Field Crop Abstr. 16, 1-11. Hozumi, K., Koyama, H., and Kira, T. J. 1955. J . Inst. Polytech. Osaka City Unio. Ser. D. 6, 121-130. Iwaki, H. 1958. Japan. J. Botany 16, 210-226. Jantti, A., and Kramer, P. J. 1956. Proc. 7th Intern. Grassland Congr. Palmerston New Zealand 1956 pp. 33-44. Jarvis, R. H., Hanley, F., and Ridgman, W. J. 1958. J. Agr. Sci. 51, 229-233. Jones, M. G. 1933. J. Roy. Agric. SOC. Engl. 94, 21-41. Jorge, N. H., and Staniforth, D. W. 1961. Agron. J. 53, 1-5. Karper, R. E. 1929. J. Am. SOC. Agron. 21, 344-354. Kasanaga, H., and Monsi, M. 1954. Japan. J. Botany 14, 304-324. Kennedy, W. K. 1958. Cornell Unio. Agr. Expt. Sta. Bull. 935. Kiesselbach, T . A., and Weihing, R. M. 1933. J. Agr. Res. 47, 399-416. Kira, T., Ogawa, H., and Sakazaki, N. 1953. J. Inst. Poly. Osaka City Unio. 4, 1-16. Koyama, H., and Kira, T. 1956. 1. Inst. Polytech. Osaka City Unio. Ser. D. 7, 73-91. Kramer, J. J., and Clarke, W. S. 1947. Plant Physiol. 22, 51-57. Kurtz, T., Melsted, S. W., and Bray, R. H. 1952a. Agron. J. 44, 13-17. Kurtz, T., Melsted, S. W., Bray, R. H., and Breland, H. L. 1952b. Soil Sci. SOC. Am. Proc. 16, 282-285. Lang, A. L., Pendleton, J. W., and Dungan, C. H. 1956. Agron. J. 48, 284-289. Larson, W. E., and Willis, W. 0. 1957. Agron. J. 49, 422-426. Lee, J. A. 1960. Evolution 14, 18-28. Linehan, P. A., and Lowe, J. 1960. Proc. 8th Intern. Grassland Congr. Reading, England, 1960 pp. 133-137. Mann, H. H., and Barnes, T. W. 1953. Ann. Appl. Biol. 40, 566-572. Mather, K. 1961. Symp. Soc. Exptl. Biol. 15, 264-281. Middleton, G. K., and Chapman, W. H. 1941. 1. Am. SOC. Agron. 33, 361-366. Milne, A. 1961. Symp. Soc. Exptl. Biol. 15, 40-61. Milthorpe, F. L. 1961. Symp. SOC. Exptl. Biol. 15, 330-355.

COhfPETITION AMONG CROP AND PASTURE PLANTS

117

Mitscherlich, E. A. 1919. Landwirtsch. Jahrb. 53, 341. Monsi, M., and Saeki, T. 1953. Japan. 1. Botany 14, 22-25. Morrow, C. E., and Hunt, T. F. 1891. Illinois Agr. Erpt. Sta. Bull. 13. Mott, G. 0. 1943. Soil Sci. SOC.Am. Proc. 8, 276-281. Mouat, M. C. H., and Walker, T. W. 1959. Plant Soil 11, 41-51. Myers, L. F., and Lipsett, J. 1958. Australian J. Agr. Res. 9, 1-12. Nelson, C. E., and Robins, J. S. 1956. Agron. J. 48, 99-102. Nelson, 0. E., and Ohlrogge, A. J. 1957. Science 125, 1200. Oka, H. I. 1960. Japun. J. Breed. 10, 61-68. Papadakis, j. S. 1941. 1. Am. Soc. Agron. 33, 504-511. Pendleton, J. W., and Dungan, G. H. 1953. Agron. J. 45, 442-444 Pendleton, J. W., and Dungan, G. H. 1958. Agron. J. 50, 341-343. Pendleton, J. W., and Seif, R. D. 1962. Crop Sci. 2, 154-156. Pendleton, J. W., Jackob, J. A., Slife, F. W., and Bateman, H. P. 1957. Agron. J . 49, 44-48. Penman, H. L. 1948. Proc. Roy. SOC. A193, 120-145. Perekaljskii, F. M. 1951. Selek Semenouod 18, 70-73; 1952. Field Crop Abstr. 5 (1036). Petersen, M. L., and Bendixen, L. E. 1961. Agron. J. 53, 45-49. Pickworth Farrow, E. 1917. J. Ecol. 5, 155-172. Porter, K. B., Jensen, M. E., and Sletten, W. A. 1960. Agron. J. 52, 431-434. Prendergast, J. J. 1959. J. Brit. Grassland Soc. 14, 238-243. Puckridge, D. W. 1962. Thesis, Dept. Agronomy, Univ. of Adelaide, Adelaide, Australia. Reestman, A. J., and de Wit, C. T. 1959. Neth. J. Agr. Sci. 7 , 257-268. Riceman, D. S., Donald, C. M., and Evans, S. T. 1940. Australia, Council Sci. Ind. Res. Pamphlet 96. Roberts, J. L., and Olsen, F. R. 1942. J. Am. SOC. Agron. 34, 695701. Robinson, P. H., and Sprague, V. G. 1947. J. Am. SOC. Agron. 39, 107-116. Rossiter, R. C. 1959. Australian 1. Agr. Res. 10, 305321. Russell, M. B., and Danielson, R. E. 1956. Agron. J. 48, 163-165. Saeki, T. 1960. Botan. Mag. (Tokyo) 73, 55-63. Sakai, K. I. 1955. Cold Spring Harbor Symp. Quant. Biol. 20, 137-157. Sakai, K. I. 1961. Syrnp. SOC. Erptl. B i d . 15, 255-263. Sakai, K. I., and Goto, K. 1955. 1. Heredity 46, 139-143. Salter, P. J. 1961. J. H o d . Sci. 36, 241-253. Santhirasegaram, K. 1962. Thesis, Waite Inst., Univ. of Adelaide, Adelaide, Australia. Sears, P. D. 1960. Proc. 8th Intern. Grassland Congr. Reading, England, 1960 pp. 130-133. Shinozaki, K., and Kira, T. 1956. 1. Inst. Polytech., Osaka City Unio. D7, 35-72. Slatyer, R. 0. 1955. Australian J. Agr. Res. 6 , 365-377. Smith, D. 1955. Agron. J. 47, 201-205. Smith, H. F. 1937. Australia, Council Sci. Ind. Res. Bull. 109. Sowell, W. F., Ohlrogge, A. j., and Nelson, 0. E. 1961. Agron. J. 53,25-28. Sprague, H. B., and Ferris, H. F. 1931. J . Am. SOC.Agron. 23, 516-533. Stahler, L. M. 1948. I . Am. SOC.Agron. 40, 490-502. Staniforth, D. W., and Weber, C. R. 1956. Agron. J. 48, 467-471. Stapledon, R. G . 1928. J. Ecol. 16, 71-104. Stem, W. R. 1960. Thesis, Dept. Agron., Univ. of Adelaide, Adelaide, Australia.

118

C. M. DONALD

Stem, W. R., and Donald, C. M. 1961. Nature 189, 597-598. Stem, W. R., and Donald, C. M. 1962a. Australian 1. Agr. Res. 13, 599-614. Stem, W. R., and Donald, C. M. 1962b. Australian J. Agr. Res. 13, 615-623. Stringfield, G. H. 1959. Agron. 1. 51, 472-473. Suneson, C. A. 1949. J. Am. SOC.Agron. 41, 459-461. Suneson, C. A. 1956. 1. Am. SOC. Agron. 48, 188-191. Sylvkn, N. 1937. Imp. Bur. Herb. Plonts Bull. 21. Trumble, H . C., and Shapter, R. E. 1937. Australia, Council Sci. Ind. Res., Bull. 105.

van den Bergh, J. P., and Elberse, W. T. 1962. 1. Ecol. 50, 87-95. 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. Wame, L. G . G. 1951. 1. H o e . Sci. 26, 84-97. Wassermann, V. D. 1963. Thesis, Dept. of Agron., Univ. of Adelaide, Adelaide, Australia. Watkin, B. R. 1954. 1. Brit. Grassland SOC.9, 35-46. Watson, D. J. 1947. Ann. Botany (London) [N.S.] 11, 41-76. Watson, D. J. 1952. Advances Agron. 4, 101-145. Watson, D. J. 1958. Ann. Botany (London) [N.S.] 22, 37-54. Watson, D. J., and Witts, K. J. 1959. Ann. Botany (London) [N.S.] 23, 431-439. Weber, C. R., and Staniforth, D. W. 1957. Agron. J. 49, 440-444. Went, F. W. 1957. In “The Experimental Control of Plant Growth,” Chronica Botanica, Waltham, Massachusetts. Wiggans, R. G. 1939. 1. Am. SOC. Agron. 31, 314-321. Willoughby, W. M. 1954. Australian 1. Agr. Res. 5, 157-180. Wilson, J. W. 1960. Proc. 8th Intern. Grassland Congr. Reading, England, 1960, pp. 275-279.

CHEMISTRY OF THE MICRONUTRIENT ELEMENTS IN SOILS

. .

. .

J F Hodgson

U S Deportment of Agriculture. Ithoco. New York

Page I . Introduction ................................................ I1 . Ccochemistry of Micronutrients and Its Relation to Soils . . . . . . . . . . A . Distribution in Igneous Rocks .............................. B. Redistribution during Sedimentation ........................ C . Amounts of Micronutrients in Rocks and Soils . . . . . . . . . . . . . . . . 111. Forms of Micronutrients in Soils ............................... A . Precipitates of Iron and Manganese ......................... B. Surface Adsorption of Heavy Metals ........................ C. Reactions of Boron and Molybdcnum with Soil Surfaces . . . . . . D . Reactions with Organic Matter ............................ E . Occlusion of Micronutrients ............................... IV . Distribution of Micronutrients in Soils .......................... A . Geographical ........................................... B. Movement . . . . . . . . . . . . . . . . . ...................... C . Profile .................... ...................... V. Factors Affecting the Availability of Micronutrients . . . . . . . . . . . . . . . A . pH .................................................... B. Organic Matter . . . . .................... C. Microbiological .... .................................. D . Oxidation and Redu .................... E . Seasonal Variation . ........... F . Rhizosphere ............................................. VI . Needs for Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References ..................................................

119 120 120 122 123 124 125 126 129 131 135 136 136 136 138 140 141 144 145 148 149 151 152 154

I. Introduction

A large number of elements are required for the growth and reproduction of plants and animals. Of these nutrients only a few are required in large amounts for agricultural production . Deficiencies of those remaining elements which are required in lesser amounts are most frequently related to specialized crops or certain types of soil. But as cropping systems become more intensive. changes in soil management practices frequently alter micronutrient availability. and depletion of 119

120

J. F. HODGSON

nutrients not added in fertilizers becomes more rapid. As the demand for higher yields increases and the plant’s requirement for major elements is more efficiently met, other nutrients are more likely to become limiting. With the emphasis on high yields must come an awareness of possible concomitant changes in crop quality. From these various considerations one might well anticipate a renewal of interest in the role micronutrient elements may play in restricting maximum production and optimum quality of farm crops. The present review is undertaken to summarize that portion of soil chemistry dealing most directly with the micronutrients of greatest agronomic importance: Fe, Mn, Cu, Co, Zn, B, and Mo. It is intended to supplement, but in no way to replace, the excellent examination of the subject by R. L. Mitchell (1955). Rather than duplicate the coverage of the latter, the writer will tend to emphasize a slightly different viewpoint, bring out developments in thought that have been introduced since 1955, and provide a broader introduction to the literature. Only a small cross section of the literature is cited, however. In many cases if cross references are provided, only one of a series of papers is cited. The terms micronutrient and trace element will be used here with restricted meaning. Use of the first term will be confined to those elements mentioned in the preceding paragraph. The words “trace element” will be used for all elements occurring in the soil in minute amounts without regard to their requirement by plants or animals. II. Geochemistry of Micronutrients and Its Relation to Soils

Geochemical principles provide a basis for an understanding of the distribution of elements in soil parent material. In this way they contribute to the prediction of areas of micronutrient sufficiency and deficiency. More important to this discussion, they provide basic principles of micronutrient association that must be considered in examining the chemical behavior of these elements in soils.

A. DISTRIBUTION IN IGNEOUS ROCKS As magmas cool forming igneous rocks, a sequence of rocks is produced, each bearing characteristic mineral species. As these minerals are formed, micronutrient elements become distributed between the newly formed solid and residual magmas. This section will attempt to set forth rules and observations that describe this distribution. Differentiation of magmas is generally believed to proceed from the early formed basic rocks, such as basalt (extrusive) and gabbro (intrusive), to acid rocks, rhyolite (extrusive) and granite (intrusive)

MICRONUTRIENTS IN SOILS

121

(Mason, 1960). The most common minerals might be expected to form in the order olivine, pyroxenes, amphibol, hornblende, plagioclase, biotite, orthoclase, and muscovite. Micronutrient elements form a host of naturally occurring minerals, but only rarely do these minerals represent an important source of the elements from an agronomic point of view. Of greater interest is the distribution of micronutrient elements in the commonly occurring soilforming minerals. According to Goldschmidt (1954), the tendency of elements to replaoe one another in crystal lattices depends on charge and ionic radius. The higher charge is preferred, and for ions of the same coordination number, the smaller ion is preferentially selected. Ringwood (1955) introduces the added factor of electronegativity. The less electronegative a cation, the greater the tendency to enter an ionic crystal lattice such as that of the silicate minerals. The group of elements, Mg, Fe, Co, Ni, and Mn, are all found as the divalent ion in the octahedral positions of silicate minerals. These elements are commonly concentrated in the first minerals formed and decrease as the magmatic cooling sequence progresses. The variations that occur within this group are proportional to the ionic radius of the divalent ion so that Mg with the smallest radius is most highly concentrated in the early members of the cooling sequence. Manganese, with the largest radius of the group, is least concentrated in the early minerals formed. The ratio of Mg:Mn may vary from around 100 in the first igneous rocks to form to 1 in the last. Cobalt, on the other hand, despite the fact that its ionic radius is very nearly that of divalent Fe, follows the distribution of Mg more closely than it does Fe (Carr and Turekian, 1961). This is in contrast to its behavior in many soils where it is most closely associated with Fe (Kubota, 1963). Zinc, Cu, and even Mo follow a pattern similar to the ferromagnesium group. Zinc is one of the more uniformly distributed among rocks formed at various stages of development, remaining approximately proportional to the sum of Fe and Mg. According to Goldschmidt (1954), Zn has a greater tendency to be associated with sulfides than most of the preceding group, an association that gives rise to the most common zinc mineral, sphalerite. Copper has an even greater tendency to become associated with sulfides and, presumably because of its high electronegativity, is to some degree excluded from silicates. This results in an inverse relationship of Cu to Si and a direct relationship of Cu to S, reported by Sandell and Goldich (1943). Like Cu, Mo most often occurs as the sulfide, MoS2. Some substitution of Mo as the molybdate ion occurs in silicate minerals, especially during later stages of differentiation (Goldschmidt, 1954).

122

J. F. HODGSON

The geochemistry of iron in magmas is principally a function of the oxidation state of the magma. Unless the magma has come in contact with air, the iron is largely Fe++.As such it is distributed among the ferromagnesium minerals as described earlier, but during later stages of magmatic differentiation it combines with sulfide to form a separate phase. Early oxidation of the molten rock can result in separation of magnetite or other oxides of Fe+++. The boron content of igneous rocks is characteristically low and evenly distributed, The more acid igneous rocks have a slightly higher concentration of B, which is often found as the highly resistant mineral tourmaline. B. REDISTFUBUTION DURING SEDIMENTATION Just as magmatic differentiation gives rise to a typical sequence of rocks and minerals, the process of sedimentation results in deposits having a characteristic sequence of particle sizes with associated mineral transformations and carbonate deposits. During sedimentation silicates commonly reorganize to form hydrolyzed alumino silicates with the typical layered structure associated with clay minerals. As this proceeds, metals are thought to be adsorbed on the developing surfaces and finally become occluded or isomorphously substituted in the new mineral species as the development of the lattice proceeds (Krauskopf, 1956). This process may take place during the transport and sorting of sedimentary particles or during the diagenesis of alumino silicate minerals at the site of sedimentation. In any case, it results in a concentration of Cu, Co, Zn, Mo, and B in the fine sediments. According to the tabulations of Krauskopf, there are nearly ten times as much Co and Zn in shales as sandstones. Mo and Cu are slightly less concentrated in shales unless the organic matter content is high. Under these conditions, Mo may concentrate spectacularly. Black shales may contain 1000 times as much Mo as is in the earth's crust generally. This value may be compared with maximum concentration factors of 77, 14, and 8 for Zn, Cu, and Co,respectively. The distribution of Fe and Mn in sediments is more a function of the oxidation potential of the environment during sedimentation than of particle size. Conditions are sufficiently constant under marine conditions to have led Goldschmidt ( 1954) to conclude that effective separation of Fe and Mn from A1 is the exception rather than the rule in most sediments. The geochemistry of B is probably more affected by the process of sedimentation than any other element considered. Boron, which remains relatively constant in igneous rocks in the range of 5 to 10 parts per

123

MICRONUTRIENTS IN SOILS

million (p.p.m.) varies from less than 3 to over 300 pap.m. in sediments. Certain bauxite and kaolin sediments, along with carbonates, are usually low in B, whereas marine shales and glauconitic sandstones are commonly high. Carbonate rocks are low in other micronutrients besides B. Krauskopf gives values for Co, Zn, Cu, and Mo in limestones and dolomites that are equivalent to 1-10%, 3-15%, 7-300/0, and 1050% of the amount found in the earth's crust, respectively. The weathering of these rocks to soils results in a concentration of the micronutrients, so that residual soils from carbonate rocks are not commonly ddcient in micronutrients. C. AMOUNTSOF MICRONUTRIENTS IN ROCKSAND SOILS The concentration of micronutrient elements in the crust of the earth (lithosphere) and in certain rocks and soils is given in Table I. Broad generalizations regarding changes in concentration of elements, as rocks weather into soil, are perhaps unwarranted, but certain consistencies do appear. Of the elements considered in Table I, it is apparent that only the metals Co, Cu, and Zn are generally of lower concentration in soils than in the material from which the soils are formed. TABLE I Micronutrients in Soils and Rocks (in p.p.m. )

Ele-

Earth's

Earth's

ment

crust5

crustb

B Mn

Fe co cu Zn Mo 0

*

c

d

10 1,000 50,000 40 70 80

3

1,000 50,000

2.3

23

45 65 1

Sedimentary rocksc

Basic rocksc

Acid rockso

10

15 600

12 670

27,000

5

2,000

86,000 45 140 130 1.4

Soilsc

Soilsa

10 1,000

33,000 23

10 850 38,000 8

30

57

20

20

60 1.9

80 2

58

40 1

2

-

8

Goldschmidt (1954).

Mason (1960). Vinogradov ( 1959 ) Swaine (1955).

.

From the standpoint of their soil chemistry, the micronutrients listed in Table I can be considered in three groups. The cations Co, Cu, and Zn, listed above, comprise a group of heavy metal cations that are held in soils principally on organic or inorganic surfaces or substituted as accessory constituents in common soil minerals. Fe and Mn enter into many of the same reactions as the previously described group. The divalent form of Fe and Mn is less strongly held by soil sur-

124

J. F. HODGSON

faces than Co, Cu, and Zn, but the property of being oxidized to higher valence states, which can form very insoluble oxides and phosphates, renders these elements much less available to processes of leaching. It is this property that explains their concentration during the formation of soils from rocks of the earth's crust as indicated in Table I. Boron and Mo represent what will be considered in this treatment as a third group, which is characterized as being held in soils as the anion. It is at first strange that these last two elements are conserved during the soilforming process. Boron, added as fertilizer, is more readily leached from soils than any of the group, but B is present in soil parent material principally as extremely resistant minerals, such as tourmaline, which are only slowly weathered to liberate B. Molybdenum toxicity is frequently associated with seepage areas in alkaline regions, an observation that reflects a high degree of mobility under these conditions. But where leaching plays a major role in soil formation, the soils are acid, and under acid conditions Mo combines very strongly with sesquioxides and clay minerals. Furthermore, biological recycling plays a large part in conserving both these elements from leaching. 111. Forms of Micronutrients in Soils

There are five general ways in which an dement can be bound in soils: ( 1 ) it can be associated with soil surfaces, either organic G I inorganic; ( 2 ) it can become occluded during development of new solid phases in which it is not a principal constituent; ( 3 ) it can precipitate with other soil components, forming a new phase; (4) it can occupy sites in soil minerals either as an original constituent or by entering the crystal lattice through solid state diffusion; or ( 5 ) it can become incorporated in biological systems and their residues in the soil. A knowledge of which of these types of reactions predominates in controlling the distribution of elements between soil matrix and soil solution is fundamental to an understanding of the soil chemistry of all nutrients. Unfortunately, the distinction between these forms is not always clear cut. The demarcation between surface adsorption and precipitation reactions, for example, is intrinsically diffuse. The surface itself may not be discretely defined. Gibbs and Marshall (1952) have shown that the surface of feldspars in contact with an aqueous solution alters to give a porous zeolite-like structure that grades into the normal feldspar crystal with increasing distance from the surface. Tiller (1961) has shown that removing this layer from feldspars by treatment with base markedly reduces the adsorption of Co and Zn. The same treatment

MICRONUTRIENTS I N SOILS

125

did not affect the adsorption of these elements by clay minerals not expected to have this relatively amorphous layer. Whether the entrance of an ion into such structures is looked upon as solid state diffusion or surface adsorption is largely a matter of definition. A. PRECIPITATES OF IRON AND MANGANESE Iron and Mn under conditions of satisfactory aeration are bound in most soils principally as precipitates of oxides and phosphates. It is the way the oxides of these metals respond to changes in pH, oxidation potential, and the presence of soluble complexing agents that for the most part governs the movement and availability of these elements to plants. The attraction of Fe for soil surfaces results in the distribution of Fe oxides as coatings on mineral particles, at least in humid temperate regions. It is well known that the color of most subsoils is derived from the state of oxidation and hydration of these coatings. The occurrence of at least six Fe oxides; magnetite, ilmenite, limonite hematite, goethite, and lepidecrocite have been recorded in soils (Brown, 1953), but the conditions leading to the formation or conversion of these different oxides are poorly defined and are beyond the scope of this review. A knowledge of Mn oxides in soils comes as much from the use of extractants for availability tests as from standard minerological approaches. Leeper (1935) initially considered five forms of Mn. By the use of extractants, including 2% hydroquinone in normal ammonium acetate, he separated Mn++ and four degrees of insoluble Mn02. Dion et al. (1947), using hydroxylamine hydrochloride, in place of hydroquinone, attempted to relate the reducibility of Mn oxides to their crystalline structure. They found that pyrolucite was more easily reduced than manganite and hausmanite. Jones and Leeper (1951) checked the plant response to these crystalline forms and found that oats could utilize Mn from pyrolucite and manganite, but not from hausmanite. These authors call attention to the importance of particle size and degree of crystallinity on crystal solubility. The degree of crystallinity may account for the solubility of manganite in their results in contrast to those of Dion et nl., since hausmanite was the only mineral they used that was highly crystalline. The ease with which Mn02 accepts hydrogen from a variety of reducing agents, including microbiological systems, was demonstrated by Mann and Quastel (1946). Of perhaps greater significance is the attention these authors call to the presence of trivalent Mn in soils and the fact that M n + + + is known to dismutate spontaneously to Mn++

126

J. F. HODCSON

and Mn+ + + + , From these considerations, the authors were led to postulate a Mn cycle in soils in which MnOa is formed through the biological oxidation of Mn++ to Mn+++ with subsequent dismutation of the latter, and is reduced directly to Mn++ by biological reduction. OF HEAVY METALS B. SURFACEADSORPTION

The soil chemistry of the remaining micronutrient cations appears to be regulated largely by reactions with mineral and organic surfaces. Coulombic forces giving rise to base exchange reactions certainly attract heavy metals as they do all soil cations, but the interactions of Cu, Co, and Zn with organic and clay mineral surfaces involve additional forces of attraction. In 1936, Jones et al., reported that a portion of the Zn added to several Florida soils could not be recovered by ammonium acetate extraction. Hibbard ( 1940) found similar results in California soils, adding that the strongly bound Zn could be replaced by using acid extractants. Even very sandy soils combined with Zn and also Cu in such a way that the metals could not be removed with neutral salt solution. Peech (1941) showed that Cu was bound even more strongly than Zn. Both these metals had a pH-dependent fraction that could not be removed by 1N NaCl. It is apparent from Brown’s results (1950) that a large part of the Zn added to a soil is extractable with neutral salt immediately after addition, but that the form extractable only by acid slowly increased with time. Zende (1954)also showed the presence of both a slow and relatively rapid fixation process for Co added to soils. The reactions of heavy metals with mineral surfaces has been the center of some attention since Elgabaly and Jenny (1943) demonstrated that the reaction of Zn with montmorillonite gave the same pattern as with soils. A certain fraction could not be removed with neutral salts. They concluded that the nonextractable Zn had entered the octahedral layer of the crystal lattice. Subsequent work has shown that most of the Zn these authors believed had entered the crystal could be easily removed by acid or by w e n less destructive means. But even these treatments leave a discrete fraction that later authors looked upon as being absorbed into the crystal structure (Hodgson, 1960). In addition to naturally occurring clay minerals containing Cu and Zn as major constituents of the octahedral layer, Caillbre et al. (1958) were able to incorporate Co into octahedral positions. In view of the isomorphic substitution of Cu, Co, and Zn in layer silicates and the relatively open crystal structure of these minerals, solid state diffusion would not be at all surprising, but to date there has not been any conclusive evidence presented in support of this contention.

MICRONUTRIENTS IN SOILS

127

The more predominant reaction of heavy metals with soils and clay minerals involves surface adsorption. Banerjee d al. (1953) found that the exchange capacity of soils and clays decreased when Co was added by an amount equal to the Co that was extractable only by acid. These authors also noted the increase in strongly bound Co with increasing pH. The fraction of adsorbed Co that is extractable only with acid was later found to increase at low concentrations of the metal (Spencer and Gieseking, 1954). Nelson and Melsted (1955) made similar observations for Zn bound to montmorillonite. They also noted that the d e s q tion of strongly bound Zn followed first-order chemical kinetics. DeMumbrum and Jackson (1956a) take the position that a certain fraction of the exchange capacity is specific for Cu and Zn. With a similar view, Hodgson (1960) defined a fraction of Co combining with montmorillonite as that which was essentially unaffected by the presence of Ca in the system. By reacting Co with montmorillonite in the presence of an excess of Ca, this fraction was characterized as having variable bonding strength with surface coverage, being pH sensitive, having a slow rate of reaction, and approaching first-order desorption kinetics. The major part of the Co was exchangeable with certain other heavy metals, but not with Ca, Mg, or NH4. It is interesting that Cu is the most effective ion in displacing Co and is probably the most strongly bound to silicate mineral surfaces; yet because of its high electronegativity, it is supposedly less disposed to enter octahedral positions than many similar ions ( Ringwood, 1955). Heydemann ( 1959) reports that adsorption of Cu by clays and quartz follows the Freundlich adsorption isotherm. Mineral surfaces vary considerably in their reactivity with heavy metals. Elgabaly (1950) extended his earlier work to many minerals and reported that vermiculite, brucite, and talc combined with especially large amounts of Zn. Unfortunately, pH was not controlled as is necessary in a critical comparison of the reactivity of different mineral species. Mitra and Prakash (1955) claim that in acid systems, montmorillonite adsorbed more Cu than kaolin did. Muscovite, biotite, and vermiculite were reported to adsorb relatively little Cu. According to Richardson and Hawkes (1958), even quartz reacts with Cu. Tiller and Hodgson (1962) studied the effect of pH on the sorption of C o and Zn by a group of mineral species and concluded that the reaction was very similar for the two ions and related in part to mineral stability. They gave the following order of reactivity for unground minerals: hectorite > vermiculite > nontronite > montmorillonite > halloysite > kaolinite. For ground minerals, the order was: muscovite > phlogopite > talc > biotite = vermiculite > pyrophyllite. Grinding had a serious d e c t on

128

J. F. HODCSON

the reactivity of mineral surfaces for Co and Zn (Tiller and Hodgson, 1962; Elgabaly, 1950), as it has been shown to alter exchange capacity (Kelley and Jenny, 1936; Dragsdorf et al., 1951; and many others). Most of the earlier work describing the reactions of heavy metals with soils was carried out at relatively high concentrations of the reacting ion. DeMumbrum and Jackson (1956b), in attempting to repeat some of these conditions, found that precipitation was a serious factor. While precipitation should be a major concern in laboratory investigations, it is an unlikely natural occurrence in acid soils with Cu, Co, and Zn at least. Even with Fe++ f, adsorption is apparently favored over precipitation as evidenced by the way iron oxides become distributed over soil surfaces. Whittig and Page (1961) have shown experimentally that adsorption of Fe+ + + by montmorillonite will precede precipitation over a range of pH values. Gibbs and Marshall (1952) found that Cubearing minerals, including oxides, are weathered in the presence of clay surfaces, Cu being adsorbed onto the silicates from its original form. They considered the bonding of Cu to clay to be as strong as that of Cu to organic matter. DeMumbrum and Jackson (1956a) repeated this approach, using Cu and Zn hydroxides and phosphates as the source of cation. Jamison (1943), in checking the possibility of Cu and Zn forming insoluble phosphates, found that inorganic soil surfaces form stronger bonds with these elements than is found in the phosphate precipitates. From considerations of solubility products (Table 11), even Cu could maintain a concentration of free ion of 600 parts per billion without precipitating as the oxide in soil solutions at pH of 7. The reactivity of heavy metals with soil surfaces makes this concentration of Cu highly TABLE I1 Hydrolysis and Solubility Constants for the Micronutrient Cationsa Cation Mn+ + Feff Fe+++ CO++ CU++ Zn+ + O K ,

=

log

-

O K ,

-

8.3 2.15 - 12.2 - 7.6 - 9.7

-

log K,, - 12.8 - 15.1

-37.8 - 15.2 - 19.0 - 16.6

(MOH+) ( H + ) (M++-)

K,, = ( M + + ) ( O H - ) ? Taken from Bjerrum et d. (1958). Only values that were corrected to zero conccntration at 22-25°C. were averaged except for the solubility product of F e + + + which is given for 18°C. a

MICRONUTRIENTS I N SOILS

129

unlikely even in acid soils. Little information is available, however, on soil solution concentration of micronutrients in either acid or alkaline soils. The likelihood of these elements being present as organic complexes in the soil solution makes the estimation of free ion concentration di5cult. The specificity of the reaction for heavy metals can best be accounted for by the formation of chemical bonds in the adsorption process. Results soon to be published by the writer indicate that the adsorption of Co by montmorillonite is endothermic and can be reversibly eliminated by treatment with F-. The reversible nature of the F- effect leads to the conclusion that the Co to clay bond is through surface hydroxyls that are replaceable by F-. These results are in agreement with those of DeMumbrum and Jackson (1956b), who found that the adsorption of Cu by montmorillonite altered an infrared peak they attributed to structural OH- groups. Hodgson and Tiller (1962) conclude from autoradiographs and other techniques that the location of these bonds is on the basal surfaces of clay minerals and arises from defects in the crystal structure that may originate from chemical or physical action on the surface. DeVore (1955) had previously suggested that growth surfaces, imperfections, dislocations, and intercrystal interfaces were particularly active in combining with trace elements. Formulation of a detailed model of the adsorption process is di5cult when even the nature of the reacting ion is not well understood. The lack of equivalence between different ions in estimating exchange capacity and observations of anion exchange led many early workers to suspect that metal complex ions of acetate or chloride were involved in the adsorption process (Elgabaly and Jenny, 1943). An associated depression in pH gave rise to the view that hydrolysis promoted adsorption, It would now appear that the formation of complex ions is of major significance only when the Concentration of the anion is very high (Page and Whittig, 1961). Many reports can be found relating to the involvement of the hydrolyzed ion. The writer reviews these in a separate paper (Hodgson et aE., 1963). One conclusion is evident, however, from the list of hydrolysis constants given in Table 11. Hydrolysis cannot explain the selective adsorption of heavy metals since in acid and neutral soils the basic ion for all those metals reported, except iron, is present in smaller amounts than the unassociated ion.

C. REACTIONSOF BORONAND MOLYBDENUM WITH SOIL SURFACES Despite the chemical dissimilarity, B bears a remarkable resemblance to the heavy metals in the way it combines with clay surfaces. Eaton

130

J. F. HODGSON

and Wilcox (1939) had considered three possible mechanisms for the chemical combination of B with soils: ( 1) anion exchange, ( 2 ) chemical precipitation, and ( 3 ) molecular adsorption. Philipson ( 1953) later made a similar classification substituting complex formation with acid groups in clays for molecular adsorption. This last suggestion may well be the most important in explaining the effect of lime on B fixation and provides the closest analogy to the heavy metal systems. The effect of lime on B retention in soils is in many cases very marked. Bobko et al. (1936) and Naftel (1938) originally interpreted this in terms of a stimulation of microbes in the soil, but Midgley and Dunklee (1939) found that microbial activity was not necessary to account for the effect of lime on B, and they argued in favor of reaction with mineral surfaces. The adsorption of B by hydrated ferric oxide, kaolin, and other clays does increase with pH, although at least in one case it goes through a minimum at pH of about 5 ( Barbier and Chabannes, 1953). Olson and Berger (1946) demonstrated that the Ca ion was not required to bring about fixation; NaOH was just as effective as lime. Calcium most likely does have a specific physiological effect in reducing B uptake by plants, however (Parks, 1944). The suggestion was also made at this time that B may enter the crystal lattice. Parks and Shaw (1941) and later Parks (1944) interpreted data on the effect of drying as favoring the substitution of B for A1 ions in aluminosilicate structures. This thesis has had recent support as a result of studies on the B content of clay sediments. Harder (1961) was unable to recover B added to illite clay, and observed that equilibrium had not been reached in 153 days after B was added to the mineral. He concluded that B entered the tetrahedral position of the clay, substituting for A1 or Si. As in the case of the heavy metals, the suggestion that B is entering the crystal lattice through solid state diffusion is strong, but conclusive evidence is lacking. Unlike the work with soils, Harder (1961) found that the uptake of B by detrital clay minerals was most pronounced at low pH values. He further characterized the sorption process as increasing with temperature. Mica minerals were found to take up more B than montmorillonite or kaolinite. Recent studies have indicated that B adsorption by soils can be partially described by the Langmuir adsorption equation. Hatcher and Bower (1958) compared adsorption and desorption of B by three soil clays in a column operation and found that the results compared favorably with the predicted behavior in each case. Biggar and Fireman (1960) extended the generalization to two other soils, but found that

MICRONUTRIENTS IN SOILS

131

at least in one case the reaction of B behaved like a precipitate rather than surface adsorption. Molybdenum, like the other micronutrients, may be present in soils in a number of different combinations. Dobritskaya (19Sl) recognizes three forms of Mo in soils: ( 1 ) part of the mineral structures of soils, (2) anion adsorbed by soil minerals, and ( 3 ) bound with organic matter. Of these, the second is probably the most important as it is affected by variables in the soil and as it influences plant uptake. Because Mo adsorption is particularly important on iron oxide surfaces, occlusion of the element in developing iron coatings should also be an important consideration. Surface reactions of Mo with mineral systems have been studied extensively by Jones (1956, 1957). He found that hydrous femc oxide adsorbs Mo much more strongly than other systems studied, followed by aluminum oxide, halloysite, nontronite, and kaolinite, in that order. The removal of iron oxides from a soil reacting strongly with Mo seriously reduced the Mo retention of the soil. The pH sensitivity of the Mo-iron oxide reaction follows the same pattern as that found in soils, and is believed to be due to anion exchange of the molybdate ion with surface hydroxyls. Jones (1957) calls attention to the analogy between the reactions of Mo and P in this respect in acid soils. However, there is no analogy between Mo and the precipitation of P by Ca in alkaline soils, so that Mo continues to increase in solubility as the pH increases above 6.5. L. H. P. Jones (1956) has also shown that the adsorption of Mo by aluminum and iron oxides in the laboratory can lead ultimately to the formation of A1 and Fe molybdates. Whether such precipitates can form in naturally occurring soils was not definitely established, but it would not be altogether surprising in view of the claim by Ishibashi d al. (1958) that iron mdybdates can form in sea water. D. REACTIONSWITH ORGANIC MATTER The role or organic matter in the reactions of micronutrients has been studied and emphasized by many workers. The high capacity of organic soils to fix micronutrients, especially copper, has focused attention on the significance that organic matter may have in the soil. But with all this attention, its role in mineral soils remains somewhat obscure. Basically, four ways can be used to assess the contribution of organic matter to the chemistry of micronutrients in soils: ( 1 ) the association of organic matter content with the distribution and availability of micronutrients in soils, ( 2 ) the effect of organic matter removal on the reactivity of soils with micronutrients, ( 3 ) a direct

132

J. F. HODGSON

attempt to assess the amount of an element present in the organic form, and ( 4 ) characterization of organic matter and its reaction sites. Organic soils are among those most commonly deficient in several micronutrient elements. Their content of Cu, for example, is frequently low, and their capacity to fix Cu is high. Agerberg (1959) estimated that one-half of the organic soils in northern Sweden were deficient in Cu for good plant growth, while only one-fifth of the mineral soils of that area were Cu deficient. Copper additions of 50-200 pounds per acre are common on muck soils, whereas they may be as low as 2% to 10 pounds per acre in certain Cu-deficient, sandy mineral soils (Teakle, 1942). There may also be considerable variation in the Cu content of a given peat or muck deposit. Since the Cu content of a peat comes almost entirely from the surrounding rock formations, those organic deposits occurring in granite or sandstone areas can be expected to be particularly low in Cu, and as Knott (1938) observed, a decrease in the Cu content of peat can be found with increasing distance from the old shore line. Of more general interest is the contribution of organic matter to the chemistry of micronutrients in mineral soils. Most micronutrients have been shown to be related to the organic matter distribution in many soils. Jensen and Lamm (1961) found a correlation coefficient of 0.81 between Zn content and organic matter distribution in different soils. Malyuga (1960) found a similar relationship between Mo and organic matter, and several authors describe such a relationship with Cu ( Agerberg, 1959). Berger and Truog ( 1945) observed a sufficiently close relationship between B and organic matter in Wisconsin soils, that those soils having less than 2% organic matter could generally be considered deficient in B. Correlative studies of this kind, while indicative of the contribution of organic matter in soils, may not necessarily give an accurate picture of the part it may play in a given soil. There is always the possibility that some other soil characteristic such as drainage may be bringing about the similar distribution of organic matter and the mineral element. Molybdenum in the West is found concentrated in seepage areas (Kubota et al., 1961), which are also high in organic matter. Organic matter may have a direct bearing on Mo distribution, but it is probably not the factor leading to the close correlation of these two characteristics. The removal of organic matter often results in a decrease in reactivity of heavy metals in soils. Baughman (1956) found that destroy-

MICRONUTRIENTS IN SOILS

133

ing the organic matter of a surface soil allowed time to recover almost all of an added souroe of Zn by extracting with dithizone. Recovery amounted to only 50 to 75 per cent when the organic matter was allowed to remain. Himes and Barber (1957) verified the complexing action of the organic matter by titrating added silver in a soil with and without the destruction of organic matter. Tiller et al. (1963) were unable to discern any effect of subsoil organic matter on the retention of Co, however. The writer has attempted to devise a means of measuring that fraction of heavy metals associated with organic matter in the soil, In an unpublished approach to the problem, a continuous leaching of the soil with a solution of 30% hydrogen peroxide and 0.25N CaC12 at controlled temperature was employed in hopes of removing organic Cu and Co without any readsorption of the metals by the soil colloids. This extraction was compared with a similar extraction with CaC12 alone. It had previously been noted that if the extraction was carried out in a beaker, significant readsorption could occur. Even the continuous flow system did not give the desired separation as attested by certain isolated subsoils, which released rather large amounts of Cu or Co to the peroxide extractant although they were essentially devoid of organic matter. A study of the reactions that can occur between organic matter and micronutrients reveals a relatively large capacity to combine very strongly with certain elements, notably Cu. Three classes of systems can be distinguished for this discussion. The high molecular weight organic compounds such as lignins are essentially immobile and serve to immobilize those elements associated with them. Short-chain organic acids and bases serve to promote solubility and movement, particularly of the heavy metals. Other complexing agents appear to be soluble, themselves, but form insoluble salts with heavy metals. Soluble complexing agents that form insoluble salts with heavy metals were first removed from soil by Bremner et al. (1946). They used pyrophosphate to remove an organic fraction that could be purified by dialysis and precipitated by the addition of heavy metals. It was later found that many complexing agents could extract an organic fraction of this description which was found principally in the subsoil and poorly drained soils (Martin and Reeve, 1957; Evans, 1959). It is likeIy that such substances are formed in well-drained, as well as poorly drained soils, but do not accumulate because of the relatively rapid microbial decomposition. Results of Kee and Bloomfield ( 1961, 1962), however, suggest that anaerobic decomposition of organic matter may

134

J. F. HODGSON

actually result in greater production of soluble organic complexing agents. Organic matter contributes significantly to the cation exchange capacity of many soils. Because the functional groups involved are weak acids and their configurations offer opportunity for chelation, these groups bind heavy metals very strongly. There is also some indication that Cu is bound to some functional groups that are not involved in the retention of alkali or alkaline earth metals. Broadbent (1957) eluted Cu and Ca separately from organic matter, using soil as an exchange column to separate the various fractions eluted. He found that Cusaturated organic matter gave four separate peaks, whereas Ca-saturated organic matter gave only two. Broadbent and Bradford (1952) had previously shown by methylation with diazomethane and dimethyl sulfate that carboxyl and phenolic groups and H+ attached to heterocyclic compounds seemed to be most important in supplying exchange groups. These groups are pictured by the authors as occurring on certain lignin side chains, amino acids, and tannins, and the heterocyclic N in nucleic acids. Himes and Barber (1957) also used methylation to implicate these functional groups in the reaction of Zn with organic matter. Dawson and Nair (1950) found that sulfhydryl groups were also effective in complexing Cu in peat. Russian workers from Tyulin (1940) to Manskaya et aZ. (1960) mostly interpret tbe binding of Cu in terms of polymerized humic and fulvic acids. Lack of knowledge of functional groups in humic and fulvic acids makes interpretation of this type of work difficult. Northmore (1959) has found a close relationship in Kenya soils between the iodine uptake and the potential of organic matter to combine with Cu. Iodine uptake does not vary with pH, but the ratio of I to Cu uptake by soils decreased linearly as pH increased and approached 1 when treated with ammonia. Northmore interprets his results as a breakdown of lignin chains to expose additional phenolic groups with increasing pH, but a direct competition of H for Cu in the reaction might offer as satisfactory an explanation. One of the most systematic investigations of organic functional groups in soils is offered by Wright and Schnitzer (1959). These authors used various means to measure the content of different oxygen-containing functional groups in a poorly drained podzol. They found that carbonyl groups were in greatest abundance in the surface soil, while carboxyls, which concentrated strikingly in the subsoil, were the predominant form in the Bh horizon. No attempt was made by these authors to estimate the relative contribution of the various functional groups to the binding of heavy metals in soils, but the method they used may provide the means,

MICRONUTRIENTS IN SOILS

135

not just of assessing the relative contribution of different organic functional groups, but of establishing the relative significance of these groups to inorganic bonding sites in mineral soils. Several methods have been used to estimate the stability constants of Cu and Zn bound to soil organic matter. Coleman et al. (1956) found a mean value of 3.2 x lo6 for the formation constant of Cu with peat from electrometric titration and direct estimation. Himes and Barber (1957) found values just slightly lower for Zn combining with soil organic matter. Results of this sort are valuable in providing an estimate of the order of magnitude of the binding constant. From them, we may gain some appreciation of the distribution of ions between soil matrix and soil solution. But the binding sites are almost assuredly too diverse to be described by a single formation constant. The relative strength of different reactive groups and the effect of the organic residue to which a reactive group is attached on the strength of bond remains to be examined for soil organic matter. Appreciable amounts of Co in soil may be present in the particularly strong chelate vitamin B12 (Duda et al., 1957; Afrikyan and Bobikyan, 1959). The Co in this molecule is present in the trivalent form and is essentially nonexchangeable with free Co in solution. (Diehl and Voigt, 1958.) Although the molecule is very large, it is soluble in water and there is some indication that it may be taken up by plants (Gray and Daniel, 1959). Boron, probably even more than the heavy metals, is associated with soil organic matter, but it is not altogether clear how the two become combined. In the case of heavy metals, Cu for example, can readily combine with any free complexing agent that may be available. Boron, on the other hand, may require metabolic processes to become incorporated in many of the organic forms with which it is associated. Several workers have examined the relation of B to polysaccharides. Isbell d a2. (1948) and Parks and White (1952) consider that B forms a diol with carbohydrate which is responsible for bonding B to soil organic matter. Clapp (1957) reviews the fairly extensive literature on this type of reaction.

E. OCCLUSION OF MICRONUTRIENTS Secondary iron oxide and siliceous minerals forming during the weathering process must surely present reactive surfaces for the adsorption of micronutrients. Ions adsorbed in this way are largely occluded as the precipitate continues to develop. Although the occlusion of micronutrients in all likelihood does occur, the nature and extent of its occurrence has not been studied so far as the writer is aware.

136

J. F. HODGSON

IV. Distribution of Micronutrients in Soils

A. GEOGRAPHICAL Maps outlining areas of reported micronutrient deficiency in the United States have been prepared by Beeson (1948, 1959), and recently by Berger ( 1962). In addition, factors iduencing micronutrient distributions in soils are discussed by a number of reviewers including Goldschmidt (1954), Mitchell ( 1955), Swaine ( 1955), La1 and Rao ( 1954), and Gilbert (1957). In view of this coverage, the present discussion will be limited to a few pertinent observations. The geographical distribution of micronutrients in soils appears to be more closely related to composition of parent material than to any other single factor. Soils may even exhibit less variation in micronutrient content than their parent rocks. Hibbard (1940) felt that soil-forming agencies tend to promote a similar content of Zn and other trace metals in soils. This type of behavior could well be expected for those elements held principally by surface adsorption if the strength of bond varies with surface coverage, as has been observed for Mo and the cations, Cu, Co, and Zn. Small amounts would be held very strongly against the forces of leaching, while larger amounts would be held less strongly. The clay and sesquioxide content, as it provides adsorbing surfaces, would be the important consideration in furthering the uniformity that Hibbard has predicted. Many successful attempts have been made to relate the micronutrient content of soil to its clay content (Atkinson et al., 1953; Jordan and Powers, 1946; Kabata, 1955). In one of these, Wahhab and Bhatti (1958) graphed the Cu, Zn, Mn, and Co contents of 32 Pakistan soils against the clay content and found a marked curvilinear relationship for Cu and Zn, and a linear relationship for Co and Mn. Cobalt was the most closely associated with clay content. B. MOVEMENT Mobility of an element in soils is no more than a reflection of its solution concentration as it is affected by the movement of water through the profile. As such, any factor that affects the solubility of an element must in the same way affect its movement. These factors will be considered in more detail as they affect the availability of micronutrients to plants, but one factor, the presence of soluble substances leached from organic residues, has received particular attention as it influences the movement of sesquioxides during podzolization. Soluble organic compounds are thought to be dective in promoting

MICRONUTRIENTS IN SOILS

137

the movement of Fe in two ways: (1) by stabilizing hydrosols of Fe in the soil solution, or ( 2 ) by forming strictly soluble organic complexes. Starkey and Halvorson as early as 1927 concluded that soluble complex formation was involved in the movement of iron. Although the body of relevant data has grown past the point where it can be satisfactorily treated in this discussion, disagreement continues as to which of the two effects predominate. Schnitzer (1957) and Bloomfield (1955a, 1958) have recently provided concise summaries of the two opposing points of view. The ability to quantitatively relate the movement of other heavy metals to that of iron in temperate regions will depend in part on which of the two alternatives predominates. There is little doubt that organic chelates can promote the movement of iron and aluminum along with other heavy metals. Artificial soil profiles have been produced in the laboratory by leaching soil and sand columns with synthetic chelates as well as leaf extracts (Bloomfield, 1955b; Wright and Levick, 1956; Thorp et al., 1957). Leaf extracts were found to solubilize Zn, Co, and other cations, as well as Fe, in proportion to the stability of the complexes formed (Titlyanova et al., 1959). Jones et al. (1957) showed that Co and Zn, when added to soil columns, moved downward only when either lucerne extract or COZ was added. From the standpoint of soil fertilization, micronutrient amendments do not generally move far in the soil profile. Scharrer and Hofner (1958) found very little movement of added Zn even in sandy soils. Lundblad et al. (1949) added up to 250 kg. of Cu per hectare (220 pounds per acre) to peat and found no more than 0.2 per cent moved out of the top 5 cm. of soil. In unpublished studies at U. S. Plant, Soil and Nutrition Lab., Albanl has shown that Co added to a coarse-textured soil of New England did not move out of the surface soil in 11 years. Of the micronutrients, Cu is probably the least mobile under the majority of conditions. Reuther et al. (1952) studied the relative movement of Cu, Zn, and Mn in pineapple soils and found that they were immobilized in the order given. Boron is the notable exception to the above comments. Katalymov (1951) observed a f34 to 76 per cent loss of B from a podzol after a few days’ leaching with water. There was a 30 to 52 per cent loss from a chernozem soil. This is an extreme example, but by no means an isolated one. Winsor (1952) found that B added in herbicidal amounts is lost from the surface 8 inches of a sandy loam within 6 months. As Scharrer and Hofner (1958) indicate, the presence of high pH and high clay content may seriously reduce the movement of B in soil profiles. Kubota et al. (1948) and Wilson et al. (1951) provide further examples of the 1

Present address: Oregon State University, Corvallis, Oregon.

138

J. F. HODGSON

way clay content can alter B movement in soils. From this type of study, it seems apparent that B in amounts toxic to most plants is removed from the majority of soils quite rapidly, but the leaching of B becomes sluggish when the element is present in smaller amounts, even though these quantities may still be toxic to the more sensitive plant species. C. PROFILE As rocks weather into soils, the micronutrient elements, like all elemental constituents, are subjected to the leaching waters passing through the soil. Plant roots, adsorbing surfaces, and coprecipitating ions compete with the leaching soil solution, with the result that each element, responding to these forces in different ways, becomes distributed in the soil in a characteristic manner. Such distributions, as summarized from Wright et ul. (1955), and Vinogradov (1959) are given for near modal podzols in Fig. 1. This approaches what might be considered a classical

FIG. 1. Distribution of micronutrients in soil profiles of near modal podzols, as summarized from Wright et al. (1955) and Vinogradov (1959).

distribution of micronutrients in the organic acid leached soils. It is presumably promoted by two principal factors. The first is an enrichment of the surface brought about by the deposition of micronutrient rich organic debris from the vegetative cover (Goldschmidt, 1937). The second is a concentration in the B horizon resulting from the downward translocation of the trace elements in the profile. The movement may come about as a result of association with mobilized clay or as a free ion or organic complex that is in part intercepted in its passage through the profile by precipitation or adsorption onto the surfaces of the lower horizons. Frequently, especially as conditions favoring true podzol formation become more remote, the above pattern becomes less evident. Swaine and Mitchell (1960), after an extensive study of Scottish soils, found that trace elements are distributed quite uniformly from horizon to horizon. The reason they give for this uniformity is that most of the trace elements are bound up in crystal lattices. The distribution of 2% per cent acetic acid-extractable trace elements was found to exhibit a

MICRONUTRIENTS IN SOILS

139

much greater variation from one horizon to another, reflecting many of the characteristics of the distribution in Fig. 1, especially for Co and Mn on well-drained soils. The data of Vinogradov (1959) and Wright et al. (1955) also show more uniformity with depth for nonpodzolic soils. Many authors have called attention to the increased uniformity in the micronutrient content of relatively arid soil profiles (Biswas, 1953; Leeper, 1947; Iyer and Satyanarayan, 1959). These results promote the idea that a high degree of profile development or an advanced stage of weathering is necessary before variations in micronutrient patterns are evident. It has been seen that soluble organic substances contribute to the downward translocation of the heavy metals, and presumably to Mo and B, as well. Immobilization then results from flocculation of stabilized hydrosols, if this mode of transport is significant, or by precipitation and adsorption of the metal from solution if present as a complex or the free ion. The immobilization is most commonly related to the clay content of the soil. Hoon and Dhawan (1943) showed that Mn followed clay concentration in Indian soil profiles. Randhawa et a2. (1961) confirmed the above observation for b t a l Mn, but found water-soluble, exchangeable, and easily reducible Mn was concentrated at the surface. Simonson et al. (1957) noted an association of Fe with clay in welldrained soils. Kubota (1963) has observed a close relationship between total Co, total Fe, and clay content within the profile of well-drained soils in the Eastern United States, but as with Mn, extractable forms of Co were concentrated at the soil surface. According to Hill et al. (1953), the clay fraction of a soil may have seven times as much Co as the sand. Whetstone et al. (1942) indicated that variations in B content of the profiles studied could be eliminated by expressing the B content in terms of concentration in the colloidal fraction. Like the other micronutrients, total B may be concentrated in either the surface or lower horizons, but water-soluble B is almost invariably concentrated in the surface of well-drained soils (Haas, 1944; Coleman, 1945; Burkser and Dopler, 1955; Yang, 1960; and others). With Co, there is some question whether the clay fraction or the iron oxides have the greatest direct significance. Fujimoto and Sherman (1951) and Burriel and Gallego (1952) have also called attention to the way Co follows Fe in the profile. Perhaps the best testimonial to the effectiveness of iron oxide surfaces in adsorbing Co lies in Kubota’s findings (1963) of a concentration of C o in Fe concretions that was almost as great as the concentration of Fe itself. But taken together with the findings of Hill et al. regarding the concentration of Co in clay and the observation of Kubota (1963) that Co moves slightly farther than

140

J. F. HODGSON

the iron in the profile, it appears that both iron and clay contribute to the concentration of CO. One of the most important factors affecting the mobilization and immobilization of micronutrients in soils is drainage. The manner and degree to which soil aeration alters the chemistry of Fe and Mn is well known. But other micronutrients are also influenced by drainage pattern. It has been seen that while the total amount of an element may be concentrated in the lower horizons, the exchangeable or acid-soluble form of the element in the same soil may be principally at the surface. This apparently holds true for well-drained soils, but for their poorly drained counterparts, even the extractable forms are concentrated in the lower horizons. This has been shown for Co, Cu, Zn, and Mo, as well as Fe and Mn (Sedletskii and Ivanov, 1941; Tiller, 1958; Swaine and Mitchell, 1960; Kubota, 1963). On the other hand, Butler (1954) observed the concentration of B, Cu, and Zn in the lower horizons of poorly drained soils, but found little relationship between this distribution and the degree of poor drainage. More will be said about the effects of poor drainage in the section on availability. Micronutrient distribution in soils must also be influenced by the presence of insoluble organic matter, by the amount of the element present and by the kind of clay mineral present (Tiller, 1958), but the relative importance of these effects in comparison with those already mentioned has not been determined. V. Factors Affecting the Availability of Micronutrients

The degree of availability of micronutrients in soils is a function of their partition among different forms-a partition that is influenced by many factors. The different forms are summarized for cations in the accompanying schematic diagram, where dotted lines indicate lack of proof for a given transformation (see opposite page). Each form is related directly or indirectly to the soil solution through some pseudo equilibrium distribution that is a function of the pH, oxidation potential, and supply and activity of individual soil constituents. Viets (1962) prefers to view groups of individual forms in terms of pools. Successive pools represent varying degrees of availability from ions in the soil solution to those remaining in primary minerals. Each pool, except for the most unavailable, corresponds to the forms of an element which are subject to removal by different types of extraotants. By this view, increases or decreases in availability are represented by shifts in the element from one pool to another. Organization in these terms is of limited value from a mechanistic point of view, but aids

MICRONUTRIENTS IN SOILS

141

substantially in visualizing the manner in which various factors contribute to the availability of micronutrients . to plsants and in testing the effect of variables by measuring the shift from one pool to another. Primary

mineral

I

weathering uptake

complex with insoluble organic matter

0

M**-MCh:

I trre

I ncorporotion

ionclompkx!

into microbial tissue oxidation reduction

MX -------FoccIusion surface in develo ing adsorption precipitoks

It

precipitotion of oxides or phosphates for Fe and Mn

\

a

solid state diffusion into soil minerols

A. pH The importance of pH to micronutrient availability can best be illustrated by McHargue's observation in 1923 that Mn added without lime was toxic to plants, but the same amount of Mn increased yields when added after liming, A change from toxicity to deficiency due to changes in pH is certainly an extreme example, but the availability of all the micronutrients being considered here is altered by liming or other pH changes. The relative effect of pH on the availability of micronutrients as determined by chemical extractions of the soil and by plant uptake is considered in Fig. 2. This representation is patterned after an earlier graph by Truog (1946). It should be emphasized that Fig. 2 represents a generalized relationship for widely varying soil conditions as taken from many different references. Many factors alter the dependency of availability on pH so that pH response does not always conform to that depicted. In addition to moderating the adsorption or precipitation of micronutrients in soils, pH may alter the plant uptake of an element through an efFect on micro-

142

J. F. HODGSON

bial activity, a change in the ability of the plant roots to absorb ions or to transport them to the tops once absorbed, variations in the stability of soluble and insoluble organic complexes, a change in the solubility of antagonistic ions, or an alteration of any rhizosphere effects that may be present. In addition, different plant species respond dissimilarly to

FIG.2. The effect of pH on the uptake of micronutrients by plants (outer limits of enclosed areas) and on micronutrients extractable by various solvents ( darkly shaded areas )

.

changes in pH and the influence of pH varies with the amount and forms of an element present in soils. The effect of pH on the availability of different ions is more varied than might be supposed from Fig. 2. Christensen et al. (1950) found that liming a soil from pH 4.6 to 6.5 decreased the exchangeable Mn 20 to 50 times, Fujimoto and Sherman (1948) observed nearly as great variation in exchangeable Mn upon the addition of 4 tons of lime to certain Hawaiian soils. These are extreme examples, but they serve to

MICRONUTRIENTS IN SOILS

143

illustrate the large effect of pH on the availability of Mn. In contrast, the uptake of Cu by plants is only slightly influenced by pH. In many cases, such as the studies of Piper and Beckwith (1949) and French d al. (1957), no differences in Cu uptake by plants could be observed with variation in soil reaction. Comparing these with Fig. 2 serves to emphasize the degree to which environmental conditions influence the effect of pH on heavy metal absorption. The differences between the way plant uptake of Co, Cu, and Zn are modified by pH are small by comparison. In the case of Cu, Zn, and Co,the amounts of the elements extracted with chemical solvents vary more with soil pH than the amounts removed by plants. This is presumably due to an increased efficiency in the process of plant uptake at higher pH values. Stewart and Leonard (1956) found that maximum uptake of ZnS04 added to soil was at pH 4, but the maximum uptake of applied Zn EDTA was at pH 7. The results could be accounted for by the predominance of Zn fixation at pH values above 4 where ZnS04 was added, but where EDTA was added, Zn was not fixed and there was more effective utilization by the plant at the higher pH values. Camp (1945) suggests that the relative insensitivity in plant uptake of Zn to changes in pH may be due to the complexing effect of soluble organic matter. This is likely true for other heavy metals as well. Wain d al. (1943) found that Mn added to calcareous soil was converted to completely unavailable forms within 7 days. The presence of organic matter allowed the plant to make use of the added Mn throughout the growing season. As with most such observations, those mentioned above are open to more than one interpretation. It should not be concluded that ion uptake and transport by plants are necessarily more efficient at higher soil pH values. The reverse may easily be true in many cases, as shown in studies of iron chlorosis by Ortega (1953), Brown (lWO), and many others. Brown (1961) discusses this aspect fully in his review of the iron chlorosis problem. The amount and form of an element in the soil are important factors in the response of that element to changes in soil reaction. Beeson et al. (1948) observed that CaCO, did not alter the uptake of natural Co from soil but greatly reduced the uptake of added Co. A similar pattern was observed for Cu and Mn, but for these elements the uptake of naturally occurring forms was also somewhat reduced. It appears that when the elements are present in larger amounts or are added to the soil in more readily available forms, they respond to changes in pH to a greater degree than indicated in Fig. 2. The observation of Leyden and Toth (1960) that a rise in pH increased the proportions of Zn utilized from native sources compared with fertilizer sources, and a

144

J. F. HODGSON

report by Delas et a2. ( 1 W ) that increasing pH reduced Cu toxicity, contributes to this conclusion. The degree to which pH may influence the uptake of applied sources of microelements can be seen in the results of Wear (1956), who found that 92 per cent of the variation in Zn uptake from applied fertilizers could be accounted for by changes in pH. Molybdenum is the only principal micronutrient element that increases in availability with increasing pH. Its response to pH changes is second only to Mn. Agricultural significance ranges from deficiencies for plants which are essentially restricted to acid soils, to toxicities for animals, a problem sometimes encountered on neutral or alkaline soils.

B. ORGANIC MATTER As pointed out in the previous section, the presence of organic matter

may promote the availability of certain elements, presumably by supplying soluble complexing agents that interfere with their fixation. At the same time, the soils that are most commonly deficient in certain heavy metals, notably Cu, and fix the greatest quantities of these elements, are organic in nature. The relative influence of soluble and insoluble organic compounds and their relation to inorganic soil constituents in the average mineral soil is not at all clearly defined. There are certainly differences in the ability of organic and mineral soils to combine with heavy metals (Lundblad et al., 1949; Schlichting and Wiklander, 1956; Steenbjerg and Boken, 1948; Henriksen, 1957), but alleged variations in Cu requirement are partly due to the way the results are reported. For example, Stenberg d al. (1948) showed that organic soils in Sweden require 20 to 25 p.p.m. Cu, whereas mineral soils need only 8 to 10 p.p.m. But the authors point out that on an area basis, the same mineral soils have a slightly higher requirement of 20 pounds Cu per acre compared to 16 for the organic soils. The effect of organic matter on Mn in soil is particularly pronounced. Higher oxides of Mn do not commonly occur in organic soils, as pointed out by Heintze (1957). Boischot and Durroux (1950) took the position that Mn defkiencies in organic soils result from biochemical action rather than precipitation. DeGroot ( 1956) claims that such deficiencies commonly disappear in marine and estuary soils as the C:N ratio decreases. As might be expected from this, the addition of organic matter to mineral soils has been found to increase the exchangeable Mn (Christensen et al., 1950) and electrodialyzable Mn (Prince and Toth, 1938). Reducing agents have been shown to give the same results (Fujimoto and Sherman, 1948). Sanchez and Kamprath (1959) also found that peat added to acid soil increased exchangeable Mn, but they further

MICRONUTRIENTS IN SOILS

145

noted that it decreased acid-extractable Mn when the soil was limed. The effect of organic matter on Mn transformations in soils must take at least three forms: (1) the production of complexing agents that effectively reduce the activity of the free ion in solution, ( 2 ) a decrease in the oxidation potential of the soil, either directly or indirectly through increased microbial activity, and (3) a stimulation in microbial activity that results in incorporation of Mn in biological tissue. Iron responds in much the same way as Mn to the presence of organic matter in soils, although the effect is not quite as pronounced. The way that Mn responds to induced changes in oxidation potential provides a contrasting behavior to heavy metal micronutrients other than Fe. This is most evident from results of Barbier and Trocm6 (1950), who found that slewage effluent used in irrigation brought about a striking depletion in the Mn content of the soil while producing a concomitant accumulation of Zn. Addition of organic supplements to soils, while commonly increasing the extractable form of an element, may not necessarily increase its availability to plants. Miller and Ohlrogge ( 1958a,b) have observed that water extracts of manure and other organic residues solubilized Zn in soil, but at the same time reduced its uptake by plants. They also found a depression of Zn and Fe uptake from nutrient solution and Cu uptake from soil resulting from additions of manure extracts. Plant uptake of Mn was increased, however. Atkinson et al. (1958) found that manure additions to a soil increased water-soluble B while reducing the uptake of B and Mn by clover. Exchangeable and easily reducible Mn were not altered. The uptake of Cu and Zn was unaltered and the uptake of Mo increased in these experiments. Unfortunately, the effects of complex formation, microbiological stimulation, and changes in oxidation potential were further complicated by increases in pH in the last studies reported. Gulyakin and Yudintseva (1960) observed that organic matter decreases the uptake of Co by wheat, peas, and oats. Many soluble complexes, both natural and artificial are capable of being adsorbed by the plant, so that soluble organic matter commonly increases plant uptake of heavy metals as well as B (Berger and Truog, 1945), and Mo (Davies, 1956). There are many reviews on this subject, the most recent and comprehensive being that of Wallace (1963).

C. MICROBIOLOGICAL Alexander (1962) gives six ways that microorganisms may affect the availability of nutrient elements in the soil. The five of these that apply to the micronutrient elements are summarized as follows:

146

J. F. HODGSON

1. The release of inorganic ions during the decomposition of organic materials. 2. Immobilization of ions by incorporation into microbial tissue. 3. Oxidation of an element, generally to a less available form. 4. Reduction of an oxidized form of an element under conditions where oxygen is limited. 5. Indirect transformations; changes in pH or oxidation potential. The sixth process suggested by Alexander involves changing the amount of an element present as in the fixation of N. The first method might be broadened to include the liberation of nutrient elements as soluble organic complexes as well as the free inorganic ion. By far the most widely studied and probably the most important microbiological effects on the availability of micronutrients involve the oxidation and reduction of Fe and Mn. There are in fact indications that microorganisms control the oxidation state of at least Mn and that changes in oxidation potential and pH have their effect only through microbiological activity. Cell poisons such as toluene (Pronin, 1933) and azide (Gleen, 1950) inhibit oxidation of Fe++ and Mn++ (TrocmB and Barbier, 1950). Soil sterilization is known to increase available Mn (Timonin and Giles, 1952; Martin, 1953; Forsee, 1954) and increase Fe+ + in soils ( Ignatieff, 1941 ) . Beijerinck (1913) first observed that soil organisms could oxidize Mn. When soil was added to an agar medium containing MnC03, he found that concretions of MnOz were produced. Beijerinck and later Bromfield and Skennan (1950) identified a number of soil bacteria and fungi that are effective in oxidizing Mn+ +, Of particular interest is the role of bacteria in gray speck disease of oats. This disease is commonly interpreted as resulting from a deficiency of Mn. Samuel and Piper (1929) showed that if oats were supplied with less than 14 p.p.m. Mn, gray speck symptoms developed. But later, Gerretsen (1935) found that if the system was sterile, the Mn level could drop to 5 p.p.m. without Mn deficiency being observed. If the solutions were inoculated with fresh soil, gray speck appeared, but if the soil was sterilized first, it did not. The general conclusion from this and other studies (Vlasyuk and Butkevich, 1957), is that organisms attracted to the rhizosphere oxidized and precipitated Mn as it approached the area of the root. Timonin (1946) further showed that oat varieties particularly susceptible to gray speck had a higher proportion of Mn-oxidizing bacteria in their rhizosphere than did nonsusceptible varieties. There is greater difficulty in demonstrating the role of microorganisms

MICRONUTRIENTS IN SOILS

147

in oxidizing iron since spontaneous oxidation of Fe++ in an organism free system more readily occurs. Gleen (1950) was able to show that a FeS04 solution, stabilized at pH 3, was oxidized when passed through a soil column. The oxidation process followed the typical sigmoid-shaped curve characteristic of biological processes and was poisoned by azide. Therefore, it is established that biological oxidation can occur even under conditions that do not favor chemical oxidation of the element, but certainly in most soils biological activity is not necessary to account for the oxidized state of the iron. The development of the ferruginous and manganiferous nodules, on the other hand, cannot usually be fully accounted for by nonbiological processes. In at least one study, biological processes were implicated in the oxidation of iron in nodules (Gerei et al., 1960). The microbial decomposition of organic complexing agents that serve to stabilize reduced forms of Fe and Mn undoubtedly provides indirect means of promoting oxidation of these elements. Such decomposition can also serve to convert other elements to less available forms. Oxidation of Fe and Mn in soils is, of course, a reversible process. As drainage becomes impeded and the oxidation potential ( E h ) approaches 0.2 volts, oxides of Fe+++ and M n + + + + can be reduced. The reduction of these elements is an energy-requiring conversion so that the importance of microbiological activity in these transformations can be demonstrated by the addition of easily decomposible organic matter (Rao, 1956; Hochster and Quastel, 1952). The effect of microorganisms on other micronutrients is principally through decomposition of organic bound forms or through direct competition for nutrients. Zinc deficiency in fruit trees (little-leaf disease) is thought to be aggravated by certain microorganisms in much the same way as Mn deficiency in oats. The disease can be eliminated by sterilization or redeveloped by inoculating with a small amount of diseased soil. Yet the disease is readily cured by the addition of Zn. Whether the microorganisms produce this effect through a competition for Zn or through the production of nonabsorbable Zn complexes, or whether there is a pathological interaction, is not well established. Sterilization of a soil has many undefined effects, among them the increase in availability of micronutrients. In addition to the effects on Mn and Fe, some of which have been discussed, sterilization is reported to increase available Cu (Mulder, 1940; Millikan, 1942; Steenbjerg, 1940; Dalton and Hunvitz, 1948) and Zn (Millikan, 1942) and to mobilize Co and B (Bronsart, 1947). These results may be explained in part by Allison’s observation ( 1951 ) that ethylene oxide sterilization increased water-soluble organic matter. Possibly enzymes liberated during the

148

J. F. HODGSON

sterilization process continue to degrade relatively complex organic compounds forming, among other things, soluble complexing agents. Allison also noted an increase in pH on sterilization. For further discussion of microbiological activity relating to micronutrient transformations and the microorganisms induced, see Alexander ( 1962).

D. OXIDATION AND REDUCTION The effect that the oxidation potential, Eh, has on the availability of Fe and Mn was discussed in the previous section as it is related to microbial activity. The important role that microorganisms play in the oxidation conversion of these two elements should not obscure the close relationship between the resultant state of oxidation of the element and the Eh of the soil. Yarilova in 1940 found that the change in Eh of a soil per unit pH was very close to the theoretical value based on the reaction. MnOz 4H+ 2e+ Mn++ 2 H 2 0

+

+

+

How general this close relationship may be is not well defined. Hemstock and Low (1953) used the half-cell oxidation potentials to calculate the equilibrium constant for this reaction. Their results would give extremely low values for Mn++ in solution. The error term in half-cell measurements when converted to the above equilibrium constant gives considerable room for variations, however. Ponnamperuma ( 1955) has provided a basis for quantitatively evaluating the effect of oxidation potential on Fe oxidation and has extensively reviewed the work relating these factors. In addition to the aforementioned relationships, changes in drainage pattern influence ions that are not thought to go through changes in oxidation states, as well as those that do. Hill d aZ. (1953) observed that Co was higher in plants grown on poorly drained soil than those on well-drained soil. Mitchell (1955) and Swaine and Mitchell (1960) found that acetic acid-extractable Co was much higher for poorly drained horizons of a soil. The difference in Co uptake between well drained and poorly drained soils was further emphasized by Alban and Kubota (1960), who found a distinctly different correlation between Co uptake by plants and Co extracted by an acid dithizone system for the two drainage classes. In a recent greenhouse experiment, Kubota et al. (1963) showed that high moisture content can indeed increase Co in the soil solution and to a lesser extent in the plant. These authors also showed that Mo was increased in the soil solution, but Cu was essentially unaffected by the high moisture treatment. There is, however, some

MICRONUTRIENTS IN SOILS

149

indication that Cu may be affected in a manner similar to Co under some conditions. Rogers d al. (1939) found that Cu was higher in soils with poor drainage, and only those soils gave a residual effect from Cu fertilization. The influence of drainage on the distribution of these ions in soils has already been discussed. The reasons for the above observations are not at all clear. It is very unlikely that Co++ + with an oxidation potential of -1.84 exists in soils except as the extremely strong and water-soluble complex, vitamin BIZ. Amin and Joham (1958) have suggested that Mo may undergo an oxidation and reduction cycle in the soil, but have offered little evidence in support of this contention. Even if such a cycle existed, the reduced form of Mo should be less soluble and would give rise to results just the opposite of those observed by Kubota et al. ( 1 x 3 ) . A better explanation of the drainage effect on micronutrient solubility may be found in the influence of anaerobic conditions on microbial decomposition of organic matter. Kee and Bloomfield (1961, 1962) have shown that under reducing conditions, soluble organic products resulting from the decomposition of plant material were much more effective in solubilizing heavy metals than corresponding products found under aerobic conditions. A more detailed understanding of these relationships will be necessary before the unusually large effects of drainage on Co and Mo availability can be explained. E. SEASONAL VARIATION The availability of many elements in soils varies considerably from one part of the year to another. The majority of information on this subject comes from seasonal fluctuations of plant composition. Unfortunately from the standpoint of interpretation, changes in weather patterns bring about many differences in plant composition besides those related to the availability of the nutrients in soils. Light, stage of growth, rate of growth and many other factors contribute to the relative uptake of nutrients and production of dry matter. There is considerable divergence of conclusions as to the effect of season on soil availability of micronutrients. This is not surprising when it is realized that seasonal changes are accompanied by variations in microbial activity, moisture, and temperature, and that the effect of these factors is moderated by a host of other considerations, from organic matter supply to the evapotranspiration rate. It is therefore possible only to point out certain trends in seasonal and weather fluctuations. Manganese exhibits the most pronounced seasonal variation in availability, probably due to microbially induced oxidation and reduction.

150

J. F. HODGSON

McCool (1934) first noted that water-soluble Mn was high during the summer months. Dorph-Petersen (1950) found 5 to 10 times as much exchangeable Mn in summer as compared to winter. Nozdrunova et al. (1958) observed the same pattern in Russian soils. Sherman and Harmer (1942), on the other hand, claim that winter favors manganous and summer favors manganic forms of the element. This conclusion was based on work with more alkaline soils than the earlier reports. It is possible that a rising winter water table results in higher availability of Mn in this season, whereas, in the absence of a high water table, Mn availability would be highest in summer. Some authors emphasize short-term fluctuations in availability over seasonal ones. DeLong et al. (1940) could find no seasonal trends, but they noted that the Mn extracted with 0.2 N acetic acid increased following each rainfall. Kosegarten (1956) went a step further, observing that exchangeable Mn increased following rainy periods and easily reducible Mn underwent a corresponding decrease. The increased availability in the summer is commonly reflected in plant response. Chlorosis from Mn toxicity has been associated with summer months (Fujimoto and Sherman, 1946). High Mn uptake was associated with high soil temperature and low moisture by Mederski and Wilson (1955). But Wain et al. (1939) point out that Mn in hay may decrease in warm weather due to rapid growth of the plant. Other elements are also affected by seasonal fluctuation. Askew and Maunsell (1937) report that Co content of forage is high in spring, decreasing as the season advances. McNaught and Paul (1939) showed that fall and winter growth may be even higher in Co than spring growth. Cu and Zn are reported to be high during rainy cool weather ( Bambergs, 1958). This is consistent with Millikan’s findings ( 1946) that flax seeded in June suffered most from Zn deficiency. Mo is said to be high in autumn (Ferguson et al., 1943). Seiffert and Wehrmann (1957) also report a fall high for Cu in forages. Warm weather uptake of Cu can be traced to an increase in soluble Cu in summer (Sarata, 1938), probably due to microbial decomposition of insoluble organic matter. This activity is acting in opposition to the effect of drying, which is reported to decrease available Cu (Steenbjerg, 1940). Variations in B uptake by plants are more striking from year to year or within a season than from season to season. Outbreaks of B deficiency are commonly associated with dry years, but there are reports that drying soils may increase B availability (Malquori et al., 1952), and that B deficiency may be worse in wet seasons ( White-Stevens, 1942). The last observation is probably due to excessive leaching of B on sandy soil.

MICRONUTRIENTS IN SOILS

151

F. RHIZOSPHERE Variations in nutrient content of different plant species growing in the same media are commonly associated with the absorption and transport system of the particular plant. There are indications, however, that plants may alter their nutrient uptake in still another way. Plant roots are known to exude a great variety of compounds in quantities sufficient to alter markedly the availability of nutrients in their environment (Rovira, 1962). The effect of these compounds on microbial activity is very marked and has been extensively investigated over a period of 50 years, but the effect they may have on the chemical properties of the rhizosphere has received much less attention. Root exudates may alter the chemical environment of the root either indirectly through their influence on microorganisms or through a direct interaction with elemental soil constituents. The indication that rhizosphere bacteria reduce the availability of Mn to oats has already been mentioned. Timonin’s observation ( 1946) that oats susceptible to Mn deficiency had a particularly high density of Mn-oxidizing bacteria around their roots would suggest that the root exudate pattern of the different varieties of oats, by influencing the rhizosphere organisms, can ultimately lead to a Mn deficiency in one species, while a second remains healthy. The oxidizing properties of some rhizosphere organisms can sometimes be beneficial. Vlasyuk and Butkevich (1957) were able to reduce Mn toxicity by inoculating with rhizosphere bacteria. Beckwith suggested in 1953 that organic substances given off by the root might form complex ions with Mn++ and Mn+++, but it was not until 1958 that Bromfield reported that root washings were capable of dissolving Mn02. In two papers ( 1958a, b ) , Bromfield demonstrated the effect of root exudates on insoluble Mn02 and was able to separate chromatographically the complexes from uncomplexed Mn+ +. Bromfield was unable to find any complexes formed when the exudate was added to Mn++. Possibly, weaker complexes were formed with Mn++ which dissociated in the acid chromatographic solvent he used, but stronger complexes of higher valence Mn were responsible for the dissolution of the solid Mn02. The nature of the root exudation may be influenced by the nutrient supply. Brown et al. (1961) found that strongly fluorescent compounds were exuded by soybeans when the iron supply was low, which did not appear when the Fe supply was sufficient. These compounds could be found in the exudate of a variety resistant to iron chlorosis, but not in the exudate of a susceptible variety.

152

J. F. HODGSON

These authors emphasize the reductive capacity of the root. They noted that soybean varieties resistant to Fe deficiency were capable of reducing Fe+ + + in solution. A susceptible variety on the other hand was not able to induce this transformation. Roots are reported to promote the dissolution of macronutrients, as well. A more general discussion of how the root may alter its chemical environment is given by Reuszer (1962). VI. Needs for Future Research

A persistent demand can be found for increased use of morphological and chemical concepts in developing tests for availability. Several successful attempts have been made in the past few years to improve the predictability of soil extractants in assessing the supply of an element for plants, by including in the assessment certain soil characteristics that affect the supply. In this way Nelson et al. (1959) were able to predict the occurrence of Zn deficiencies from the combined results of 0.1 N HC1-extractable Zn and titratable alkalinity. Massey ( 1957) improved the correlation of dithizone-extractable Zn with plant uptake by considering the effect of soil pH. Walker and Barber (1960) included consideration of organic matter content with chelate-extractable Mn and thereby increased the ability to predict the supply of available Mn. Alban and Kubota (1960) improved the correlation of Co extraction with plant uptake by dividing the soils on the basis of drainage. These examples will likely be followed by greater reliance on the factors altering soil availability in methods for assessing the supply of micronutrients to plants. For intelligent application of many of these concepts, the soil chemist must develop a better understanding of their significance under field conditions as well as in the laboratory: 1. The significance of the nature of mineral surfaces in their reactivity with micronutrients is an example. Different minerals selected from geologic deposits are found to vary considerably in their reactivity with micronutrients in the laboratory, but these differences may be much less pronounced when the minerals are taken from soils (Tiller ct al., 1963). The significance of these differences under field conditions is less well understood. 2. The reactivity of organic matter and mineral surfaces has been demonstrated many times, but the relative significance of organic and inorganic binding sites in the retention of micronutrients in mineral soils remains to be investigated under either laboratory or field conditions.

MICRONUTRIENTS IN SOILS

153

Many micronutrient amendments are rendered completely unavailable to plants after a year or more in the soil, but the long-term equilibrium reactions that are involved are only poorly defined. Understanding the rate of solid state diffusion and conditions promoting the occlusion of micronutrients in secondary minerals are of particular interest in this connection. Of special importance in the writer’s view is a need for an accurate estimation of the nature and amounts of micronutrients in the soil solution. A determination of the total amount of any given micronutrient should not be difficult, but it is likely that even in mineral soils most micronutrients in solution are associated with a soluble organic fraction. The nature of the micronutrient in soil solution is as vital as is the amount in interpreting most laboratory experiments in terms of field conditions. In addition to characterizing the soil mass in a generalized way, the writer feels that greater attention should be given to localized differences within the soil. Steenbjerg and Boken (1950) observed marked variation in available Cu within distances of 20 cm., but variations probably occur on an even smaller scale. Of particular interest are differences thought to be introduced through the action of plant roots. If the root alters its chemical environment to the extent suggested by the recent results discussed, then our knowledge of the soil chemistry of micronutrients will be far from complete until these rhizosphere effects are resolved. Attention in this review has been restricted to the more intensively studied micronutrients, but serious consequences can result from deficiencies and toxicities of other elements commonly taken up by plants and animals in small amounts. The recent discovery that trace amounts of Se and V will correct certain animal problems introduces a need for a better understanding of the transformations of these elements in soils. Se has already been studied extensively as it occurs in toxic amounts. The chemistry of iodine in soils is poorly understood even though the element has been known for over a century to be required for animals. Questions involving Si and A1 transformations are among the most fundamental to our knowledge of soil genesis. Chlorine is of less concern, since so far as the writer is aware, no indications have been reported of C1 deficiencies of plants in the field. Animals must receive supplements to maintain adequate C1 levels in most cases anyway. Perhaps more important than any of the preceding problems is the impending need to express variations in plant uptake, availability, and soil distribution of micronutrients in terms of measurable parameters. A qualitative understanding of the &ect of soil characteristics on micro-

154

J. F. HODGSON

nutrient behavior is emerging, The foregoing discussion attempts to highlight areas the author feels need further attention, but to achieve the efficiency of production required by an ever increasingly intensive agriculture, further steps are necessary. Attempts must be made to describe variables in measurable terms and relate them more quantitatively. Soil systems will have to be defined more precisely, especially in carrying laboratory results to field conditions. Soil scientists are obviously a long way from describing in quantitative terms such factors as the role or organic matter in the soil chemistry of any of the micronutrients, but as such goals are accepted the means of achieving them will sureIy follow. Only through these means will soil chemistry of the micronutrients achieve the degree of predictability the future may be expected to demand. ACKNOWLEDGMENT The writer wishes to thank Mrs. Ann O’Brien for assistance with the literature review. REFERENCES Afrikyan, E. K., and Bobikyan, R. A. 1959. Dokl. Akad. Nauk A m . S S R 29 (2), 89-92. Agerberg, L. S. 1959. Kgl. Skogs Lantbruksakad. Tidskr. 98, 343-351. Alban, L. A., and Kubota, J. 1960. Soil Sci. SOC. Am. Proc. 24, 183-185. Alexander, M. 1962. “Introduction to Microbiology.” Wiley, New York. Allison, L. E. 1951. Soil Sci. 72, 341-351. Amin, J. V., and Joham, H. E. 1958. Soil Sci. 85, 156-160. Askew, H. O., and Maunsell, P. W. 1937. New Zealand J . Sci. Technol. 19, 337342. Atkinson, H. J., Giles, G. R., and MacLean, A. J. 1953. 3. Sci. Agr. 33, 116124. Atkinson, H. J., Giles, C. R., and Desjardins, J. G. 1958. Plant Soil 10, 32-36. Bambergs, K. 1958. Latuiias PSR Zinatnu Akad. Vestis pp. 59-64. Banejee, D.K., Bray, R. H., and Melsted, S. W. 1953. Soil Sci. 75, 421-431. Barbier, G., and Chabannes, J. 1953. Ann. Agron. 4, 27-43. Barbier, G., and TrocmB, S. 1950. Compt. Rend. Acad. Agr. France 36, 247-252. Baughman, N. M. 1956. Ph.D. Thesis, Purdue Univ., Lafayette, Indiana. Beckwith, R. S. 1953. Australiun Conf. Soil Sci. Adelaide 1, 2.17, 6. Beeson, K. C. 1948. Fertilizer Reu. 23, 11-14. Beeson, K. C. 1959. “Mineral Nutrition of Trees,” Duke Uniu., School Df Forestry, BUZZ. 15, 71-80. Beeson, K. C., Gray, L., and Hemner, K. C. 1948. J. Am. SOC. Agron. 40, 553-562. Beijerinck, M. W. 1913. Ve~520gAkad. Wetenschappen 22, 415-420. Berger, K. C. 1962. 3. Agr. Food C h m . 10, 178-181. Berger, K. C.,and Truog, E. 1945. Soil Sci. SOC. Am. Proc. 10, 113-110. Biggar, J. W., and Fireman, M. 1960. Soil Sci. SOC. Am. Proc. 24, 115-120. Biswas, T. D. 1953. J. Zndian SOC. Soil Sci. 1, 21-31. Bjerrum, J., Schwarzenbach, G., and Sillen, L. G. 1958. Chem. SOC. ( L o d o n ) Spec. Publ. 7, 1-131. Bloomfield, C. 1955a. J . Sci. Food Agr. 6, 641-651.

MIQ7OhTJTRLENTS IN SOILS

155

Bloomfield, C. 195%. Chern. Ind. (London) pp. 1596-1597. Bloomfield, C. 1958. Chern. Ind. (London) pp. 259-260. Bobko, E. V., Matveeva, T. V., Dubachova, T. D., and Filippov, A. I. 1936. Ann. Agron. 6, 691-701. Boischot, P., and Durroux, M. 1950. Ann. Agron. I, 551-554. Bremner, J. M., Mann, P. J. C., Heintze, S. G., and Lees, H. 1946. Nature 158, 790-791. Broadbent, F. E. 1957. Soil Sci. 84, 127-131. Broadbent, F. E., and Bradford, C. R. 1952. Soil Sci. 74, 447-457. Bromfield, S. M. 1958a. Plant Soil 9, 325-337. Bromfield, S. M. 1958b. Plant Soil 10, 147-159. Bromfield, S . M., and Skerman, V. B. D. 1950. Soil Sci. 69, 337-348. Bronsart, H. v. 1947. Nachrbl. Deut. Pflanzenschutzdienst (Berlin) 1, 115-116. Brown, A. L. 1950. Soil Sci. 69, 349-358. Brown, C. 1953. J. Soil Sci. 4, 220-225. Brown, J. C. 1960. Soil Sci. 89, 246-247. Brown, J. C. 1961. Aduan. Agron. 13, 329-369. Brown, J. C., Holmes, R. S., and Tiffin, L. 0. 1961. Soil Sci. 91, 127-132. Burkser, E. S., and Dopler, T. F. 1955. Dopooidi Akad. Nauk Ukr. RSR pp. 474-477. Burriel, F., and Callego, R. 1952. Anales Edafol. Fisiol. Vegetal (Madrid) 11, 569-600. Butler, J. R. 1954. 1. Soil Sci. 5, 156-166. Cailkre, S., HCnin, S., and Esquevin, J. 1958. Clay Minerals Bull. 3, 232-237. Camp, A. F. 1945. Soil Sci. 60, 157-164. Carr, M. H., and Turekian, K. K. 1961. Geochim. Cosmochim. Acta 23, 9-60. Clapp, C. E., Jr. 1957. Ph.D. Thesis, Cornell Univ., Ithaca, New York. Christensen, P. D., Toth, S. J., and Bear, F. E. 1950. Soil Sci. SOC. Am. Proc. 15, 279-282. Coleman, N. T., McClung, A. C., and Moore, David P. 1956. Science 123, 330331. Coleman, D. 1945. Better Crops Plant Food 29 ( 6 ) , 11-13, 42-43. Dalton, F. H., and Hurwitz, C. 1948. Soil Sci. 66, 233-238. Davies, E. B. 1956. Soil Sci. 81, 209-221. Dawson, J. E., and Nair, C. K. N. 1950. In “Copper Metabolism” (W. 0. McElroy and B. Class, eds.), pp. 315-335. Johns Hopkins Press, Baltimore, Maryland. deGroot, A. J. 1956. Trans. Intern. Congr. Soil Sci. 6th Congr. Paris, pp. 531-539. Delas, J., Delmas, J., and Demias, C. 1960. Compt. Rend. 250, 3867-3869. DeLong, W. A., Sutherland, A. J., and Salisbury, H. F. 1940. Sci. Agr. 21, 89-91. DeMumbrum, L. E., Jr., and Jackson, M. L. 1956a. Soil Sci. 81, 353-357. DeMumbrum, L. E., Jr., and Jackson, M. L. 1956b. Soil Sci. SOC. Am. Proc. 20, 334-337. DeVore, G. W. 1955. 1. Geol. 63, 159-190. Diehl, H., and Voigt, A. 1958. Iowa State Coll. J. Sci. 32, 471-473. Dion, H. G., Mann, P. J. G., and Heintze, S. G . 1947. J . Agr. Sci. 34, 17-22. Dobritskaya, Y. I. 1961. Vestn. Sel’skokhoz. Nauki Vses. Akad. Sel‘skokhoz. Nauk 6 ( 6 ) , 109-116. Dorph-Petersen, K. 1950. Tidsskr. Planteaol 53, 650-677.

156

J. F. HODGSON

Dragsdorf, R. D., Kissinger, H. E., and Perkins, A. T. 1951. Soil Sci. 71, 439-448. Duda, J., Maliliska, E., and Pedziwilk, Z. 1957. Acta Microbiol. Polon. 6, 355-365. Eaton, F. M., and Wilcox, L. V. 1939. U.S. Dept. Agr. Tech. Bull. 696. Elgabaly, M. M. 1950. Soil Sci. 69, 167-174. Elgabaly, M. M., and Jenny, H. 1943. J. Phys. Chem. 47, 399-408. Evans, L. T. 1959. J . Soil Sci. 10, 110-118. Ferguson, W. S., Lewis, A. H., and Watson, S. J. 1943. J . Agr. Sci. 33, 44-51. Forsee, W. T., Jr. 1954. Soil Sci. SOC. Am. Proc. 18, 475-478. French, C. E., Smith, C. B., Fortmann, H. R., Pennington, R. P., Taylor, G. A., Hinish, W. W., and Swift, R. W. 1957. Penn. State Univ. Agr. Expt. Sta. Bull. 624, Fujimoto, C. K., and Sherman, D. G. 1946. Hawaii Agr. Expt. Sta. Rept. pp. 54-55. Fujimoto, C. K., and Sherman, D. G. 1948. Soil Sci. 66, 131-145. Fujimoto, G., and Sherman, D. G. 1951. Agron. J . 43, 424-429. Gerei, L., MBtk, F., and Benedek, J. 1960. Agrokem. Talujtan 9, 491-494. Gerretsen, F. C. 1935. Trans. 3rd Intern. Congr. Soil Sci. Oxford A35 1, 189-191. Gibbs, 0. E., and Marshall, C. E. 1952. Missouri Agr. Expt. Sta. Res. Bull. 518. Gilbert, F. A. 1957. “Mineral Nutrition and the Balance of Life,” Univ. Oklahoma Press, Norman, Oklahoma. Gleen, H. 1950. Nature 166, 871-872. Goldschmidt, V. M. 1937. 1. Chem. SOC. pp. 655-673. Goldschmidt, V. M. 1954. “Geochemistry” (A. Muir, ed.), pp. 1-730. Oxford Univ. Press (Clarendon), London and New York. Gray, L. F., and Daniel, L. J. 1959. J. Nutr. 67, 623-634. Gulyakin, I. V., and Yudintseva, E. V. 1960. Zzv. Timiryazev. Sel‘skokhoz. Akad. 5, 114-122. Haas, A. R. C. 1944. Soil Sci. 58, 123-137. Harder, H. 1961. Geochim. Cosmochim. Acta 21, 284-294. Hatcher, J. T., and Bower, C. A. 1958. Soil Sci. 85, 319-323. Heintze, S. G. 1957. J . Soil Sci. 8, 287-300. Hemstock, G. A,, and Low, P. F. 1953. Soil Sci. 76, 331-343. Henriksen, A. 1957. Tidsskr. Plunteavl 61, 685-717. Heydemann, A. 1959. Geochim. Cosmochim. Actu 15, 305-329. Hibbard, P. L. 1940. Soil Sci. 49, 63-72. Hill, A. C., Toth, S. J., and Bear, F. E. 1953. Soil Sci. 76, 273-283. Himes, F. L., and Barber, S. A. 1957. Soil Sci. SOC. Am. Proc. 21, 368-373. Hochster, R. M., and Quastel, J. H. 1952. Arch. Biochem. Biophys. 36, 132-146. Hodgson, J. F. 1960. Soil Sci. SOC. Am. Proc. 24, 165-168. Hodgson, J. F., and Tiller, K. G. 1982. “Clays and Clay Minerals,” Proc. 9th Natl. Conf. Clays Clay Minerals, 1960, pp. 404-411. Pergamon Press, New York. Hodgson, J. F., Tiller, K. G., and Fellows, M. 1963. Soil Sci. SOC. Am. Proc. (in press). Hoon, R. C., and Dhawan, C. L. 1943. Zndiun J . Agr. Sci. 13, 601-608. Ignatieff, V. 1941. Soil Sci. 51, 249-263. Isbell, H. S., Brewster, J. F., Holt, N. B., and Frush, H. L. 1948. J . Res. Natl. Bur. Std. 40, 129-149. Ishibashi, M., Fujinaga, T., and Kuwamoto, T. 1958. Nippon Kagaktl Zasshi 79, 1496-1499. Iyer, J. G., and Satyanarayan, Y. 1959. J . Biol. Sci. 2, 110-115.

MICRONUTRIENTS IN SOILS

157

Jamison, V. C. 1943. Citrus I d . 24 ( l o ) , 3-5, 12-13. Jensen, H. L., and Lamm, C. G. 1961. Acta Agr. S c a d . 11, 63-81. Jones, G. B., Riceman, D. S., and McKenzie, J. 0. 1957. Australian J. Agr. Res. 8, 190-201. Jones, H. W., Gall, 0. E., and Barnette, R. M. 1936. Florida Agr. Expt. Sta. Bull. 298. Jones, L. H. P. 1956. Science 123, 1116. Jones, L. H. P. 1957. J. Soil Sci. 8, 313-327. Jones, L. H. P., and Leeper, G. W. 1951. Plant Soil 3, 141-153. Jordan, J. V., and Powers, W. L. 1946. Soil Sci. Soc. Am. Proc. 11, 324-331. Kabata, A. 1955. Roczniki Nauk Rolniczych Ser. A. 70, 609-615. Katalymov, M. V. 1951. Dokl. Akad. Nauk SSSR 80, 413-416. Kee, N. S., and Bloomfield, C. 1961. Geochim. Cosmochim. Actu 24, 206-225. Kee, N. S., and Bloomfield, C. 1962. Plant Soil 16, 108-135. Kelley, W. P., and Jenny, M. 1936. Soil Sci. 41, 367-382. Knott, J. E. 1938. Ohio Vegetabb Growers’ Assoc. Proc. 23, 10-16. Kosegarten, E. 1956. 2. Pflanzenernaehr. Dueng. Bodenk. 73, 25-39. Krauskopf, K. B. 1956. Econ. Geol. 50, 411-463. Kubota, J. 1963. Soil Sci. (in press). Kubota, J,, Berger, K. C., and Truog, E. 1948. Soil Sci. SOC. Am. Proc. 13, 130134. Kubota, J., Lazar, V. A,, Langan, L. N., and Beeson, K. C. 1961. Soil Sci. SOC. Am. Proc. 25, 227-232. Kubota, J., Lemon, E. R., and Allaway, W. H. 1963. Soil Sci. SOC. Am. Proc. (in press). Lal, K. N., and Rao, M. S. S. 1954. “Micro-element Nutrition of Plants.” Banaras Hindu Univ. Press, Banaras, India. Leeper, G. W. 1935. Proc. Roy. SOC. Victoria 47, 225-261. Leeper, G. W. 1947. Soil Sci. 83, 79-94. Leyden, R. F., and Toth, S. J. 1960. Soil Sci. 89, 223-228. Lundblad, K.,Svanberg, O., and Ekman, P. 1949. Plant Soil 1, 277-302. McCool, M. M. 1934. Contrib. Boyce Thompson Znst. 6, 147-164. McHargue, J. S. 1923. J. Agr. Res. 24, 781-793. McNaught, K. J., and Paul, G. W. 1939. New Zeuhnd J. Sci. Technol. 21B, 95- 101. Malquori, A., Stradaioli, G., and Perici, E. 1952. Ann. Sper. Agrar. (Rome) 6, 571-589. Malyuga, D. P. 1960. Tr. Biogeokhim. Lab. Akad. Nauk SSSR 11, 197-207. Mann, P. J. G., and Quastel, J. H. 1946. Nature 158, 154-156. Mansknya, S. M., Drozdova, T. V., and Emel’yanova, M. P. 1960. Tr. Biogeokhim. Lob. Akad. Nauk S S S R 11, 65-69. Martin, A. E., and Reeve, R. 1957. J . Soil Sci. 8, 268-278. Martin, J. P. 1953. Am. Fruit Grower 73 ( 1 2 ) , 7, 18. Mason, B. 1960. “Principles of Geochemistry,” 2nd ed. Wiley, New York. Massey, H. F. 1957. Soil Sci. 83, 123-129. Mcderski, H. J,, and Wilson, J. H. 1955. Soil Sci. SOC. Am. Proc. 19, 461-464. Midgley, A. R., and Dunklee, D. E. 1939. Soil Sci. SOC.Am. Proc. 4, 302-307. Miller, M. H., and Ohlrogge, A. J. 1958a. Soil Sci. SOC. Am. Proc. 22, 225-228. Miller, M. H., and Ohlrogge, A. J. 1958b. Soil Sci. SOC. Am. Proc. 22, 228-231. Millikan, C. R. 1942. Proc. Roy. SOC. Victoria 54, 145-195.

158

J. F. HODGSON

Millikan, C. R. 1946. 1. Dept. Agr. Victoria 44, 69-73, 88. Mitchell, R. L. 1955. In “Chemistry of the Soil” (F. E. Bear, ed.), ACS Monograph No. 126, pp. 253-286. Reinhold, New York. Mitra, S. P., and Prakash, D. 1955. Proc. Natl. Acad. Sci. India A24, 11, 176-181. Mulder, G. 1940. Z. Pflanzenkrankh. Pflanzenschutz 50, 230-272. Naftel, J. A. 1938. Am. Fertilizer 89 ( 7 ) , 5-8. Nelson, J. L., and Melsted, S. W. 1955. Soil Sci. SOC. Am. Proc. 19, 439-443. Nelson, J. L., Boawn, L. C., and Viets, F. G. 1959. Soil Sci. 88, 275-283. Northmore, J. M. 1959. Nature 183, 1309-1310. Nozdrunova, E. M., Rytikova, M. N., and Shemyakina, A. F. 1958. Dokl. Mosk. Sel‘skokhoz. Akad. Nauchn. Konf. 34, 155-159. Olson, R. V., and Berger, K. C. 1946. Soil Sci. SOC.Am. Proc. 11, 216-220. Ortega, E. 1953. Mem. Congr. Cient. Mex. IV Centenario Univ. Mex. 2, 260-262. Page, A. L., and Whittig, L. D. 1961. Soil Sci. SOC. Ann. Proc. 25, 282-286. Parks, R. Q. 1944. Soil Sci. 57, 405-416. Parks, R. Q., and Shaw, B. T. 1941. Soil Sci. SOC. Am. Proc. 6, 219-223. Parks, W. L., and White, J. L. 1952. Soil Sci. Soc. Am. Proc. 16, 298-300. Peech, M. 1941. Soil Sci. 51, 473-486. Philipson, T. 1953. Acta Agr. Scand. 3, 121-242. Piper, C. S., and Beckwith, R. S. 1949. Proc. Spec. Conf. Plant Animal Nutr. Melbourne, Australia, 1949 pp. 144-147. Ponnamperuma, F. N. 1955. Ph.D. Thesis, Cornell Univ., Ithaca, New York. Prince, A. L., and Toth, S. J. 1938. Soil Sci. 46, 83-94. Pronin, M. E. 1933. Severo Kavkazrk. Z e m v o i Inst. ( N . Caucasian Grain Inst.) Collection Sci. Papers 1, 107-113. Randhawa, N. S., Kanwar, J. S., and Nijhawan, S. D. 1961. Soil Sci. 92, 106-112. Rao, H. G. G. 1956. J . Indian SOC. Soil Sci. 4, 225-231. Reuszer, H. W. 1962. Soil Sci. 93, 56-61. Reuther, W., Smith, P. F., and Specht, A. W. 1952. Soil Sci. 73, 375-381. Richardson, P. W., and Hawkes, N.E. 1958. Geochim. Cosmochim. Acta 15, 6-9. Ringwood, A. E. 1955. Geochim. Cosmochim. Acta 7, 189-202. Rogers, L. H., Gall, 0. E., Gaddum, L. W., and Barnette, R. M. 1939. Florida Agr. Expt. Sta. Tech. Bull. 341. Rovira, A. D. 1962. Soik Fertilizers 25, 167-172. Samuel, G., and Piper, C. S. 1929. Ann. Appl. Biol. 16, 493-524. Sanchez, C., and Kamprath, E. J. 1959. Soil Sci. SOC. Am. Proc. 23, 302-304. Sandell, E. B., and Goldich, S. S. 1943. 1. Geol. 51, 99-115, 167-189. Sarata, U. 1938. Japan J . Med. Sci. 11. 4, 199-202. Scharrer, K., and Hofner, W. 1958. 2. Pfinzeneraehr. Dueng. Bodenk. 81 ( 3 ) , 201-212. Schlichting, E., and Wiklander, L. 1956. Kgl. Lantbruks-Hogskol. Ann. 22, 93-99. Schnitzer, M. 1957. Chem. Ind, (London) p. 1594. Sedletskii, I. D., and Ivanov, D. 1941. Compt. Rend. A c d . Sci. URSS 30, 51-53. Seiffert, H. H., and Wehrmann, J. 1957. 2. Pflanzenernaehr. Dueng. Bodenk. 79, 142-154, Sherman, G. D., and Harmer, P. M. 1942. Soil Sci. SOC. Am. Proc. 7, 398-405. Simonson, G. H., Prill, R. C., and Riecken, F. F. 1957. Proc. Iowa Acad. Sci. 64, 385-392. Spencer, W. F., and Gieseking, J. E. 1954. Soil Sci. 78, 267-276. Starkey, R. L., and Halvorson, H. 0. 1927. Soil Sci. 24,381-402.

MICRONUTRIENTS IN SOILS

159

Steenbjerg, F. 1940. Tidsskr. Planteuol 45, 259-369. Steenbjerg, F., and Boken, E. 1948. Tidsskr. Plnnteuul 52, 375-459. Steenbjerg, F., and Boken, E. 1950. Plant Soil 2, 195-221. Stenberg, M., Ekman, P., Lundblad, K., and Svanberg, 0. 1948. N o d . Jordhrugsforskn., Proc. 7th Congr. 111, pp. 689-700, 709. Stewart, I., and Leonard, C. D. 1956. U . S. At. Energy Comm. TID-7512, 245-251. Swaine, D. J. 1955. Commonwealth Bur. Soil Sci. ( C t . Brit.) Tech. Commun. 48. Swaine, D. J., and Mitchell, R. L. 1960. J. Soil Sci. 11, 347-368. Teakle, L. J. H. 1942. J. Australion Inst. Agr. Sci. 8, 70-72. Thorp, J., Strong, L. E., and Gamble, E. 1957. Soil Sci. SOC. Am. Proc. 21, 99-102. Tiller, K. G. 1958. J. Soil Sci. 9, 225-241. Tiller, K. G. 1961. Ph.D. Thesis, Cornell Univ., Ithaca, New York. Tiller, K. G., and Hodgson, J. F. 1962. Clays Clay Minerals Proc. 9th Natl. Conf. Clays Clay Minerals, 1960, pp. 393-403. Pergamon Press, New York. Tiller, K. G., Hodgson, J. F., and Peech, M. 1963. Soil Sci. 95, 392-399. Timonin, M. I. 1946. Soil Sci. SOC. Am. Proc. 11, 284-292. Timonin, M. I., and Giles, G. R. 1952. J. Soil Sci. 3, 145-155. Titlyanova, A. A., Tyuryukanov, A. N., and Makhonina, G. I. 1959. Dokl. Akad. Nauk S S S R 126, 1346-1349. TrocmC, S., and Barbier, G. 1950. Compt. Rend. 230, 572-574. Truog, E. 1946. Soil Sci. SOC. Am. Proc. 11, 305-308. Tyulin, A. F. 1940. Pedology ( U S S R ) 3, 9-22. Viets, F. G . 1962. J. Agr. Food Chem. 10, 174-178. Vinogradov, A. P. 1959. “The Geochemistry of Rare and Dispersed Chemical Elements in Soils.” Consultants Bur., New York (translated from the Russian). Vlasyuk, P. A., and Butkevich, K. P. 1957. Dokl. Vses. Akad. Sel‘skokhoz. Nauk 22 ( 5 ) , 3-9. Wahhab, A., and Bhatti, H. M. 1958. Soil Sci. 86, 319-323. Wnin, R. L., Hunt, I. V., and Marsh, G. C. 1939. J. Southeastern Agr. Coil. Wye Kent 44, 114-119. Wain, R. L., Silk, B. J., and Wills, B. C. 1943. J. Agr. Sci. 33, 18-22. Walker, J. M., and Barber, S. A. 1960. Soil Sci. SOC. Am. Proc. 24, 485-488. Wallace, A. 1963. Soil Sci. SOC. Am. Proc. 27, 176-179. Wear, J . I. 1956. Soil Sci. 81, 311-315. Whetstone, R. R., Robinson, W. O., and Byers, H. G. 1942. U . S . Dept. Agri. Tech. Bull. 797. White-Stevens, R. H. 1942. Better Crops Plnnt Food. 26 ( 2 ) , 6-9, 42-46. Whittig, L. D., and Page, A. L. 1961. Soil Sci. SOC. Am. Proc. 25, 278-281. Wilson, C. M., Lovvorn, R. L., and Woodhouse, W. W., Jr. 1951. Agron. J. 43, 363-367. Winsor, H. W. 1952. Soil Sct. 74, 459-466. Wright, J. R., and Levick, R. 1956. Trans. Intern. Congr. Soil Sci. 6th Congr. Paris E pp. 257-262. Wright, J. R., and Schnitzer, M. 1959. Nature 184, 1462-1463. Wright, J. R., Levick, R., and Atkinson, H. J. 1955. Soil Sci. SOC. Am. Proc. 19, 340-344. Yang, T. T. 1960. Rept. Taiwan Sugur Expt. Sta. (Taiwan) 21, 1-16. Yarilova, E. A. 1940. Trans. Dokuchaeo Soil lnst. ( U S S R ) 24, 309-350. Zende, C. K. 1954. J. IndiQn SOC. Soil Sci. 2, 67-72.

This Page Intentionally Left Blank

IMPACT OF CHEMICAL WEED CONTROL ON FARM MANAGEMENT PRACTICES

. .

. .

. .

W B Ennis. Jr., W C Shaw. L L Danielson. D L. Klingman. and F. L Timmons*

.

U . S. Department

.

of Agriculture. Beltsville. Maryland

I. Introduction ................................................ A . Losses Caused by Weeds ................................. B. Technological Advances .................................. C . Cost and Benefits of Chemical Weed Control . . . . . . . . . . . . . . . . D . Impact on Labor Requirements ............................ E . Weed Control. a Total Farm Problem ....................... I1. Production of Cultivated Crops ................................ A . Choice of Crop and Crop Rotation .......................... B . Choice of Variety ........................................ C . Seedbed Preparation ..................................... D . Method of Seeding or Planting ............................ E . Seeding Rates and Row Spacing ........................... F. Fertilization ............................................. G . Cultivation ............................................. H . Irrigation ............................................... I. Harvesting .............................................. J . Erosion Control and Soil Fertility .......................... K . Chemical Fallow ........................................ L . Diseases ................................................ M . Future Trends .......................................... I11. Production of Deciduous Fruits and Tree Nuts . . . . . . . . . . . . . . . . . . . A . Planting and Maintenance ................................. B. Harvesting .............................................. IV. Production of Pastures and Rangelands .......................... A . Establishment of New Seedlings ............................ B. Pasture Renovation ....................................... C . Control of Brush on Grazing Lands ......................... D . Control of Herbaceous Weeds ............................. E . Effects of Weeds on Animals and Animal Products . . . . . . . . . . . . F. Water Conservation ...................................... V . Changes in Management of Farm Water Systems . . . . . . . . . . . . . . . . . A . Design and Construction of Water Channels. Roadways. and Structures ..............................................

' U . S.

Department of Agriculture. Laramie. Wyoming .

161

Puge 162 163 165 166 168 168 169 170 172 172 173 174 176 178 180 181 182 182 182 183 185 185 191 192 192 194 195 197 199 200 200

201

162

VI.

W. B. ENNIS, JR. ET AL.

B. Chemical versus Mechanical Methods ....................... 202 206 C. Improved Management of Water Systems .................... D. Choice of Irrigation and Cropping Practices . . . . . . . . . . . . . . . . . . 207 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 209 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Introduction

The native vegetation in the natural environment was not the kind to support large populations of livestock or man. As the human and livestock populations increased, it became necessary to replace the native vegetation in many areas with more productive, more efficient, and more economical plants for food and fiber. The sequence of plant successions that occurs when native vegetation is disturbed or cultivated fields are abandoned is influenced by intensity of cultivation and livestock grazing, incidence of fire, competitiveness of introduced species in comparison to native species, and other factors, Annual weeds usually comprise an early stage of plant successions that terminate with a form of stabilized vegetation best adapted to the area (Fig. 1 ) (Shaw et al., 1960). The fundamentals of plant ecology emphasize that plant succes-

dI- ’

L

z w I-

0

a > k

> l-

0 3

n 0

u

a n

0 0

~~

CLIMAX VEGETATION

WOODY SHRUBS

GRASSES AND WEEDS

LOW VALUE CROPS

-

HIGH VALUE CROPS

VEGETATIVE COMPOSITION SCIENTIFIC -NATURAL

TECHNOLOGY SUCCESSION

FIG.1. Relation of plant ecology to control of weeds and food production where forest is the climax vegetation.

IMPACT OF CHEMICAL WEED CONTROL

163

sions often occur in the direction of less productive vegetation rather than toward the growth of more productive economic crop plants. In efforts to produce crop plants, man attempts to utilize all available technology to stabilize the vegetation at a highly productive level and to prevent it from returning to a less productive level. The control of weeds is a basic, essential, and important aspect of this fundamental ecological process. The efficient management of plants, soils, and animals to improve product yields and quality and conserve our basic resources is as important as improving their potentials through breeding, fertilization, and other techniques. To manage crop and livestock production efficiently with maximum soil and water conservation, an understanding of the limiting factors in production, as well as a command of the latest technological advances, is essential ( Shaw, 1960). It is well known that limitations in farm operations may include one or more of the following: unadapted varieties, inefficient insect control, poor crop stands, inadequate supplies of water, presence of diseases, soils with poor physical properties, impeded drainage, mineral nutrient deficiencies, weather hazards, and excessive weed competition. When one or more of these factors become limiting, crop yields may be low. Highest crop yields are obtainable only when none of the production factors is limiting. The use of efficient, economical, and effective chemical weed control methods in farm operations has a marked impact on other farm practices.

A. LOSSES CAUSED BY WEEDS Control of weeds is one of the most important production practices in farm management, Weed competition with crop plants and the costs of their control constitute some of the highest costs in the production of food, feed, and fiber crops. Weeds also present costly control problems in farm ditches and ponds, on industrial sites, and on other noncrop areas. They adversely affect the health of millions of people by causing allergies and other ailments. Cost of controlling weeds and losses from failing to control them are borne directly or indirectly by many segments of private and public life. Although much progress has been made in improving weed control methods, the weed problem still represents a major challenge to optimum efficiency in farming operations. The annual national loss in agricultural production due to weeds and the cost of weed control has been estimated as more than four billion dollars annually. Weeds are among the most serious problems facing agriculture. The introduction of new

164

W. B. ENNIS, JR.ET AL.

herbicides has made possible the development of efficient chemical methods for attacking the nation’s weed problem. Weeds cause damage in many ways.

1 . Losses in Productivity and in Efficiency of Land Use The yields of crops and livestock are reduced and quality is impaired by weeds. In many instances the presence of particular weeds in fields determines the choice of crops to be planted. Weeds reduce the value of agricultural land. They increase production and harvesting costs, impede production mechanization, and reduce efficiency of equipment operation. Excessive tillage to control weeds frequently damages the crops and soil. Brush on rangelands causes losses in the production of forage, and increases the cost of handling livestock, and prevents efficient management.

2. Losses in Product Quality Weeds in cotton, hay, and leafy and other vegetable crops reduce their quality. In addition, wild onion ( Allium canadense L.), wild garlic ( A . vineale L.), and bitterweed (HeZenium tenuifolium Nutt.) impart off-flavors to milk so that frequently it cannot be sold. The presence of burs in wool is a serious source of loss in the wool industry. Contamination by noxious weed seeds greatly reduces values of crop seeds and sometimes prevents their sale. Weeds and weed debris in grains and other farm products reduce their sales value or cause spoilage in storage. Certain serious seed contaminants, such as crotalaria ( Crotalaria speciabilis Roth and C. mucronata Desv.) in corn or soybeans, dodder (Cuscuta spp.) in alfalfa or clovers, and wild vetch (Vicia spp.) in cereals, necessitate costly cleaning operations before marketing. 3. Reduction in Efficiency of Water Management Aquatic weeds reduce markedly the flow of water in irrigation and drainage canals. Obstruction to flow causes high water levels in canals and streams that result in ( a ) flooding, ( b ) seepage into adjoining areas or poor drainage, ( c ) breaks in canal banks, and ( d ) inadequate delivery of irrigation water to farms located at a distance from the water source. In addition reduced velocity of flow causes increased siltation and reduced carrying capacity and necessitates more frequent mechanical cleaning. Aquatic weeds that break loose and floating weeds obstruct weirs, gates, and other structures. Algae and fragments of plant material clog sprinkler-irrigation equipment. Aquatic weeds form breeding grounds for obnoxious insects such as mosquitoes. They reduce recre-

IMPACT OF CHEMICAL WEED CONTROL

165

ational values by interfering with fishing, swimming, boating, hunting, and navigation on streams. Weeds on watershed areas and flood plains utilize large amounts of water and thereby deny its use for the production of food, feed, and fiber,

4. Losses in Human E@ciency Weeds such as poison ivy (Rhus radicans L.) and ragweeds (Ambrosia spp. ) that cause hayfever and other debilitating allergies contribute markedly to human illness and make necessary the annual expenditure of large sums for medical attention and for travel to noninfested areas during the hayfever seasons. In addition, many hours of manpower are lost. 5. Losses Due to Harboring of Insects and DiseaseProducing Organisms Propagation and survival of many harmful insects are favored by weeds on which they can breed and feed. For example, Russian thistle (Salsola kali L. var. tenuifolia Tausch) is an important food plant for migrating beet leafhoppers, which transmit curly top virus to sugarbeets and vegetables. Weeds also serve as alternate hosts for many plant parasitic organisms that cause diseases such as wheat rust and as hosts for parasitic nematodes which cause tremendous losses to agriculture each year. 6. Damage from Brush and Other Wecds on Public and Private

Noncrop Areas Weeds on ditchbanks, on highway, railroad, and utility rights-of-way, in fence rows, and on other areas adjacent to cropland produce large quantities of seeds and thus serve as reservoirs for invasion of fields through movement of seeds by wind, water, man, and animals, or by vegetative spread. Weeds in lawns, parks, golf courses, playgrounds, cemeteries, and other public and privately owned areas are unsightly and increase maintenance costs by millions of dollars each year.

B. TECHNOLOGICAL ADVANCES Efficiency of farm production in the United States has increased in proportion to advances in farming technology. The introduction of power equipment for land preparation, tillage, planting, fertilization, and harvesting of crops provided one of the most significant advances in farm production in the United States. The discovery and development of new herbicides and principles of weed control provided American

166

W. B. ENNIS,

p.ET AL.

farmers with opportunities further to increase production efficiency. Among the important discoveries or developments were ( 1) the selective properties of certain organic herbicides such as 2,4-dichlorophenoxyacetic acid ( 2,4-D), ( 2 ) low-gallonage, low-pressure spraying techniques, ( 3) preemergence application of herbicides, ( 4 ) preplanting techniques, (5) use of granular formulations, ( 6 ) chemical renovation and fallowing, ( 7 ) residual organic herbicides such as the substituted phenylureas, benzoic acids, phenylacetic acids, and triazines, (8) selective grass killers, such as the carbamates and 2,2-dichloropropionic acid (dalapon), ( 9 ) the role of plants and soils in converting inactive chemicals to herbicides and vice versa-providing a unique basis for selective herbicidal action, ( 10) precision subsurface application and placement of herbicides in soils, and (11) advances in surfactant and formulation technology. The use of chemical weed control provides an unusual opportunity for achieving greater mechanization and efficiency of farm production. Improved weed control measures provide for production of high quality farm products that meet consumer demands. In addition the use of efficient chemical weed control practices ensures fewer crop failures due to weeds than the use of conventional weed control methods. An arsenal of effective and safe selective organic herbicides is available to farmers and home owners. With so many effective herbicides available, farmers can be well armed against a wide variety of weeds such as those that reduce growth of agronomic and horticultural crops, others that stifle grasses and legumes in pastures and rangelands, those that clog drainage ditches and irrigation systems, and weed pests in lawns and gardens and on noncrop land. C. COSTAND BENEFITSOF CHEMICAL WEEDCONTROL Chemical weed control is increasing rapidly on farms in the United States. About 53 million acres were treated with herbicides in 1959 in 41 States (Table I ) at a cost of $128 million. The acreage treated in 1959 was about dduble that treated in 1949 (Table 11). During this same period the number of new organic herbicides available to farmers increased from about 20 to nearly 60. These new chemicals possess selective properties for controlling weeds in many crops and under many soil and climatic conditions. The use of herbicides markedly reduced labor needs, improved crop quality, and yields, and improved other farming operations. In 1962 an estimated 85 million acres of agricultural land and over 30 million acres of other land were treated with herbicides at a cost of approximately $200 million to the consumers for the chemi-

TABLE I Estimated Extent and Cost of Chemical Weed Control on Farms in the United States, 1959 ~

Combined cost of herbicides and-applicationb

Acreage treated Crop

Preemergence

Cotton Soybeans Small grains Rice Peanuts Sugarbeets Sorghum Forage seeds Vegetables Fruits and nuts Strawbemes Ornamentals Lawns Hay Pastures Rangeland Noncrop land

Total

Preemergence

1,OOO acres

-

Corn

Postemergence

2,235.4 1,000.6 545.5 -

-

31.9 81.6 8.0

-

71.9

-

2.0 0.2 2.6

-

30.1

-

27.2 4,037.0

Postemergence

Total

$1,000

17,816.5 552.7 10.0 20,723.3 502.0 3.0 42.6 2,085.0 281.8 204.1 5.4 3.3 2.2 57.3 272.4 2,370.0 2,011.0 1,943.5

20,051.9 1,553.3 555.5 20,723.3 502.0 34.9 124.2 2,093.0 281.8 276.0 5.4 5.3 2.4 59.9 272.4 2,400.1 2,011.0 1,970.7

8,226.3 3,221.9 2,296.6

48,886.1

52,923.1

30.1 2,596.2

29,753.6 1,406.1 17.5 37,094.7 888.5 9.0 197.2 6,463.5 1,868.3 835.6 42.8 20.4 43.1 809.0 1,691.6 5,759.1 6,173.8 17,141.7

37,979.9 4,628.0 2,314.1 37,094.7 888.5 116.2 624.8 6,511.5 1,868.3 1,417.7 42.8 55.6 45.1 1,489.1 1,691.6 5,789.2 6,173.8 19,737.9

$18,253.2

$110,215.6

$128,468.8

-

107.2 427.6 48.0

-

582.1

-

35.2 2.0 680.0

-

5 From U. S . Dept of Agriculture A R S 34-23 ( 1961). Estimates for Alaska, California, Delaware, Hawaii, New Jersey, New York, Ohio, Oklahoma, and Washington not included. b Cslculated from average costs incurred by farmers and other landowners in the States reporting.

2

ti

8

E!

ij

c

i8 3

I-

3

168

W. B.

ENNIS, JR.ET AL.

cals alone. The benefits derived from chemical weed control continue to stimulate interest in safe and efficient herbicides. TABLE I1 Estimated Acreages Treat-d for Weed Control in the United Statesa Corn

Small grains

Year

( 1,000 acres)

( 1,000 acres)

1949 1952 1959"

4,559 8,150 20,051

18,751 16,792 20,723

Grazing lands All other ( 1,000acres) ( 1,000 acres)

-

-

2,192 4,411

2,629 7,738c

Total ( 1,000 acres)

23,310 29,763 52,923

a Estimates for 1949 and 1952 from U . S. Dept. of Agriculture Statistical Bullettin 156, April 1955; 1959 estimates from U . S. Dept. of Agriculture ARS 34-23 ( 1961 1. b Estimates from Alaska, California, Delaware, Hawaii, New Jersey, New York,

Ohio, Oklahoma, and Washington not included. 0 Estimated 2 million acres noncrop land treated included.

D. IMPACXON L A B ~REQUIREMENTS R The adoption by farmers of satisfactory chemical weed control methods has made possible a marked reduction in the man hours required to produce many crops. In Mississippi between 1954 and 1956 the use of herbicides for weed control in cotton reduced labor needs by 16 to 35 man hours per acre. Cost of labor for weeding of crops such as strawberries decreased from as high as $200 an acre to as little as $16 to $20 an acre through the use of selective herbicides. The chemical control of winter weeds in spinach in Mississippi resulted in benefits of $162 per acre. Hand and machine requirements for removal of aquatic weeds from ditchbanks and irrigation canals have been markedly reduced through the use of herbicides. Chemical weed control is having an impact on the design of equipment required in the production of crops. In the past, rows were spaced primarily to accommodate machines of limited or fixed wheel spacings. Today, closer row spacing can be utilized in many crops because of reduced need for cultivation to control weeds. Thus, farm equipment including harvesters can now be developed with the versatility to fit closer spaced rows.

E. WEEDCONTROL, A TOTALFARM PROBLEM Selective herbicides to control weeds in many crops grown in rotation are providing farmers for the first time with effective methods to control weeds on the farm as a unit. Weeds must not be permitted to produce seeds year after year and spread by seeds or vegetatively. Farmers can no longer afford to control weeds effectively in one crop while permitting them to grow and reproduce in other crops grown in rotation on the same field. Weeds should not be permitted to reproduce and spread on

IMPACT OF -MEAL

WEED CONTROL

169

ditchbanks, along fence rows, and on other noncrop land adjoining crop land because they will spread into crop land. Practices must be developed to utilize chemical and other improved methods to control weeds in each crop in the rotation. There is a broader choice of chemicals for controlling different weeds and the grower can kill more species of weeds and thus prevent resistant species from gaining dominance. In many instances rotational schemes that provide for use of different herbicides on the same crop grown in succession and in the production of different crops on the same field so as to achieve maximum weed control without the build up of harmful herbicide residues in the soil are needed. Weed research scientists must be continually alert to gain understanding of soil characteristics in relation to choice of herbicides for use on particular crops, to the pattern of rainfall, and to the management of irrigation water. Chemical weed control has had and will continue to have considerable impact on farm-management practices. In the following sections of this paper are cited examples of improvements in production practices brought about by the use of chemical weed control and changes in farmmanagement practices that may be required to take full advantage of chemical weed control practices. 11. Production of Cultivated Crops

Field and horticultural crops, not including pastures and rangelands, were grown and harvested on approximately 320 million acres in 1959 in the United States. The major field crops are grown on approximately 312 million acres, and the remaining acreage is devoted to horticultural crops. The U. S. Agricultural Research Service (1954) estimated weed losses and the costs of controlling weeds in 58 field and horticultural crops at about $3.3 billion per year. The acreage of horticultural crops, though comparatively small, requires large amounts of hand labor and mechanical cultivation for weed control at costs ranging from $50 and upward per acre. Farmers are much interested in chemical weed control to reduce these losses. A survey by the U. S. Agricultural Research Service and the Federal Extension Service (1961) showed that farmers treated more than 40 million acres of field crops in 1959 (Table I ) . Eleven major field crops were grown on 80 per cent of the total agricultural land treated with herbicides. In the production of cereals, rice, and many other crops, the use of herbicides has become increasingly necessary for efbcient weed control. The use of herbicides for the control of weeds does not eliminate

170

W. B. ENNIS, JR. ET AL.

the need for sound cultural practices and good management. On the contrary, the chemical control of weeds places a premium on skilled farm management, but management practices may be drastically altered. Important advances have also been made in cultural, mechanical, and biological methods of weed control. Research and farm practices have shown that greatest weed-control efficiency results when combination methods are used. A. CHOICEOF CROPAND CROPROTATION All crops grown on the farm are subject to weed competition. In attempts to control heavy infestations of weeds, especially perennials, greater consideration must be given to weed control on the farm as a farm unit rather than to limiting control techniques to weeds in a single crop. The cost of controlling heavy infestations of perennial weeds often appears high because only the crop in which they currently appear is being considered rather than the total weed problem in all crops grown on the farm over a period of years. Most weed species can be controlled by one or more herbicides currently available. However, some important weeds cannot be selectively controlled in some crops because the crop does not possess sufficient tolerance to the herbicide. Therefore, the choice of a crop or a crop rotation to facilitate control of specific weeds, especially perennial weeds, is of utmost importance. The crop should be tolerant to a herbicide which is effective in controlling the weed population. The crop should also provide maximum competition with the weed population. High seed viability, rapid germination, early emergence, high seedling vigor, rapid early growth, and formation of a foliage canopy to provide shade are excellent crop characteristics to supplement the use of herbicides to control weeds. The importance of choosing an appropriate crop or crop rotation for the control of weeds was effectively demonstrated by Hodgson (1958) in a six-year investigation at Bozeman, Montana. Sixteen different combination and single weed control treatments involving cropping practices, crop rotations, cultivations, and application of herbicides in successive years for the control of Canada thistle (Cirsium aruense (L.) Scop.) were studied. The results showed that 2,4-D was more effective in controlling Canada thistle when used with competitive crops such as wheat, pasture grasses, or silage corn than when used alone. Combination treatments involving the use of 2,4-D with spring wheat gave effective control of Canada thistle and resulted in a large increase in yield of wheat during the six years of the study (Fig. 2 ) .

I M P A m OF CHEMICAL WEED CONTROL

171

Several crop production practices and management methods to determine their effectiveness for controlling Canada thistle were investigated ( Hodgson, 1958). The practices and methods evaluated included: spraying spring wheat with 2,4-D;planting of forage crops plus the use of 2,4-D and mowing; combining intensive cultivation, wheat crop-

FIG.2. Canada thistle control by combination herbicide-cultural practices. Left, area planted to wheat, highly fertilized with nitrogen and treated with 2,4-D. Right, planted to wheat and fertilized only.

ping, and the use of 2,4-D; and rotating crops and using herbicide treatments. In these investigations, combination methods involving crop competition, cultivation, and chemical methods were more effective than any of the methods alone. The density and composition of weed populations should determine the choice of crops and the manner in which they are rotated. For example, in soybeans Canada thistle is difficult or nearly impossible to control with presently available herbicides. However, Canada thistle may be controlled rather easily in wheat treated with 2,4-D. These are but a few of many examples that indicate the importance of selecting a crop rotation for combination treatments involving the use of herbicides and crop competition for weed control.

172

W. B.

ENNIS, JR. ET AL.

B. CHOICEOF VARIETY The control of weeds may influence the choice of a crop variety in at least two ways. Differential varietal tolerance to herbicides exists in most crop species. Notable examples are the differential responses of hybrid corn varieties to 2,4-D and of flax varieties to trichloroacetic acid ( TCA ) and 2-methyl-4-chlorophenoxyaceticacid ( MCPA ). In addition, varieties vary greatly in their competitiveness with weeds. It. two varieties of a particular crop are equal in all other attributes, priority should be given to the variety possessing tolerance to herbicides likely to be used for controlling weeds in the crop, In addition, varieties that possess high seedling vigor, emerge rapidly, and develop a foliage canopy are usually better competitors with weeds than varieties lacking these attributes (Shaw et al., 1955;Smith, 1958). C. SEEDBEDPREPARATION Thorough seedbed preparation is generally regarded as an aid to weed control. Yet, little definitive information is available on the importance of thoroughness of seedbed preparation in relation to effectiveness of the technique as a weed control practice. The crop to be planted and the cultivation practices to be used after it emerges, rather than the weed population to be controlled or the herbicide to be used for its control, usually determine the type of seedbed preparation. Seedbed preparation is known to influence the performance of preemergence herbicides. Some herbicides perform most efficiently on a fine, thoroughly prepared seedbed, whereas others, especially if volatile, seem to be little influenced by the thoroughness of seedbed preparation, especially when incorporated with the soil. The increased use of soil-applied herbicides for weed control as preplanting soil-incorporated treatments and as preemergence treatments indicate that seedbed preparation will become important in the use of such herbicides in the future (McWhorter and Wooten, 1961; Wooten and McWhorter, 1961). The simultaneous planting of cotton and the subsurface incorporation of ethyl N,N-di-n-propylthiolcarbamate ( EPTC ) in the soil to control annual weeds as well as nutsedge (Cypews rotundus L.) and johnsongrass (Sorghum halaperne (L.) Pers.) without crop injury have appeared promising, The further development of this technique will require a better understanding of the chemical and biological effects of the herbicide and advances in equipment engineering technology. This method when fully perfected could revolutionize seedbed preparation and provide an opportunity to control perennial weeds by pre-

IMPACT OF CHEMICAL WEED CONTROL

173

emergence treatments ( Holstun, 1962; McWhorter and Wooten, 1961; Wooten and McWhorter, 1961).

D. METHODOF SEEDING OR PLANTING New herbicides including N,N-dimethyl-2,2-diphenylacetamide( diphenamide ) and certain thiocarbamates to which tomatoes are tolerant give excellent control of weeds when applied as a preemergence treatment on direct-seeded tomatoes. The application of highly selective preemergence herbicides may permit the direct field seeding of this crop and greatly reduce the time and labor involved in growing tomato plants in propagation beds and transplanting them for field production. Similar possibilities exist in tobacco culture. The choice of preemergence herbicides has an important bearing on method and depth of seeding. All field and horticultural crops should be seeded at maximum depths consistent with optimum emergence and stand establishment to give protection against injury from preemergence or soil-incorporated herbicides that may be leached downward by heavy rains after planting. The effect of depth of seeding on the tolerance of rice to herbicides was effectively demonstrated in investigations in Arkansas. In these studies, rice seeded one-half inch deep was severely injured by isopropyl N - ( 3-chlorophenyl ) carbamate ( CIPC ), but little or no injury of rice seeded 1 to 2 inches deep occurred (Smith, 1960). Seeding rice in flooded fields is practiced on more than 95 per cent of the rice acreage in California to control barnyardgrass (Echimchloa crusgulli (L.) Beauv.) and other weeds. The weed grass germination is inhibited while the rice grows relatively well in water. Seeding rice in flooded fields requires extensive seedbed preparation involving expensive leveling of land to a uniform gradient between levees and the proper construction of levees to permit a uniform depth of water of 6 to 8 inches ( Fig. 3). These operations require heavy, mechanized equipment ( Smith, 1958, 1960, 1961). The development of efficient chemical methods of controlling weed grasses, broadleaved weeds, aquatic species, and other weeds in rice may have a profound effect on present seeding methods. For example, aerial equipment is required for seeding rice in flooded fields. Effective chemical methods of controlling weeds will make it possible to drill or broadcast rice with ground equipment. The practice of seeding in soil rather than in water will permit a broader choice of irrigation practices for production of rice. Efficient chemical weed control practices in rice (Fig. 3 ) will also aid in the conservation of water and provide wider flexibility in irrigation practices (Smith, 1958, 1960, 1961). Chemical weed control practices will also influence the seeding of other crops.

174

W. B. ENNIS, JR.ET AL.

FIG. 3. Control of bamyardgrass in rice by herbicide treatment. Foreground, treated with postemergence application of 3,4-dichloropropionanilide ( DPA ) . Background, untreated rice heavily infested with bamyardgrass.

E. SEEDING RATESAND Row SPACING The rates of seeding and row spacing of many field and horticultural crops have evolved through research and farm practice on the assumption that weeds would be present and compete with the crop plants. Many crops are seeded at excessive rates in an attempt to give the crop a competitive advantage over weeds. Chemical methods of controlling weeds are having an impact on seeding rates, row spacing, and plant populations per acre. Two examples may be used to indicate the impact of chemical weed control on these practices. Before development of chemical weed control methods in cotton, the crop was drilled in the row. This practice required large amounts of seed per acre. Hand-hoeing was also necessary to reduce the drilled cotton to an appropriate stand. The development of preemergence chemical methods of controlling weeds in cotton made possible the development of machinery for dropping the seed in hills, thus the necessity for thinning to a stand by hand-hoeing or cross-cultivating was eliminated (Fig. 4 ) . Studies by Harris (1960) indicate that the development of satisfactory full-season methods of chemical weed control has greatly altered cultural practices in cotton production.

IMPACT OF CHEMICAL WEED CONTROL

175

A similar problem exists in the production of sugarbeets. Normally this crop is drill seeded in rows followed by blocking and thinning. Two developments will have an impact on the mechanized production of sugarbeets. The development of monogerm seed of adapted varieties has permitted planting sugarbeets to the desired stand. However, if full potential of this practice is to be realized, chemical weed control methods to replace hand-weeding methods are required. The use of monogerm varieties and preemergence chemical treatments, such as TCA and

FIG. 4. Chemical control of weeds in cotton. Right, cotton received preemergence herbicide treatment. Left center, two rows not treated. Chemical treatments can eliminate need for hand-hoeing or cross-cultivating to thin stands of cotton and remove weeds, and also reduce the number of cultivations.

mixtures of TCA and 7-oxabicyclo-( 2.2.1) heptane-2,3-dicarboxylicacid ( endothall), shows promise of eliminating blocking, thinning, and handhoeing (Fig. 5 ) . Recent investigations involving a number of crops indicate that, as satisfactory chemical methods of controlling weeds are developed, seeding rates may be reduced, row spacing may be closer, and thus plant populations per acre may be increased. A trend toward closer row spacing and higher plant populations per acre is occurring in the production of peanuts, soybeans, sorghum, vegetable leaf crops, and others as chemical methods of weed control are improved (Fig. 6).

176

W. B. ENNIS,

JR ET AI,.

FIG.5. Chemical weed control in sugarbeets. Left, treated preemergence with a combination of sodium trichloroacetate and endothall. Right center, untreated area heavily infested with weeds.

F. FERTILIZATION In many crops the weed population may influence the type of fertilizer, its placement in relation to the crop plant, and time of application. For example, in the production of rice, the time of applying phosphate is very important because this mineral nutrient when applied before seeding rice in a dry seedbed stimulates the growth of barnyardgrass and other weeds. When rice fields are heavily infested with barnyardgrass and other weed grasses and broadleaved weeds, competition may be reduced by applying phosphate to a crop other than rice in the rotation or by delaying applications until just before the rice is inundated

IMPACT OF CHEMICAL W E D CONTROL

177

for the first time. This fertilizer practice has evolved in an effort to reduce weed grass competition with rice. However, applying phosphate to a crop other than rice in the rotation or delaying its application is not efficient for maximum rice yields, but represents a compromise to reduce barnyardgrass competition with rice. The development of CIPC, 3,4dichloropropionanilide ( DPA ), and the phenoxy compounds for fullseason control of weed grasses and broadleaved weeds in rice will permit the application of phosphate at an optimum time for maximum response of the rice plant (Smith, 1958, 1960, 1961).

FIG.6. Chemical weed control in soybeans. Left, preemergence treatment with 3-amino-2,5-dichlorobenzoicacid ( amiben ). Right, untreated. Chemical weed control provides opportunities to space rows closer, increase plant populations, reduce cultivations, and eliminate hand-hoeing.

The time of applying nitrogen to fields of rice heavily infested with weed grasses is also very important. Barnyardgrass and other weeds are greatly stimulated by nitrogen applied before the rice is seeded. When rice fields are heavily infested with barnyardgrass, delaying the nitrogen application until heading of the weed reduces its competition with rice. If the nitrogen is applied while the barnyardgrass is vegetative, this weed species uses most of the nitrogen and its competitiveness with rice is enhanced. When barnyardgrass infestations are high, the yields of rice have been almost doubled by delaying nitrogen application until heading

178

W. B. ENNIS, JR. ET AL.

of the weed. The delay is undesirable from the standpoint of optimum time of nitrogen application for the rice plant. However, this practice has evolved as a necessity in rice fields heavily infested with barnyardgrass. The use of CIPC, DPA, and the phenoxy herbicides for the control of barnyardgrass and other grasses, sedges, and broadleaved weeds in rice will essentially eliminate the weed problem and permit the timely application of nitrogen that results in optimum yields of rice (Smith, 1958, 1960, 1961).

G. CULTIVATION The average cost of tillage is estimated at 16 per cent of the value of the crops produced. At least half of the tillage required is necessary to control weeds. Thus farmers lose about 8 per cent of the value of the products they produce annually because of the tillage needed to control weeds. In many crops in which weeds are controlled by chemical methods the additional cultivations needed to control weed populations may also damage crops and reduce yields. American farmers cultivate an estimated 80 million acres of corn an average of at least three times each year. Two of these cultivations are necessary for the sole purpose of controlling weeds. At an average cost of $1.50 per acre per cultivation, farmers spend about 240 million dollars each year in cultivating corn to control weeds. Investigations on the control of weeds in corn have shown that if weeds are controlled by chemicals, more than one cultivation of corn results in reduced corn yields (Fig. 7 ) . In experiments on soil types with good physical properties, maximum corn yields have been obtained by the use of chemicals to control weeds without cultivation. In studies on the control of weeds in soybeans in Missouri, there were no significant benefits from more than one cultivation when the weeds were effectively controlled with a preemergence herbicide. Cultivations alone gave poor weed control and lower yields than combinations of preemergence herbicide treatments and cultivations. When weeds were not controlled in soybeans by a preemergence herbicide, two cultivations gave significantly increased yields and better weed control than one. Three cultivations were not significantly superior to two (Peters et al., 1959, 1961). Sixteen weed control practices involving combinations of hand-hoeing, a preemergence herbicide, postemergence chemical treatments, and flaming treatment in hill-dropped and cross-cultivated cotton were evaluated for weed control effectiveness at Stoneville, Mississippi (Holstun et al., 1W). The direct cost of controlling annual weeds by hoe labor and cultivation ranged from $8 to $26 per acre and required

IMPACX OF CHEMICAL WEED CONTROL

179

14 to 53 hours per acre of hoe labor. Through the judicious use of a preemergence herbicide, postemergence oil application, flaming, and cross-cultivation, either singly or in various combinations, hoe-labor requirements can be greatly reduced with appreciable savings in weedcontrol costs. With some of the combination treatments, hoe labor necessary for the control of weeds was practically eliminated.

FIG.7. Weed control in corn typically obtained by combination preemergence treatment with atrazine and postemergence treatment with 2,4-D. Left, untreated. Right, treated.

In a ten-year study on the control of weeds in cotton in Mississippi (Harris, 1960), an average of six to eight cultivations per year were required in combination with hoeing to control weeds. However, when weeds were satisfactorily controlled in cotton by pre- and postemergence herbicide treatments, the cost of the herbicide for weed control was less than the cost of hand-hoeing and the number of cultivations required was greatly reduced. Weed control with herbicides returned $6 more per dollar invested for weed control than hand-hoeing and cultivation. The cost of controlling weeds in cotton by the use of combination herbicide treatments was 2% cents per pound of lint cotton produced and the

180

W. B. ENNIS, JR. ET AL.

return was $17 per dollar spent. The cost of hand-hoeing to control weeds was 3% cents per pound of lint cotton produced and the return was only $11 per dollar invested. The cost of weed control by the use of cultivation alone was 2 s cents per pound of lint cotton produced and the return was $8 per dollar spent. These investigations also indicate that when weeds are controlled in cotton with herbicides, the number of cultivations can be reduced from an average of eight to an average of about two cultivations per year. On many soil types, when herbicides were used to control weeds in cotton, no benefits were obtained from cultivation. At an average cost of $1.50 per acre per cultivation, using herbicides would result in an average saving of $9 per acre per year. The savings from reduced cultivation alone would more than pay for the cost of the herbicide treatments. The heavy seeding rates and associated large populations of many vegetable row crops, including greens, salad crops, vegetable legumes, and a number of root and bulb crops, prevent the effective use of mechanical equipment for the control of weeds between crop plants in the row. Herbicides now in use are providing effective, economical methods of controlling weeds in many of these crops. Herbicides now in the early stages of development will bring continued rapid improvement in methods of vegetable crop production. H. IRRIGATION The impact of chemical weed control in irrigation practices is illustrated in the culture of rice (see Section 11, D) , cotton, and certain other crops. To obtain optimum weed control with certain commercial preemergence herbicides in the western United States, it has been necessary to irrigate cotton fields so as to ensure proper moistening of the soil. Without moistening soil to the top of the rows, the herbicides become localized on the soil surface and fail to kill germinating weed seeds in the subsurface zone of the row. To achieve needed uniformity in moisture penetration of the row requires special land leveling and preparation of rows for planting. Sprinkler irrigation has also proved valuable after application of preemergence herbicides to ensure consistency of weed control. In addition, uniformity in effectiveness of postemergence herbicides depends on proper timing and uniformity of application of irrigation water. Thus, irrigation practices may need to be altered in some cases to make the most effective use of herbicides.

181 In the western cotton-producing states, irrigation practices are highly important as they influence weed populations in cotton fields. Early irrigations are designed to reduce early-weed competition. However, late-season weeds become serious in irrigated cotton. The use of lay-by herbicide treatments followed by timely irrigation improves the effectiveness of lay-by chemical treatments and results in good control of late-season weeds (Miller & al., 1961). The control of dodder in alfalfa with CIPC depends on proper irrigation. The effectiveness of CIPC is greatly increased if it is applied when soil is moist and subsequent irrigations are delayed to prevent emergence of additional dodder seedlings. IMPACT OF CHEMICAL WEED CONTROL

I. HARVESTING Weed-free crops are much more efficiently harvested than those containing weeds. Late-season weeds in cotton greatly increase the difficulty of harvesting and normally result in lower grade cotton. In the river bottom soils of the Corn Belt, heavy weed populations that dcvelop in corn after the last cultivation cause lodging, decrease the efficiency of corn harvesting equipment, cause breakage of equipment, and reduce yields. Vine-type weeds are particularly troublesome in combining harvesting of grain crops. Weed seeds in crop seeds greatly increase the cost of harvesting and cleaning. Certain weed seeds are also difficult and costly to remove. Crotalaria in soybeans and Canada thistle blossom buds in canning peas are difficult to remove with present cleaning equipment. Weed seeds in crop seeds result in increased dockage and lower net profits per acre. Harvesting and processing of such crops as sweetpotatoes, white potatoes, and peanuts are very difficult when heavy weed infestations occur in the fields. Weeds that become entwined in machine parts greatly reduce the efficiency of harvesting machinery. Rhizomes of nutsedge, quackgrass ( Agropyron repens ( L. ) Beauv. ) , and bermudagrass (Cynodon dactylon (L.) Pers.) often grow into potato tubers and reduce their market quality and value. The removal of seeds of weed grasses and broadleaved weeds from seeds of forage grasses and legumes is expensive and often ineffective. Use of 3-(3,4-dichlorophenyl)-l,l-dimethylurea( diuron ), 2,3,6-trichlorophenylacetic acid (fenac), and other chemicals for weed control in small-seeded forage crops grown for seed production offers excellent possibilities for reducing the cost of removing weed seeds from seeds of these crops.

182

W. B.

ENNIS, JR.ET AL.

CONTROL AND SOIL FERTILITY J. EROSION Development of improved chemical methods of controlling weeds will greatly enhance the practicality of minimum tillage and mulching practices in crop production. Preplanting soil-incorporated and preemergence herbicides prevent weed competition and provide an environment in which the rate, time of application, and placement of fertilizer can be based on optimum conditions for use by crop plants. The use of herbicides for weed control in pineapple in Hawaii has increased the effectiveness and value of mulching paper and plastics used in the culture of the crop. Herbicides have also increased the value of grain-straw mulches used in the culture of many horticultural crops by controlling weeds and volunteer grain crops. The use of herbicides to increase the effectiveness of mulch seeding combined with minimum tillage will improve soil structure, increase moisture penetration, reduce water runoff and evaporation, and reduce soil losses from water and wind erosion. K. CHEMICAL FALLOW In the Great Plains States, herbicides are showing increasing promise for the control of weeds in the fallow year of the crop rotation without the use of cultivation. In a wheat-sorghum-fallow rotation, herbicides are being used effectively to control weeds in sorghum and wheat and in the fallow year with minimum tillage operations. The minimum tillage made possible by the effective use of combination herbicide treatments should greatly aid in moisture conservation and in reducing wind erosion (Phillips, 1961; Phillips and Timmons, 1954).

L. DISEASES The chemical control of barberry in the United States has greatly reduced losses from stem rusts of wheat, oats, barley, and certain other grasses. By controlling barberry, an alternate host of the stem rust fungus, early-season stem rust epidemics have been retarded and the number of physiological races of stem rust kept at a minimum. Losses have been reduced accordingly. In the production of peanuts in the southeastern United States, a disease complex consisting of bacteria and fungi may cause serious reductions in peanut yields when deep frequent cultivations for weed control move soil onto the foliage of peanuts (Fig. 8). Plant injury by cultivation greatly increases the incidence of the soil rot fungus (Rhizoctonia solani) and the southern blight fungus (Sclerotium rolfsii), and serious reductions in peanut yield may occur if special precautions are

IMPACT OF CHEMICAL WEED CONTROL

183

not observed (Boyle et al., 1958). A combination of 4,6-dinitro-o-secbutylphenol ( DNBP ) and sodium 2,4-dichIorophenoxymethyl sulfate (sesone) applied just as the peanuts emerge combined with minimum

FIG.8. Chemical control of nutsedge and other weeds in peanuts. Center, ethyl N,N-di-n-propylthiolcarbamate ( EPTC ) applied as a preplanting soil-incorporated treatment. Right, untreated peanuts with heavy nutsedge infestation. Chemical treatments eliminate the need for moving soil onto the foliage of peanuts during cultivation. This practice has decreased the incidence of certain serious diseases of peanuts and has increased yields.

tillage has given excellent control of weeds and reduced losses due to the disease complex.

M. FUTURETRENDS In attempting to evaluate the impact of efficient chemical weed control methods on crop production, it must be recognized that farmers have entered a new era in crop production which involves the utilization of various chemicals to control weeds and other pests and for crop production and crop protection in a wide variety of ways. As new and more efficient chemical weed control methods are discovered through research, other cultural practices involved in crop production will need to be reevaluated in the light of the new technological advances in weed control.

184

W. B. ENNIS, JR. ET AL.

Current trends indicate that mechanical equipment will be used less in the future to control weeds. Instead, herbicides will be used to achieve more effective and economical control of weeds than that by conventional mechanical methods. This does not imply that the need for mechanical equipment will be reduced. Quite the contrary. It means greater emphasis on specificity and accuracy in placing chemicals in the exact amount, in the exact place, and at the exact time that they may provide maximum and efficient weed control. All crops grown on the farm are subject to weed competition. To obtain a better balanced and more efficient weed control program, emphasis must be placed on the rotational use of herbicides on all tolerant TABLE I11 Hypothetical Rotations Illustrating the Principle of Keeping Maximum Pressure on the Weed Population by Using a Series of Herbicides Differing in their Effectiveness in Controlling Various Weed Speciesa

A. Rotation of different herbicidesb on the same crop Chemical weed control treatments

Year

Crop sequence

Preemergence

First Second Thud Fourth

Corn Corn Corn Soybeans

Simazine Atrazine CDAA 2,3,6-TBA Amiben

'

Postemergence 2,4-D 2,4,5-T 2-( 2,4-DP) Linuron

+

B. Use of various herbicidesb on all crops in a rotation Year

Crop sequence

First Second Third Fourth

Soybeans Corn Peanuts Cotton

Chemical weed control treatments Preplanting Preemergence

-

Amiben Atrazine

EPTC EPTC

Diuron

EPTC

-

Postemergence Linuron (directed) 2,4-D DNBP-2,CDEP ( a t emergence) Prometryne (directed)

a This procedure reduces the chance of a species tolerant to a specific herbicide becoming dominant and reduces the chance of an accumulation of herbicide residues in the soil. IJ Amiben3-amino-2,5dichlorobenzoic acid Atrazine2-chloro-4-ethylamino-6-isopropylamino-s-triazine

CDAA-2-chloro-N,N-diallylacetamide Diuron--3-( 3,4-dichlorophenyl)-1,l-dimethylurea L i n u r o n 3 - ( 4-dichlorophenyl) -1-methoxy-1-methylurea Prometryne-2-methoxy-4,6-bis ( isopropylamino) -s-triazine Simazine-2-chloro-4,6-bis( ethylamino ) -s-triazine 2,4-DEP-tris ( 2,4-dichlorophenoxyethyl) phosphite 2,4,5-T-2,4,5-trichlorophenoxyacetic acid 2- ( 2,4-DP)-2- ( 2,4-dichlorophenoxy)propionic acid EPTC-ethyl N,N-di-n-propylthiolcarbamate

IMPACT OF CHEMICAL WEED CONTROL

185

crops throughout the rotation. Rotating different herbicides on the same crop grown continuously and full utilization of different herbicides on tolerant crops throughout a crop rotation make use of the sound principle of preventing undesirable shifts in weed populations, avoiding an accumulation of herbicide residues in the soils, and providing for the full use of chemicals in combination with cultural practices to reduce weed seed populations in the soil (Table 111). 111. Production of Deciduous Fruits and Tree Nuts

The production of deciduous fruits and tree nuts, such as strawberries, cranberries, cane fruits, blueberries, pome fruits, stone fruits, almonds, pecans, walnuts, and others, constitutes an important source of farm cash income and farm family subsistence. Chemical weed control has given new hope to the producers of fruits and tree nuts in the present period of mounting production costs and labor scarcity. The growers make large initial investments to establish these perennial crops. The continuing high maintenance costs are a financial burden resulting from the purchase and application of fertilizers, lime, insecticides, fungicides, and fruit thinners and stickers. Labor for pruning, weeding, and harvesting is costly and scarce. The mechanization of field operations and the development of chemical methods of controlling weeds have proceeded concurrently under the stimulus of this economic pressure. Recent advances in weed control technology including the development of herbicides, the improvement of mechanical weeding equipment, the adoption of new cultural practices, and the use of combinations of these methods provide important practical weapons for the battle to control or reduce production costs. A. PLANTING AND MAINTENANCE Many factors, including length of season, maximum and minimum temperatures, availability of moisture, soil composition, soiI-nutrient level, and the presence of nematodes and disease-producing organisms in the soil, must be correctly evaluated in planning the establishment of fruit and tree nut plantings. Management decisions made at this time are critical because of the high investment required during the initial growing period before the first commercial harvest. Maintenance methods for use throughout the long life of many perennial crops should be carefully planned to include the adoption of the new chemical weed control programs that will minimize cost and supervisory time. Selected examples of established methods of using herbicides and the results of continuing weed research are cited here to show the great impact of

188

W. B. ENNIS, JR. ET AL.

chemical weed control on farm-management practices in deciduous fruit and tree nut crops. 1, Nursery Plantings Weeds are a critical problem in nursery plantings because of the large numbers of plants grown in a very limited space. Weeds between the rows can be removed by mechanical cultivation, but in the past those between plants in the row were removed by hand. Hand-weeding costs have therefore represented large segments of the total nursery production costs. These high weeding costs have been the stimulus for the evaluation of many chemical and mechanical methods for weed control. Steam sterilization of greenhouse and field propagation beds was one of the earliest effective methods discovered. Costs of equipment and fuel and the time required for operation of the equipment limited the use of steam sterilization. Soil-fumigating chemicals including chloropicrin, formaldehyde, and ally1 alcohol largely displaced steam sterilization for a time, but these chemicals in turn have been displaced by more recent discoveries. Recently developed soil-fumigating chemicals, including methyl bromide, 3,5-dimethyltetrahydro-1,3-5,2H-thiadiazine-2thione ( DMTT), and sodium-N-methyldithiocarbamate ( SMDC ), control many weed species. Many additional herbicides are available for control of weeds that emerge after stock is set in the nursery. These include simazine, atrazine, sesone, 3-( p-chlorophenyl ) -1,l-dimethylurea ( monuron ) , diuron, CIPC, 2,2-dichloropropionic acid ( dalapon ) , DNBP, aromatic solvents, and many others from which selections are made for control of specific weeds in specific crops. A number of herbicides are available in granular formulations, and these can be used on established plantings to prevent emergence of weeds in the rows with greater safety to the crop than the use of postemergence sprays which may leave harmful residues on the crop plant foliage. These various chemicals and methods of use have greatly reduced the economic problem of controlling weeds in nurseries. Estimated annual costs of $500 to $1000 per acre for hand-weeding nursery crops are reduced to approximately $50 to $75 per acre when herbicides are used. 2. Small Fruits The control of weeds is the greatest single cultural problem in small fruits including strawberries, grapes, cane fruits, cranberries, and blueberries. The widely differing growth habits of these crops and their diverse soil, climatic, and cultural requirements contribute to complex weed problems that traditionally require highly specialized mechanical

IMPACT OF C2EMICAL WEED CONTROL

187

equipment, cultural techniques, and large amounts of hand labor for weed control. The perennial nature of small fruits necessitates a continuous weed control program both summer and winter in many areas. Chemical methods of weed control have reduced production costs by minimizing equipment needs, fertilizer requirements, and the number of mechanical field operations per season. a. Strawberries. Careful mechanical cultivation is safe and effective for a short period after new strawberry plantings are established. The growth of stolons, or runners, soon prevents close cultivation. Handhoeing or pulling of weeds is necessary when herbicides are not avail-

FIG. 9. Center, untreated strawberries overgrown with knotweed. Right, knotweed controlled in strawberrics by winter treatmcnt with sesone.

able. Cultivation and hand-weeding costs can be reduced more than 50 per cent by applying sesone 2 weeks after planting and thereafter at monthly intervals following clean cultivation for control of summer annual weeds including germinating knotweed ( Polygonum aviculare L. ), pigweed ( Anzaranthus rettroflexus L. ), lambsquarters ( Chenopodium album L. ) , carpetweed ( Mollugo verticillata L. ) , crabgrass ( Digitaria spp. ), goosegrass ( Eleirsine i d i c a ( L. ) Gaertn. ), and others. A critical spring infestation of knotweed in strawberries and the benefits of chemical weed control are shown in Fig. 9. Small-grain straw mulches often carry large numbers of cereal seeds that germinate during late winter

188

W. B. ENNIS, JR. ET AL.

and overgrow the strawberries (Fig. 10). Autoclaving or methyl bromide treatment will kill cereal seed in straw mulches, but the cost of these treatments is often prohibitive. A combination of CIPC and sesone applied immediately after mulching kills germinating cereal, chickweed ( SteZluria media ( L. ) Cyr. ), henbit ( Lamium ampZexicauZe L. ), annual bluegrass (Poa annuu L.), and knotweed seeds. Thus, growers

FIG. 10. Left, strawberries heavily infested with oats as a result of using straw mulch. Right, sesone and isopropyl N-( 3-chloropheny1)carbamate ( CIPC ), applied immediately after spreading straw mulch, controlled volunteer oats.

can use straw mulches without fear of a costly weed problem developing from the germination and establishment of cereal seeds. The annual cost of controlling weeds by hand and mechanical cultivation is $200 per acre in certain production areas. The approximate annual cost of the herbicide treatments mentioned is about $30 per acre and such treatments result in a saving of about $100 per acre. Improved herbicides and techniques of using herbicides are now in the developmental stages, and these can be expected to give improved weed control at even greater savings in strawberry production. Effective weed control methods may also make possible the planting of strawberries in wide beds because cultivation to control weeds would be unnecessary. b. Grapes. Grapes are grown commercially in a number of geograph-

IMPACT OF CHEMICAL WEED CONTROL

189

ical areas involving a wide variety of soils, climatic conditions, and cultural practices. Weeds that grow under the arbor and under the canopy of foliage interfere with pest and disease control and reduce yield and quality. Specialized mechanical equipment such as the grape hoe has been developed, but its use is time-consuming; careless use can damage grape stalks and roots; and weeds under the arbor are

FIG. 11. Vineyard heavily infested with johnsongrass in San Joaquin Valley of California. Annual and perennial weeds in the row of vineyards are difficult to control by hand and mechanical methods. Perennial weeds such as johnsongrass can now be effectively and economically controlled by dalapon treatment. (U. S. Bureau of Reclamation photograph.)

not controlled. Herbicides are highly efficient and are convenient and economical to use. Single applications of diuron after spring cultivation or split applications after spring cuItivation and after the Iast cultivation in the fall control many annual broadleaved weeds and weed grasses. Band applications of a mixture of diesel fuel, CIPC, and DNBP in the spring control many annual weeds. Perennial grasses, including johnsongrass and bermudagrass, are critical weeds that can now be controlled with directed applications of dalapon. The critical problem caused by johnsongrass in grape production is illustrated in Fig. 11.

190

W. B. ENNIS, JR. ET AL.

c. Blueberries. New plantings of blueberries are quickly overgrown by annual and perennial weeds unless effective control measures are used. Intensive repeated cultivations and hand-weeding are helpful, but their costs are prohibitive. Weeds in established and producing plantings are controlled with diuron applied in late fall or early spring. Diuron combined with DNBP kills the tops of growing weeds quickly and gives long-term control of germinating weeds. Most perennial grasses in blueberries can be controlled by adding dalapon to the diuron spray. Brush weeds are particularly difficult to control, but careful spot sprays with 3-amino-1,2,4triazole (amitrole) will kill them. Broadleaved perennial weeds are controlled with the amine salt of 2,4-D. Established and germinating annual winter weeds such as chickweed and henbit are controlled by a combination of 3,5-dinitro-o-cresol (DNC) and CIPC. The efficient control of weeds in blueberries increases production and reduces labor requirements substantially. d . Cranberries. Brush, sedges, bracken fern, and grasses are the most serious weed problems in cranberries. Weed control is especially difficult because cultivation is impossible and the only mechanical control is mowing. Chemical control of many of these weeds is now possible with selected aromatic solvents, copper sulfate, kerosene, and amitrole. A number of new chemicals including simazine, 2,6-dichlorobenzonitrile ( dichlobenil), and 2,3,5,6-tetrachloroterephthalicacid hold promise for further rapid advances in the control of weeds in cranberries. The effective use of chemicals to control weeds in cranberries will facilitate mechanical harvesting, reduce the hand labor presently used at processing plants to remove weed debris, and improve the quality of the fruit because of reduced handling. e. Cane fruits. Mechanical control of weeds in cane fruits including raspberries, blackberries, currants, and gooseberries requires repeated cultivations throughout the growing season. Great care must be exercised in cultivation to avoid injury to the shallow root system of cane fruits. Cultivation is ineffective in controlling weeds between plants in the row and under the spreading foliage as the growing season advances. The use of herbicides makes possible the control of weeds under the canopy of foliage without the root injury that usually results from mechanical weed control measures. Usable herbicides include DNBP, 2,4-D, diuron, CIPC, sesone, simazine, and dalapon, selected and applied for the control of specific weeds. 3. Tree Fruits and Nuts Tree fruit and nut crops, including apples, pears, peaches, plums, prunes, citrus, walnuts, almonds, pecans, and tung, are grown by using

IMPACT OF CHEMICAL WEED CONTROL

191

clean cultivation, sod culture, or combinations of these methods with sod or cover crop strips between the trees combined with narrow-strip cultivation under the trees. Close cultivation injures shallow roots and cultivating equipment carelessly used often injures the bark on the lower trunks, permitting entrance of disease-producing organisms and reducing general vigor of the trees. Close cultivation does not control all weeds under the trees and the remaining weeds serve as hosts for nematodes, disease-producing agents, and insects and provide cover for rodents that damage the tree roots and trunks. For example, borers feeding on curly dock (Rumex crispus L . ) under apple trees in contact with the lowhanging boughs have been observed migrating and feeding on apple fruits. Large populations of European red mite have been observed on weeds growing under peach trees. Many types of orchard crops are grown on steep slopes where cultivation is difficult and causes severe erosion. Sod culture is possible and preferable in these locations because chemicals can now be used to control brush. Herbicides including diuron, monuron, simazine, amitrole, dalapon, petroleum oils, 2,4-D, and 2,4,5-T are being used to control weeds in the various types of orchard culture and continuing research is developing new chemical methods. Clean cultivation is being supplanted or accompanied by applications of monuron, diuron, simazine, and petroleum oils on flat lands. Sod culture is facilitated by controlling broadleaved weeds with 2,4-D, perennial grasses with dalapon, and brush with amitrole.

B. HARVESTING The development of effective chemical methods of controlling weeds is closely linked to the rapid advances in the mechanization of harvesting of fruits and nuts. Mechanical aids such as elevators and supports for fruit containers assist in the handpicking of large fruits including apples, pears, and peaches. A smooth soil surface, wide spacing of trees, and good weed control are necessary in the use of mechanical equipment in general. Mechanical shakers combined with catching or pick-up equipment are used extensively in harvesting small tree fruits and nuts. Herbicides provide the necessary control of weeds and brush under the trees to ensure optimum performance of harvesting equipment. The use of shaking and pick-up equipment for harvesting prunes, walnuts, and blueberries is an excellent example of the interdependence of chemical weed control and mechanization in the fruit and nut industry. Hand harvesting of tree fruits and small fruits, including cranberries,

192

W. B. ENNIS, JR.ET AL.

blueberries, cane fruits, and strawberries, is often impeded by the presence of poisonous or thorny weeds. Poison ivy is one of the most widely distributed weeds among tree and small fruits. Spiny pigweed (Amarunthus spinosus L. ), horsenettle ( Solanurn curolinense L. ) and sandburs (Cenchms spp.) are also found in certain of the fruit crops in some areas. Hand pickers often object and sometimes refuse to work in infested areas. Poison ivy in apple orchards is easily controlled with amitrole without injury to the trees. Herbicides now being developed will do much to alleviate further the problem of controlling poisonous and thorny weeds in these crops. Efficient, economical control of weeds is obtained with herbicides, and such control reduces damage from diseases, nematodes, insects, and rodents, and facilitates harvesting of many crops. IV. Production of Pastures and Rangelands

In continental United States 526 million acres are “improved and “open pasture” (Wooten and Anderson, 1957), and many of these are weedy. It is estimated that brush removal would improve grazing on 240 million acres (Anonymous, 1952). In addition, millions of acres are infested with weed grasses and other herbaceous weeds that displace valuable forage species and reduce production and quality of forage, cause poisoning and mechanical injury to livestock, and reduce the quality of milk and meat products from the grazing animals.

A. ESTABLISHMENT OF NEW SEEDLINGS Weeds that cause reduced stands, weak plants, and failures in the establishment of forage crops can now be controlled by specific selective herbicides. These herbicides offer outstanding promise as aids in improving forage crops on vast areas of agricultural lands now largely unimproved. New chemicals, such as 4- ( 2,4-dichlorophenoxy ) butyric acid [4-( 2,4-DB) 1, that can kill selectively broadleaved weeds in grasses and legumes are available. Herbicides such as dalapon control grasses in certain legumes and a few herbicides even remove weed grasses from some other grass crops. Also, certain chemicals kill selectively unwanted legumes in other legume crops. These herbicides offer tremendous potentials for reducing the hazards in establishing forage species and increasing the production and quality of the forage. Before selective herbicides were available, fall planting of forage crops was widely recommended in the humid and semihumid areas of the United States. Fall planting, in many areas, was mainly advantageous because weeds emerging with the legume or grass species were soon

IMPACT OF CHEMICAL WEED CONTROL

193

killed by the first fall frost. The established forage plants then would attain sufficient growth to compete successfully with weeds that emerged in the spring. Thus better stands of legumes and grasses often resulted even though favorable rainfall distribution was usually less reliable in the fall. However, a serious problem was frequently encountered in fall plantings from devastating damage caused by insect pests such as grasshoppers. Small seedlings of forage plants were attacked in the spring less frequently. The improved chances for successful establishment of forage crops such as birdsfoot trefoil seeded in the spring by using selective postemergence herbicides have been demonstrated. For example, combinations of dalapon and mowing resulted in the establishment of twice as many birdsfoot trefoil plants as the best mowing treatments (Kerr and Klingman, 1960). More important was the tremendously greater yield in the year after treatment. The best combination of dalapon and mowing treatments during the seedling year resulted in 3860 pounds of birdsfoot trefoil dry matter per acre in the following year. Birdsfoot trefoil in plots receiving the best mowing treatment without supplemental dalapon yielded only 140 pounds per acre. Dalapon effectively controlled grasses without noticeable effect on birdsfoot trefoil. More recently postemergence sprays of 4-(2,4-DB) have been better than mowing for broadleaved weed control in that none of the seedling legume is removed .by the weed control treatment. In New York, use of dalapon plus 4-( 2,4-DB) resulted in a yield of 2100 pounds per acre greater than that of the check in the year of seeding the birdsfoot trefoil. Equally promising for weed control is mixing EPTC in the soil before alfalfa, red clover, birdsfoot trefoil, and other legumes are planted ( Shafer, 1958; Peters and Davis, 1958). This herbicide controls grasses for 6 weeks, and also many broadleaved weeds, without injury to most legume species. The use of improved selective herbicides makes feasible the seeding of legumes in the spring where such planting is desirable. Good yields of high-quality forage can be harvested the seeding year. To help control weeds, companion crops have been planted with spring seedings of forage grasses and legumes. Recent research shows that companion crops such as oats and barley injure the slow-starting perennial forage species as much as the weeds that almost always grow if a companion crop is not planted (Peters, 1961).Since selective herbicides have been developed, present cultural practices, including recommended use of companion crops for establishing new seedlings of forage crops, should be reevaluated in many areas. Chemical weed control hastens establishment of coastal bermudagrass

194

W. B. ENNIS, JR.ET AL.

in pastures. Without the use of herbicides stands of newly sprigged pastures may be reduced or even lost because of severe competition from annual grasses and broadleaved weeds. In addition, the forage production from the poor stands is low, A number of herbicides control selectively these weeds and allow a better turf to become established more quickly. In Mississippi, for instance, coastal bermudagrass sprigged in May and sprayed immediately with simazine at 4 pounds per acre yielded 3.75 tons per acre dry matter and no weeds, as compared with the yield of 1.9 tons per acre bermudagrass and 1.6 tons weeds from the untreated check. The following year the treated area yielded 9.5 tons per acre, whereas the untreated yielded only 6.0. Obviously this weed control practice makes possible better utilization of fertilizer, earlier grazing, and higher quality forage. It is clear that herbicides have altered practices in establishing stands of forage crops on prepared seedbeds. Herbicides also are important in upgrading grazing areas not suited to plowing. B. PASTURE RENOVATION Herbicides are promising substitutes for plowing on many pastures that are too steep, too rocky, or too poorly drained to be adapted to plowing in the preparation of seedbeds. Yet, some of these areas are highly productive of improved forage when the fertility needs are met and improved species established. Sprague ( 1959) reported that disking, reseeding, and subsequently fertilizing hilly grazing lands have increased yields 2 to 6 times: “Truly, renovation of our humid area pastures represents agriculture’s greatest undeveloped resource.” Three requirements for stand establishment were outlined: ( 1 ) killing all undesirable plants and preventing their vegetative reestablishment, ( 2 ) making soil corrections by intelligent use of lime and fertilizer, and (3) providing good seed-soil contact at the proper depth for good emergence. One further requirement (suppressing competing vegetation while forage seedlings are becoming established) should be added. Chemical renovation removes competing plants and permits surface seeding or seeding with a minimum of soil disturbance. After 3 years of evaluation in New Jersey, TCA plus 2 diskings was equal to plowing to convert pasture to more desirable species and to increase yield of highquality forage. Both plowing and chemical renovation treatments were considerably better than 12 diskings for seedbed preparation, as evidenced by the higher yield of forage. However, the use of all three methods more than doubled the forage yields from the untreated area.

IMPACT OF CHEMICAL WEED CONTROL

195

The potential for chemical renovation has increased with the development of newer herbicides more efficient than TCA. Combination of amitrole at 1 pound per acre and dalapon at 4 pounds per acre is particularly promising when applied to closely grazed sods during the growing season and 3 weeks before seeding. The control of weeds during establishment of small-seeded forage species is equally important whether the new seedlings are on prepared seedbeds or on soils that have been given chemical renovation. Postemergence sprays of 4- ( 2,4-DB) control broadleaved weeds. Dalapon plus 4-( 2,4-DB) controls both broadleaved and grass weeds in new seedings of alfalfa or birdsfoot trefoil.

C. CONTROL OF BRUSII ON GRAZING LANDS Successful use of herbicides has converted millions of acres of brushinfested rangelands to productive grasslands. Additional extensive brushinfested rangelands can be improved only by removing the brush canopy and reducing the competition that brush species provide. For instance, in eastern Texas, Oklahoma, Kansas, Missouri, and Arkansas, ranchers kill blackjack and post oaks (Quercus spp.) with 2 or 3 annual aerial sprayings of an ester of 2,4,5-T (Elwell et al., 1945; Darrow, 1956). Where there is an understory of grasses, the control of brush increased production of forage 2 to 7 times (Fig. 12). The rate of recovery of these grasses is almost unbelievable if they are protected from grazing for one or two growing seasons. Similar responses of forage species occur in other geographic areas after removal of other brush species. Chemical control of 88 per cent of shinnery oak ( a low-growing Quercus sp.) resulted in an increase in oven-dry forage of 1680 pounds per acre near Woodward, Oklahoma. Here and in other western States forage production was increased severalfold from spraying various sagebrush species ( Artemisiu) with 2,4-D (Alley and Bohmont, 1958; Hyder and Sneva, 1956), from spraying 2,4,5-T or mesquite (Prosopis) and associated species on the arid ranges of the Southwest (Cable and Tschirley, 1961; Fisher et al., 1959; Parker and Martin, 1952), and from control of chaparral species on mountain slopes. Good management of livestock is virtually impossible on rangeland extensively infested with large mesquite or other trees. Controlling the brush with modern herbicides and other techniques facilitates herding the cattle when necessary, regularly inspecting and treating the herd for disease and insect disorders, and orderly culling and marketing of the excess cattle. Increased income from the improved care of livestock made possible by controlling the brush pays for a large part of the cost of spraying.

196

W. B. ENNIS, JR.ET AL.

Control of brush also serves as land preparation for seeding more valuable forage species. Crested wheatgrass is often successfully drilled without further preparation after sagebrush control. In California, seedings of smilograss (Oryzopsis rniliacea (L.) Benth. and Hook.), and Harding grass (Phalaris tuberosa var. stenoptera ( Hack. ) Hitchc. )

FIG.12. Control of post and blackjack oaks on rangeland in Oklahoma by aerial spraying with 2,4,5-T. An ester form of 2,4,5-T was applied at 2 pounds per acre in 1952 and again in 1953. Although the range was grazed in 1953, recovery of the grass understory was evident when photograph was taken in September 1953. ( Leonard and Harvey, 1956), or Lehmann lovegrass ( Eragostis Zehmanniana Nees) in Arizona were successful after burning and where chaparral sprouts were controlled by spraying with 2,4-D and 2,4,5-T. In Louisiana, controlling brush by spraying 2,4,5-T, fertilizing with basic slag, and seeding white clover, serecia lespedeza, and bahiagrass resulted in a good stand of improved forage species and increased the

IMPACT OF CHEMICAL WEED CONTROL

197

forage eaten by cattle from 2.1 tons per acre on the untreated area to 3.9 tons per acre on the treated. Equally important, protein content of the forage on the treated area was 10.5 per cent as compared with 6.4 on the untreated. Similarly in Missouri, seeds of Korean lespedeza were broadcast in a sparse stand of Kentucky bluegrass where individual hickory ( C a y a sp.) and associated trees had been killed by spraying their bases with an ester of 2,4,5-T in a solution of diesel fuel oil. Cattle ate 0.2 ton per acre of air-dry forage on the untreated areas and 1 ton on areas where trees had been basally sprayed. Fertilization and supplemental weed control would have increased the response observed.

D. CONTROL OF HERBACEOUS WEEDS Every pound of herbaceous weeds grown in pastures reduces the production of more desirable forage plants by about an equal amount. The reduction in forage attributable to weeds varies; it depends upon moisture conditions and the weed and pasture species involved. Losses due to weeds in pastures are tremendous. For instance, in one eastern Nebraska pasture over two-thirds of the total production of plant material consisted of unpalatable weeds. The 3-year average yield of dry matter in the pasture was 2,.1tons. Of this, 1.15tons was broadleaved weeds and 0.35 ton was weed grasses (Klingman and McCarty, 1958). Weeds are controlled in pastures more efficiently than ever before. The potentialities of 2,4-D alone or in combination with other agronomic practices for controlling weeds in pastures were clearly shown in studies at several locations in humid and semihumid areas. In Nebraska 2,4-D was much more efficient than mowing for controlling broadleaved weeds and contributed to successful management and reseeding of parts of the pasture after plowing (Klingman and McCarty, 1958). Three annual sprayings with an ester of 2,4-D at 1 pound per acre reduced the stand of weeds 70 per cent; mowing reduced these same weeds only 34 per cent; plowing and reseeding plus annual sprayings with 2,4-D controlled more than 90 per cent. Cattle were attracted to grazing the treated areas because unpalatable weeds were greatly reduced and the density of desirable grasses was increased (Fig. 13).Control of weeds increased the amount of vegetation eaten by cattle. In untreated pasture areas cattle ate an average of only 0.55 ton of dry matter per acre. In areas mowed for weed control cattle ate 0.65 ton; in the areas sprayed with 2,4-D, 0.85 ton; and in the reseeded and sprayed areas, 1.4 tons. Cattle ate about 20, 50, and 150 per cent more forage, respectively, in the treated than in the untreated. The ecology and vegetational composition of pastures are vastly

FIG.13. Comparative weed control in pasture attained by mowing and spraying with 2,4-D. Above, mowed early in June each of two years and grazed continuously; little grass remained for grazing and weeds were 8 to 24 inches tall. Below, sprayed early in June each of two years with an ester form of 2,4-D; more weeds were controlled and more grass was available for grazing than in mowed area. 198

IMPACT OF CHEMICAL WEED CONTROL

199

changed by use of selective herbicides. In Mississippi spraying with 2,4-D decreased weed plants from 650 to 1 per square yard and increased grass plants from 35 to 100 and legume plants from 10 to 90 per square yard ( Harris, 1952). The 2,4-D controlled tarweed (Helenium sp. ), boneset ( Eupatoriuni perfoliatum L. ), Cassia spp., evening primrose (Oenotheru sp.), and others. The herbicide did not injure the grasses or permanently damage legumes. Late spring applications gave the best results. More recently the development of the technique of winter treatment provided farmers with opportunities to use herbicides for bitterweed control in pastures in the South whereas they were reluctant or unable to do so earlier because of hazard to cotton in adjacent fields. Such annuals as bitterweed are controlled by preemergence treatments with 2,4-D applied in late winter without hazard to cotton crops. Grazing and quality of forage were greatly increased on summer range for cattle and sheep on a mountain meadow in northeastern California where 2,4-D controlled plantain leaf buttercup ( Ranunculus sp. ). Meadow vegetation was increased in density and yield concurrently with decreases of weeds susceptible to 2,4-D ( Cornelius and Graham, 1953). Grass and lespedeza production were increased threefold on an overgrazed pasture in Oklahoma by controlling weeds. Sprays of 2,4-D killed ragweed and other troublesome weeds where lespedeza was growing without causing severe injury to the lespedeza (Elder, 1951). OF WEEDSON ANIMALSAND ANIMALPRODUCTS E. EFFECTS

Losses due to weeds are not limited to the reduction in palatable forage available to livestock. They cause losses through imparting undesirable flavors to dairy products and causing mechanical injury to grazing animals and death from poisonous species. Losses in animal production caused by sublethal poisoning often go undetected. For instance, a bizarre abortion problem in beef and dairy cattle grazing on unimproved, lowland pastures went unsolved for at least fifty years in Wisconsin (Simon et al., 1958).Research showed that abortions were caused by weeds that contained sublethal amounts of nitrate. Spraying these weeds with 2,4-D solved the abortion problem (Sund and Wright, 1957). Losses of cattle from larkspurs (Delphinium spp. ) poisoning averages about 5 per cent of herds grazing on heavily infested ranges. Use of herbicides is beginning to reduce these losses. 2- ( 2,4,5-Trichlorophenoxy )propionic acid (silvex) and 2,4,5-T applied to the vegetative growth just before the bud stage effectively control tall larkspur (D. barbeyi Huth) (Hervey and Klinger, 1961). Bigpod larkspur is con-

200

W. B. ENNIS, JR. ET AL.

trolled by spraying 2,4-D on plants in the rosette stage in Oregon. Other species may be controlled similarly. Controlling larkspur species reduces manpower required to herd cattle away from infestations and makes possible the efficient use of ranges where formerly limited grazing was necessary. Wild onion (Allium canadense L.) and wild garlic ( A . vineale L.) cause off-flavors in milk that reduce its quality and value, and they sometimes cause off-flavors in meat when livestock graze them just prior to slaughter. Spraying pastures annually with 2,4-D can virtually eliminate this problem.

F. WATERCONSERVATION The selective chemical control of vegetation on some ranges has increased water yield from springs and stream flow (Biswell and Schultz, 1957). For instance, conversion of large acreages of California brushland to grass resulted in an increase in water available. Such increase is important for livestock water and possibly as a source of irrigation water. Equally important is the increased amount of water stored in the soil. Sonder and Alley (1961) reported significantly more soil moisture in July in Wyoming where sagebrush ( Artemisia sp. ) was controlled. V. Changes in Management of Farm Water Systems

Weeds that grow in or along irrigation and drainage canals or ditches and in or around farm ponds and reservoirs are among the problems most costly to the farmer. They hinder and often prevent adequate delivery of irrigation water to, or drainage of excess water from, the land (Fig. 14), and they greatly reduce the usefulness of farm ponds. In water-deficient regions emersed aquatic weeds, bank weeds, and certain phreatophytes, plants growing with their roots in wet soil, waste large quantities of precious water by transpiration into the dry air. Mowing and other mechanical methods are costly, slow, and of only limited effective use. Recently, the development of effective and economical herbicides for controlling aquatic and bank weeds made possible much more efficient use of canals and ditches and aided greatly in delaying or preventing the natural process by which farm ponds are converted into swamps and finally into wet waste land. A recent survey showed that in 1957 weed control practices, consisting mostly of chemical methods, resulted in estimated reductions in water and other losses amounting to 7.75 million dollars on irrigation and drainage systems in the 17 western States (Timmons, 1960). Because of the recent increased interest of commercial and public agencies in research on control of

IMPACT OF CHEMICAL WEED CONTROL

201

FIG. 14. Weed problems in western irrigation canals. Above: American pondweed causes water to overflow canal banks. Below: reed canarygrass reduces carrying capacity of small irrigation lateral. (Bureau of Reclamation photographs. )

aquatic and bank weeds, the immediate future promises greatly improved herbicides for controlling these weeds. New improved chemical methods probably will further revolutionize the management of water channels and water storage systems on farms. A. DESIGN AND CONSTRUCTION OF WATER CHANNELS, ROADWAYS,AND S T R U ~ ~ U R E S The progress made thus far in developing effective herbicides for aquatic and bank weeds emphasizes the need for access roads on one or both sides of irrigation and drainage canals to permit the use of herbi-

202

W. B. ENNIS, JR. ET AL.

cides over the entire right-of-way. A need has developed also for cattle guards or gates through cross-fences and for bridges over cross-ditches to permit unhindered or continuous travel of vehicles along each channel. Bridges, checks, drops, and other structures frequently interfere with the operation of mowing, chaining, and certain other mechanical methods of removing aquatic and bank weeds. These structures cause much less interference with the use of most chemical methods of weed control. Thus, more extensive and effective weed control can be practiced on irrigation and drainage systems than was previously possible or economical. Most channels have been constructed to carry specific volumes of water, usually without adequate allowances for the effects of weeds. The volumes of flow in many of these channels have been reduced much below the designed capacities by weed growth and the resulting silt deposits. Control of weeds by dependable chemical methods helps to restore the flow to designed capacities in existing waterways and to eliminate the necessity for reconstructing larger channels to allow for expected weed growth. Also, there is less need for constructing larger new channels to allow for reduced flow capacities caused by weeds. In recent years the use of membrane and other low-cost linings in irrigation canals, ponds, and reservoirs for the prevention of seepage losses has increased rapidly. An important problem encountered is rupturing of the lining by the shoots of vigorous perennial weeds such as johnsongrass, nutgrass, and woody plants. The problem has been partially solved by chemical treatments to eliminate these weeds and by the application of soil-sterilant herbicides before installation of the linings. Chemical control of submersed aquatic weeds eliminates the necessity for use of mechanical methods of weed control which frequently damage the canal linings. Future development of improved herbicides promises to facilitate further the use of membrane and other low-cost linings in irrigation channels and farm ponds.

B. CHEMICAL VERSUS MECHANICAL METHODS During the past ten years, chemical methods of controlling aquatic and bank weeds in and along irrigation and drainage channels and farm ponds have replaced mechanical methods to a considerable extent. The chemical methods usually are more effective, more convenient, less time consuming, and often less expensive (Fig. 15). Generally, control of submersed weeds with aromatic solvents or acrylaldehyde ( acrolein ) is less expensive than mechanical methods such as chaining or dragging in canals with capacities of 70 cubic feet per second (c.f.s. ) or less. Where

IMPACT OF CHEMICAL WEED CONTROL

203

I 15. ( A ) An irrigation canal with a heavy growth of pondweed, anacharis, and ,.,gae before treatment. ( B ) Eight days after treatment with aromatic solvent. (Bureau of Reclamation photographs.)

204

W. B. ENNIS, JR. ET AL.

sprinkler irrigation is extensively practiced, chemical control of aquatic weeds may be economically used to replace mechanical methods in larger canals with capacities up to 200 c.f.s. Masses and fragments of aquatic weeds dislodged by mechanical devices cause frequent clogging of sprinkler heads, valves, and pumps (Fig. 16). Aquatic weeds killed by chemicals disintegrate slowly and do not cause clogging. In the Columbia Basin of Washington new techniques have been used recently for con-

FIG.10. Removing submersed aquatic weeds from an irrigation canal by chaining, Above, two crawler tractors drag a ship ancher chain to break weeds loose from bottom of canal. Below, the loosened waterweeds float down the canal and must be removed by hand raking or a mechanical device such as a dragline. (Bureau of Reclamation photographs.)

IMPACT OF CHEMICAL WEED CONTROL

205

trol of submersed waterweeds with xylene in canals with capacities as great as 300 to 500 c.f.s. The use of 2,4-D in oil-water emulsion, amitrole, and dalapon for control of cattails (Typha spp.), tules, and other emersed aquatic weeds in channels and ponds has largely superseded mechanical methods, such as mowing and draglining. Timely and regular use of herbicides de-

FIG. 17. Hand-weeding and mowing require more manpower than chemical methods. Above: hand-scything pondweeds in a canal. These loosened weed mats must then be removed by hand at a point further down the canal. Below: after mowing reed canarygrass, the mowed grass must be removed from the canal by hand. ( Courtesy Bureau of Reclamation, Region 1, Boise, Idaho. )

206

W. B. ENNIS, JR.ET AL.

creases the cost of controlling weeds and results in less frequent need for removing silt deposits by mechanical methods. The Bureau of Reclamation reported (Timmons, 1958) that the average cost of chemical control of cattail ranges from $25 to $43 per mile of channel as compared with $407 to $418 per mile for draglining. Spraying with 2,4-D, silvex, and 2,4,5-T reduced, but did not eliminate, mowing, burning, bulldozing, and other mechanical methods of controlling bank weeds and phreatophytes along channels, around ponds, and on flood plains. Chemical methods usually are more convenient and effective. However, mechanical methods often remove the vegetation more rapidly, are sometimes less expensive, or may be less hazardous to crops on nearby farmland. Furthermore, the development of less expensive or more selective herbicides in the future undoubtedly will extend the replacement of mechanical methods. One of the most important advantages of controlling aquatic and bank weeds with herbicides is the considerable saving in manpower on farms and in irrigation and drainage districts that serve the farms. Hand and mechanical methods of controlling such weeds usually require much more manpower than do chemical methods (Fig. 17). In areas and seasons of labor shortage, the saving in man hours and the more rapid accomplishment of weed control by the use of herbicides are important considerations aside from the relative costs. Considerable manpower and machinery are released for other work. Despite the many advantages of chemical methods, mechanical methods are still better adapted for use in some situations (Boyle and Suggs, 1960) and must be depended upon to a large extent in areas where cotton, grapes, tomatoes, and other crops highly sensitive to 2,4-D, silvex, 2,4,5-T and other phenoxy-type herbicides are extensively grown. However, amitrole, dalapon, and aromatic or fortified weed oils can be used for control of weedy grasses in such areas. Some mechanical methods such as chaining and underwater mowing provide effective control of emersed weeds like cattail as well as submersed weeds. Chaining or dragging removes or smoothes down silt bars or hummocks and compacts bottoms and sloping banks of irrigation canals and thus reduces seepage losses of water. However, the compacting and sealing action does not occur where rocks or coarse gravel are present.

c.

IMPROVED

MANAGEMENT OF WATER

SYSTEhZS

The effective herbicides now available for control of submersed aquatic weeds are, in most situations, too expensive for use in irrigation and drainage canals with flow greater than 200 c.f.s. An exception is

IMPACT OF CHEMICAL WEED CONTROL

207

the economical use of copper sulfate in California and Washington for control of algae in canals with capacities as large as 4000 c.f.s. In recent laboratory and field experiments several chemicals when applied to the bottoms of canals have shown promise of giving satisfactory control of submersed aquatic weeds for one year or more. Such chemicals, applied on an area rather than a volume basis, might prove economical for use in large canals where submersed aquatic weeds are causing, in many localities, an increasingly serious threat to the adequate and timely delivery of irrigation water to farmland. The use of selective herbicides such as 2,4-D, silvex, and 2,4,5T for control of broadleaved weeds and brush along channels and around farm ponds greatly aids the establishment of desirable low-growing grasses for stabilizing banks and preventing erosion and sloughing. The dense stands of grasses also prevent or reduce the encroachment of undesirable bank weeds. A recent study in Washington indicates that regular use of copper sulfate at frequent intervals for control of algae in irrigation canals encourages the spread of a dwarf species of water plantain (AZismu grumineum K. C. Gmel.), which competes with and eliminates or reduces the troublesome rank-growing pondweeds ( Potamogeton SPP. ) . The extensive and intensive use of herbicides on irrigation and drainage systems has greatly increased the economy and efficiency of management in many areas. An example of such results is reported by the Imperial Valley Irrigation District in California (Timmons, 1958). The more complete and regular weed control by herbicides has reduced the time required for inspection and for cleaning clogged structures, removing weed jams, or repairing ditch breaks and similar maintenance work. Even more important is the more adequate and timely delivery of irrigation water to crops and the more prompt drainage of excess water from farmland. Where chemicals such as xylene and acrolein are properly applied, the treated water may be used as furrow irrigation without injury to crops. Thus, chemical treatment causes no loss of irrigation water or delay in irrigation such as that necessary when water is turned out of canals for several days to control submersed aquatic weeds by drying.

D. CHOICEOF IRRIGATION AND CROPPING PRACTICES The use of herbicides for controlling submersed weeds in irrigation canals makes possible the use of sprinkler irrigation without risk of clogging the pump valves and sprinkler heads with floating masses or fragments of aquatic weeds such as those dislodged by chaining or other mechanical operations. Chemical weed control has been an im-

208

W. B. ENNIS, JR.ET AL.

portant factor in the rapid increase in the use of sprinkler irrigation whereby expensive land leveling and ditching are avoided. Dependable weed control on irrigation systems with herbicides may eventually result in the acceptance of devices for automatic measuring and delivering of water for crop production with a considerable saving in time now required for operating the turnout gates. The principal deterrent to this manpower-saving practice is the risk of clogging the measuring devices and automatic gates by floating weeds and the irregular flow of water caused by weed growth. The general use of herbicides for controlling weeds on irrigation and drainage systems and in farm ponds in some areas has had important effects on the management of crop lands. The more ample and regular supplies of irrigation water have permitted greater latitude in the choice of crops and cropping practices. In other areas, the more efficient drainage of excess water through weed-free drainage ditches has also permitted a wider choice of crops and cropping practices. On the other hand, the differential tolerance of crops to herbicides used for control of bank weeds sometimes determines which crops may be safely grown adjacent to irrigation or drainage channels or, conversely, which herbicides may be safely used on ditches adjacent to certain crops. Consideration must also be given to crops that may be safely irrigated with water from canals or ponds treated with herbicides for weed control. In some areas more complete and timely control of bank weeds along channels and around ponds by herbicides has eliminated or greatly reduced the production of weed seeds which may be carried to cropland by wind or irrigation water. Destroying these important sources of weed seeds has reduced the need for weed control on cropland. VI. Summary

Chemical weed control is having a far-reaching impact on all phases of crop production. New chemical, cultural, mechanical, biological, and combination methods of weed control will affect crop choice; the variety to use; seedbed preparation; method of seeding; seeding rates; row spacing; planting spacing in the row; plant populations; fertilizer practices including type, time of application, and placement; cultivation; irrigation practices; harvesting; seed-cleaning operations; erosion control; fallow practices for weed control; disease and insect control practices; pasture renovation; pasture and range management; clearing new lands for crops or pasture; forest management; the utilization of farm water resources for irrigation and recreation; and the maintenance of drainage ditches, ditchbanks, irrigation canals, and farm roadsides.

IMPACT OF CHEMICAL WEED CONTROL

209

The use of herbicides to control weeds is removing one of the most important obstacles to completely mechanized crop production. Chemical weed control is no longer a supplemental technique, but constitutes a sound farm practice which is replacing obsolete hand and cultural methods. Chemical weed control practices possess great potential to reduce man hours, permit minimum tillage, reduce equipment horsepower requirements, and reduce the cost of crop production. In 1962 more than 85 million acres of agricultural land in the United States were treated with herbicides to control weeds at a cost of more than 200 million dollars for the chemicals alone. Chemicals are being used increasingly to control weeds, with a corresponding decrease in conventional methods of weed control. Present trends indicate that the use of herbicides will continue to increase. Chemical weed control methods will greatly affect the farm equipment industry and necessitate some redirection of equipment engineering objectives. Centuries-old hand methods of weed control which required large labor forces are being replaced by chemical methods. These new methods are on the threshold of wide-scale use. Skilled equipment operators with better training to understand the application of potent herbicides in minute quantities will be needed to replace the hand laborers of the past century. The rapid advance of weed control technology and the acceptance of chemical weed control practices by farmers will have profound effects on crop production and soil and water management practices. These significant changes will necessitate a continuing review of agricultural research and educational program objectives and direction. REFERENCES Alley, H. P., and Bohmont, D. W. 1958. Wyoming Agr. Expt. Sta. Bull. 354. Anonymous. 1952. Armour’s Analysis. 1, 3. Biswell, H.H.,and Schultz, A. M. 1957. Calif. Agr. 11( lo), 3-4,10. Boyle, L. W., Hauser, E. W., and Thompson, J. T. 1958. Weeds 6, 481-484. Boyle, W. D., and Suggs, D. D. 1960. U. S. Bur. Reclamation Mimeo. Publ. 1-18. Cable, D. R., and Tschirley, F. H. 1981. J . Range Management 14, 155-159. Cornelius, D. R., and Graham, C. A. 1953. J . Forestry 51, 631-634. Darrow, R. A. 1956. Proc. Southern Weed Conf. 9, 109-112. Elder, W.C. 1951. Oklahoma Agr. Expt. Sta. Bull. B-369,11. Elwell, H. M., Elder, W. C., Klingman, D. L., and Larson, R. E. 1954. Proc. N . Central Weed Control Conf. 11, 91-95. Fisher, C. E., Meadow, C. H., Behrens, R., Robinson, E. D., Marion, P. T., and Morton, H.L. ISSO. Texas Agr. Expt. Sta. Bull. 935. Harris, V. C. 1952. U . S. Dept. Agr. Mimeo. News Release 978-52,1. Harris, V. C. 1980. Weeds 8, 818-624.

210

W. B. ENNIS, JR. ET AL.

Hervey, D. F., and Klinger, B. 1961. Research Rept., Western Weed Control Conf. p. 12. Hodgson, J. M. 1958. Weeds 6, 1-11. Holstun, J. T., Jr. 1963. Weeds 11, 190-194. Holstun, J. T., Jr., Wooten, 0. B., Jr., McWhorter, C. C., and Crowe, G. B. 1960. Weeds 8, 232-243. Hyder, D. N., and Sneva, F. A. 1956. J . Range Management 9, 24-38. Kerr, H. D., and Klingman, D. L. 1960. Weeds 8, 157-167. Klingman, D. L., and McCarty, M. K. 1958. U . S. Dept. of Agr. Tech. Bull. 1180, 1-49.

Leonard, 0. A., and Harvey, W. A. 1956. Calif. Agr. Expt. Sta. Bull. 755. McWhorter, C. G., and Wooten, 0. B. 1961. Weeds 9, 45-49. Miller, J. H., Kempen, H. J., Wilkerson, J. A., and Foy, C. L. 1961. Weeds 9, 273-281.

Parker, K. W., and Martin, S. C. 1952. U . S. Dept. Agr. Circ. 908. Peters, R. A. 1961. Agron. J. 35, 195-198. Peters, E. J., and Davis, F. S. 1958. Proc N . Central Weed Control Conf. 15, 116. Peters, E. J., Klingman, D. L., and Larson, R. E. 1959. Weeds 7, 449-458. Peters, E. J., Davis, F. S., Klingman, D. L., and Larson, R. E. 1961. Weeds 9, 639-645.

Phillips, W. M. 1961. U. S. Dept. Agr. Tech. Bull. 1249, 1-30. Phillips, W. M., and Timmons, F. L. 1954. Kansas Agr. E x p t . Sta. Bull. 366, 1-40. Shafer, N. E. 1958. Proc. N . Central Weed Control Conf. 15, 115. Shaw, W. C. 1960. Better Crops Plant Food Spec. Issue pp. 26-31. Shaw, W. C., Willard, C. J., and Bernard, R. L. 1955. Ohio Agr. Expt. Sta. Bull. 761, 1-24.

Shaw, W. C., Hilton, J. L., Moreland, D. E., and Jansen, L. L. 1960. U . S . Dept. Agr. ARS 20-9, 119-133. Simon, J,, Sund, J. M., Wright, M. J., Winter, A., and Douglas, F. D. 1958. I . Am. Vet. Med. Assoc. 132, 164-169. Smith, R. J., Jr. 1958. Rice J. 61:18-19, 26-27. Smith, R. J., Jr. 1960. Weeds 8, 256-267. Smith, R. J., Jr. 1961. Weeds 9, 318-322. Sonder, L. W., and Alley, H. P. 3961. Weeds 9, 27-35. Sprague, M. S. 1959. Farm Chem. 112, 44-45. Sund, J. M., and Wright, M. J. 1957. Agron. J . 49, 278-279. Timmons, F. L. 1958. Proc. Western Weed Control Conf. 16, 41-46. Timmons, F. L. 1960. U . S. Dept. Agr. ARS 34-14, 1-21, illustrations. ( A joint publication of Agricultural Research Service, U. S. Dept. of Agriculture and Bureau of Reclamation, U. S. Dept. of the Interior.) U. S. Agricultural Research Service. 1954. U. S. Dept. Agr. ARS Spec. Rept. 20-1, 1-199.

U. S. Agricultural Research Service and Federal Extension Service. 1961. U . S. Dept. Agr. ARS 34-23, 1-65. Wooten, H. H., and Anderson, J. R. 1957. U . S . Dept. Agr. Agr. Inform. Bull. 168, 102.

Wooten, 0. B., and McWhorter, C. G . 1961. Weeds 9, 36-41.

THE PHYSICS OF WIND EROSION AND ITS CONTROL'

.

.

W S. Chepil and N P. Woodruff

.

United States Deportment of Agriculture. Manhattan Kansas

I. I1. I11.

IV .

V.

VI .

VII . VIII . IX

.

Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Surface Wind . . . . . . . . . . . . . . . . . . ...................... A. Drag Velocity and Drag on Noneroded Surfaces . . . . . . . . . . . . . B . Effect of Soil Movement on the Surface Wind . . . . . . . . . . . . . . . . Equilibrium Forces on Soil Grains ............................. A . Forces at Threshold of Soil Movement ....................... B . Forces during Soil Entrainment ............................ The Cycle of Wind Erosion ................................... A . Soil Destabilization and Stabilization Processes . . . . . . . . . . . . . . . B. Conditions Resulting from Destabilization and Stabilization Processes ............................................... C . Soil Erosion and Its Resulting Conditions .................... Soil Properties That Influence Wind Erosion ..................... A. Cohesive and Dispersive Forces of Water and Raindrops . . . . . . . B. Mechanical Stability and Abradability of Soil Structural Units . . C. Relative Importance of State and Stability of Soil Structure . . . . D . Seasonal Influences on Soil Structure and Erodibility ......... E. Basic Soil Factors ....................................... Wind Erosion Control ...................................... A. Vegetative and Other Types of Cover ...................... B. Soil Clods and Ridges .................................... C . Windbreaks and Wind Barriers ............................ D . Crop Strips and Crop Rows ............................... The Wind Erosion Equation .................................. A . General Framework ...................................... B . Data Needed to Estimate Potential Soil Loss . . . . . . . . . . . . . . . . . Needed Research ............................................ Conclusion ................................................. References ..................................................

Page 211 214 216 216 221 222 222 227 229 231 235 242 249 250 251 259 259 261 270 274 279 283 289 291 291 295 296 298 299

Notation

Basic dimensions of terms used are indicated in parenthesis by dimensional symbols m. Z. t. and f. denoting mass. length. time. and weight or 1 Contribution from Soil and Water Conservation Research Division. Agricultural Research Service. USDA. with Kansas Agricultural Experiment Station cooperating . Department of Agronomy Contribution No . 795.

211

212

W. S. CHEPIL AND N. P. WOODRUFF

force, respectively. Units of English or metric system are given after the dimensional symbols. Pages where each term is used first also are listed. Log values in this review are all to base 10, unless otherwise indicated. Symbol

Page 220 Empirical constants, coefficients, or exponents in various equations; 262 their values differ in different equations (dimensionless) 262 Proportion of fractions 0.84 mm. (as determined by standard dry 292 seiving) in a soil (dimensionless); per cent Weight of soil abraded per unit weight of abrader blown by wind of 251 velocity u (dimensionless) 219 Aerodynamic surface roughness ( I ) ; cm., in. Percentage of nonerodible clods > 0.84mm. in diameter in a soil 263 ( dimensionless ) ; per cent Local climatic factor in the wind erosion equation (dimensionless); 292 per cent 246 Dust concentration at 6 feet above ground (f/P); mg./cu. ft. Total dust load in the atmosphere up to 1 mile high ( f l 1 3 ) ; tons/cu. mi. 247 246 Dust concentration at height z ( f / I 3 ) ; mg./cu. ft. 223 Center of gravity (dimensionless) 224 Diameter of a soil particle or grain ( I ) ; cm., mm. Equivalent diameter of a soil particle or grain ( 1 ) ; cm., mm. 245 Distance (along prevailing wind erosion direction) of full protection from wind erosion afforded by a surface barrier adjoining a field ( I ) ; ft. 292 Distance across a field (along prevailing wind erosion direction) ( I ) ; ft. 292 27 1 Distance between surface wind barriers ( 1 ) ; ft. Zero displacement height-a vertical displacement of a wind velocity gradient by vegetation or other roughness elements of the ground surface ( I ) ; cm., in. 218 Potential average amount of erosion or soil loss by wind ( f / P / t ) ; tons/acre/annum 294 Ratio of mean drag and lift per unit area on the whole soil bed to mean drag and lift per unit area on the top grain moved by wind (dimensionless) 224 Angle of repose of a grain on the ground with respect to the mean drag level of the wind (degrees) 224 Threshold drag on the top grain on a soil bed ( f ) ; dynes 223 Threshold drag per unit horizontal area occupied by the top grain on a soil bed ( f / P ) ; dynes/sq. cm. 224 Mechanical stability of a surface soil crust (dimensionless); per cent 292 Percentage of clay in the soil (dimensionless) ; per cent 262 Acceleration of gravity ( Z / @ ) ; cm./sec./sec., ft./sec./sec. 224 Height of wind barrier or projection on the ground surface ( 1 ) ; ft., in., cm. 271 Soil erodibility index. It is equal to X , / X , in which X , is the quantity eroded from an area not exceeding 30 feet in length along wind erosion direction when the soil contains 60 per cent of clods > 0.84 mm. in

>

THE PHYSICS OF WIND EROSION AND ITS CONTROL

Symbol

213 Page

diameter, and X, is the quantity eroded under the same set of conditions and time from soil containing any other proportion of clods > 0.84 mm. in diameter (dimensionless) Soil erodibility (potential average annual soil loss from an unsheltered, I wide, and isolated field with a smooth, bare, and noncrusted surface under climatic conditions like those at Garden City, Kansas ( f / P / t ) ; tons/acre/annum Height above 2, where the forward average wind velocity above a k noneroding surface is zero ( I ) ; cm., in. k' Height above 2, where wind velocity above an eroding surface is constant no matter how strongly the wind blows (1); cm., in. Soil surface roughness expressed as ridge roughness equivalent ( 1 ) ; K in., cm. Orientation of vegetative cover factor in the wind erosion equation K O ( dimensionless ) Total surface roughness. It is equal to D, A,. ( I ) ; cm., in. K* Equivalent field width in the wind erosion equation ( I ) ; ft. L Lo Lift on the top grain on the ground at the threshold drag F , on the grain ( I ) ; dynes ratio of water content of a soil to water conM e Equivalent moisture-a tent at 15 atmosphere percentage ( dimensionless) M Average moisture of a soil surface Number of numbers in a statistical analysis (dimensionless) N Kinematic viscosity of air ( . @ / t ) ;sq. cm./sec. V P Mean pressure of lift and drag on the top grain on the ground ( f / 1 2 ) ; dynes/cm.2 P-E Effective precipitation of Thomthwaite ( 1931) (dimensionless) Rate of soil flow (weight of soil moved past a unit width normal to 4 direction of flow and of unlimited height, per unit time) ( f / l / t ) ; g./cm./sec. Q Maximum wind velocity for a specified unit of time ( l / t ) ; cm./sec., mi./hr. Q, Average of the maximum wind velocities for a specific continuous period ( l / t ) ; cm./sec., mi./hr. 0, Wind velocity equaled or exceeded on an average of once in t years, called recurrence interval ( Z/t); cm/sec., mi./hr. Three-year running average wind velocity ( l / t ) ;cm./sec., mi./hr. Q 7 Resistance of discrete soil grains against the force of wind, due to cohesion of water film on the surface of the grains. Like wind drag, it is expressed in units of force acting parallel with the ground per unit area of the ground ( f / . @ ) ; dynes/sq. cm. R Quantity (oven-dry weight times 1.2) of vegetative cover per unit area of ground ( f / F ) ; Ib./acre Density (weight per unit volume) of air ( f / P ) ; g./cc. P P / g Mass density of air ( m / B ) ;g./cm.,/sec./sec. Difference between the density (weight per unit volume) of soil grain P' and air, known as immersed density of the grain ( f / 1 3 ) ; g./cc. Density (weight per unit volume) of dry erodible soil grains (f/Zs); Pe g./cc.

+

246

292 217 222 292 292 219 292 223 237 292 232 238 226 234

245 233 233 232 233

250 292 220 220 224 237

214

W. S. CHEPIL A N D N. P. WOODRUFF

Symbol

Page

Density (weight per unit volume) of dry nonerodible soil aggregates ( f / Z 3 ) ; g./cc. Kind of vegetative cover factor in the wind erosion equation (dimensionless ) Standard deviation (expressed in same units that the population to which it refers is expressed) Recurrence interval of a specific climatic event ( t ) ; years Turbulence factor-a ratio of “maximum” to mean drag and lift on the top grain on the ground, assuming the ratio to be equal for both drag and lift. It is taken as ( p 3a)/P (dimensionless) Mean wind drag per unit horizontal area of the ground surface ( f / Z * ) ; dynes&. cm. Threshold drag per unit horizontal area of the ground surface, where zC is uniform ( f / P ) ;dynes/sq. cm. Mean threshold drag per unit horizontal area of the entire ground surface ( f / F ) ; dynes/sq. cm. Equivalent quantity of vegetative cover in the wind erosion equation ( f / P ) ;lb./acre Daytime visibility distance of a dusty atmosphere ( 1 ) ; mi. Drag velocity above a noneroding ground surface ( Z / t ) ; cm./sec., mi./hr. V’, Drag velocity above an eroding ground surface ( Z / t ) ; cm./sec., mi./hr. V,, Threshold drag velocity, i.e., minimal drag velocity required to initiate soil movement ( Z / t ) ; cm./sec., mi./hr. k‘ above an eroding- surface ( V t ) ; cm./sec., Wind velocity at height mi./hr. Wind velocity ( Z / t ) ; cm./sec., mi./hr. Threshold wind velocity, i.e., minimal velocity at some specified height z required to initiate soil movement ( Z / t ) ; cm./sec., mi./hr. Wind velocity u at height z ( Z / t ) ; cm./sec., mi./hr. Quantity of soil material removable by wind (of specified drag velocity) from the surface of not more than 30 feet of ground along wind erosion direction ( f / l * ) ; tons/acre An expression in the Gumbel (1941,1945) equation indicating frequencies of rain or windstorms of various intensities at a given location. It is equal to lo&[- lo&( 1 - l / t ) ] with t expressed in years Percentage of water-stable particles < 0.02 mm. and > 0.84mm. in diameter as determined by wet sieving of soil (dimensionless); per cent Mean aerodynamic surface-a surface above which the turbulent air flow is unrestricted or “free” compared with restricted, sometimes laminar flow (as though vegetation) below 2, (dimensionless) Height above 2, ( I ) ; cm., in., ft.

+

273 292 226 232

224 220 224 225 292 246 218 222 242 222 251 243 218 245 233 263 217 218

1. Introduction

In many countries throughout the world, wind erosion has depleted the fertility of the soil, and in some it has transformed the fertile soils into sandy deserts. Substantial portions of central Asia, the Middle

THE PHYSICS OF WIND EROSION AND ITS CONTROL

215

East, and North Africa were once fertile lands supporting prosperous populations, but through improper land use and soil exhaustion they changed to their present barren state. The downfall of ancient civilizations such as those of Greece and Rome is a story of depletion of grasslands and forests, soil erosion, and soil ruin. In North America, relatively little wind erosion occurred while land was under natural vegetation. It accelerated after man began to overgraze and overcultivate the land. It became worse in the Great Plains, the semiarid and subhumid area that extends almost from the Mississippi River to the Rocky Mountains and from the Gulf of Mexico into the Prairie Provinces of Canada. The first, and probably the last, serious wind erosion in the Great Plains occurred during the 1930s. The general realization of the great economic losses caused by wind erosion during that period helped to stimulate serious attention to its basic causes, effects, and remedies. Soil surveys of the Great Plains were initiated to aid in stabilizing agriculture in that area. Emergency wind erosion control programs were established and administered by the various States and by Federal agencies. Special wind tunnels were developed and used to study the wind erosion problem continually, not just when it occurred in the field. Numerous papers and bulletins were published on wind erosion and control. The publications on the subject, though voluminous, have been fragmentary and somewhat lost in the literature of argiculture and related fields. This review is the first attempt to bring the research information together into an analysis of the subject as a whole. The subject deals with movement and abrasion of soil by wind. Movement is initiated when the pressure of the wind against the surface soil grains overcomes the force of gravity on the grains. The grains are moved aIong the surface of the ground in a series of jumps known as saltation. The higher the grains jump, the more energy they derive from the wind. The concentration (number per unit volume) of saltating grains increases with distance downwind till, if the eroding field is large enough, it becomes the maximum that a wind of a particular velocity can sustain. The impacts of the saltating grains initiate movement of larger and denser grains and of smaller dust particles. The saltating grains collide against massive materials and other grains and cause disintegration of all involved. The disintegrated units exhibit different degrees of mobility and sort into different erosion products, such as lag sands, lag gravels, dunes, and deposited dust (loess), Wind erosion occurs only when soil grains capable of being moved in saltation are present in the soil. Comparatively few saltating grains jump higher than a few feet above the ground. Over 90 per cent gen-

216

W. S. CHEPIL AND N. P. WOODRUFF

erally do not rise higher than 1foot. Therefore, wind erosion is essentially a surface phenomenon extending to saltation height. Dust clouds are merely the result of movement in saltation. The above-mentioned processes and products of wind erosion constitute only part of the physics of wind erosion and its control. The subject includes the intricate processes and conditions that cause erosion and the counteracting processes and conditions that suppress it. The severity of wind erosion depends on equilibrium conditions between soil, vegetation, and climate. Wind erosion is accelerated by processes that cause surface soil structural disintegration and depletion of vegetative cover. Conversely, wind erosion is hindered by stabilization processes such as soil consolidation and aggregation and by vegetation and vegetative residue developing on the surface. The speed or intensity of all the processes fluctuates considerably with vagaries of the weather and with various land uses. The subject includes causes, effects, and remedies of wind erosion. Processes of soil destabilization, soil erosion, and soil stabilization must be understood to design effective and lasting methods of wind erosion control. To design suitable methods of wind erosion control, soil conservationists must know the conditions that influence wind erosion and how ko evaluate the relative significance of each condition. Procedures have been developed to supply them with the so-called wind erosion equation which can be used to estimate the potential amount of wind erosion from measured conditions of the field. Conversely, the equation may be used to estimate the conditions needed to reduce wind erosion to any degree, The procedures are outlined briefly in Section VII of this review. It. The Surface Wind A N D DRAG ON NONERODED SURFACES A. DRAGVELOCITY

Wind structure near the ground directly influences the movement of soil by wind. A wind strong enough to produce soil movement is always turbulent; that is, its flow is characterized by eddies moving at variable velocities and in all directions. The average forward velocity, generally regarded as velocity, of a turbulent wind near the ground increases with height according to an exponential law. Zero velocity is somewhere above the average roughness elements of the surface. The taller the roughness elements of the ground, or the taller and less air-permeable the vegetative cover, the higher level at which zero velocity is found. From this level upward, the velocity increases very rapidly at first, then less rapidly as we go up, as shown on the left-hand side of Fig. 1.

217

THE PHYSICS OF WIND EROSION AND ITS CONTROL

The change in velocity with height is known as the velocity gradient. It will be noted from the left-hand side of Fig. 1 that the estimated zero velocity is at height 2, k in which 2, is the so-called aerodynamic surface and k is height above 2, where the velocity is zero. Usually k is so limited that 2, k is approximately the same as 2, when plotted on an arithmetic scale as shown on the left-hand side of Fig. 1.

+

+

280 GIn W 040

-.s,eo Y

w

rJ

p

10

k 8

2 6

0 $ 4 t

g"

:*

W

>

8 a

5

(3

X

t .B .6 A

0 200 400 WIND V E U C I T Y (CM. PER SECOND)

600

FIG. 1. Wind velocity distribution determined simultaneously and, therefore, for the same wind above ( a ) sorghum stubble with maximum height of 53 cm. no matter how strong the wind blew and ( b ) growing wheat which was 5cm. high when no wind blew and lower when it did. The velocity distribution for a different wind but the same surface as in ( b ) is shown by curve ( c ) . (Unpublished data of Chepil and Siddoway. )

It is important to note that the aerodynamic surface 2, is often quite indistinct. It is estimated by plotting (on an arithmetic scale) the velocity above the surface projections against the height above the average grotind surface and projecting the curve, thus obtained, to the ordinate ( a t which velocity is actually or presumably zero, as shown by continuous line on the left-hand side of Fig. 1). The velocity at height 2, k is in fact zero if the surface is impervious, such as the ground surface. Over a porous vegetation-covered surface, however, the

+

218

W. S. CHEPIL AND N. P. WOODRUFF

velocity at 2, is somewhat greater than zero, indicating that some air movement is taking place through the vegetation (shown by discontinuous line on the left-hand side of Fig. 1, curve (a).The height Z, roughly separates the two types of air flow near vegetation-covered surfaces-the relatively fast-moving, so-called “free flow” above the vegetation and the slow-moving, so-called “restricted flow” below the tops of vegetation. Z,, as determined, usually is found somewhat below the maximum height of vegetation or vegetative residue. Thus, for sorghum stubble having a maximum height of 53 cm., the estimated 2, was about 31 cm., as shown on the left-hand side of Fig. 1. For growing wheat 5 cm. high, Z,, was found to be nearly at the ground surface, probably because young wheat bends considerably in the wind and is quite porous. When wind velocity within the free-flowing, fully turbulent zone up to about 5 feet (1.6 meters) above the surface projections, is plotted against the log of height above the mean aerodynamic surface, the velocity curve is a straight line (as shown on the right-hand side of Fig. 1).This shows that velocity at any height z above the mean aerodynamic surface 2, conforms with the Prandtl (Brunt, 1944) and von Karman (1934) equation x

v , = 5.75v. log -

k

in which V. is the so-called drag velocity and k is the height above the mean aerodynamic surface Z, , at which height the wind velocity is zero, or presumably zero. The rougher the aerodynamic surface, the greater is the value of k, so that k may be considered as an index of the aerodynamic surface roughness. Neither the aerodynamic surface roughness nor height k has any relationship to the height of vegetation or other roughness elements of the surface, but only to the variation in height, density, or spacing, flexibility, and other characteristics. Height of 2, above the average ground surface, on the other hand, is determined primarily by the height of vegetation. The distance between Z, and the average level of the ground surface is the vertical displacement of the velocity gradient by the vegetation or other roughness elements of the surface and often is referred to as the zero displacement height Dh (Fig. 2 ) . Geiger (1957) calls it the roughness height. For restricted-flow zone below the tops of vegetation or other roughness elements of the surface, no simple velocity-versus-height relationship has been found.

THE PHYSICS OF WIND EROSION AND ITS CONTROL

219

Transposing Eq. (1) for free-flow zone

v. =

0,

z

5.75 log -

k

The drag velocity, V., is an index of the rate of increase of velocity with the log of height. The stronger the wind the greater the drag velocity, but for a given inflexible surface the values of 2, and k remain the same no matter how strong the wind blows (Fig. 1 ) . Moreover, the drag velocity up to at least 5 feet above the surface projections for a

FIG.2. Diagrammatic representation of the relative positions of the ground and vegetative roughness elements above the ground (marked by slanting lines). K, = D , A, in which K , is total surface roughness, D , is the zero displacement height, and A, is the aerodynamic surface roughness.

+

given wind in a given geographic region remains the same no matter what type or how rough the surface. The drag velocity within 5 feet above the aerodynamic surface 2, therefore can be used as an index of the general atmospheric wind force. The velocity alone at any given height is meaningless unless the values of 2, and k are known. For many years the power law of Hellmann (1915), as reported by Geiger (1957), has been used to characterize the distribution of wind velocity with height above the ground. More recently, however, the exponential law of Prandtl (Brunt, 1944) and von Karman (1934) has

220

W. S. CHEPIL AND N. P. WOODRUFF

been accepted as more nearly describing the velocity distribution near the ground. The power equation might be valid for heights greater than 5 feet, but below 5 feet and for fully turbulent flow as would occur when wind erosion is in progress, the exponential law of Prandtl and von Karman is generally valid. For such flow, air temperature differences with height above ground vanish, but the increase of velocity with log of height is its primary characteristic. For rough pipes and relatively smooth soil surfaces it was found by von Karman (1934), Nikuradse (Rouse, 1950), and Zingg (1953a) that

-

t =p

(V.)2

(3)

where ;is the mean wind drag per unit horizontal area of the ground surface and p is the density of air, that is, the weight per unit volume of air (about 0.0012 g./cmsS). If VI is expressed in cm./sec., then t is in dynes/cm.2. The force t is dynamic, acting generally in the direction of flow. Like wind velocity, it fluctuates greatly in all directions. For rough, vegetation-covered surfaces, Sheppard ( 1947), and Chepil and Siddoway (Table I ) found that the mean drag, t, for a given natural wind varies significantly with surface roughness. The rougher the surface, TABLE I Measured Drag on Field Surfaces of Different Degrees of Roughnessa

0

Time period (variable)

Sorghum stubble 20 inches high (dynes/cm.2)

Growing wheat 9 inches high (dynes/cm.2)

1 2 3 4 5 Average

4.9 14.3 9.0 6.9 8.2 8.7

2.5 7.9 6.6 5.5 4.5 5.4

Unpublished data of Chepil and Siddoway.

the greater is the drag. Typical data presented in Table I show that Eq. (3), developed from experiments with rough pipes, cannot be used to compute the surface drag over soil and vegetation-covered surfaces. Roughness of pipes did not exceed 1.5 mm., but that of vegetationcovered surfaces was much greater, as shown in Fig. 1. Consequently, for vegetation-covered surfaces, Eq. ( 3 ) must be modified to

t

= up (V*)2

(4)

in which a is a drag coefficient which varies with aerodynamic surface roughness as influenced by type and height of vegetation.

THE PHYSICS OF WIND EROSION AND ITS CONTROL

221

It should not be construed from Table I that because a rough surface takes up a greater drag, it is more erodible than a smooth one. This might be true if a rough soil surface were composed only of erodible fractions, but if the roughness elements are composed of nonerodible clods and erodible fractions, as they usually are, the erodible fractions are moved down and trapped in the depressions and the clods then take up most of the drag. Conversely, if the soil is covered with anchored vegetation or vegetative residue, much of the drag is taken up by the vegetative matter and only the residual drag is taken up by the soil. OF So= MOVEMENT ON B. EFFECT

THE

SURFACE WIND

Over an eroding soil surface the velocity gradient was found by Bagnold (1936) and Chepil and Milne (1941a) to undergo a considerable change to which Eq. (1) of Prandtl and von Karman does not apply. They showed that sand and soil movement in saltation reduces the momentum and, therefore, the surface velocity of the wind, as shown

FIG. 3. Wind velocity gradients over an eroding and noneroding surface in a portable field tunnel. Broken lines denote gradients over an eroding surface, and continuous lines over the same surface "fixed" with a fine spray of water. V,= 4.8 mph. or 214 cm./sec., k = 0.02 cm., and k' = 0.5 cm. (Chepil, 1941.)

222

W. S. CHEPIL AND N. P. WOODRUFF

in Fig. 3. The solid lines of Fig. 3 indicate typical wind gradients over a “fixed surface over which no soil movement took place, and the dashed lines indicate the velocity gradients when soil movement was in progress. The surface was fixed by spraying it with water. The drag velocity curves over a fixed surface confrom with Eq. ( 1 ) . An eroding soil surface, on the other hand, reduces wind velocities to considerable height; consequently, new drag velocity curves are established, all of which have a common velocity at height 2, k‘ as shown in Fig. 3. The new wind velocity distributions conform with the Eq. (5)

+

2

v, = 5.75 V.‘ log -

k’

+ Vt

(5)

in which V.‘ is a drag velocity above an eroding surface, K is height (,above 2,) to which all drag velocity curves V.‘ converge, and V t is velocity at height k’. V t remains constant no matter how strong the wind blows. Therefore,

V.’ =

v, -Vt

5.75 log

z -

k’

It will be evident from Eq. ( 5 ) and Fig, 3 that the higher the drag velocity V.’, that is, the stronger the wind blows, the lower is the velocity below height k’. This seemingly illogical condition apparently is due to greater concentration of saltating soil grains with strong winds, which tends to lower the wind velocity below height k’. Height k’ was found to be considerably below the average height of saltation. Field measurements by Chepil and Milne (1941a) indicated, too, that the lowering of wind velocity due to soil movement vanes directly with soil erodibility; that is, the more erodible the soil the greater the concentration of moving soil grains and the greater is the reduction of wind velocity near the ground. 111. Equilibrium Forces on Soil Grains AT THRESHOLD OF SOILMOVEMENT A. FORCES A moving fluid such as air or water exerts three types of pressure on a soil grain resting on the ground (Einstein and El-Samni, 1949; Ippen and Verma, 1953; Chepil, 1959b). One is a positive pressure against that part of the grain facing into the direction of fluid motion. This pressure results from the impact of the fluid against the grain and is called the impact or velocity pressure. The velocity pressure causing

THE PHYSICS OF WIND EROSION AND ITS CONTROL

223

the initiation of movement of a soil grain varies directly as the square of the fluid velocity, and its magnitude is the force per unit of crosssectional area of the grain normal to the direction of fluid motion. The second type is a negative pressure on the lee side of the grain, known as viscosity pressure. Its magnitude depends on the fluid’s coefficient of viscosity, density, and velocity. The third type of pressure is a negative pressure on the top, as compared to the bottom, of the grain, caused by the Bernoulli effect. The Bernoulli law states that wherever the fluid velocity is speeded up, as at the top of the soil grain, the pressure (measured transverse to the general direction of fluid motion) is reduced. This is called the static, isotropic, or internal pressure. The impact or velocity pressure on a soil grain lying on the ground is known as form drug, and the pressure due to viscous shear in the fluid close to the surface of the soil grain is called skin friction drug. The sum of the two forces is the total drug. Separation of the two kinds of drag appears unnecessary in determining the equilibrium forces on the soil grain. The total drag in this review is referred to as drug. The drag, F,, on the top grain at the threshold of its movement is due to the pressure difference against its windward and leeward sides. The arrow marked by F , in Fig. 4 indicates the general direction and the average level at which it acts. LC WIND DIRECTION

LEVEL,

zo

FIG.4. Forces of lift, drag, and gravity acting on a soil grain in a windstream at the threshold of movement of the grain. Net moment opposing F, is (0.52gD3p’ - L C ) tan for a spherical grain (Chepil, 1959b).

+

A decrease in static pressure at the top of the grain as compared to that at the bottom causes a lift on the grain. It is determined by, but is not the same as, the pressure difference against the top and the bottom halves of the grain. The arrow marked L, in Fig. 4 indicates the general direction in which it acts. It acts through the center of gravity, c.g.

224

W. S. CHEPIL AND N. P. WOODRUFF

The minimum mean drag and lift forces, known as the threshold drag and lift, required to move the top soil grains by wind are influenced by the diameter, shape, and immersed density of the grains; by the angle of repose + of the grains with respect to the mean drag level of the wind; by the closeness of packing q of top grains on the sediment bed; and by the impulses of wind turbulence T associated with drag and lift (Jeffries, 1929; White, 1940; Kalinske, 1943; Einstein and El-Samni, 1949; Chepil, 1958)* From Fig. 4,the threshold drag F, acting on a spherical grain is

F, = (0.52 gD3p‘ - L,) tan +

(7) in which g is the acceleration due to gravity, D is the diameter of the grain, p’ is the immersed density of the grain, and + is the angle of repose of the grain with respect to the average drag level of the fluid. In this equation the expression 0.52 go3$is the immersed weight of the spherical grain. From experiments, Chepil (1959b) found that L, is equal to about 0.75 F , for any size of spherical elements, such as soil grains on the surface, and for any wind velocity within the range required to move different sizes of soil grains. Therefore, by substitution, transposing, and factoring, Eq. ( 7 ) becomes

F, = 0.52 gD3p’ tan +/( 1+ 0.75 tan +)

(8)

Equation ( 8 ) indicates the threshold drag F , required to move the top grain of diameter D. But the threshold drag F’, per unit crosssectional horizontal area occupied by the grain is equal to F,/0.7854D2 in which 0.7854D2 is the largest cross-sectional horizontal area occupied by the spherical grain. Then, by substitution and simplification (Chepil, 1959b) F’, = 0.66 gDp‘ tan +/( 1 0.75 tan +) (9)

+

Drag and lift per unit horizontal area occupied by the topmost grains are much higher than drag and lift per unit area on the whole bed, because the topmost grains which take up most of the drag and lift occupy only a portion of the bed area. If q is the ratio of drag and lift on the whole bed to drag and lift on the topmost grain moved by the fluid, then Eq. ( 9 ) becomes t,

= 0.66 gDp’ tan +q/( 1

+ 0.75 tan +)

(10)

in which t, is the threshold drag per unit horizontal area of the whole bed. Equation (10) is applicable to wind of uniform velocity. Since the airstream of a velocity required to move the top soil grains is not uni-

THE PHYSICS OF WIND EROSION AND ITS CONTROL

22s

form, movement of the grains is facilitated by the maximum lift and drag impulses of turbulent flow. Therefore, for turbulent flow, Eq. (10) should be modified to

-

tc = 0.66 gDp’

+ 0.75 tan 4 ) T

tan +y/( 1

(11) in which z‘, is the mean threshold drag per unit horizontal area of the whole soil bed and T is the ratio of maximum to mean lift and drag on the soil grain exerted by the turbulent wind. By estimating wind velocities at different heights above the mean level of the sediment beds, Chepil (1959b) found that the drag on the topmost grains on the bed acts at an average level of about one-third of the grain diameter below the top of these grains. On the basis of these experiments, he found that the angle of repose + of the topmost grains with respect to the mean level of drag is about 24 degrees. Therefore, tan + is equal to about 0.45. Assuming that all the drag is taken up by the topmost grains on a bed, White (1940) determined the coefficient q by counting the grains lying on top of the bed, computing the largest horizontal cross-sectional area of the grains, and dividing this area by the total horizontal area of the bed. He found that coefficient q, so determined, had a value of about 0.1. However, exact determination of the coe5cient in this manner is impossible since all sorts of gradations between complete exposure and virtually complete embedding of the surface grains occur. Chepil ( 1959b) therefore determined the coefficient from actual measurements of pressure on topmost spherical bodies, such as soil grains, and from drag on the whole surface computed from the threshold drag velocity of the wind in accordance with Eq. ( 3 ) . He found that the coefficient determined from those two conditions has a value of about 0.2. In studying the motion of sediment particles in water, Ippen and Verma (1953) concluded, “Nothing is known in detail as yet concerning the turbulent pressure fluctuations near the bed.” Lack of such knowledge was largely due to lack of suitable instruments for measuring turbulence. Chepil and Siddoway ( 1959), therefore, devised a strain gage anemometer for this purpose. Analysis of oscillograms obtained with this instrument, and shown by example in Fig. 5, indicated that pressure of both lift and drag at a level of the topmost grains is distributed statistically according to a somewhat skewed normal error law. Hence, from a statistical standpoint, the maximum pressure of lift and drag has no definite limit and, therefore, the ratio of maximum to mean cannot be given. The standard deviation, however, completely describes the spread of pressure of lift and drag around the mean. Analysis showed that the

226

W. S. CHEPIL AND N. P. WOODRUFF

standard deviation of the pressure distributions varies directly with the drag velocity of the wind and that the ratio of mean pressure to standard deviation is nearly constant at the position of the topmost grains of a size range eroded by wind. Nearly all, or 99.73 per cent, of the pressure range is included within p f 30 in which p is the mean pressure of drag Pressure 2 dynedcm

FIG.5. Oscillograph records of pressure of drag and lift on spherical gravel at

bed level for a drag velocity of 47cm. per second: ( a ) Drag on 3.2-mm. gravel, ( b ) drag on 8.4-mm. gravel, ( c ) lift on 3.2-mm. gravel, ( d ) lift on 6.4-mm. gravel. Time interval, 1 second (Chepil, 1959b).

and lift and a is the standard deviation of the pressure around its mean. The turbulence factor, T , therefore, was taken as (P+ 3a)/P which assumes that the “maximum” pressure is P + 3a. On the basis of this assumption, the turbulence factor for both lift and drag at the position of the topmost grains on a soil bed was found to be approximately 2.5. Oscillograms of Fig. 5 show that the smaIlest-scale cycles of pressure, although irregular as to both magnitude and duration of occurrence, have a period of about 1/80 to 1/120 second. Duration of this primary period varied little with all wind velocities used. Pressures equal to or greater than F + 3a for lift and drag occurred two to three times per second, depending somewhat on drag velocity of the wind. Assuming that tan + = 0.45, = 0.2, and T = 2.5, Chepil (1959b) for computed, on the basis of Eq. ( l l ) ,the mean threshold drag various sizes and immersed densities (differences in bulk density be-

k

227

THE PHYSICS OF WIND EROSION AND ITS CONTROL

tween the grain and the air) of soil and sand grains and compared them with the actual mean threshold drag determined by wind tunnel tests. The threshold drag determined in a tunnel agreed reasonably well with the threshold drag computed in accordance with Eq. ( l l ) ,as shown in Table 11. This seems to confirm the general validity of Eq. (11) and the approximate values of the parameters that it embodies. TABLE I1 Concordance of Computed with Actual Threshold Drag for Soil Grainsa Minimum grain diameter D (cm.)

Immersed grain density

0.015 0.025 0.030 0.042 0.059 0.010 0.015 0.018 0.025 0.042 0.059 0.084 0.119 0.200 0.336 0.475 0

Q’

Computed threshold drag (dynes/cm.Z)

Actual threshold drag (dynes/cm.Z)

2.65 2.65 2.65 2.65 2.65 2.09 1.96 1.94 1.91 1.91 1.80 1.78 1.74 1.65 1.52 1.55

0.69 1.16 1.39 1.95 2.74 0.37 0.51 0.61 0.84 1.40 1.86 2.62 3.63 5.78 8.94 12.90

0.59 0.85 1.11 1.60 2.32 0.27 0.35 0.48 0.69 1.08 1.51 2.33 3.59 5.15 10.30 14.00

Data from Chepil (1959b).

B. FORCESDURING SOIL ENTRAINMENT Lift and drag on soil grains change rapidly as the grains move up from the surface of the ground. Lift decreases with height and becomes hardly detectable a few grain diameter heights above the ground. This height is considerably less than the height to which many grains rise in saltation. The greater the ground roughness and drag velocity of the wind, and therefore the steeper the velocity gradient, the higher lift extends. Lift is caused, apparently, by a steep wind velocity gradient near the ground. Drag, on the other hand, increases with height just as wind velocity increases with height, and apparently is due mainly to the direct pressure of the wind against the grain. A diagrammatic representation of lift and drag on a small sphere, such as a soil grain, is given in Fig. 6. It is shown that lift in this case

228

W.S. CHEPIL

AND N. P. WOODRUFF

almost ceased to exist at about 2.5 cm. height. Drag on the sphere, on the other hand, continued to increase all the way up to the height of measurement, just as velocity increased with height. The drag on the grains is generally much greater than the lift. After being shot into the air, the grains rise to various heights, and because WIND DIRECTION

/

J

\

I

FIG.6. Pattern of approximate pressure differences between position 1 on top of the sphere and other positions on the sphere at various heights in a windstream. Length of lines in the shaded areas outside the circular line (sphere) denotes the relative differences in air pressures. The sphere is 0.8cm. in diameter and the drag velocity is 98 cm. per second (Chepil, 1961). of the force of gravity, fall at an accelerating velocity. There is at the

same time a horizontal acceleration of the falling grain due to the force of drag. The downward and forward accelerations are proportioned uniformly so that the inclined path of the falling grain is almost a straight line. However, the average force of drag is much greater than the force of gravity, and therefore the angle of descent is only about 6 to 12 degrees from the horizontal. If the ground were perfectly smooth and there were no lift, the angle of ascent (expressed as deviation from the horizontal) should be the same as of descent. However, grains in saltation rise vertically or nearly SO. From measurements of pressure on suspended spheres, such as soil grains, Chepil (1961) concluded that the essentially vertical rise must be due in some measure to the presence of lift near the ground but that lift alone could not possibly be the sole factor involved. Another factor,

THE PHYSICS OF WIND EROSION AND ITS CONTROL

229

apparently, is the surface obstructions from which the saltating grains rebound (Fig. 7). The obstructions are usually spherical or nearly spherical soil aggregates or other grains resting on or creeping along the ground. The topmost grains that compose the eroding surface occupy on the average about 0.1 of the total surface and therefore are spaced about three diameters apart (White, 1940), as shown in Fig. 7. The saltating grains descend at an average angle of about 9 degrees from the

t

---

- 7-----

GRAIN

r! k

FIG. 7. Diagrammatic representation of a saltating grain striking a stationary grain at an average impact point A and rebounding in a vertical direction A’. POSsible extreme points of impact are B and C with rebound directions B’ and C’ ( Chepil, 1961 ) ,

horizontal, strike the top portions of spherical ground objects, and then rebound predominantly in a vertical or nearly vertical direction. Because of the particular angle of descent and configurations of the ground surface, as shown diagrammatically in Fig. 7, the rebounds should be generally vertical even if lift did not exist. Lift merely contributes to the vertical rise of soil grains. The vertical momentum of saltating grains carries some of them upward and above the zone of lift. IV. The Cycle of Wind Erosion

All the processes of soil stabilization, destabilization, and wind erosion, including conditions resulting from those processes, may be termed the cycle of wind erosion (Chepil, 1961).This cycle is a part of the much broader cycle of weathering, which is defined as all physical and chemical changes produced in rock and soil materials by the elements of the weather and which result in disintegration, decomposition, movement, and sorting of the materials (Polynov, 1937). Each process associated with the cycle of wind erosion yields a specific product and each product leads to or causes another process.

230

W. S. CHEPIL A N D N. P. WOODRUFF

All the processes go only in one direction, and therefore form a cycle as shown in Fig. 8. The cycle of wind erosion is characterized by a continual transition among the different processes and their associated products or resulting conditions. The cycle may be viewed as an equilibrium between soil

0 Soil stobilization

conditions

0

= Process or group of processes

erosion

0Prr;;cttor =

products

FIG.8. The cycle of wind erosion (Chepil, 1962~).

stabilization processes on the one hand and soil destabilization processes on the other. Always associated with soil destabilization processes are the processes of soil erosion and their resulting products or conditions. The conditions resulting from soil destabilization, erosion, and stabilization cannot be defined explicitly because they vary greatly with the intensity of the processes that produce them. Therefore, what is an erodible condition under one set of climatic or weather processes may not constitute an erodible or an equally erodible condition under another. The processes and products associated with the cycle of wind erosion embody the causes, effects, and remedies of wind erosion and may be listed as follows. PROCEsSEs

Soil destubilization processes-basic causes of erosion Climatic causes High wind velocity and turbulence Low precipitation High temperature Soil structural breakdown Soil weathering Improper and excessive cultivation

PRODUCTS OR CONDITIONS P r i m r y causes of erosion

Increased surface wind velocity Erodible soil conditions Dry, small, loose, and light soil p i n s Erodible surface conditions Smooth, bare, unsheltered, large, and improperly oriented fields

THE PHYSICS OF WIND EROSION AND ITS CONTROL

Erosion processes--ejeects of basic causes

Initiation of soil movement Soil transportation Movement by saltation Movement by surface creep Movement by suspension Avalanching Detrusion

231

Products and conditions resulting from erosion

Expanded eroding area Progressively smoother and more erodible conditions to leeward of eroding area Residual soil materials Lag sands and gravels Dunes

Sorting

Loess

Abrasion Breakdown of soil structure Destruction of vegetation Soil stabilization processes-remedies erosion Soil deposition (stilling of erosion) Sedimentation Trapping

Increased quantity of erodible soil fractions of

Soil consolidation and aggregation Proper tillage and cropping practices Revegetation Proper tillage and cropping practices Proper field orientation and management Climatic and weather influences Decreased wind velocity and turbulence Increased precipitation Decreased temperature

Loose and bare soil conditions Conditions resulting from stabilization Nonerodible soil conditions Moist and firm surface soil Soil aggregates large and dense enough not to be moved by wind Nonerodible surface conditions Rough and covered surface Surface sheltered by barriers Restricted width of erosion-susceptible field Broad sides of field or field strip oriented at right angles to prevailing wind direction Reduced surface wind velocity

Associated with the cycle of wind erosion are the processes of soil formation and soil removal (Fig. 8 ) . The primary objective of the soil conservationists is to modify the processes that affect soil removal so that the rate of soil removal does not exceed the rate of soil formation. A.

SOIL DESTABILIZATION AND STABILIZATION PROCESSES

The opposing processes of soil destabilization and stabilization form an equilibrium; therefore, it is convenient to discuss them together. Soil destabilization processes may be considered as the basic causes

232

W. S. CHEPII, AND N. P. WOODRUFF

of wind erosion, whereas the products or conditions resulting from these processes may be termed the primary causes. The basic causes of accelerated wind erosion are associated with the equilibrium between climate, soil, and vegetation. Accelerated wind erosion in many parts of the world developed after man began to interfere unduly with the natural equilibrium between climatic, soil, and vegetative environment ( Sears, 1935). Burning, overgrazing, and overcultivation have been the chief means of disturbing this equilibrium. The problem of excessive wind erosion, therefore, is associated principally with the way the farmer uses his land (Bennett, 1939). Relatively little wind erosion occurs on grassland or woodland, but even here overgrazing and overclearing often cause accelerated erosion by wind and water. Accelerated erosion occurs chiefly on cultivated land. The equilibrium between climate, soil, and vegetation embodies the following opposing processes: ( 1) increased vs. decreased wind velocity, temperature, and precipitation, ( 2 ) soil loosening and deaggregation vs. soil consolidation and aggregation, and ( 3) devegetation vs. revegetation.

1. Increased us. Decreased Wind Velocity, Temperature, and Precipitation Extended periods of low precipitation, high temperature, and high wind velocity often contribute to the severity of wind erosion (Bennett, 1939; Zingg, 1953b, 1954). Wind erosion becomes progressively more serious as the sequence of dry years continues; this is so because conditions become progressively more erodible, Conversely, the severity of wind erosion was found to diminish only after the return of at least two consecutive years of favorable moisture and vegetative growth ( Zingg, 195313). Great variations in precipitation, temperature, and wind velocity exist in continental climates throughout the world. These variations are controlled generally by the normal probability law, just as are floods. Gumbel (1941, 1945) and Potter (1949) devised a workable method for determining the probability of occurrence of different flood magnitudes. Zingg (1949, 1950) has applied this method to analysis of intensity-frequency of occurrence of winds of various velocities in different regions. The method is based on a relationship Qm = Q.

+ a( 0.7797~-0.45005)

(12)

where Q,,, = wind velocity equaled or exceeded on the average of once in t years, where t is termed the recurrence interval, i.e., time in years.

THE PHYSICS OF WIND EROSION AND ITS CONTROL

233

= average of the maximum wind velocities, Q, in miles per hour for a specified continuous period. CI = standard deviation of maximum wind velocities for a specific continuous period, or N CI = Qu2) where Q is maximum wind veN--l locities for a specific continuous period (based on one event for each specified unit of time).

Qa

{-

(T-

y = log,

[-log, (1 -f ) 1.

N = number of events for a period of record (number of numbers in the analysis). Chepil et al. (1962) applied the method of Gumbel and Potter to analyze the intensity-frequency of a combination of major climatic conditions that influence wind erosion. An example of normal distribution of climatic conditions that influence wind erosion is given in Fig. 9.

RECURRENCE INTERVAL (YEARS)

FIG.9. Intensity-frequency data for three-year running average of wind velocity Q’, corrected to that at 30-foot height, divided by the three-year average of P-E index at the Branch Agricultural Experiment Station, Garden City, Kansas, for 1920 through 1960 (Chepil et al., 1962).

234

W.

S . CHEPIL AND N. P. WOODRUFF

These climatic conditions include a three-year running average of wind velocity, Q, which affects wind erosion directly, and a three-year running average of P-E index of Thornthwaite (1931), which affects wind erosion inversely (Chepil et al., 1962). The P-E index is an index of soil moisture and is influenced by monthly precipitation and temperature. A three-year running average was taken because it is known (Zingg, 1953b) that wind erosion generally becomes most or least intense after two or three consecutive years of drought or adequate moisture, respectively. All values of the wind erosion climatic factor (Q/P-E) falling on a straight line on the loglo probability scale of Fig, 9 would fit the normal curve perfectly. A longer period of record no doubt would have produced a better fit. Figure 9 indicates that during the period of record beginning in 1920, the most severe climatic conditions that influenced wind erosion at Garden City, Kansas, occurred in the 1930’s, and slightly less severe conditions in the 1950’s. Wind erosion was actually most thereby subsevere in the 1930’s and next to most severe in the 1950’~~ stantiating the validity of the wind erosion climatic factor as an index of severity of wind erosion, other factors remaining the same. The general frequency of occurrence of periods of high wind and temperature and low precipitation can be predicted from past records for any given location, but unfortunately the time when these periods will occur cannot be predicted. Owing to these climatic variations, occasionally some wind erosion will occur in severely affected regions even under virgin conditions. With sufficient resourcefulness, however, man can modify the intensity of wind erosion so that it is insignificant in amount. Atmospheric turbulence also contributes to the severity of wind erosion. It tends to increase the velocity near the earths surface and therefore to increase the frictional force on the ground (von Karman, 1934). Parkinson (1936) observed that the presence of dust storms in central United States generally is associated with instability, or turbulence, of the air masses. Over these, man presently has no control.

2. Soil Loosening and Deaggregatim vs. Soil Consolidation and Aggregation In some regions subject to frost, the spring season is potentially the most hazardous from the standpoint of wind erosion. Frost action on moist soils during the winter tends to loosen and break down soil clods, therefore increasing the erodibility by wind. In the summer an increase in cementing substances somewhat dispersible in water tends to cement the soil mass, to increase the proportion of nonerodible clods, and to decrease the erodibility by wind.

THE PHYSICS OF WIND EROSION A N D ITS CONTROL

235

Excessive and improper tillage often causes excessive soil loosening and pulverization and increases the hazards from erosion by wind (Woodruff and Chepil, 1956; Woodruff et al., 1957). Suitable tillage in regions where wind erosion is a hazard is necessary, especially to kill weeds and to conserve moisture. Moisture must be conserved to reduce the risks from wind erosion. Therefore, proper tillage and cropping practices are required if weeds are to be controlled, moisture conserved, and erosion curtailed. Wetting by rain, compaction, and activities of soil microorganisms greatly influence soil consolidation and aggregation and, consequently, erodibility by wind ( McCalla, 1950; Chepil, 1956, 1958). On the other hand, repeated wetting and drying, and especially freezing and thawing, of the surface soil tend to soften and disintegrate the surface crust and aggregates and to enhance wind erosion. Because of these counteracting processes, maximum degree of soil consolidation and aggregation occurs usually below the depth of 3 to 4 inches (Chepil, 1954b). Tillage to bring the consolidated soil material (clods) to the surface reduces erosion by wind if the vegetative cover buried is negligible. However, the effects of tillage are temporary because the forces of the weather tend to break the clods to sizes small enough to be moved by wind (Chepil, 1954b). As the clods at the surface are broken down, clods below the surface are being formed, making repeated tillage necessary in maintaining a cloddy surface indefinitely.

3. Devegetation us. Revegetation The most important basic cause of wind erosion is depletion or destruction of vegetation or vegetative residue on the land. Drought, at times, reduces or stops vegetative growth, but drought alone is seldom the cause of severe wind erosion. For example, little erosion occurred in semiarid regions of North America during drought periods when the land was protected by natural vegetation, but serious extensive erosion occurred during the drought periods after man began to burn and overgraze the vegetation and later to bury it by excessive and improper cultivation ( Sears, 1935; Jacks and Whyte, 1939; Malin, 1946). Vegetation must thrive and keep pace with soil weathering, cultivation, decomposition, and other denudation processes if wind erosion is to be kept in check.

B. CONDITIONS RESULTING FROM DESTABILIZATION AND STABILIZATIONPROCFSSES The above-mentioned processes associated with climate, weather, and human activities tend to create conditions that increase or decrease

236

W. S . CHEPIL A N D N. P. WOODRUFF

wind erosion ( Chepil, 1958 ) , They are equilibrium conditions, like the processes that produce them. They are as follows: 1. Increased vs. decreased surface wind velocity 2. Dry vs. moist soil particles 3. Light vs. heavy soil particles 4. Loose vs. consolidated soil particles 5. Smooth vs. rough surface 6. Bare vs. covered soil surface 7. Unsheltered vs. sheltered soil surface 8. Large vs. small eroding field 9. Improperly vs. properly oriented fields, crop strips, crop rows 1. lncreased vs. Decreased Surface Wind Velocity

Although atmospheric turbulence tends to increase the surface velocity of the wind and contributes greatly to movement of soil and snow by wind, control of this has not been possible. It has been possible, however, to slow down the surface velocity and therefore to reduce wind erosion by roughening the surface and by establishing windbreaks and barriers in the path of the wind (Woodruff, 1954; Woodruff and Chepil, 1956; Woodruff et al., 1957). These and other principles of wind erosion control will be discussed in Section VI of this review.

2. Dy us. Moist Soil Particles Only dry soil particles are readily moved by wind. Soil particles that have been oven-dried and those that have been air-dried are about equally erodible by wind. Damp and moist soil particles, due to cohesion of the water films, are virtually stable. The force of cohesion between erodible soil particles varies directly with moisture content. Seldom is TABLE I11 Influence of Equivalent Moisture, Me, of a Silt Loam Soil on the Rate of Soil Erosion Under Different Wind Velocities at 6-Inch Heighta Rate of soil erosion under Equivalent 20-m.p.h. velocity moisture (mg./cm. width/sec. )

0.01 0.25 0.29 0.34 0.71 1.03 a

315 295 235 230 68 2

Data from Chepil (1956).

28-m.p.h. velocity (mg./cm. width/sec.) 605 630 590 540 290 49

32-m.p.h. velocity (mg./cm. width/sec. 1 820 780 710 640 390 40

THE PHYSICS OF WIND EROSION AND ITS CONTROL

237

a natural wind strong enough to overcome the cohesive force of moisture at about 15-atmosphere percentage, which corresponds approximately to per cent water at permanent wilting point of plants, A relatively great increase in wind velocity is required to produce movement of discrete soil grains when their moisture content is increased slightly above the 15-atmosphere percentage. The 15-atmosphere percentage has an equivalent moisture, M e , of 1 (Table 111). The equivalent moisture is a ratio of the water content in question to water content at 15-atmosphere percentage. It is equal to w/w' in which w is the amount of water held by the soil grains and w' is the amount of water held by the same soil at a 15-atmosphere percentage.

3. Light us. Heavy Soil Particles Lighter particles are more erodible than heavier ones, but only if the diameter of the particles is greater than about 0.1 mm. Both size and density determine the weight, and therefore the erodibility, of the individual particles. Size is designated usually by diameter as determined by dry sieving. Density is defined as the weight in grams per cubic centimeter volume of a discrete soil grain or aggregate, including any air spaces within the grain or aggregate. It is convenient to express size and density together by what is known as equivalent diameter. Equivalent diameter is approximately equal to peD/2.65 in which pe is the bulk density of the erodible soil particles and D is their diameter as determined by dry sieving. The most erodible particles of 2.65 density are about 0.1mm. in diameter. They require the minimal drag velocity, known as threshold drag velocity (designated as V, t ) of about 15cm. per second to initiate movement. This is equivalent to about a 10-mile-per-hour wind velocity at 1 foot above a smoothed soil surface. Sizes greater and smaller than 0.1mm. equivalent diameter are less erodible by wind. The dividing point between erodible and nonerodible particles is not distinct for it varies with the drag velocity of the wind, the equivalent size range of erodible particles, and the proportion of so-called erodible and nonerodible fractions. Relatively few particles greater than 0.5 mm. in equivalent diameter are moved by common erosive winds, although a few LIP to 2mm. equivalent diameter may be moved by exceedingly high winds. For most mineral soils, the 0.5mm. equivalent diameter of a soil grain corresponds to about 0.84 mm. actual diameter. The 0.84 mm. is one of the sizes in the sieve series of the United States Bureau of Standards. This size of square sieve openings has been used, therefore, to separate the so-called erodible from the nonerodible soil fractions (Chepil and Bisal, 1943; Chepil, 1952).

238

W. S . CHEPIL AND N. P. WOODRUFF

Soil clods or aggregates that are just large enough not to be moved by wind are most effective in protecting the erodible soil particles (Chepil, 1958). This is because a unit volume of the smallest clods has the greatest surface for protecting the erodible particles. Density of the nonerodible fractions has no direct bearing on erodibility (Chepil, 1958). Only their size and to some degree their shape are the determining factors. Dust particles, especially those less than 0.02 mm. in diameter, are highly resistant to movement by direct force of wind. Moreover, they hinder the movement of the larger particles mixed with them. The more fine dust present in the wind-eroded soil, the greater is the threshold drag velocity of the wind required to initiate soil movement. Loose particles smaller than 0.01 mm., if not mixed with coarser particles and if placed in a bed that is thoroughly smoothed, are not moved even by an exceedingly strong wind. The high resistance of the fine dust particles to movement by wind is partly due to cohesion among the particles. More particularly, when the bed is thoroughly smoothed, the particles are too small to protrude above the viscous, nonturbulent layer of air, known as the laminar layer, close to the surface. It is known (Goldstein, 1938) that particles of height D would be submerged in the laminar layer as long as the Reynolds number of the form VID/v is less than 3.5. The kinematic viscosity, v, for air is approximately 0.15. If, on the other hand, the Reynolds number is greater than 3.5, the particles behave as obstructions in the path of the wind, throw off eddies to their lee sides, and disrupt the laminar layer. Under a force of wind equal to or greater than that required barely to move fine dust particles, the particles will disrupt the laminar layer if they are greater than 0.05mm. in diameter (Chepil, 1945b). If the surface composed of dust particles is roughened to a degree where the surface projections are at least 0.05mm. in height, movement of the particles takes place under a relatively low velocity of wind. In such cases the projections composed of many dust particles clinging together are broken off and moved bodily by the wind. Movement ceases as soon as the projections are leveled down to less than 0.05mm. in height. Under field conditions the surface roughness elements usually are much greater than 0.05mm. The dust particles cling to the larger grains, and therefore are moved readily with them.

4. Loose vs. Consolidated Soil Particles The most erodible soil condition exhibits no cohesion among the individual soil particles and aggregates. This condition is brought about by stirring or tilling the soil in a dry condition. Wetting a loose soil

THE PHYSICS OF WIND EROSION AND ITS CONTROL

239

bed followed by drying produces a certain degree of cementation (consolidation ) among the various individual particles and aggregates and tends to reduce erodibility by wind (Chepil, 1958). The degree of cementation is greatest at the surface of the ground, especially if the ground is exposed to impacts of raindrops (Fig. 10). On the other hand,

FIG.10. Surface crust on clay soil partly destroyed by abrasion with dune sand for 5 minutes with a wind velocity of 28 miles per hour at 1-foot height, exposing a more erodible soil beneath (Chepil, 1958).

frost action, tillage of dry soil, and abrasive force of windblown sand and soil particles all tend to loosen the bond between individual soil particles and aggregates and to increase the erodibility by wind.

5. Smooth vs. Rough Surface A smooth soil surface is generally more erodible by wind than is a rough one (Table I V ) . This is because it is less effective in slowing down the velocity of wind near the ground. A smooth surface reduces wind turbulence, but whatever effect the decreased turbulence has in reducing wind erodibility usually does not compensate entirely for the increased surface velocity (Chepil and Milne, 1914b). Roughening is not always effective in reducing wind erosion. If the soil is composed mostly of erodible fractions, roughening the surface does little good because the roughness elements continue to erode with

a0

W. S. CHEPIL AND N. P. WOODRUFF

the wind. But if the roughness elements, such as ridges, are composed of erodible and nonerodible fractions (as they usually are) the erodible fractions move from the ridges into the furrows where they are trapped, and the ridges soon become stabilized with a mantle of soil aggregates too large to be moved by the wind. TABLE IV Initial Rates of Erosion (Soil Flow) over Rough and Smooth Surfaces of Hatton Fine Sandy Loam under Different Wind Ve1ocitiesa.b ~

Wind velocity at 12-inch height (m.p.h.)

Smooth surface (g./cm. width/sec.)

Rough surface ( g./cm. width/sec. )

17 22 30

0.32 0.88 2.10

0.10 0.19 0.70

Rate of soil flow on

Data from Chepil and Milne (1941b). Rough surface was composed of ridges 2.5 inches high, 9 inches apart, at right angles to wind direction. a b

The aerodynamic surface roughness, A,, (Fig. 2) described in Section 11, A, is only one element of surface roughness that influences wind erosion. The other element, which is far more influential in reducing wind erosion, is the zero displacement height, Dh. Between the ground surface and height Dh the air is usually stagnant or is slow-moving and often laminar ( Geiger, 1957). The aerodynamic surface roughness plus the zero displacement height is the total surface roughness, which, in this review, is referred to simply as surface roughness. The greater the surface roughness, the lower is the wind velocity against the ground and the lower is the rate of erosion.

6. Bare vs. Covered Soil Surface The greatest frequency and magnitude of wind erosion occur on soils that have been partly or completely denuded of vegetation or vegetative cover. Bare, aggregated soils may exhibit resistance to erosion, but generally temporarily, because aggregates exposed to the weather usually disintegrate to erodible particles (Slater and Hopp, 1951; Chepil, 1954b). Covers other than vegetative also reduce or eliminate wind erosion if they are sufficiently resistant and durable against the force of wind. Those include various dust palliatives such as asphaltic, resinous, and latex films, gravel pavements, and chemical dust binders. Virtually all vegetative covers include both elements of roughness and cover and tend to reduce wind erosion on both counts. Pound for pound, a standing crop or stubble is more effective in controlling wind

THE PHYSICS OF WIND EROSION A N D ITS CONTROL

2/41

erosion than is flattened vegetation because it has greater total roughness. Pound for pound, a tall crop or stubble is more effective than a short one for the same reason (Chepil, 1944). On the other hand, fine vegetation or vegetative matter (such as wheat stubble) is a more effective cover than coarse (such as sorghum stubble) because it has a greater protective surface. Grass affords one of the best protective covers because it is finer than most cultivated crops, has a relatively great protective surface both above and below the surface of the ground, and is well anchored. Grass that is easily bent by wind is less effective in controlling erosion than grass that is not. More details on the relative effectiveness of different kinds and orientations of vegetative matter will be given under Section VI, A.

7. Unsheltered vs. Sheltered Soil Surface A surface may be bare and the soil finely divided, yet the soil may not be eroded if it is sufficiently sheltered from the force of wind. Sheltering is afforded on the lee sides of natural wind barriers such as shrubs, trees, hills, or mountains, or by artificial wind barriers such as walls, picket fences, hedges, crop strips, and crop rows. The extent and degree of sheltering afforded by the barriers vary with their spacing, height, width, shape, and air penetrability (Bodrov, 1935; Bates, 1944; Woodruff, 1954; Staple and Lehane, 1955; Caborn, 1957).

8. Large vs. Small Eroding Area The larger the unprotected field, the more it is erodible by wind, up to a certain limit. This is because the rate of soil movement (flow) increases with distance downwind across the wind-eroded field or adjoining fields until it reaches a maximum that a wind of a given velocity Several contiguous can sustain ( Chepil and Milne, 1941a; Chepil, 1957~). erodible fields are in effect as erodible as if they were one. Sometimes whole communities become as one eroding field-a condition that makes wind erosion control an exceedingly difficult problem.

9. Improperly vs. Properly Oriented Fields, Crop Strips, and Crop Rows Because the rate of soil flow increases with distance along the prevailing wind erosion direction, it follows that fields or field strips with their broad sides at right angles to, and their narrow sides parallel with, the prevailing wind will have the minimum overall rate of erosion ( Chepil, 1 9 5 7 ~ )Field . orientation is of little consequence where erosive winds blow equally from all directions. Under such conditions, small, nearly square fields are less subject to wind erosion and will trap the greatest amount of snow.

242

W. S. CHEPIL AND N. P. WOODRUFF

Orientation of crop rows with respect to prevailing wind direction is even more important than orientation of unprotected fields (Zingg et al., 1952). This is because crop rows are usually so close together that sheltering of the soil by the rows rather than reducing soil avalanching becomes a dominant influence. When rows are oriented as nearly as possible at right angles to prevailing erosive wind, the soil is better protected from the wind and more snow is accumulated in the sheltered areas. C. SOIL EROSION AND ITS RESULTING CONDITIONS

1 . Initiation of Soil Movement If the wind is increased gradually, a velocity is reached that starts the most erodible grains in motion. This velocity is known as the minimal fluid threshold velocity. A further increase in velocity causes heavier (denser or larger) grains to be set in motion. Further increases cause movement of still heavier grains. Ultimately, for soils containing only erodible fractions a velocity is reached which is just high enough to move all sizes. This velocity is known as the maximal fluid threshold velocity (Chepil, 1945b). Usually soils are composed of erodible and nonerodible fractions, so that no velocity is available that will remove all the fractions. For such soils no maximal fluid threshold velocity exists. The first type of movement of soil grains is in a series of jumps known as saltation (Fig. 11).The higher the grains jump, the more energy they derive from the wind. The impacts from the most erodible grains moving in saltation cause the movement of the larger, denser, and smaller particles. In the field, knolls, ridges, and other more exposed or more erodible spots first start to erode. Once erosion has started, it spreads fanwise to leeward and the bombarding action of the particles in saltation causes the movement of other particles. The threshold velocity under the bombarding action from the most exposed and most erodible grains is known as the impact threshold velocity (Bagnold, 1943). Both the fluid and the impact threshold velocities are the same for the most erodible grains, but the impact threshold velocity becomes increasingly lower than the fluid threshold for grains of increasingly greater size and density. The fluid and impact threshold drag velocities for dry grains greater than 0.1mm. in diameter vary as the square root of the product of equivalent diameter of the grain and the density relationship of the fluid and the grain (Bagnold, 1943). This square root law may be expressed by

THE PHYSICS OF WJND EROSION AND ITS CONTROL

243

in which D is the diameter of the grain, g the gravity constant, p’ is the immersed density of the grain, p is the density of the fluid, and a is a coefficient whose value depends on the range of equivalent size of particles present on the eroding surface and on whether movement is

FIG. 11. Photograph of paths of wind-blown soil particles moved primarily in saltation. Wind direction is from left to right. The photographed area is 1 inch high and 1.5 inches wide. Exposure is 1/200 of a second. Under magnification the paths appear as distinct spirals, indicating that particles spin as they fly through the air. The rate of spinning ranges from 1,200 to 60,000 revolutions per minute.

initiated by direct pressure of wind or by pressure and bombardment from the most exposed and most erodible grains. The relation between the threshold velocity vf at any height Z, equivalent diameter of the soil particles, and the roughness of the surface as exemplified by the value of k can be expressed by

As shown from Eq. (14), the greater the value of k, which varies with roughness of surface, the lower the velocity (at some fixed height) required to move the particles. This relation applies only to a condition where the roughness elements are the soil fractions moved by the wind. It means that the larger the erodible particles or the higher they are

244

W. S. CHEPIL AND N. P. WOODRUFF

perched on a rough surface, the higher they will protrude into the airstream and the greater the force of wind that would contribute to their movement, other factors being equal. On the other hand, where the roughness elements or the surface projections or barriers are nonerodible, the threshold law expressed by Eqs. (13) and (14) still applies, but the value of coefficient a is increased considerably. Under such a condition much of the surface drag is dissipated against the nonerodible fractions and only the residual drag contributes to the movement of erodible particles. If the soil material is composed only of erodible particles of a limited size range, such as an increment of dx commonly obtained by dry sieving, the value of coefficient a of Eqs. (13) and (14) based on centimeter-gram-second units is equal to about 0.1 for particles greater than 0.1 mm. in equivalent diameter. However, natural soil materials have a much wider range in size of fractions and therefore are associated with values of coefficient a larger and smaller than 0.1. If a soil, such as a commonly occurring dune material, is composed only of erodible fractions ranging from the largest down to the smallest erodible particles, the value of coefficient a of Eqs. (13) and (14) is only about 0.085. For such materials the threshold drag velocity varies as the square root of the average equivalent diameter of all the component particles (Chepil, 1958). Thus, the threshold drag velocity for a mixture of different equivalent sizes of erodible particles is lower than that required to erode only the largest of the particles. Movement of the larger particles is facilitated by bombardment received from the smaller particles moving in saltation. 2. Soil Movement Movement in saltation causes two other types of movement-the rolling and sliding of coarser grains along the surface of the ground, known as surface creep, and the floating of fine dust particles through the air, known as suspension. The presence of coarse grains and fine dust particles in the soil hinders the movement in saltation (Chepil, 1958). Coarse grains hinder the movement by sheltering the finer, more erodible grains from the wind. Dust hinders the movement by cohering to the grains and to other dust particles. Dust is readily kicked up by grains moving in saltation in the same manner as dust is kicked up by traveling vehicles, animals, etc. Once kicked up in the air, dust particles can be lifted high in the atmosphere by upward velocity of eddies of erosive wind. The upward eddies of erosive wind have a velocity of at least 2 or 3 miles per hour, sufficient to lift silt and some very fine sand to an

THE PHYSICS OF WIND EROSION AND ITS CONTROL

245

indefinite height. Dust clouds often extend 2 and 3 miles high and are the most visible and therefore the most dramatic aspects of “dust storms.” The proportion of the three types of soil movement varies greatly for different soils. In the cases examined, between 50 and 75 per cent of the weight of the soil was carried in saltation, 3 to 40 per cent in suspension, and 5 to 25 per cent in surface creep (Chepil, 1945a). a. Rate of soil mouement. If the wind velocity is greater than that required barely to move the soil grains, then, according to Bagnold (1943) for dune sands and Chepil ( 1 9 4 5 ~ )for dry soils

q = a v D-, -P ( V . ’ ) 3 g in which q is the rate of soil movement (total weight of soil material moved past a unit width normal to the direction of movement and of unlimited height per unit time). Equation (15) shows that the rate of soil movement varies directly as the cube of the drag velocity V’. and as the square root of the average equivalent diameter D, of the soil particles moved by wind. The coefficient a varies greatly with different conditions. It varies with the size distribution of the erodible particles (Chepil, 1941; Bagnold, 1943), the proportion of fine dust particles present in the mixture (Chepil, 1941, 1945c), the proportion and size of nonerodible fractions (Chepil, 1941, 195Ob), position in the field (Chepil and Milne, 1941a), and the amount of moisture in the soil ( Chepil, 1956). All these factors, and perhaps many more, affect the rate of soil movement and hence the value of coefficient a. Equation (15) applies equally well to movement in saltation, suspension, and surface creep (Chepil, 1945a). b. Quantity of erodible soil. The rate of movement of cultivated soils is seldom constant; it changes with the surface conditions of the soil, which, in turn, change with the duration of exposure to the wind and with the erosional history of the field. For that reason the weight of soil material removable from the surface by the wind under some conditions is a more accurate measure of erodibility of dry cultivated soils than the rate of soil removal. The weight of soil material X that is removable from a given area by the wind may be expressed in terms of drag velocity of the wind by X=a(V*’)5 (16) where the coefficient a varies with many factors. The quantity of erodible soil for a given drag velocity varies in great measure with the degree of soil abrasion as influenced by the characteristic length of the eroded area. For that reason it is better to express

246

W.S. CHEPIL AND N. P. WOODRUFF

erodibility in dimensionless form applicable to any size of eroding area, direction of wind, or units of measure by

I , = a( V1’)6

(17) In Eq. (17) I , is the soil erodibility index, which is equal to X2/Xl where X I is the weight of soil material removable per unit area from a “small” area, such as in a wind tunnel, where the soil contains 60 per cent of clods greater than 0.84mm. and X 2 is weight removable under the same set of conditions from soil containing any other proportion of clods greater than O.84mm. in diameter. “Small” area has a width not exceeding 30 feet along the direction of the wind (Chepil and Woodruff, 1959). c. Dust concentration and visibility. The concentration of dust (weight of dust in unit volume of air) was found by Chepil and Woodruff (1957) to vary with height in accordance with an empirical power equation of the form a c, = Zb

in which C, is concentration of dust at height z above ground and a and b are constants. Constant a varies with the intensity of erosion and b is approximately equal to 0.28. Equation (18) agrees reasonably well with the basic formula of Schmidt as reported by Vanoni (1946) in his experiments with water. TOTAL DUST LOAD C,

.04.06 .I

.2

,4

.a

I

2

(TONSICU. MILE )

4

6

OUST CONCENTRATION C. AT 6 FEET

I0

20

4 0 6 0 100

IM G I C U . FT.)

FIG. 12. Relation between daytime visibility and dust load and concentration (Chepil and Woodruff, 1957).

THE PHYSICS OF WIND EROSION AND ITS CONTROL

247

A quick estimate of an approximate total quantity of dust load suspended in the lower atmosphere can be obtained from daytime visibility data of Chepil and Woodruff ( 1957) and Langham-et al. ( 1938), using Fig. 12. The total dust load C, refers to the square. mile against the earths surface and 1 mile high. Dust clouds have been observed to extend up to more than 2 miles in height. The dust concentration above the 1-mile height may be estimated for any intensity of erosion by assuming the relationship given in Eq. (18). Such an estimate would be only approximate. 3. Avalanching On an unprotected eroding field, the rate of soil flow is zero on the windward edge and increases with distance to leeward until, if the field is large enough, the flow becomes the maximum that a wind of a particular velocity can sustain. The acceleration of soil flow with distance downwind over an unprotected field is known as soil avalanching ( Chepil, 1 9 5 7 ~ ) . Maximum rate of flow is approximately the same for all soils and is about equal to that of dune sand. However, most fields are not large enough for development of maximum rate of soil flow. The distance required for soil flow to reach a maximum on a given soil is the same for any erosive wind. It varies only and inversely with erodibility of a field surface. That is, the more erodible the surface, the shorter the distance in which maximum flow is reached (Table V). TABLE V Relation between Wind Tunnel Erodibility of a Field Surface and Distance for Soil Flow to Reach a Maximuma

a

Wind tunnel erodibility, I , ( dimensionless)

Distance in the field for soil flow to reach a maximum (feet)

920.0 300.0 50.0 39.0 19.0 7.5 6.1 5.1

180 300 1,100 1,600 2,200 3,900 4,100 5,200

Data from Chepil ( 1959a).

Any factor that influences the erodibility of the field surface influences the rate of soil avalanching. The lower the proportion of nonerodible soil clods or the less the amount of crop residue, the greater is the erodibility of a field surface, the higher is the rate of soil avalanching and the

248

W. S. CHEPIL AND N. P. WOODRUFF

narrower the field or field strips have to be to keep the rate of erosion down to some tolerable rate. 4. Sorting The wind moves the finer and lighter particles faster than the coarser and denser ones (Udden, 1898; Fly, 1935; Moss, 1935; Daniel, 1936, Chepil, 1957a). The finer the eroded particles the greater is their speed, height, and distance of travel. The finer particles have greater mobility despite the fact they are less erodible. The wind separates the soil into several distinct grades as follows: ( a ) Residual soil m a t e k l s : Nonerodible clods and massive rock materials that remain in place. ( b ) Lag sands, lug gravels, and Zag soil aggregates: Semierodible grains that have been moved primarily by surface creep. ( c ) Sand and clay dunes: Accumulations of highly erodible grains that have been moved primarily in saltation. ( d ) Loess: Dust lifted off the ground by saltation and carried high in the air and deposited in uniform layers both near and far from dunes. Dust is carried in true suspension. The composition of freshly deposited dust is like the composition of the loess laid down in the Pleistocene age (Swineford and Frye, 1945; PBwB,1951; Warn and Cox, 1951; Chepil, 195%). Huge deposits of loess in many regions of the world show the great importance of wind as a geologic mover of dust. There are no distinct demarcations of size between the various grades of wind-sorted materials. The size limits of one grade overlap considerably with size limits of another grade ( Chepil, 1957b). In some cases, wind erosion virtually removes the surface soil (Zingg, 1954; Chepil, 1957a,b). This nonselective remooal by wind is associated primarily with loess which was already sorted and deposited from the atmosphere during past geologic eras. Another type of soil removal is selective removal. It occurs on soils developed from glacial till, residual material, mountain outwash, and sandy soils of various origins. On these soils the wind tends to remove the silt and clay and to leave the sands and gravels behind. This process often causes the surface soil to become progressively more sandy and therefore more erodible and less productive.

5. Abrasion Abrasion by impacts of particles transported along the surface by wind is an important phase of the wind erosion process on all soils ( Chepil, 1945d).

THE PHYSICS OF WIND EROSION AND ITS CONTROL

249

Soils usually are covered with a thin crust that is somewhat resistant to wind erosion. As soon as some soil particles are loosened and moved by wind, their abrasion against the surface causes the crust to disintegrate and exposes a more highly erodible soil (Fig. 12). Also, the nonerodible clods gradually break under impacts of saltating grains. The longer erosion continues, the greater is the quantity of erodible material formed by abrasion and the higher the rate of soil flow. The materials detached from clods and surface crust by abrasion accumulate on the leeward side of fields or, if they are fine, are carried far through the atmosphere. Abrasion caused by wind-blown soil grains is also extremely injurious to plants. For example, yield of winter wheat forage at maturity was reduced 78 per cent and of grain 86 per cent after only one 10-minute exposure of young plants to soil movement caused by a 28-mile-per-hour wind at 6-inch height (Woodruff, 1956). The greater the intensity and the more prolonged the soil movement, the greater is the destruction of vegetation on the lee of an eroding area. By the same token, the lower the rate of erosion, the more vegetation tends to cover and stabilize the soil. The equilibrium between climatic environment, soil conditions, and vegetative growth shifts continually with the seasons and with the way man uses his land. V. Soil Properties That Influence Wind Erosion

Chepil (1958) found that soil properties or conditions that influence wind erosion directly may be grouped into the following four categories: ( a ) Stability of soil against erosion as influenced by cohesive and dispersive forces of water and raindrops. ( b ) State of soil structure, such as size, shape, and density of erodible and nonerodible soil fractions. ( c ) Stability of soil structure against breakdown by mechanical agents, such as tillage, abrasion from windblown materials, and direct force of wind. ( d ) Stability of soil structure against breakdown by natural causes, such as wetting and drying and freezing and thawing. The above-mentioned properties are known as the primary properties because they influence erodibility directly. Most of them in turn are influenced by basic factors inherited by or imparted to the soil. Until the influence of the primary soil properties or conditions is thoroughly understood and expressed, it will be difficult or impossible to evaluate the importance of the basic soil factors that affect the primary properties and erodibility by wind. The basic factors will be discussed in Section V, E, 1 4 .

250

W. S. CHEPIL AND N. P. WOODRUFF

The most important primary soil conditions that influence wind erosion have been described in Section IV, B, 24.In this section, we shall indicate where these conditions exist in the soil, how they are created, and how best they may be maintained. A. COHESIVE AND DISPERSIVE FORCXS OF WATERAND RAINDROPS Soil erodibility by wind is a function of the cohesive force of the adsorbed water films surrounding the discrete soil particles. The resistance, r, due to cohesion of the water films among the discrete soil grains and to the force of gravity on the grains, must be overcome by the wind before erosion can occur, The values of resistance, 7, were where M e is the equivalent moisture. found to be equal to 6( Since for “smooth soil surfaces (with surface roughness elements not exceeding 1 inch in height) V’. is equal to dq,the rate of movement of moistened erodible particles on such surfaces, utilizing Eq. (15), may be expressed by

and the relative quantity of moistened soil material removable from a limited area before soil movement ceases, utilizing Eq. (17), may be expressed by

Equations (19) and (20) apply only to conditions where moisture has been added to originally loose, dry soils. They do not apply to soils that have been moistened and then dried to various degrees, thereby causing a substantial degree of cementation of the originally discrete soil fractions-a cementation due to shrinkage of the water films on fine particles by drying. Wetting and drying cause little cementation of drifted soil materials, such as those accumulated in drifts by wind, but they cause considerable cementation of most other soil materials. The drifted materials that cover much of the surface of eroded fields are composed essentially of waterstable grains devoid of fine dust particles required to bind them together. Only the impacts from a few grains moving in saltation are needed to separate the water-stable grains and to start them again in motion by the wind. On the other hand, when a loose soil other than drifted material is wetted and dried, the fine particles tend to bind the whole soil body to

THE PHYSICS OF WIND EROSION A N D ITS CONTROL

251

form a somewhat compact mass more resistant to wind than was the originally loose soil. Then, too, a surface crust almost invariably is formed, owing to impacts of raindrops on the ground. Except at the immediate surface, the primary (water-stable) aggregates and the secondary aggregates, or clods, usually undergo little transformation by individual wetting from rain and drying. A greater change occurs in the degree of compactness and cementation among the various recognizable aggregates. This type of cementation has an important influence on erodibility by wind, but the degree of cementation generally is too weak to be detectable by wet or dry sieving. Thus, wet or dry sieving, or elutriation in water or air, does not measure directly some important phases of soil structural stability that influence the erodibility by wind. In addition to the above-mentioned conventional methods of structural analysis, other methods must be used if erodibility is to be determined fully. One of these methods is a direct measure of stability, or resistance, of the various structural units to breakdown by abrasion from windborne soil particles, as will be described next.

B. MECHANICAL STABILITYAND ABRADABILITY OF SOILS T R U ~ R A UNITS L Resistance of a dry soil to breakdown by mechanical agents, such as tillage, force of wind, or abrasion from windborne materials, is known as mechanical stability. It is due to coherence of the soil particles. Mechanical stability has been determined conveniently by dry sieving and repeated dry sieving on a rotary sieve (Chepil, 1951). Mechanical stability of the various phases of field structure is a relative measure of the resistance to disintegration by abrasion to which the soil is subjected when it is eroded by wind. The relative resistance of the soil to abrasion by windborne soil particles has been expressed as the cwicient of abrasion (Chepil, 1955a). It is the quantity of soil material abraded off a soil aggregate per unit weight of abrader blown against the aggregate by a 25-mile-perhour windstream. Since the amount of abrasion varies as the square of wind velocity, the coefficient of abrasion a can be expressed by the equation 2

a =A,,(:) in which A, is the weight abraded per unit weight of abrader blown at wind velocity D expressed in miles per hour. The coefficient of abrasion (abradability ) of the different structural units of the soil varies inversely with their mechanical stability, as determined by repeated dry sieving (Table VI). Furthermore, modulus of rupture, a measure of cohesive

252

W. S. CHEPIL AND N. P. WOODRUFF

strength of soil briquets as determined by the method of Richards ( 1953), varies inversely with the coefficient of abrasion and inversely with the diameter of mechanical soil particles from which a briquet is formed. Owing to abrasion, soil structure breaks down progressively as wind TABLE VI Relation between the Coefficient of Abrasion and Mechanical Stability of Different Phases of Field Structure of Soil5 Mechanical Phase of Soil Coefficient stability (%I field structure class of abrasion 2.94 3.0 Sandy loam Consolidated fraction 1.92 8.8 < 0.42 mm. from Silt loam 2.03 Silty clay loam 5.0 fresh drifts Clay 2.45 4.6 Consolidated fraction < 0.42 mm. from residual soils

Sandy loam Silt loam Silty clay loam Clay

1.48 0.61 1.14 1.25

17.0 28.1 27.3 17.4

Consolidated residual soils

Sandy loam Silt loam Silty clay loam Clav

0.46 0.14 0.23 0.32

40.8 54.4 65.4

Clods % to % inch diameter from residual soils

Sandy loam Silt loam Silty clay loam Clay

0.011 0.003 0.005 0.002

83.8 91.7 90.6 92.8

5

56.0

Data from Chepil (1958).

erosion continues. The amount of breakdown depends on mechanical stability of the structural units. Initiation of perceptible soil movement for the first time in the field generally requires a much higher drag velocity than for succeeding windstorms, since the soil usually is covered with a transient surface crust that is initially resistant to wind erosion. As soon as some soil particles are loosened and moved by wind, their abrasive action against the surface causes the crust to disintegrate and expose a more highly erodible soil (Fig. 10). Then, too, the nonerodible clods gradually become broken down by impacts of saltating grains. The erodible fractions are being sorted continually from the less erodible fractions and usually are piled in hummocks in the vicinity of the eroded area. The longer erosion continues, the more erodible material accumulates on the leeward side of an isolated field and the lower is the velocity of wind required to initiate erosion.

T HE PHYSICS OF WIND EROSION AND ITS CONTROL

253

Therefore, a range of threshold drag velocity for any soil depends on the previous erosional history of the field. This range varies from the original threshold velocity of the previously noneroded field to the threshold velocity of dry dune materials. The lowest threshold velocity for a dry dune material is about 13 miles per hour 1 foot above smooth ground (Chepil, 194%). The highest threshold velocity for a previously noneroded soil is indefinite; some fields do not erode no matter how high the natural wind. Surface soil, such as exists in the field after wetting and drying, is not homogenous, although often it appears to be so. It is composed of various types of structural units cemented together in varying degrees (Chepil, 195313). The strength of cementation and, consequently, the abradability when the soil is dry vary greatly for different soils and different structural units of the soil. Two types of soil cements seem to be responsible for consolidation of the soil in different structural units: ( 1 ) water-insoluble and ( 2 ) water-soluble or water-dispersible. These cements appear to be responsible for the following types of structural units with distinct degrees of mechanical stability and abradability by wind: ( 1) primary (water-stable) aggregates; ( 2 ) secondary aggregates, dry aggregates, or clods; ( 3 ) fine materials among the secondary aggregates; and ( 4 ) the surface crust. These phases of field structure in cultivated soils are shown in Fig. 13. Each secondary aggregate in Fig. 13 is designated by a line surrounding a number of primary aggregates, of which the secondary aggregate is composed.

I

---z

Surfacecrust

@

Primary aggmgates

Secondary aggregate *$

Maierlals among the secondary

aggregates

FIG.13. Diagrammatic representation of structure of cultivated soil after wetting

by rain and drying (Chepil, 1953b).

254

W.

S. CHEPIL AND N. P. WOODRUFF

1 . Water-Stable Aggregates These primary aggregates, which seldom exceed l m m . in diameter in cultivated dryland soils, are held together by water-insoluble cements composed of clay particles and irreversible or slowly reversible inorganic and organic colloids ( McCalla, 1950). The water-stable granules possess high strength of coherence and stability against the disintegrating forces of the weather (Chepil, 1951, 1953a, 1954a). Since they are the most stable structural units of the soil, they represent the units into which the secondary aggregates ultimately disintegrate, both by forces of weather and by abrasive action of wind-eroded soil particles, The waterstable aggregates are readily separated from the other soil fractions by the wind and usually are accumulated in drifts or mounds within and outside the eroded fields. Particles finer than the water-stable aggregates are removed in the form of dust, while the coarser fractions (clods, gravel, and rocks) remain behind as residual soil materials. The drifted particles are principally individual water-stable aggregates or discrete sand grains. The drifted sand grains and clay aggregates exhibit the greatest mechanical stability, whereas those of intermediate texture exhibit a somewhat lower mechanical stability (Table VII ). Without appreciable quantities of fine dust, the windblown grains tend to remain as discrete units, giving the soil materials a characteristically TABLE VII Mechanical Stability of Different Structural Units and of Fine Materials among the Structural Units of Wind-Eroded and Residual Soil Materialsa Mechanical stability Sandy loam

Silt loam

Silty clay loam

Clay

(%I

(%I

(%I

(%I

Particles > 0.42 mm. from fresh drifts (chiefly water-stable)

97.6

95.5

95.0

97.0

Dry aggregates or clods > 0.42 mm. obtained by dry sieving

83.8

91.7

90.6

93.8

Surface crust %- to %-inch thick on residual soil

60.2

73.3

69.3

58.5

Particles < 0.42 mm. from residual soils after consolidation0

17.0

28.1

27.3

17.4

3.0

8.8

5.0

4.6

Structural units

Particles < 0.42 mm. from fresh drifts after consolidationb

Datafrom Chepil (1958). Consolidation was accomplished by spraying the dry soil material in a column 2 inches high with 1 inch of water, followed by drying. 'j

THE PHYSICS OF WIND EROSION AND ITS CONTROL

255

mellow structure commonly referred to as “good tilth.” An ideal structure from the standpoint of resistance to erosion by wind and of other desirable features is a soil that has a substantial proportion of water-stable aggregates greater than lmm. in diameter. Any treatment that would achieve this condition would aid greatly in establishing lasting resistance to wind erosion. Presently, dryland soils are virtually devoid of waterstable aggregates large enough to resist movement by wind. Their resistance to wind erosion has been enhanced by formation of secondary aggregates known commonly as clods.

2. Secondary Aggregates or Clods Secondary aggregates are next in order of mechanical stability, depending on soil class, depth, and tillage treatment. They are held together in a dry state primarily by water-dispersible cements acting under pressure from depth and time. The cements are composed mainly of water-dispersible particles smaller than 0.02 mm. in diameter (Table VIII ) . When these fine particles are removed by repeated decantation TABLE VIII Relation between Dry Clod Formation and Percentage of Particles < 0.02 mm. Dispersed in W a t e p Clods

Soil material and treatment Dry sieve fraction < 0.42 mm., consolidatedb

Soil textural class Sandy loam Silt loam Silty clay loam Clay

Particles < 0.02 mm. dispersed in water

> 0.42 mm. after dry sieving

(%I

(%I

10.2 19.3 18.2 9.8

17.0 28.1 27.3 17.4

0 0 Dry sieve, fraction < 0.42 mm. Sandy loam Silt loam 0 0 from which particles < 0.05 mm. 0 .09 were removed by shaking and Silty clay loam repeated decantation in water, Clay 0 .23 and then consolidated* a Data from Chepil (1958). ‘1 Consolidation was accomplished by spraying dry soil material in a column 2 inches high with 1 inch of water, followed by drying.

after shaking in water, the water-stable aggregates to which the clods disintegrate after being shaken in water are much like sand grains in that they fail to cohere to each other after a layer of them has dried ( Chepil, 1958). Fine water-dispersible particles are necessary to bind the water-stable aggregates together to form clods.

256

W. S. CHEPIL AND N. P. WOODRUFF

Many clods maintain their identity for some time after repeated wetting and drying in the field. Individual rains have little influence on the form or compactness of clods below the surface, even after they lose their visible identity after the soil is wetted and dried. Only within a narrow zone of the immediate surface where the soil mass assumes a structure distinctly different from that below do the clods become appreciably disintegrated by impact of raindrops. Abrasive tests have indicated that after repeated wetting and drying the clods become merely embedded in the fine, loosely consolidated portion of the soil. The strength of cementation between the clods is generally much lower than within the clods; this is why blocks of soil abrade unevenly when exposed to impacts of windborne soil grains. In some extreme cases, however, a surface soil may become completely cemented into a single, seemingly homogeneous mass.

3. Materials among the Clods The cohesive forces that exist among the clods after the soil has been wetted and dried vary greatly, as within the clods, depending on the number and the nature of wettings, on the depth and consequent pressure exerted against the soil, and on the physical-chemical nature of the soil. The degree of cementation that holds the clods together after the soil has been wetted and dried is due in large measure to the quantity of particles of the size of silt and clay dispersible in water (Chepil, 1958). Wetting apparently causes either some water-soluble or waterdispersible cements to become released from the originally discrete structural units; on drying, the cements cause a certain degree of cementation between the units. The greater the quantity of fine particles dispersible by water, the greater the degree of cementation among the structural units and the greater is the resistance of the soil to breakdown by mechanical forces after it has been wetted and dried. Pressure likewise increases the cementation among the clods and other structural units. The greater the depth, the greater the pressure exerted on the soil and the greater the degree of cementation and mechanical stability among the structural units, until the whole soil mass, at a certain depth, may become strongly cemented together. This condition often is referred to as a massive structure. Tillage breaks the massive structure to various sizes of blocks referred to as clods. Tillage, if suitable, may bring the clods to the surface to resist erosion by wind. The fine particles that tend to cement the clods and other structural units together are composed of silt, clay, and various materials of organic and inorganic origin. Dispersed silt, although usually not considered as a soil cement, acts as a weak cement of sufficient strength to resist

257

THE PHYSICS OF WIND EROSION A N D ITS CONTROL

considerably the force of wind (Chepil, 1955a). Silt particles are dispersed by water much more readily than clay particles. The presence of large quantities of dispersed silt particles in a soil appears to cause the formation of a compact, massive structure, which, while quite resistant to wind erosion, may present a serious structural problem otherwise. Bradfield and Jamison (1938) concluded that hard and intractable soils were usually those largely composed of fine silt having a single-grain structure when dispersed in water.

4. The Surface Crust Because of impacts of rain, the soil material at the surface becomes more dispersed than the soil below. On drying, the dispersed soil forms a thin surface crust that is more compact and mechanically stable than some parts of the soil below. The crust often does not exceed onesixteenth inch in thickness, but occasionally it may reach a thickness of one-fourth inch or more. The crust is easily recognizable by its dense, platy structure. This type of structure becomes less distinct with depth, until it merges with the soil below. Medium-textured soils containing a high proportion of silt are most subject to dispersion in water and, therefore, these soils produce the thickest and most compact crust (Table IX) . This property contributes to the usually high resistance of the medium-textured soils to erosion by wind. Sandy soils generally are less subject to surface crust formation, because they do not contain a high proportion of silt and clay. That property contributes considerably to the high erodibility of sandy soils TABLE IX Relation between Mechanical Stability of the Surface Crust and Percentage of Particles < 0.02mm. Dispersible in Watefl Particles

Mechanical stability of crust

< 0.02 mm.

(%I

(%I

44.7 60.2

6.2 10.4

dispersed in water

Soil textural class Sandy loam

Soil material Drifted Residual

Silt loam

Drifted Residual

60.8 73.2

10.2 16.8

Silty clay loam

Drifted Residual

59.7 69.3

10.4 15.4

Clay

Drifted Residual

38.1 58.5

4.9 9.6

0,

Data from Chepil ( 1958).

258

W.

S . CHEPIL AND N. P. WOODRUFF

by wind. Clay soils are highly variable with respect to wind erosion. Those that contain a high proportion of fine water-dispersible particles tend to puddle and resist erosion by wind. On the other hand, some clays are not subject to a high degree of dispersion; consequently, the surface crust and the clods tend to remain as fine granules, some of which are moved readily by wind. Rain often carries some of the finely dispersed and water-soluble cementing materials downward, leaving the coarser particles, such as sand or water-stable aggregates, at the top. Some of these coarser particles remain loose on the surface and often contribute to the initial stage of wind erosion. Being on the surface, they dry rapidly. Consequently, these coarser particles may be moved by wind soon after a rain, even before the drying of the surface has become apparent. Abrasion from these particles tends to wear down the surface crust, to hasten the drying of the surface, and to accelerate the soil movement as long as the wind that is strong enough to move the soil material continues. Small showers often tend to smooth the soil surface, to loosen some of the surface particles, and, if the field is large, to accelerate rather than alleviate soil movement by wind. On many soils the rate of soil movement is slow at the beginning, but it accelerates as the surface crust is worn through and a weakly consolidated soil beneath it is exposed to the wind (Chepil, 1953b, 1 9 5 7 ~ )The . nature of the surface crust and its relation to erosion by wind perhaps can be interpreted best from its appearance as it is destroyed by abrasion with dune sand (Fig. 10). The surface crust proved completely stable under the same wind of 28 miles per hour without the abrader.

5. Order of Mechanical Stability and Abradability Susceptibility of the soil to abrasion by impacts from windborne soil material varies inversely with its mechanical stability (Table V I ) . The order of mechanical stability from highest to lowest, and hence the order of abradability from lowest to highest, for the different structural units in a dry state is as follows: ( 1) water-stable aggregates, ( 2 ) secondary aggregates or clods, ( 3 ) surface crust, and ( 4 ) fine materials among the clods cemented together and to the clods after the soil has been wetted and dried. The last of the structural units at some depth below the surface may possess mechanical stability and abradability approaching that of clods. Mechanical stability tends to reduce wind erosion by resisting the breakdown of nonerodible units to smaller erodible particles.

THE PHYSICS OF WIND EROSION AND ITS CONTROL

259

C. RELATIVE IMPORTANCE OF STATEAND STABILITYOF SOIL STRUCTURE Erodibility of the soil depends on ( 1 ) size, shape, and density of the structural units, and ( 2 ) mechanical stability of the structural units. The first may be referred to as the state of structure, and the latter as the stability of structure. Both phases of structure are measurable by elutriation, dry sieving, and repeated dry sieving. The relative importance of the state and stability of dry structure with respect to erodibility by wind varies with the area of the field, the roughness of the surface, and many other factors. If the area of the field is small, the amount of abrasion from erosion is small and erodibility of the field is determined primarily by the state of structure, or specifically by the proportion of discrete granules small enough to be moved by wind. If, on the other hand, the field is large, mechanical stability of the structural units is the more important factor. In such case, if the soil structural units lack mechanical stability, the presence of even a small quantity of loose, erodible material on the surface is usually sufficient for substantial disintegration of the structural units by abrasion from windborne material and for consequent intense erosion of the loosely cemented soil (Chepil, 1951). The relative importance of the state and stability of structure of different soils is shown in Table X, based on wind-tunnel tests. A surface crust formed by spraying the soil with water and then drying (condition b ) reduced greatly the quantity of soil material eroded by wind. However, when the soil that was wetted and dried was subjected to impacts of soil particles blown in from the outside (condition c ) , the crust soon was worn through, the rate of soil removal was increased considerably, and erosion continued as long as the stream of sand passed over the soil. The amounts of erosion occurring under condition b are comparable to those obtained in small, isolated fields where abrasion is limited; the amounts of erosion occurring under condition c, on the other hand, are applicable to those on the leeward sides of large, open fields where the intensity of abrasion from eroded particles is relatively great.

D. SEASONAL INFLUENCES ON SOILSTRUCTURE AND ERODIBILITY Biological activities and alternating wetting and drying and freezing and thawing have a strong influence on soil structural conditions and erodibility (Peele, 1940, 1941; McCalla, 1942, 1945, 1950; Martin, 1946). The structural conditions and erodibility fluctuate in accordance with the varying influences of the seasons. Soil cloddiness and mechanical stability of clods are decreased and erodibility increased in winter in cases where the soil is moistened at

TABLE X The Influence of State of Structure and Stability of Structure on Erodibility of Soil by Wind= Clods > 0.5mm. equivalent diameter

Degree of cementation between the clods after consolidation

Soil class

(%I

Sandy loam Silt loam Silty clay loam Clay

39.8 32.3 42.1 12.1

b

I I"

Amount or rate of soil erosionb

(%I

Condition a ( tons/acre )

Condition b (tons/acre)

Condition ( tons/acre/min. )

17.0 28.1 27.3 17.4

3.4 4.5 2.9 9.5

0.4 .2 .3 3.4

13.0 5.6 9.4 11.0

C

Data from Chepil (1958). Conditions: a-Exposure to wind of well-mixed, loose, and dry soils till movement ceased. &Exposure to wind after consolidating the soil by spraying with 1 inch of water and drying. Exposed till movement ceased. c-Exposure to wind and a stream of windborne sand after consolidating the soil. Rate of sand flow was 1,000 g. per minute per 8-inch width.

#

8 5

1:

?

3 8

rrl

261

THE PHYSICS OF WIND EROSION AND ITS CONTROL

least occasionally (Table XI). Also, the changes are greatest at or near the surface of the ground and least, if any, at a 6-inch depth, The changes in cloddiness and stability of clods with depth vary with the seasons and with soil texture. The effects of tillage are temporary, because the forces of the weather, especially freezing and thawing of moist soil during the winter, tend to break the clods to sizes small enough to be moved by wind. As the clods at the surface are broken TABLE XI Influence of Seasons on Some Phases of Soil Structure and Wind Erodibility at Various Depthsajb

(%I

Amount eroded in tunnel until movement ceased ( tons/acre )

65.0 46.7

87.8 72.7

0.40 1.50

Fall Spring

71.9 58.1

87.8 80.0

0.24 0.80

Fall Spring

80.5 80.5

88.8 90.6

0.06 0.09

clods

> 0.84 Depth ( inches )

Season

m. (%)

Otol

Fall Spring

1 to 3 3 to 6 a

b

Mechanical stability

of

Data from Chepil (1954b). Averages for Cass loam during a three-year period at Manhattan, Kansas.

down, however, clods below the surface are being formed. Hence, repeated tillage of a proper type is useful in maintaining a cloddy surface indefinitely. The degree of cloddiness that can be maintained varies with the nature of the soil and with the depth and nature of tillage.

E. BASICSOIL FACTORS The basic soil factors affect wind erosion indirectly. The most important of these are soil texture, water-stable structure, organic matter, soil microorganisms and various products of organic matter decomposition, moisture as it influences wind erosion indirectly, calcium carbonate, water-soluble salts, and nature of the soil colloids. Some of these factors, such as soil moisture, affect erodibility directly by affecting the resistance to the forces of erosion and indirectly by influencing the state of structure such as size, shape, and density of the water-stable aggregates and clods.

261,

W. S. CHEPIL AND N. P. WOODRUFF

1 . Soil Texture The relationship between the soil erodibility index I , and the amount of clay, on the average, conforms with the equation

I,, = aGb'C

(22)

in which G is percentage (by weight) of clay in the soil and a, b, and c are constants (Chepil, 1952). In general, the higher the proportion of silt (0.002 to 0.05 mm. ) in the soil, the higher the percentage of nonerodible clods and the lower the soil erodibility, On the other hand, the higher the proportion of sand (0.05 to 0.5 mm. ) in the soil, the lower the percentage of nonerodible clods and the higher the erodibility. These relations do not apply to all soil classes, nor especially to the finer textured soils. This is because clay, rather than silt or sand, is a predominant factor influencing erodibility by wind. Nevertheless, silt is a factor that tends to decrease, whereas sand tends to increase, the erodibility. In all soil textures, erodibility is inversely associated with the proportion of nonerodible clods as determined by rotary dry sieving. Sand grains have little or no cohesive property, are readily loosened by force of impact of windborne materials, and execpt for the coarse grains 0.5 to 1mm. in diameter, are easily carried by wind. Silt and clay, on the other hand, cohere after wetting and drying, and therefore seldom exist as individual particles but act as binding agents in the formation of nonerodible clods. The relative effectiveness of silt and clay as binding agents depends somewhat on their proportion to each other and to the sand fraction. The first 5 per cent of silt or clay mixed with sand is about equally effective in creating cloddiness, but the quality of the clods is different. Those formed with clay and sand are harder and less subject to abrasion by windborne sand than those formed from silt and sand. For proportions greater than 5 per cent and up to 100 per cent the silt fraction creates more clods, but the clods are softer and more readily abraded than those formed from clay and sand. The greatest proportion of nonerodible clods exhibiting a high degree of mechanical stability and low abradability is obtained in soils having from 20 to 30 per cent of clay, 40 to 50 per cent of silt, and 20 to 40 per cent of sand (Chepil, 1955a). 2. Water-Stable Structure The effects of water-stable aggregates and fine water-dispersible particles on the state and stability of soil clods and on erodibility by wind were discussed in Section V, B, 1-3. This section is devoted only

THE PHYSICS OF WIND EROSION AND ITS CONTROL

263

to the influence of the total water-stable structure on soil cloddiness and erodibility by wind. Russell (1938) in his review on soil structure referred to the waterstable particles ( as conventionally determined by sedimentation, elutriation, or sieving of soil in water) as the building blocks of field structure of soils. Some of these particles are primary particles of sand, silt, and clay and some are water-stable aggregates, often referred to as primary aggregates. Few primary particles or aggregates exist individually in soils. They usually are grouped into secondary aggregates commonly referred to as clods (Fig. 13). Chepil (1943) indicated that both coarse ( > 0.42 mm.) and fine ( < 0.02 mm. ) water-stable particles increase cloddiness and decrease erodibility by wind. He (Chepil, 1953a) further indicated that each unit per cent change in water-stable particles > 0.84 mm. influences the proportion of nonerodible clods and erodibility about equally (Table XII). The relations between the dry nonerodible clods B , > 0.84 mm. TABLE XI1 Relation of Water-Stable Structure to Dry Soil Structure and Erodibility by Winda Water-stable fractions Soil class Sand Loamy sand Sandy loam Loam Silt loam Clay loam Silty clay loam Silty clay Clay b

Number of fields

3 12 13 25 35

5 10 1 8

> 0.84 mm. < 0.02 mm. > (%I

(%)

4.8 3.8 3.0 3.2 3.0 4.5 5.2 3.9 1.9

1.7 3.0 10.0 13.8 17.8 14.0 14.4 10.8 10.8

Clods 0.84 mm. ( "/o )

3.3 6.9 36.6 44.4 48.6 43.6 49.4 19.2 20.2

Amount eroded in tunnelb ( tons/acre)

49.8 29.4 5.4 2.2 1.9 1.5 1.2 3.4 6.7

Data from Chepil ( 1953a). Drag velocity of wind, 61 cm. per second.

in diameter, and the water-stable fractions conform with the simple arithmetic equation B=a(Y-b) (23) in which Y is the percentage of water-stable fractions < 0.02 and > 0.84 mm. in a soil and a and b are constants. For the first inch of soil, a and b were found to have a value of 3 and 4, respectively. The constants change with depth in soil, probably due to changes in soil compaction. The coefficient of correlation between the percentage of nonerodible clods B and the water-stable fractions < 0.02 mm. and > 0.84 mm. computed

264

W. S. CHEPIL AND N. P. WOODRUFF

from the 112 cases shown in Table XI1 is 0.70. This is very highly significant, since the required value for significance at the 1 per cent level for 112 cases (Table XII) is 0.26. The relation between the soil erodibility index I, and the percentage of water-stable fractions < 0.02 mm. and > 0.84 mm. is exponential and conforms with the equation

I , = ab-Y (24) in which a and b are equal to 1000 and 1.35, respectively. The coefficient of correlation between the log of erodibility and the percentage of water-stable fractions < 0.02 mm. and > 0.84 mm. is 0.72. This correlation is highly significant, as for Eq. (23). Chepil ( 1953a) found that the water-stable fractions < 0.02 mm. and < 0.84 mm. may vary together with or independently of each other. No single-value property, such as percentage of water-stable aggregates above or below a certain size, can be used, therefore, as an indicator of erodibility by wind. The whole size distribution of water-stable fractions must be taken into account. 3. Organic Matter General observations in Canada have indicated that high organic matter content of soil is conducive to high fertility and good tilth but facilitates erosion by wind (Hopkins, 1935; Hopkins d aZ., 1946). Preliminary experiments undertaken to verify these observations showed that wheat straw in the process of decomposition increased soil cloddiness and decreased erodibility by wind (Canada Department of Agriculture, 1949). These trends were reversed after the straw was decomposed. When samples of soil were brought together under identical climatic conditions and treatment, the black soils containing a relatively high content of decomposed organic matter (humus), contained more wind-erodible fractions than did the brown and the chestnut soils, and were more susceptible to wind erosion. Chepil (1955b) noted from field experiments on some soils of the Great Plains of the United States that during the time of rapid decomposition of vegetative matter the proportion of coarse water-stable aggregates > 0.84 mm. in diameter increased, the content of fine water-stable particles < 0.02 mm. decreased, soil cloddiness by percentage of nonerodible dry soil aggregates > 0.84 mm. increased, and erodibility, as determined by wind tunnel tests, decreased (Table XIII). The greater the quantity of vegetative matter added to soils, the more pronounced were these effects initially. When additions of vegetation to soils were stopped, these effects gradually diminished, disappeared in about a year

265

THE PHYSICS OF WIND EROSION AND ITS CONTROL

or two depending on the quantity of vegetative matter initially added, and then shifted in reverse and remained in reverse for at least 2 to 5 years depending on the original amount of vegetative matter added. The more vegetative matter added, the longer these effects lasted. Four years after additions of vegetative materials were stopped, all of the nine widely different soil types tested showed a significantly lower degree of soil cloddiness and higher erodibility by wind. TABLE XI11 Influence of Decomposing and Decomposed Vegetative Matter on Soil Structure and Erodibility by Winda Amount vegetative matter added

> 0.84 mm.

< 0.02 mm.

> 0.84 mm.

Relative amount of erosion in wind tunnel

(%I

(%)

(%I

(%I

(%I

Water-stable particles

Clods

Vegetative matter still decomposing (0.5 year after adding)

0 1 6

1.4 2.0 5.1

16.3 13.2 11.3

38.1 39.1 43.7

100 90 79

L.S.D. 1%level L.S.D. 5%level

1.2 0.9

1.7 1.3

6.7 5.1

N.S. N.S.

Vegetative matter decomposed ( 4 years after adding)

1 6

0.9 0.8 1.2

11.5 10.9 9.4

48.8 47.6 42.4

100 118 282

L.S.D. 1%level L.S.D. 5% level

0.3 0.2

1.5 1.1

3.7 2.8

180 136

0

a Data from Chepil (1955b).

The above-mentioned results substantiate, in general, results of numerous previous studies on the effects of decomposition of vegetative matter on soil aggregation and deaggregation. The literature reveals that numerous cementing substances produced by soil microorganisms as they attack the vegetative matter bind soil particles to form aggregates. The cementing substances may be divided into these major categories: ( a ) lyophyloc and lyophobic colloids consisting of decomposition products of plant residues (Myers, 1937; McCalla, 1945); ( b ) the microorganisms themselves and their secretory products such as mucus, slime, or gum (Peele, 1940; Martin, 1942; McCalla, 1950); and ( c ) polysaccharides synthesized by some microorganisms ( McCalla, 1945; Martin, 1946).

266

W. S . CHEPIL AND N. P. WOODRUFF

The aggregating effects of the initial products of decomposition are temporary ( Browning, 1944; McCalla, 1945). Aggregation declines as the products are destroyed by other microorganisms ( Martin, 1942). Incorporating the vegetative matter into the soil is not as important as leaving it on the surface, where it decomposes less rapidly and, therefore, continues to replenish the cementing products for much longer periods ( Havis, 1943; McCalla, 1945). Improved aggregation persists long after the bacterial population has declined (Peele, 1941 ) but remains only as long as the initial decomposition products exist ( McCalla, 1950). The products concentrate in and around the water-stable soil aggregates (Hide and Metzger, 1939). Although many initial cementing substances persist in the soil only a short time, others like the polysaccharides persist for long periods. It is likely that their persistence is due to combination with the mineral soil constituents that render them resistant to decomposition ( Martin, 1946). In such a combination they probably contribute to the more water-stable soil structure. Increases in soil aggregation are not discernible until after decomposition of vegetative matter has begun. The aggregating effects apparently are due to the products of decomposition and not particularly to the binding action of vegetative fibers in the soil. These sticky products of decomposition increase the size of both the water-stable aggregates and the dry (secondary) aggregates, or clods. These products are not entirely water-soluble; otherwise water-stable aggregates would not be formed. Many of the water-stable aggregates formed by decomposition of the vegetative matter are large enough to resist wind erosion. Gradually, the initial cementing materials lose their sticky property or are destroyed and replaced by secondary materials. Mechanical forces of expansion and contraction of the soil by wetting and drying and especially by freezing and thawing apparently cause the secondary cements to break up and the coarse primary and secondary aggregates to disintegrate to a more or less granulated condition. The secondary cements are more brittle and cause more granulation than do the initial products. The granules are essentially water stable. They form a friable, mellow soil which is more erodible by wind. High organic matter levels are essential to maintenance of soil fertility, and high soil fertility must be maintained to produce more vegetative matter. Continual additions of vegetative matter to the soil should tend to produce some wind-resistant aggregates and should tend to counterbalance excessive granulation and increased wind erodibility caused by the secondary products of decomposition. On the basis of information derived from studies on this subject, the benefits obtained from the

THE PHYSICS OF WIND EROSION AND ITS CONTROL

267

primary products of decomposition in augmenting resistance of soil to wind erosion are small compared to the detrimental effects from the secondary products of decomposition collectively known as “humus.” However, a high humus content of the soil must be maintained despite these detrimental effects, because the beneficial effects of humus in augmenting soil fertility far exceed the detrimental effects. Greater benefits to control erosion, no doubt, would be derived by leaving as much living or dead vegetative matter as possible anchored on top of the ground to protect the soil surface from wind. Thus, vegetative cover, not just organic matter, is the key to effective soil stabilization and conservation.

4. Calcium Carbonate Hopkins (1935) observed in Canada that soils high in free calcium carbonate (CaC03), or lime, and organic matter have been eroded severely by wind. Hardt (1936) concluded from his investigations of muck soils in Bavaria that the high content of lime, particularly in the clay fraction, appreciably increases erodibility by wind, but that organic matter, or humus, has little influence on erodibility of those soils. Chepil (1954a) carried out extensive field studies on the influence of powdered, precipitated CaCO:{ and decomposed organic matter (humus) on soil structure and wind erodibility of some soils of the central United States. For all soil textures tested, except loamy sand, addition of precipitated CaC03 to soils in the field decreased soil cloddiness and mechanical stability of clods and increased erodibility by wind, as shown by example in Table XIV. The differences in erodibility produced by adding different quantities of CaC03 were generally highly significant. Except for loamy sand, the greatest effects were obtained with about 3 per cent of CaCO,. Amounts greater and smaller than this had somewhat lesser effects, On loamy sand, the more CaCOs added, the greater was the increase in soil cloddiness and mechanical stability of clods and the decrease in erodibility by wind. This soil class is exceedingly erodible by wind, and adding even as much as 10 per cent of CaC03 left it in a condition still much more erodible than any of the other treated and untreated soils. It is expected that results with sand would have been similar to those with loamy sand. The foregoing results remained about the same for 5 years after treatment, when the experiment was terminated. It is apparent that effects from lime last as long as lime remains in the surface soil. Results obtained with high-lime and low-lime soils of similar texture and similar content of organic matter substantiated the results obtained by additions of lime to soils (Chepil, 1954a). Erodibility by wind

268

W.S. CHEPIL AND N. P. WOODRUFF

apparently was unaffected where the free lime content did not exceed 0.3 per cent. Adding CaC03 to soils had virtually no effect on the size distribution of water-stable aggregates except to increase the proportion of waterstable particles < 0.02mm. in diameter. Analysis of particle size disTABLE XIV Effect of Calcium Carbonate on Soil Structure and Erodibility by Winda Mechan-

Amount

tunnelb (tons/ acre)

CaCO, added

> 0.84

< 0.02

> 0.84

mm.

mm.

mm.

of clods

(%I

(%I

(%I

(%I

(%I

0 1 3 10

1.7 1.9 1.8 1.9

24.4 24.1 24.2 28.9

68.6 64.3 59.4 65.7

87.6 81.4 84.0 88.2

0.27 0.50 0.57 0.32

Dalhart fine sandy loam

0 1 3 10

0.9 1.2 1.0 0.9

13.4 11.8 13.8 18.8

57.7 52.8 51.6 55.3

83.5 82.2 80.4 83.2

0.62 0.86 1.00 0.63

Pratt loamy fine sand

0 1 3 10

1.2 0.8 1.6 0.9

3.9 3.5 5.7 10.5

11.0 13.5 20.5 39.4

31.2 31.4 49.7 68.1

16.40 14.00 7.00 2.10

Soil type Hastings, Keith, and Baca silt loams

a

Data from Chepil ( 1954a).

* From a 5-foot-long tray and with a drag velocity of

81 cm. per second.

tribution of the lime itself indicated that the increases in this fraction in soil were due to the lime particles added rather than to dispersion of the soil. However, when the soil contained a high proportion of humus, addition of lime decreased the proportion of water-stable particles < 0.02 mm. in diameter (Canada Department of Agriculture, 1943). Apparently the lime tended to aggregate the fine discrete particles < 0.02 mm., but only in the presence of humus. Reduction of these fine particles decreased soil cloddiness and erodibility by wind, as already indicated in Section E, 3; hence the highest erodibility was recorded for soils containing the highest quantity of both CaC03 and humus (Table XV). The effects of precipitated CaC03 on sand and loamy sand are similar in some respects to the effects of quartz silt. Probably this is because the crystals in precipitated CaC03 when shaken in water are predominantly of the size of silt. Silt is a mild cementing agent and is partly responsible for the formation of fragile secondary aggregates

269

THE PHYSICS OF WIND EROSION A N D ITS CONTROL

or clods. As shown by Chepil (1955a), small initial amounts of silt are very effective in reducing erodibility of sand. Addition of 10 per cent CaC0, to Pratt loamy fine sand raised the proportion of water-stable particles of the size of silt and clay from 3.9 to 10.5 per cent and reduced erodibility from 16.4 to 2.1 tons per acre (Table XIV) . This reduction in erodibility was virtually the same as that produced by adding the same proportion of quartz silt. TABLE XV Some Properties of Soils 2% Years after Application of Calcium Carbonate and Ground Wheat Straws. b Amount and kind of material added None 3%CaC0, 3% CaCO, and 3% straw 10%CaCO, 10%CaCO, and 10%straw

CaCO,

Organic matter

Clods > 0.84 mm.

Mechanical stability of clods

(%I

(%I

(%I

(%)

Amount eroded inwind tunnel* ( tons/acre)

0.67 3.13

2.53 2.57

65.7 57.0

62.6 59.6

0.33 0.61

3.41 9.10

2.78 2.43

44.7 53.8

50.6 59.1

1.60 0.84

9.08

4.09

35.4

53.1

2.95

Chepil (1954a). b Averages of results obtained with Baca silt loam, Lamed sandy loam, and Sutphen clay of the brown, reddish chestnut, and black soil zones, respectively. a

On other than sand and loamy sand the action of precipitated CaCOs is considerably different from that of quartz silt. In soils containing an appreciable proportion of clay, CaC0, appears to weaken the cementing strength of the clay and causes the clods to soften and granulate. It has little influence on the state of the primary or water-stable aggregates. It primarily weakens the bonds that hoId the water-stable aggregates together to form clods. This action is probably basically due to the flocculation phenomenon, which may be observed clearly when precipitated CaC0, is shaken in water. Thompson (1952) asserts that the presence in soils of large amounts of calcium, usually present in the form of CaCO,, tends toward the development of a granular soil structure. If the granules are small enough they will be eroded readily by wind. More often than not, granulation of dryland soils tends to induce wind erosion (Hopkins et al., 1946). In semiarid regions, soils are characterized generally by a layer of CaC0, accumulation, which normally lies just below the solum. This layer often is brought up to the surface by tillage implements, especially

270

W. S. CHEPIL AND N. P. WOODRUFF

where some of the soil has been removed by erosion. Exposure of the CaC03 layer increases the hazard of wind erosion. Higher ground, such as a knoll, commonly is eroded to expose this layer. Knolls frequently cause erosion of adjacent lands by serving as focuses from which CaC03 may be spread. Such areas are a serious erosion hazard to surrounding lands and should be protected specially, as with grass. In humid regions, applications of lime are required sometimes to correct soil acidity. The amounts of ground limestone applied in such cases range from 2 to 5 tons an acre, according to the degree of soil acidity. As shown from studies by Chepil (1954a), these amounts are too small to have an appreciable effect on soil structure and erodibility by wind. Moreover, the favorable soil moisture conditions in these regions almost preclude the hazard from wind erosion. VI. Wind Erosion Control

Soil stabilization proceeds under natural conditions or is accomplished

by man usually in three major successive stages: ( 1) trapping of moving soil particles; ( 2 ) consolidation and aggregation of trapped soil particles; and ( 3 ) revegetation of the surface. Trapping of eroding soil particles is known as the stilling of erosion. Trapping may be accomplished by roughening the surface, by placing barriers in the path of the wind, or by burying the erodible particles by tillage. Trapping is accomplished naturally by soil crusting resulting from rain followed by a slow but inevitable process of revegetation. The height to which particles rise in saltation has an important bearing on the most effective methods of stilling wind erosion. This is because saltation is the cause of all other forms of soil movement and is the major cause of soil abrasion. The ratio of height of rise to the horizontal equivalent of grain leap is about 1:7 for rise up to 2 inches, 1:8 for 2 to 4 inches, 1:9 for 4 to 6 inches, and 1:lO for heights above G inches. The capacity of stubble or ridged strips to trap particles in saltation is governed by the width of the trap strip and its receptiveness (Chepil, 1945a). The strip should be wide enough so particles will not jump over it and continue their movement, and it should be receptive enough so all the saltating particles that enter it will come to rest. In addition, a trap strip must be wide and high enough to allow sufficient reservoir for trapped soil so that at no time will it fill to its maximum capacity. Besides the use of trap strips, the whole eroding surface may be roughened or covered with any material that stills erosion. Methods of stilling wind erosion are known as emergency methods

THE PHYSICS OF WIND EROSION A N D ITS CONTROL

271

(Woodruff et al., 1957). Their effects are only temporary. Once erosion is stilled, plant cover must be established or plant residues must be maintained for more permanent control. Both temporary and permanent control of wind erosion employ a single principle, known as the principle of surface barriers and cover. Perhaps the simplest way to explain it is by describing the behavior of bare soils subjected to an erosive wind. In bare soils containing a mixture of erodible and nonerodible fractions, the quantity of soil removed by wind is limited by the height and number of nonerodible fractions that become exposed on the surface. If these soils are unaffected by encroachment of erodible material from the outside and if the length of the eroded area along the direction of the wind is limited, the removal of erodibIe fractions continues untiI the height of the nonerodible fractions that serve as barriers to the wind is increased to a degree that affords complete shelter to the erodible fractions. Movement then ceases (Fig. 14). The time required for movement to cease varies greatly with the soil structural conditions and the length across the field parallel to wind direction. The smaller the size of nonerodible fractions, the higher is the initial rate of soil movement q and the shorter the time required for movement to cease. The higher the proportion of erodible to nonerodible fractions, the higher is the initial rate of soil removal and the longer the time required for movement to cease. Also, the larger the field the greater the time required for removal of erodible fractions. If the soil contains a large proportion of erodible fractions, few nonerodible clods per unit area of ground become exposed by the wind. The nonerodible clods under such a condition have to reach a considerable height before soil removal will cease. If, on the,other hand, the soil contains a small proportion of erodible fractions, many nonerodible clods will be exposed on the surface by the wind and their height when soil movement ceases will be relatively low. The greater the number of clods exposed on the surface, the lower is their height when soil movement ceases. At a stage when soil removal ceases, the ratio of distance between the nonerodible barriers divided by the height of the barriers remains constant for any proportion and size of nonerodible fractions present in the soil. The ratio is known as the critical suflace barrier ratio? It is a ratio of distance between the nonerodible surface barriers, D,, to the height of the barriers, H , that will barely prevent the movement of 2 Chepil (1950a) originally called this the critical surface roughness constant. The present term is believed to be more generally appropriate.

272

W. S. CHEPIL AND N. P. WOODRUFF

FIG.14. Appearance of a silt loam composed of 92 per cent nonerodible fractions ( A ) before exposure to wind, and ( B ) after exposure for the period required for soil removal to cease. Drag velocity of the wind was 60 cm. per second and wind direction was left to right (Chepil, 1958).

THE PHYSICS OF WIND EROSION A N D ITS CONTROL

273

erodible fractions by the wind. It is equal to D,/H. On cultivated soils this ratio was found to have a value of 4 to 20, depending on the drag velocity of the wind and on the threshold velocity of the erodible soil fractions (Chepil, 1950a). The critical surface barrier ratio of 4 means that the surface barriers of height H will prevent the movement of soil within a distance of 4 H downwind of the barriers. This dominant principle of surface barriers governing the erodibility of bare, cultivated soils can be expressed by

in which Z , is the relative quantity (weight) of soil erodible from unit area of ground; e and u are densities (weights per unit volume) of erodible and nonerodible soil fractions, respectively; W, and W, are relative weights (in per cent) of erodible and nonerodible soil fractions, respectively; V1is volume of nonerodible surface barriers before exposure to wind; V2 is volume of such barriers (in cc./sq. cm.) after soil movement has ceased, and a is a coefficient which, for the units used, has a value of about 6. If the density of erodible and nonerodible soil fractions is the same, the expression p,,/pE = 1 and therefore may be dropped out. The critical surface barrier ratio is in fact an eflectiue distance between nonerodible surface barriers (measured in terms of heights of such barriers) required to reduce the quantity of erosion to zero. The effective distance, D,/H, may be expressed by

-_ H

a V. -Vet

(26)

in which V. is the drag velocity of the wind, V e t is the minimal drag velocity required to initiate movement of erodible soil particles, and a is a constant which varies with the characteristics of the surface barriers, such as their shape and porosity (air penetrability). If the nonerodible surface barriers are extremely low, as they would be for fine gravel, a relatively large number of the gravel pieces would be needed to protect the erodible fractions from the wind. The gravel pieces in such cases would protect the erodible fractions more by covering than by sheltering them from the wind. Thus, virtually all nonerodible materials placed on the surface of the ground to control wind erosion have an element of cover in addition to barriers which protect the erodible soil fractions from the wind. The principle of surface barriers and that of cover is therefore inseparable. The principle of surface barriers and cover that governs the erodibil-

274

W. S. CHEPIL AND N. P. WOODRUFF

ity of cultivated soils is clearly manifested where the eroding area is small. The larger the area the greater the time required for erosion to cease. In fact, in large fields soil removal seldom ceases .for a . given wind. On the average, about 120 hours of continuous exposure to erosive wind blowing from a single direction would be required to stabilize a one-half-mile length. Erosive winds, however, seldom blow continuously from one direction for such periods. A change in wind direction also would prolong the period required to stabilize a field. Then, too, great quantities of nonerodible fractions in large fields are converted to erodible particles by abrasion from the moving soil particles. The nonerodible surface barriers under such conditions tend to be destroyed and the rate of soil movement tends to accelerate rather than decrease, as is usual in small, isolated fields. The decrease and ultimate cessation of soil movement are possible only if the surface projections or barriers are indestructible by wind erosion. The desert pavement composed of a mantle of nonerodible gravel is one example of virtual indestructibility of a stabilized surface. The principle of surface barriers and cover extends beyond the surface roughness elements composed of nonerodible soil clods. It extends to almost all elements employed in wind erosion control, such as vegetative covers, soil ridges, windbreaks and wind barriers of various sizes and characteristics, and crop strips. All these elements of wind erosion control are designed to: ( a ) take up some or all of the wind force so that only the residual force, if any, is taken up by the erodible soil fractions; ( b ) trap the eroded soil, if any, on the lee or among surface roughness elements or barriers, thereby reducing soil avalanching and intensity of erosion and preventing erosion from spreading to other fields and farms. A. VEGETATIVEAND OTHERTYPESOF COVER The importance of vegetative protection on the land cannot be overstressed as it is the one generally applicable method for permanent and effective control. Zingg (1954)in reporting results from studies of different field surfaces stated that different amounts, types, and orientations of residues removed 5 to 99 per cent of the direct wind force from the immediate soil surface. Natural vegetative covers are perhaps the most effective, easiest, and most economical to maintain in agricultural areas. All the common crops, such as grasses, wheat, sorghum, corn, legumes, and cotton, provide cover of varying degrees of effectiveness where wind erosion is a hazard. In addition to the cover grown in place, crop residues often are placed artificially on the soil to provide temporary cover until permanent vegetation can be established. Chepil et al. (1960) reports

THE PHYSICS OF WIND EROSION AND ITS CONTROL

275

such applications, if well anchored, to be an effective, economical way to control wind erosion. Other than nonprocessed vegetative covers a-re’ used. principally on nonagricultural land where it is not feasible to obtain cover by growing and managing vegetation. Some of the nonvegetative and processed vegetative materials used are gravel and crushed rock, various surface films such as resin-in-water emulsion (petroleum origin), rapid curing cutback asphalt, asphalt-in-water emulsions, starch compounds, latexin-water emulsion ( elastomeric polymer emulsion), by-products of the paper pulp industry, and wood cellulose fiber. These materials may be used as the only cover or as temporary expedients to protect the land while a natural vegetative cover is being established. Highway and military departments are particularly interested in these materials for stabilizing highway shoulders and ditches, ammunition dumps, airfield landing strip shoulders, and other conditions resulting from construction where there is a great need for quick, effective soil stabilization. The relative effectiveness and the maintenance of different kinds of vegetative and nonvegetative covers will be discussed in this section. 1. Effectioeness Grasses and legumes once they are established are usually most effective because they provide a dense, complete cover. Wheat and other similar small grains are effective after they have passed the crucial 2 or 3 months after planting period. Corn, sorghum, and cotton are only of intermediate effectiveness, principally because they are planted in rows too far apart (U to 42 inches) to protect the soil. After the plants have completed their growth and the residue becomes the primary cover, the durability of the residues as measured by the resistance to decay by natural weathering largely determines their effectiveness. Duley (1958) in discussing some of the general characteristics of different kinds of residue indicated that legume residues tend to decay rapidly because they contain high amounts of protein which supply nitrogen for the organisms that promote decay. He found corn and sorghum stalks to be quite durable, especially when on top of the soil. He found wheat and rye straw more resistant to decay than oat straw. The importance of density and orientation of residue on effectiveness for wind erosion control is well illustrated by results of some recent research by Siddoway (Table XVI). The more erect and the finer and denser the residue, the smaller the amount of erosion. Gravel and crushed rock of any size > 2mm., if applied in sufficient quantities, have been found, generally, to provide good wind erosion

276

W. S. CHEPIL AM) N. P. M’OODRUFF

control. Chepil et al. (unpublished data, 1962) also have reported that fine, medium, and coarse gravel spread uniformly at 20, 50, and 100 tons per acre, respectively, adequately controlled wind erosion even on dune sand where no traffic was involved. TABLE XVI Average Effects of Kind and Orientation of Crop Residue on Erosion of Sandy Loam Soil by Wind of Uniform Velocitya Quantity of soil eroded in a wind tunnel Quantity

Covered with wheat residue

of crop residue Standing, 10 above soil surface inches high ( pounds/acre) (tons/acre)

0 500 1,000 2,000 3,000 6,000 a

16.0 2.8 0.1 Tb

T T

Covered with sorghum residue

Flat (tons/acre)

Standing, 10 inches high (tons/acre)

Flat (tons/acre)

16.0 8.5 2.5 0.1 T T

16.0 13.0 8.1 3.9 1.4 T

16.0 14.5 10.4 5.3 2.2 0.2

Unpublished data from F. H. Siddoway. T = trace, insignificant.

Most of the other nonvegetative materials stabilize soils against wind erosion by forming a surface film. These materials usually are dispersed in water and sprayed on the soil surface. Chepil ( 1 9 5 5 ~ )and Chepil et nl. (unpublished data, 1962) have indicated the following desirable characteristics for surface films: (1) they should be indispersible in water, durable, yet porous enough to allow percolation of water; ( 2 ) they should be weak enough for seedling penetration; ( 3 ) they must be able to maintain their sticky property indefinitely when used as permanent wind erosion control covers; and, (4) they must be easy to apply. If excessive dilution with water is required, they lose their effectiveness, and if they must be heated before application, special equipment is needed. Although none of the group of surface film covers listed here are as effective or as economical as well-anchored vegetative mulch, most provide ample protection to the soil if applied in sufficient quantity. Cutback asphalt is particularly effective. Asphalt and resin emulsions are also quite effective, especially the resin emulsions which have the properties of remaining moist for at least 3 months after application and make clods and soil surface roughness resistant to soil slaking by rain. Some of the latex emulsions are effective but very expensive. Hydrolized starches are relatively ineffective as covers for wind erosion

THE PHYSICS OF WIND EROSION A N D ITS CONTROL

277

control because they are readily washed away by rain. Unhydrolized starches, on the other hand, are ineffective as surface films. Wood cellulose fiber is reasonably effective if a binder such as asphalt of sufficient amount is mixed with the material. 2. Maintenance Excessive tillage, or tillage with improper implements, and overgrazing are the major causes of vegetative cover reduction on crop and grazing lands. Land management policies which avoid these destructive practices must be adhered to constantly if wind erosion control is to be realized from vegetative covers. On rangelands, controlled grazing should be practiced at all times. The number of animals per acre should be regulated closely in order to realize maximum use of the grass and still maintain a sufficient amount of vigorous, complete vegetative cover. Supplementary practices should include contour furrowing to reduce runoff, establishment of more productive grass species, application of fertilizers where economics permit, and supplying adequate watering sites preferably on more nonerosive locations to prevent blowouts caused by excessive animal traffic. Stubble mulching and minimum tillage or plow-plant systems of farming are all excellent methods of maintaining vegetative residues on cropland. Stubble mulching usually is defined as a year-around system of managing plant residues in which all tilling, planting, cultivating, and harvesting operations are performed to keep a sufficient amount of the residue on the surface at all times to provide protection from erosion. The practice requires use of tillage implements which generally undercut the residue without soil inversion (Fig. 15). Several types of implements are available and commonly used for this practice. Research by Anderson (1953, 196l), Siddoway et al. ( 1956), Woodruff and Chepil (1958), and Fenster ( 1960) has shown that each implement maintains on the surface a slightly different amount of residue. The average range is from 50 to 90 per cent maintained after each operation, but values range from 30 to 115 per cent (Table XVII) depending upon kind of implements, kind, height, and amount of residue, and moisture, texture, and density of soil. Maintenance greater than 100 per cent means that more residue was brought up than buried below the surface. Stubble mulching, while principally applied to wheat, also can be used in row-crop production. Greb and Black (1962) reported that maximum use of subsurface tillage on sorghum residues in a summer fallow system has preserved from 30 to 45 per cent of the original amount. The minimum tillage or plow-plant system of farming row crops has gained considerable acceptance in recent years. In this sys-

278

W. S. CHEFIL AND N. P. WOODRUFF

FIG. 15. This rodweeder with small duckfoot shovels leaves as much as 80 per cent of the wheat stubble, mostly standing above the surface. TABLE XVII Residue Maintenance with Tillage lmplementsa Average maintained after each tillage operation Type of implement Subsurface implements Blades (36 inches or wider) Sweeps (24 to 36 inches) Rodweeders, plain rod Rodweeders, with semi-chisels Mixing implements Heavy duty cultivator ( 16 to 18-inch sweeps) Heavy duty cultivator (2-inch chisels 12 inches apart) One-way disk ( 24 to 26-inch pans) Tandem or offset disks

(%) 90 90 90

85

Range maintained (%)

70 to 113 60 to 112 80 to 115 55 to 105

80 75

50 to 100

50 50

30to 90

-

-

a Data from Anderson (1953, 1961, 1962), Woodruff and Chepil (1958), and Fenster ( 1960).

THE PHYSICS OF WIND EROSION AND ITS CONTROL

279

tem, special equipment is used to till, plant, and apply insecticides, herbicides, and fertilizers all in one operation. Subsequent cultivation then is kept to a minimum. Lane and Wittmus (1961) reported that since the system leaves residues on the surface, it provides good wind and water erosion control, requires less labor, lowers tillage costs, and produces slightly higher yields in some areas of Nebraska. Maintenance problems with organic film covers develop because freezing and thawing and swelling and shrinking of soils break up the film. Chepil ( 1 9 5 5 ~ )reported that films can be maintained longer on sand and loamy soils than on clays. They also can be maintained longer on any soil if they are applied uniformly without too much dilution with water, if they do not penetrate into the top soil layer, and if they retain their sticky properties ( Chepil et al., unpublished data, 1962). Once films arc applied, maintenance consists of keeping all traffic off to avoid breaking the film. Where traffic is involved, use of materials that penetrate the top soil layer is necessary. Gravel and crushed rock afford rather permanent covers; maintenance here consists of avoiding any disturbance of the rock by mechanical action such as tillage and of using herbicides to control vegetative growth where such growth is undesirable.

B. SOIL CLODSAND RIDGES Practical application of the principle of surface barriers and cover is well exemplified by tillage with different implements. The soil surface is made cloddy and rough by: (1)regular tillage processes used to prepare suitable seedbeds and to control weeds for crop productions; and ( 2 ) special tillage practices used specifically to bring clay to the surface for possible increased cloddiness and to roughen the land to prevent wind erosion. Roughening the surface is effective only to the extent that the roughness elements are nonerodible. Ridging dune sand for example, is of little value because the ridges on sand are erodible and are soon leveled by the wind. The role of different tillage implements and operations in creating cloddiness and roughness of the land surface will be discussed in this section. 1 . Regular Tillage Much of the regular tillage in semiarid regions where wind erosion is generally most severe is on fallow. Repeated tillage of fallow is needed to kill weeds and thereby conserve soil moisture; however, repeated tillage tends to pulverize the soil and induce wind erosion. In winter wheat areas, pulverization of fallow by repeated tillage is somewhat compensated by the fact that wheat is seeded in fall and, if germination

280

W. S . CHEPIL AND N. P. WOODRUFF

and growth is favorable, provides a vegetative cover the following spring when protection from wind erosion is needed most. But in spring wheat areas, fallow presents a real problem because the only vegetative cover that may be present in spring is residue from a crop grown almost 2 years before. Therefore, fallow in spring wheat areas generally must be tilled early in spring to create a rough, cloddy soil surface before high winds occur. In winter wheat areas, severe wind erosion often occurs in spring when wheat has failed to make sufficient growth the previous fall. Under such circumstances a farmer is faced with the dilemma of whether to till and stop wind erosion, but kill the wheat, or not to till and hope that wheat will not be completely destroyed by wind. Row crops such as corn are less susceptible to damage from erosion because of the relative roughness of the land when they are seeded. Other row crops, such as sorghum and cotton, are less susceptible to damage from erosion because of their late planting date when the wind erosion problem is generally past. Cotton stubble has little resistance to wind erosion. It is usually tilled to create a cloddy surface before spring winds occur. Sorghum and corn stubble, if heavy enough, makes an effective vegetative cover and is, therefore, left standing till the windy season is over. If the stubble is too light, it is usually tilled to create the necessary clods. The greater the required cloddiness, the deeper the tillage should be. It is important in all tillage operations to avoid excessive or frequent tillage, because this can lead only to soil surface smoothing and clod pulverization. Soil moisture at time of tillage has a decided effect on cloddiness. Lyles and Woodruff (1962) have reported that different soils have differing moisture contents at which soil pulverization is most severe, and that more clods are produced if the soil is either extremely dry or extremely moist than if it is at intermediate moisture content. The type of tillage implement used also has a definite influence on soil cloddiness and surface roughness. Lyles and Woodruff (1962) working with a moldboard plow, a one-way disk, and a subsurface sweep in controlled soil moisture conditions found that influences on cloddiness due to tillage machinery lasted longer than did influences due to soil moisture. They also reported that the moldboard plow produced a rougher, more cloddy surface with higher mechanical stability of clods than did the one-way disk or subsurface sweeps. Tillage implements commonly used in stubble-mulch farming, with the exception of chisel cultivators, usually do not leave a ridged, rough surface. One-way and offset or tandem disks leave a smooth surface. Subsurface sweeps, because they do not disturb the soil surface, do not create a rough, ridged

THE PHYSICS OF WIND EROSION AND ITS CONTROL

281

soil surface, but they do create a greater vegetative roughness by allowing the vegetation to remain erect. Small sweeps and chisels produce a more cloddy condition than large sweeps and one-way disks operated at shallow depths. Listers generally provide the maximum cloddiness and surface roughness. It is important that planting and seeding equipment preserve as much protective residue as possible, keep the soil surface rough and cloddy, and at the same time place the seed in moist, firm soil to promote rapid germination. Major types of planters available for small grains include hoe, single and double disk, deep furrow drills, and seeding attachments on one-ways and cultivators. For wider spacing of row crops, such as corn, sorghum, and cotton, listers, furrow opener planters, and seeding attachments on cultivators are available. For small grains, the deep furrow disk and hoe drills provide maximum surface ridging, pass through heavy residues easily and concentrate them in the ridges, and place the seed in good moisture (Siddoway et al., 1956; Woodruff and Chepil, 1956; Zingg and Whitfield, 1957; Fenster, 1960; McCalla and Army, 1961). All of the rowcrop planting equipment generally leaves a rougher surface than does some of the other small-grain seeding equipment; however, lister planters generally provide the maximum surface roughness and offer excellent protection to small plants.

2. Special Tillage Emergency tillage to provide a rough, cloddy surface is a temporary measure and its only purpose is to create an erosion-resistant soil surface. It is usually a last resort carried out when vegetative cover is depleted by excessive grazing, drought, improper or excessive tillage, or by growing crops that produce little or no residue, or when potentially severe erosive conditions are encountered or expected soon. It should be done before blowing starts rather than after, because soils rapidly become more erodible under abrasion of moving soil particles, thus requiring more drastic measures to prevent further erosion. Woodruff et al. (1957) and Chepil et al. (1961) have indicated that various tillage implements can be used as emergency tillage tools. The most common are listers, duckfoot cultivators, and narrow-tooth chisel cultivators. The effectiveness of any of these implements, as measured in terms of the degree of cloddiness and roughness they create, depends to a great extent upon soil moisture, texture, and density. Lyles and Woodruff (1961) found that the cloddiness potential of soils could be increased markedly by increasing density; also, the cloddiness potential of soils with a high day content is greater than for sandy soils. Speed

282

W. S . CHEPIL A N D N. P. WOODRUFF

of travel, depth of tillage, spacing between tillage point carriers, and the type of tillage point also influence the degree of roughness and cloddiness. Intermediate speeds of 3.5 to 4.0 miles per hour usually provide good roughness and cloddiness; such speeds do not throw the tillage layer, which would reduce roughness and pulverize clods (Woodruff et al., 1957). Emergency tillage should be accomplished at a depth which brings up compact clods, usually 3 to 8 inches. Spacing of lister and chisel points must be governed by severity of erosion and presence or absence of crops. Close spacing with any implement will create a rougher surface than will wide spacing. However, if a crop is involved and there is a possibility of saving part of it, then wide spacings of 48 to 54 inches will provide sufficient roughness for some control and at the same time permit most of the crop to continue growing. Insofar as type of tillage point is concerned, listers and narrow chisels are most effective. Chepil et aE. (1961) have indicated that listers produce a high degree of roughness, and in extremely sandy soils where clods can be produced only by deep tillage they are the most effective tools available. Chisel cultivators are more widely used as emergency tillage implements because they require less power and destroy less crop than do listers. They vary in effectiveness, depending on the type of point. Woodruff et al. (1957) reported that wedge-shaped heavy duty chisels generally bring up more clods and leave rougher surfaces than do duckfoot shovels or narrow chisels. Direction of wind with reference to direction of tillage also influences the effectiveness of emergency tillage. It will be more effective if the wind blows across rather than parallel to the ridges. For this reason emergency tillage always should be accomplished perpendicular to the prevailing wind erosion direction. Deep plowing is another form of special tillage used to bring adequate amounts of clod-forming clay subsoil to the surface, thereby reducing wind erosion. It is accomplished with large moldboard or disk plows. Most of the plowing is done at 24- to 30-inch depths; however, some of the larger moldboard plows are capable of plowing 42 inches deep. Research in Oklahoma by Harper and Brensing (1950) has indicated that clay content of surface soils was increased from 4 per cent to 12 per cent by deep plowing. Chepil et al. (unpublished data, 1962) reported that deep plowing increased the clay in the surface soil on an average from 5 to 12 per cent in fields in Texas and Kansas. However, Chepil (1953a) has found that about 27 per cent of clay in the surface soil is required for maximum benefit to control wind erosion. Furthermore, Chepil et al. (unpublished data, 1962) concluded that increased

THE PHYSICS OF WIND EROSION AM) ITS CONTROL

283

ridging and cloddiness resulting from deep plowing of sandy soils are only temporary, particularly if wind erosion occurs. Therefore, it must be supplemented with other suitable control practices.

C. WINDBREAKS AND WINDBARRIERS One type of windbreak is a planting of trees or shrubs in 1 to 10 rows to provide a barrier of sufficient height and density to present a formidable obstacle to the wind. Other wind barriers are crops in narrow rows, snowfences, solid wooden or rock walls, and earthen banks. Windbreaks or wind barriers function as do other surface barriers in providing wind erosion control; i.e., they take up or deflect a sufficient amount of the wind force to lower the wind velocities to the leeward below the threshold required for initiation of soil movement. The effect of any barrier in reducing the rate of soil movement depends on many factors, including wind velocity and direction, and shape, width, height, and porosity of the barrier. The velocity of the unobstructed wind has an important influence on the effectiveness of a barrier. Nearly all barriers provide maximum percentage reductions in wind velocity at leeward locations near the barrier, with a gradual decrease downwind. These percentage reductions for rigid barriers generally remain constant no matter how hard the wind blows (Woodruff and Zingg, 1952). The percentage reductions for porous, resilient barriers tend to increase slightly with increased velocities (Bates, 1944; Fryrear, unpublished data, 1962). This means that for such barriers the degree of wind erosion control will be greater for lowvelocity winds than for high-velocity winds. The direction of the wind influences both the size and location of the leeward-protected area. The area of protection is greatest for wind blowing at right angles to the barrier length and is smallest or almost nil for wind blowing parallel with barrier direction. It is, therefore, important that a complete system of barriers be provided for protection from winds from all directions, or that cognizance be taken of the prevailing wind direction for a given region, if there is a prevailing direction. The shape of windbreaks characterizes the outer perimeter, or surface, which is in contact with the airstream. The data of Woodruff and Zingg (1952) have indicated that a streamlined or very abrupt vertical barrier will provide less protection than will a sloped or triangular outer surface. In barriers composed of several rows of growing plants, such as trees, the ultimate shape can be controlled by proper selection of species within the barrier rows. Porosity is an important factor influencing the effectiveness of a

284

W. S . CHEPIL AND N. P. WOODRUFF

barrier. Dense barriers provide large reductions in velocity for relatively short leeward distances, whereas porous barriers provide smaller reductions in velocity but for more extended leeward distances (Woodruff and Zingg, 1952; Woodruff, 1954; Caborn, 1957). Generally, some porosity is desirable in order to gain extended protection; however, large openings must be controlled, because too much openness causes air jetting with serious erosion in the immediate leeward zone. Height of barrier is a very important factor influencing effectiveness because it governs the limits of influence. Expressed in multiples of barrier height, the zone of wind velocity reduction on the leeward side of a barrier may extend a distance equal to 40 or 50 times the height of the barrier. Influences to these distances are, however, insignificant in terms of wind erosion control, and if complete control is desired barriers must be spaced at relatively close intervals. Actual effective limits of influence vary with open wind velocity, barrier porosity, and threshold velocities of soils. Chepil (1949) has reported that willow barriers form a protective influence extending only 6 to 7 heights in some of the highly erodible sandy regions of China. Woodruff and Zingg (1952) in wind tunnel tests indicated that full protection from a 40-mile-per-hour wind velocity was provided only for a distance equal to 9 times the height of the barrier. Woodruff, Fryrear, and Lyles (unpublished data, 1962), presenting data on effective zones for wind erosion protection for various narrow tree windbreaks, have shown the zone to depend on levels of open wind velocity and have indicated an average protected distance of about 12 times the height of the barrier for winds of 40 miles per hour measured at the %-foot elevation. Iizuka (1950) has observed that a windbreak which reduced wind velocities to 61, 69, and 77 per cent of that in the open at leeward distances of 10, 20, and 30 times the barrier height, respectively, decreased soil blowing at those distances to 0.14, 18, and 50 per cent of that in the open, respectively. This relatively limited infiuence of barriers emphasizes the need for complete barrier systems designed to provide extended protection across fields and for use of other supplementary wind erosion control practices.

1. Tree Windbreaks Plantings of trees in middle rows and shrubs in outside rows (Fig. 16) have been made for a number of years in an attempt to reduce wind velocities. This type of windbreak received special emphasis in the 1930s when there was a serious wind erosion problem in the Great PIains region of the United States. Most of the windbreaks planted

THE PHYSICS OF WIND EROSION AND ITS CONTROL

285

during that period were wide, 10 rows or more, because it was believed that wide belts were necessary to provide adequate reductions in velocity and for attainment of the so-called forest condition believed to be necessary for propagation and self-preservation of trees. The trend today is toward narrower plantings, i.e., 1-,2-, 3-, and 5-row barriers, which have been found to be just as effective as wider belts in reducing wind velocities.

FIG.16. A windbreak composed of one to several rows of trees and shrubs is effective in reducing wind velocities and controlling wind erosion for some distance to leeward from the windbreak.

The type of tree species planted in a windbreak has a considerable bearing on the effectiveness. Research by Woodruff, Fryrear, and Lyles (unpublished data, 1962) has shown that in Kansas Osage-orange was most effective, followed in order by arborvitae, Siberian elm, cottonwood, and jackpine. The rate of growth of trees also largely governs the extent of protection that can be expected in later years. In general, trees that grow rapidly provide a greater protected length than do the slower gowing trees. Combinations of different tree and shrub species planted in rows to provide windbreaks vary considerably in their abiIity to provide protection from wind erosion. The amount of protection is not directly related to the number of rows in the windbreak (Table XVIII). The seasons

286

W. S. CHEPIL AND N. P. WOODRUFF

govern porosity of deciduous species and, therefore, influence the effectiveness of the windbreaks. Tree windbreaks have very definite limitations as a general method of wind erosion control, not only because of the relatively close spacing required which is objectionable where large machinery is used, but also TABLE XVIII Effect of Number of Rows and Season on Amount of Wind Erosion Protection Provided by Tree Windbreaksa

Windbreak

Effectiveness indexb

Protected distance ( H units)d

Summer conditions 2-row %ow l-row &row l-row l-row

Mulberry Plum, cedar, mulberry, elm, olive Osage orange Cedar ( 2 ) , shrub Siberian elm Jackpine

55.9 42.6 32.4 30.8 27.1 19.7

18.0 15.0

-

11.0 9.5

-

Winter conditions 10-row Cedar ( 1), deciduous ( 9 ) &ow Plum, elm ( 2 ) , cedar, honeysuckle l-row Osage orange 7-row Ash ( 2 ) , elm, cottonwood ( 2 ) , Osage orange, coffee

-

46.6 24.9 24.9

9.2 12.0

7.2

-

Unpublished data from Woodruff et al. (1962). Effectiveness index is computed by summing the products (velocity-reduction ratio at a given leeward location times the distance of the location from the barrier expressed in H units), c Based on 40-mile-per-hour wind at 5(rfoot height and a 25-mile-per-hour threshold velocity at the %-foot height. d H = average height of trees in a single-row windbreak and average height of tallest trees in a multiple-row windbreak. a b

because of the competition they afford adjacent crops for moisture and nutrients. Greb and Black (1961)have reported a direct yield reduction in wheat and sorghum attributed to extraction of soil moisture and nitrate nitrogen by the windbreak as far as its roots extend into a field. They indicate a ratio of root length to tree height of 2.51, and that deciduous species are more competitive than conifers. Staple and Lehane (1955) from measurements during five dry years in Canada have indicated that sapping of moisture by caragana windbreaks was not appreciable and extended into the fields a distance equal to about the height of the trees, or 25 feet. The general aridity of the areas where

THE PHYSICS OF WIND EROSION AND ITS CONTROL

287

wind erosion control is most needed also limits the use of tree windbreaks as control measures. 2. Crop Barriers

Annual crops are frequently interplanted in narrow strips or rows so that one crop provides protection to the other crop (Fig. 17). Sobolev (1947) has reported their use in the U.S.S.R. for preventing wind erosion and trapping drifting snow. Sheng (1961) also reported the use of the

FIG.17. This picture, taken in early fall, indicates rows of sorghum protecting winter wheat that has just emerged. Sorghum will be harvested for grain, and its stubble will protect the wheat in the crucial wind erosion period next spring.

perennial grass, Miscanthus, planted at 15- to 30-meter intervals in rice fields along the coastal regions of Taiwan. In the United States, crop rows are used frequently for protection of vegetable crops. Schultz and Carlton (1959) reported good protection from wind erosion of asparagus located on peat soils in California by interrow planting of barley. Rows of annual crops also have been used in the Great Plains for a number of years principally to trap snow and shelter new tree plantings (Ferber, 1958). Research designed to measure the effectiveness of annual crop barriers generally has been lacking; however, some recent studies by Fryrear ( unpublished, 1962) have provided some information. Results

288

W. S. CHEPIL AND N. P. WOODRUFF

from tests on sudangrass, grain and forage sorghum, broomcorn, sunflowers, castorbeans, crotalaria, and kochia have indicated that some of the crops provide adequate barriers for protection from wind erosion if they are spaced sufficiently close (Table XIX). TABLE XIX Effectiveness of Some Annual Crop Barriers for Wind Erosion Controla

Crop

Heightb ( feet )

Effectiveness index0

Protected distance* (feet)

Kochia Sudangrass Grain sorghum Forage sorghum Broomcorn

3.5 3.0 2.0 1.5 4.0

17.0 11.4 10.7 7.9 3.9

42.0 22.5 12.0 6.0 4.0

Unpublished data from Fryrear ( 1962). Harvested height. 0 Effectiveness index equals sum of products (velocity-reduction ratio at a given leeward location times the distance of the location from the barrier expressed in H units), Based on 40-mile-per-hour wind at 5o-foot height and a 25-mile-per-hour threshold velocity at the 50-foot height. a

b

*

3. Artificial Barriers Artificial barriers such as snowfences, board walls, bamboo and willow fences, earthen banks, hand-inserted straw rows, and rock walls have been used for wind erosion control on a rather limited scale. Because of high cost of material or of labor required for their construction, their use is restricted generally to applications where high value crops are involved or in areas where overpopulation requires intensive agriculture. Sheng (1961) reports use of hand-inserted straw barriers between rows of sweet potatoes, erection of 2-meter high woven bamboo fences at intervals of 30 to 50 meters, and construction of 2-meter high rock walls at 10- to 20-meter intervals for wind erosion control along the coastal regions of Taiwan. Sneesby (1953), reporting results of studies in England, indicated that solid barriers 20 feet high provided wind erosion protection for 340 feet, and earthen banks 2 feet high provided 50- to 60-foot protected lengths. In the United States, research on and application of artificial barriers to wind erosion control has been limited. Whitfield (1938) used sign board type artificial barriers constructed from sheet metal roofing and timber frames for wind intensifiers to reduce the height of sand dunes prior to stabilization by planting grass. Snowfences constructed from lath held together with wire providing a density of approximately 40 per

THE PHYSICS OF WIND EROSION AND ITS CONTROL

289

cent have been used for protecting vegetable crops (Schultz and Carlton, 1958). These fences have not proved very effective as control measures because they provide only a relatively short zone equal to 10H, or about 40 feet of velocity reduction of sufficient magnitude to reduce wind velocities below the threshold for initiation of soil movement (Woodruff and Zingg, 1955). The close spacing thus required makes them infeasible.

D. CROP ~TRIF'S AND CROP

ROWS

Crop strips function not so much as protective barriers but as soil traps designed to reduce soil avalanching. Crop strips or strip cropping are terms used to describe a method of farming, usually involving two or more crops, whereby strips of erosion-resistant crops are planted between strips of erosion-susceptible crops. The strips are usually all the same width. Crop rows involve only one crop and will be discussed here from the standpoint of the effects of different row spacings on wind erosion.

1 . Strip Cropping Strip cropping as usually practiced does not require any change in cropping practices, nor does it remove any land from cultivation. The field is subdivided into alternate strips of erosion-resistant crops and erosion-susceptible crops or fallow ( Fig. 18). Erosion-resistant crops are

FIG.18. A strip cropping sequence of wheat-sorghum-fallow is effective in controlling wind erosion.

290

W. S. CHEPIL AND N. P. WOODRUFF

small grains and other closely seeded crops that cover the ground rapidly. Erosion-susceptible crops are cotton, tobacco, sugar beets, peas, beans, potatoes, peanuts, asparagus, and most truck crops. Corn and sorghum are intermediate in their resistance to wind erosion. Chepil ( 1 9 5 7 ~ )and Chepil et al. (1961) have indicated that strip cropping controls soil blowing by reducing soil avalanching, which increases with width of eroding field. Since the rate of avalanching varies directly with field erodibility, the actual width of strip required varies greatly with factors that influence field erodibility such as soil texture, wind velocity and direction, quantity of crop residue, degree of soil cloddiness and surface roughness. Chepil (1960a) has made detailed studies of the effectiveness of crop strips in relation to soil texture and direction of erosive winds. He has reported that directional deviation of erosive winds from the perpendicular requires correspondingly narrower strips, and that required width of strip increases as soil texture becomes finer, except for clays and silty clays subject to granulation (Table XX). Mathews ( 1954) has recommended that strips should not be wider than 16 rods in order to be sufficiently effective, or narrower than 5 rods in order to make economical use of farm machinery. In the southern Great Plains where TABLE XX Average Width of Strips Required to Control Wind Erosion Equally on Different Soil Classes and for Different Wind Directions" Width of stripsb

Soil class

Wind at right angles to strips (feet)

Wind deviating 20" from right angles (feet)

Wind deviating 45' from right angles (feet)

Sand 20 18 14 25 22 18 Loamy sand Granulated clay 80 75 54 100 92 70 Sandy loam 150 140 110 Silty clay 250 235 170 Loam 280 260 190 Silt loam 350 325 250 Clay loam 0 Data from Chepil (1980a). b For negligible surface roughness, average soil cloddiness, no crop residue, 1-foot high erosion-resistant stubble on windward, 40-mile-per-hour-wind velocity at 5o-fOOt height, and a tolerable maximum rate of soil flow of 0.2 ton per rod width per hour.

THE PHYSICS OF WIND EROSION A N D ITS CONTROL

291

cotton is grown, some special forms of wind strip cropping are employed wherein cotton in two to four rows is alternated with various numbers and sequences of rows of sorghum or of other high-residue yielding crops (Burnett al at., unpublished data, 1962). In conclusion, the chief benefit from strip cropping for wind erosion control is realized because the strips control soil avalanching and the serious damage which can result from it. Strip cropping alone will not fully control wind erosion; it must be supplemented with other practices, such as stubble mulching, to be fully effective. In combination with strip cropping, the supplementary practices need not be as intensive as they would have to be for large fields.

2. Crop Rows The relative effectiveness of different row spacings for wind erosion control has not been fully evaluated. Generally speaking, the closer the row spacing, the more effective will be the crop. Most close-spaced crops, i.e., those planted with drills with spacing ranging from 7 to 14 inches, are erosion resistant once they are established. Sorghum, corn, cotton, and other crops normally planted in 40- to 42-inch rows are not so resistant. Recent experiments have shown that some of these crops can be grown in closer-spaced rows without detrimental effects on yields. The direction of crop rows with reference to prevailing erosive winds has some effect on erosion. Siddoway (unpublished data, 1962) has shown that the relative amount of erosion from soil planted to wheat in 10-inch rows is about six times greater when the wind is blowing parallel to the rows than when the wind is perpendicular to the rows. Zingg et al. (1952) working with a portable wind tunnel with 9-inch high sorghum stubble in 40-inch rows showed soil losses three times greater with rows parallel to the wind than with rows perpendicular to the wind. VII. The Wind Erosion Equation

A. GENERAL FRAMEWORK

A wind erosion equation, with all its accompanying charts and tables, has been developed to indicate the relationships between the amount of wind erosion and the various field and climatic factors that influence erosion (Agricultural Research Service, 1961; Chepil, 1962a). The equation is being modified continually as new data become available, It is designed to serve a twofold purpose: (1) As a tool for determining the potential amount of wind erosion on any field under existing local climatic conditions.

292

W. S. CHEPIL AND N. P. WOODRUFF

( 2 ) As a guide for determining the conditions of surface roughness, soil cloddiness, vegetative cover, sheltering, or width and orientation of field necessary to reduce the potential wind erosion to an insignificant amount. The equation embodies the major primary factors that govern wind erodibility of land surfaces. These primary factors influence wind erosion directly. They have been recognized during the course of many years of accumulation of experimental data on the problem. Some of them may be grouped or converted for convenience into equivalent factors, or may be disregarded, as follows: Indioidual Primary Factors

Per cent soil fractions > 0.84mm. as determined by standard dry sieving, A Mechanical stability of the surface crust, F, Average wind velocity, o Average moisture of soil surface, M Soil surface roughness, K Distance (along prevailing wind erosion direction across field, D t ) Distance ( along prevailing wind erosion direction protected by barrier, D b ) Quantity of vegetative cover, R Kind of vegetative cover, S Orientation of vegetative cover, KO

Equivalent Factors

1

Soil erodibility, I

i

Local climatic factor, C

Transient, and therefore generally disregarded

Same

Equivalent width of field, L

Equivalent quantity of vegetative cover, V

The percentage of nonerodible dry soil fractions > 0.84 mm., A, as determined by standard dry sieving is an equivalent of their true percentage and of their stability against breakdown by tillage and abrasion from wind erosion. Sieving breaks a portion of the nonerodible clods to smaller, erodible ones. The problem is to sieve the soil with such vigor or for such period of time to neither overemphasize nor underemphasize the influence of one of these factors in relation to the other. Therefore, the method of dry sieving is standardized (Chepil, 1962a). The percentage of nonerodible dry soil fractions > 0.84mm. in diameter as determined by standard method of dry sieving is directly related to soil erodibility I . This relation was derived from three major studies :

(1) Wind tunnel experiments on the relation between soil cloddiness and wind erodibility ( Chepil, 1950b; Chepil and Woodruff, 1954, 1959).

THE PHYSICS OF WIND EROSION AND ITS CONTROL

293

( 2 ) Field measurements in the vicinity of Garden City, Kansas, during 1954-1956 on the relation between wind tunnel erodibility and natural field erodibility ( Chepil, 196Ob). ( 3) Analysis of intensity-frequency of occurrence of climatic conditions in the vicinity of Garden City, Kansas, during 1954-1956 (Chepil et al., 1962). The mechanical stability of the surface crust, F,, if the crust is present, is of little consequence in the long run. It is disintegrated readily under the action of abrasion after wind erosion has started. It is a transitory condition and has some significance only if we desire to determine erodibility of the field at the moment the estimation is made. If we are interested in average erodibility for the entire soil-drifting season or year, as we ordinarily are, this condition should be disregarded. The rate of soil movement by wind varies directly as the cube of wind velocity, v, and inversely as the cube of average soil surface moisture M . It is convenient to consider these two factors together as a local wind erosion climatic factor, C . A map has been prepared indicating the approximate value of this factor for any location in the United States and the agricultural areas of Canada (Chepil et al., 1962). The soil surface roughness, K , is expressed in terms of height of standard soil ridges (the same as ridge roughness equivalent of Zingg and Woodruff, 1951) and means that the surface, other factors being equal, will resist the wind as much as the standard soil ridges in which nonerodible clods do not exceed inch in diameter and which have a height-spacing ratio of 1:4. For example, a ridge roughness equivalent of 2 inches for a given soil surface means that the wind drag against the surface will be as great as against the surface composed of standard ridges 2 inches high and 8 inches apart running at right angles to wind direction, composed of the same proportion of erodible and nonerodible fractions as the soil, and exposed to the same drag velocity of the wind as the soil. Width of field or field strip alone does not determine how erodible it is unless the prevailing wind direction and the presence or absence of adjoining wind barriers are taken into account too. No matter how narrow the field strip might be, if wind direction is parallel to its length, the strip would be almost as erodible as a large field of a width equal to the length of the strip. Furthermore, if any barrier is present on the windward side of the field, the distance Db (along the prevailing wind erosion direction) which it fully shelters from the wind must be subtracted from the total distance D, (along the prevailing wind erosion direction) across the field to determine the unsheltered distance across

294

W. S. CHEPIL AND N. P. WOODRUFF

the field along the prevailing wind erosion direction. This is the distance L that directly determines the quantity of erosion. It may be termed the equivalent width of field. The quantity R, kind S, and orientation K , of vegetation or vegetative cover can be expressed together in terms of equivalent pounds per acre. The equivalent vegetative material is small grain stubble to which S has been assigned the value of 1. The equivalent orientation is the absolutely flat, small-grain stubble with straw aligned parallel with wind direction, for which K , has been assigned the value of 1. The kind of vegetative cover factor, S , denotes the total cross-sectional surface area of the vegetative material. The finer the material, the greater its surface area, the more it slows down the wind velocity, and the more it reduces wind erosion. The orientation of vegetative cover factor, K,,, is in effect the vegetative surface roughness factor and the two terms mean the same thing. The more erect the vegetative matter, the higher it stands above the ground, the more it slows down the wind velocity near the ground, and the lower the rate of erosion. The factors R, S, and K , are multiplied together to give what is termed the equivalent quantity of vegetative cover, V (Chepil, 1962a). The wind erosion equation then may be expressed as

which says that the potential average annual quantity of erosion, or soil loss, E, expressed in tons per acre is a function of the following factors:

Z = soil erodibility, C = local wind erosion climatic factor, K = soil surface roughness, L = equivalent width of field (the maximum unsheltered distance across the field along the prevailing wind erosion direction) , V = equivalent quantity of vegetative cover. The mathematical relationships among the factors in the wind erosion equation are complicated, but charts and tables have been prepared from which the quantity of erosion (soil loss), as influenced by each of these factors, can be read at a glance (Chepil, 1962a). Moreover, the charts and tables can be used in reverse to determine what conditions are necessary to reduce wind erosion to any degree. Space is too limited here to include these charts and tables and to indicate how they can be used to estimate the potential soil loss of a field or the conditions needed to reduce the soil loss to an insignificant amount.

THE PHYSICS OF WIND EROSION AND ITS CONTROL

295

B. DATANEEDEDTO ESTIMATE POTENTIAL SOIL LOSS Each of the individual primary factors that influence wind erosion must be determined before the potential soil loss can be estimated. They are as follows: Datum 1. Soil erodibility I in tons per acre per annum, determined from percentage of nonerodible soil fractions > 0.84mm. in diameter. The percentage of nonerodible fractions is determined by standard dry sieving (Chepil, 1962b) or from reference tables of known average cloddiness of different soils during the wind erosion season. Datum 2. Local wind erosion climatic factor C, in per cent, estimated for a particular geographic location from the wind erosion climatic map (Chepil et al., 1962). Datum 3. Soil surface ridge roughness equiuabnt, K , in inches. Usually K is equal to the average height of clods or ridges of which the soil surface is composed (Zingg and Woodruff, 1951; Chepil, 1962a). Several measurements can be made with a ruler and averaged. Widely spaced ridges, such as those used in emergency tillage for wind erosion control, have a ridge roughness equivalent less than their height. Usually, if the distance between them is increased beyond the 1:4 ratio, their ridge roughness equivalent is decreased proportionately. Thus, if the ridges are 6 inches high and the distance between them, measured along the prevailing wind erosion direction, is 48 inches, their height spacing ratio is 1:8, as compared to 1:4 for standard ridges, so that their ridge roughness equivalent is 4/8 of 6 inches, or 3 inches, if soil cloddiness remains the same as for standard ridges. Datum 4. Distance Df,in feet across the field (along prevailing wind erosion direction). This distance can be measured or computed from the width of field if the prevailing wind erosion direction is known ( Chepil, 1959a). No adequate published data on the prevailing wind erosion direction at various geographic locations are available at present ( 1962). Datum 5. Distance Db, in feet (along prevailing wind erosion direction) of full protection from wind erosion afforded by a barrier, if any, adjoining the field. This distance for standard pervious continuous barrier is about 10 times the height of the barrier (Woodruff and Zingg, 1952). Data on the effectiveness of different kinds of barriers in shielding the soil surface from

296

W. S. C€IEPIL AND N. P. WOODRUFF

erosion are meager, If height of barrier is no greater than normal height of stubble, the influence is negligible and no evaluation is made. Datum6. Quantity of vegetative cover, R, above the ground in pounds per acre. This is estimated by sampling, cleaning, drying, weighing, and computing on a pounds per acre basis in accordance with standard procedure ( Chepil and Woodruff, 1959). For some types of standing stubble, such as sorghum or corn, the quantity can be estimated roughly from height of stubble and number of stalks per unit area. Unpublished supplementary charts and tables are available to facilitate this type of estimation. All quantities of R presented in this review are based on washed, oven-dry material multiplied by 1.20. This represents approximately the average thoroughly cleaned, air-dry weights. Datum 7. Kind of vegetative cover factor, S (dimensionless), obtainable from supplementary tables ( Chepil, 1962a). Datum 8. Orientation of vegetative cover factor, K , (dimensionless), obtainable from supplementary charts ( Chepil, 1962a). VIII. Needed Research

Field and supplemental wind tunnel studies on the basic causes, effects, and remedies of wind erosion began in the severe dust storm period of the 1930's. Data have been collected and recorded continuously till the present time. The first attempt to apply some of this information as part of the wind erosion equation was published by Chepil and Woodruff in 1954. From then, general wind erosion research and research as applied to the wind erosion equation have been continued simultaneously. One is not and could not be separated from the other. Considerable information still is required on air flow, temperature, evaporation, and crop yields in the vicinity of windbreaks and other types of surface barriers such as snowfences, hedges, crop strips, crop rows, ridges, and soil clods. Part of this study is expected to be applied to classification standards for shelterbelts presently in existence in the Great Plains. Ultimately it is hoped that greater clarification may be made of the principles governing air flow patterns and soil erodibility in the vicinity of barriers ranging from the size of clods to field shelterbelts. Experiments on models in a wind tunnel are being initiated to speed up attainment of basic information on this subject. Much damage to soils and crops could be avoided if severe wind erosion conditions could be predicted a few months to a year ahead

THE PHYSICS OF WIND EROSION AND ITS CONTROL

297

of their occurrence. Such predictions might be possible in view of the fact that severe wind erosion conditions tend to occur in cycles. A prediction of severe conditions one growth season ahead of their occurrence should give farmers ample opportunity to establish special tillage and cropping practices that would be effective. Although it is known at present what soil structure approaches an ideal condition for resisting wind, little information is available on how best to create such a condition and at the same time permit the soil to absorb water freely and serve as a good medium for crop growth. None of the present cropping systems, including grasses, are entirely suitable, and some are detrimental. Studies are needed on new techniques of developing a suitable soil structure. More information is needed on the influence of moisture on soil structure as influenced by different types of tillage action. Possibilities of finding new methods and materials to develop desirable sizes of stable soil aggregates should be explored further. It is recognized that vegetative covers, alive or dead, offer one of the most effective conditions for controlling wind and water erosion. However, better implements and probably more extensive education on how best to use the present implements are needed to maintain protective crop residues on the surface, to control wind and water erosion, runoff, and evaporation, and to maintain high level of crop yields. One of the problems associated with present methods of maintaining vegetative covers is that they tend to leave the surface soil loose, fine, and highly erodible by wind. When drought occurs and vegetative covers become depleted, serious erosion sometimes occurs. Implements that improve structure of the surface soil and at the same time maintain vegetative residues on the surface need to be improved. Information on how to preserve vegetative matter above the ground or how to develop vegetative matter resistant to decomposition also is needed, Recognition, selection, and development of plant species suited for reclaiming eroding sand dune land is needed urgently. The general framework of the wind erosion equation has been developed, but many details are still lacking. These details may be filled with accessory charts and tables as more research information becomes available. Information is needed on the average soil surface roughness K for soil surfaces tilled with different implements on different soil classes, with different soil moisture contents. This information is important to determine the nature of the implements and methods of tillage that might be more suited than the present ones for permanent and emergency tillage programs for wind erosion control.

298

W. S. CHEPIL AND N. P. WOODRUFF

Information is needed on the average distance Db of full protection from wind erosion afforded by barriers of various degrees of air penetrability in various geographic regions and for various soils. This type of information for windbreaks and other barriers is presently almost completely lacking. Information is needed on the prevailing wind erosion direction for various locations. Available data needed to determine the prevailing wind erosion direction include: ( a ) average hourly wind velocity from each of the 16 points of the compass, and ( b ) per cent duration of wind from each of the 16 points of the compass. The prevailing wind erosion direction needs to be computed from the above data. A map then can be prepared for estimating the prevailing wind erosion direction on individual farms. This type of information would be valuable in determining factors D, and Db and, inversely, in determining how wide crop strips running in a certain direction should be to control wind erosion in various regions. Soil erodibility I, based on standard dry sieving procedure, needs to be determined for various soil types wherever wind erosion is a problem. Information on the values of kind of vegetative cover factor S and orientation of vegetative cover factor K , is needed for cultivated and grass crops other than those already investigated. It is expected that the wind erosion equation will become more useful as more specific information on the influence of the major primary factors I, C, K , Df,Db, R, S , and K , becomes available. IX. Conclusion

This review has been devoted to discussion of progress made in obtaining new information on wind erosion and its control. However, the solution of the problem is dependent on the overall progress made in research, testing, and extension. It is beyond the scope of this review to discuss the overall progress made in the solution of the wind erosion problem. Substantial progress apparently has been made. Probably the best evidence of this is the fact that the severity of dust storms in the Great Plains during the 1950’s was considerably less than during a period of similar climatic conditions in the 1930’s (Chepil and Woodruff, 1957; Chepil et al., 1962; unpublished data by Chepil et al.) This difference is believed to be due to better techniques, more favorable financial resources, and more earnest desire on the part of everyone to conserve the soil.

THE PHYSICS OF WIND EROSION AND ITS CONTROL

299

REFERENCES

Agricultural Research Service. 1961. U.S. Dept. Agr. Soil Water Conseru. Res. Dio. ARS Spec. Rep. 22-69. Anderson, D. T. 1953. Agr. Inst. Reo. 8, 13-15. Anderson, D. T. 1961. Can. 1. Soil Sci. 41, 99-114. Anderson, D. T. 1962. Proc. Gt. Plains Agr. Council Workshop Stubble Mulch Farming, Lincoln, Nebraska, 1962 pp. 1-14. Bagnold, R. A. 1936. Proc. Roy. Soc. A157, 594-620. Bagnold, R. A. 1943. “The Physics of Blown Sand and Desert Dunes.” Morrow, New York. Bates, C. G . 1944. U. S. Dept. Agr. Farmers’ Bull. 1405. Bennett, H. H. 1939. “Soil Conservation.” McGraw-Hill, New York. Bodrov, V. A. 1935. U.S. Forest Sero. Translation No. 164. Bradfield, R., and Jamison, V. C. 1938. Soil Sci. SOC. Am. Proc. 3, 70-76. Browning, G. M. 1944. Soil Sci. 57, 91-106. Brunt, D. 1944. “Physical and Dynamical Meteorology,” 3rd ed. Cambridge Univ. Press, London and New York. Caborn, J. M. 1957. Dept. Forestry, Edinburgh Unio. Bull. 29, 135 pp. Canada, Department of Agriculture. 1943. Can. Dept. Agr. Rept. Inuest. IM-6582. Canada, Department of Agriculture. 1949. Can. Dept. Agr. Tech. Bull. 71. Chepil, W. S. 1941. Sci. Agr. 21, 488-507. Chepil, W. S. 1943. Soil Sci. 55, 275-287. Chepil, W. S. 1944. Sci. Agr. 24, 307-319. Chepil, W. S. 1945a. Soil Sci. 60, 305-320. Chepil, W. S. 1945b. Soil Sci. 60, 397-411. Chepil, W. S. 1945c. Soil Sci. 60, 475-480. Chepil, W. S. 1945d. Soil Sci. 61, 167-177. Chepil, W. S. 1949. Agron. 1. 41, 127-129. Chepil, W. S. 1950a. Soil Sci. 69, 149-162. Chepil, W. S. 1950b. Soil Sci. 69, 403-414. Chepil, W. S. 1951. Soil Sci. 72, 465-478. Chepil, W. S. 1952. Soil Sci. SOC.Am. Proc. 16, 113-117. Chepil, W. S. 1953a. Soil Sci. 76, 389-399. Chepil, W. S. 1953b. Soil Sci. SOC. Am. Proc. 17, 185-190. Chepil, W. S. 1954a. Soil Sci. 77, 473-480. Chepil, W. S. 1954b. Soil Sci. SOC. Am. Proc. 18, 13-16. Chepil, W. S. 1955~. Soil Sci. 80, 155-162. Chepil, W. S. 1955b. Soil Sci. 80, 413-421. Chepil, W. S. 1955c. Soil Sci. SOC. Am. Proc. 19, 125-128. Chepil, W. S. 1956. Soil Sci. SOC. Am. Proc. 20, 288-292. Chepil, W. S. 1957a. Am. 1. Sci. 255, 12-22. Chepil, W. S. 1957b. Am. 1. Sci. 255, 206-213. Chepil, W. S. 1957c. Kansas Agr. Erpt. Sta. Tech. Bull. 92. Chepil, W. S. 1958. U.S. Dept. Agr. Tech. Bull. 1185. Chepil, W. S. 1959a. J . Soil Water Consero. 14, 214-219. Chepil, W. S. 195913. Soil Sci. SOC. Am. Proc. 23(6), 422-428. Chepil, W. S. 1960a. J. Soil Water Consero. 15, 72-75. Chepil, W. S. 1960b. Soil Sci. SOC. Am. Proc. 24, 143-145. Chepd, W. S. 1961. Soil Sci. SOC. Am. P ~ o c .25(5), 343-345.

300

W. S. CHEPIL AND N. P. WOODRUFF

Chepil, W. S. 1962a. Proc. Gt. Plains Council Workshop Stubble Mulch Farming, Lincoln, Nebraska, 1962 pp. 1-5. Chepil, W. S. 1962b. Soil Scl. SOC. Am. Proc. 26, 4-8. Chepil, W. S. 1962c. Trans. 7th Intern. Congr. Soil Sci. Madison, Wisconsin, 1960 1, 225-231. Chepil, W. S., and Bisal, F. 1943. Soil Sci. 56, 95-100. Chepil, W. S., and Milne, R. A. 1941a. Sci. Agr. 21, 479-487. Chepil, W. S., and Milne, R. A. 1941b. Soil Sci. 52, 417-433. Chepil, W. S., and Siddoway, F. H. 1959. J. Meteorol. 16, 411-418. Chepil, W. S., and Woodruff, N. P. 1954. J. Soil Water Conseru. 9, 257-265. Chepil, W. S., and Woodruff, N. P. 1957. Am. 1. Sci. 255, 104-114. Chepil, W. S . , and Woodruff, N. P. 1959. U.S. Dept. Agr. Agr. Res. Sero. Res. Rept. 25, 1-21. Chepil, W. S., Woodruff, N. P., Siddoway, F. H., and Lyles, L. 1960. Agr. Eng. 41, 754-755, 759. Chepil, W. S., Woodruff, N. P., and Siddoway, F. H. 1961. U. S . Dept. Agr. Farmers’ Bull. 2169, 16 pp. Chepil, W. S., Siddoway, F. H., and Armbrust, D. V. 1962. J . Soil Water Consero. 17( 4 ) , 162-165. Daniel, H. A. 1936. 1. Am. SOC. Agron. 28, 570-580. Duley, F. L. 1958. U.S. Dept. Agr. Agr. Handbook 136, 1-31. Einstein, H. A., and El-Samni, El-Sayed Ah. 1949. Rev. Mod. Phys. 21, 520-524. Fenster, C. R. 1960. Soil Sci. SOC. Am. Proc. 24, 518-523. Ferber, A. E. 1958. U. S . Dept. Agr. Misc. Publ. 759, 1-22. Fly, C. L. 1935. Oklahoma Panhandle Expt. Sta. Bull. 57. Geiger, R. 1957. “The Climate Near the Ground,” 2nd rev. Harvard Univ. Press, Cambridge, Massachusetts. Goldstein, S. 1938. “Modern Developments in Fluid Dynamics.” Oxford Univ. Press ( Clarendon ) , London and New York. Greb, B. W., and Black, A. L. 1961. J. Soil Water Conseru. 16, 223-227. Greb, B. W., and Black, A. L. 1962. Agron. J . 54, 116-119. Gumbel, E. J. 1941. Ann. Math. Statist. 12, 163-190. Gumbel, E. J. 1945. Trans. Am. Geophys. Union [l] 26, 68-82. Hardt, G. 1936. 2. Pjlanzenernaehr. Dueng. Bodenk. A45, 216-238. Harper, H. J., and Brensing, 0. H. 1950. Oklahoma Agr. Expt. Sta. Bull. 13-362, 1-28. Havis, L. 1943. Ohio Agr. Expt. Sta. Bull. 640. Hide, J. C., and Metzger, W. H. 1939. Soil Sci. Soc. Am. Proc. 4, 19-22. Hopkins, E. S. 1935. Trans. 3rd Intern. Congr. Soil Sci. Oxford 1935 1, 403-405. Hopkins, E. S . , Palmer, A. E., and Chepil, W. S . 1946. Can. Dept. Agr. Farmers’ Bull. 32, 4th rev. Iizuka, H. 1950. Meguro For. Expt. Sta. Bull. 45, 95-129. Ippen, A. T., and Verma, R. P. 1953. Reprinted from Proc. Minnesota Intern. Hydraul. Cono., Hydrodynamics Lab., Massachusetts Inst. Tech., Cambridge, Massachusetts. Jacks, G. V., and Whyte, R. 0. 1939. “Vanishing Lands.” Doubleday, New York. Jeffries, H. 1929. Proc. Cambridge Phil. SOC. 25, 272-276. Kalinske, A. A. 1943. Ann. N . Y . Acad. Sci. 44, 41-54. Lane, D. E., and Wittmuss, H. 1961. Nebraska Expt. Sta. Leaflet Ext. Circ. 61-714, 1-8.

THE PHYSICS OF WIND EROSION A N D ITS CONTROL

301

Langham, W. H., Foster, R. L., and Daniel, H. A. 1938. 3. Am. SOC. Agron. 30, 139-144. Lyles, L., and Woodruff, N. P. 1961. Soil Sci. 91, 178-182. Lyles, L., and Woodruff, N. P. 1962. Agr. Eng. 43, 150-153, 159. McCalla, T. M. 1942. Soil Sci. SOC. Am. Proc. 7, 209-214. McCalla, T. M. 1945. Soil Sci. 59, 287-297. McCalla, T. M. 1950. Trans. Kansas Acad. Sci. 53, 91-100. McCalla, T. M., and Army, T. J. 1961. Adoan. Agron. 13, 125-196. Malin, J. G. 1946. Kansas Hist. Quart. 14, 1-71. Martin, J. P. 1942. Sod Sci. SOC. Am. Proc. 7, 218-222. Martin, J. P. 1946. Soil Sci. 61, 157-166. Mathews, 0. R. 1954. U. S. Dept. Agr. Farmers’ Bull. 1797. Moss, H. C. 1935. Sci. Agr. 15, 665-675. Myers, H. E. 1937. Soil Sci. 44, 331-357. Parkinson, G. R. 1936. Bull. Am. Meteorol. SOC. 17, 127-135. Peele, T. C. 1940. 3. Am. Soc. Agron. 32, 204-212. Peele, T. C. 1941. Soil Scl. SOC. Am. Proc. 6, 176-182. P6w6, T. L. 1951. 3. Geol. 59, 399-401. Polynov, B. B. 1937. “The Cycle of Weathering.” Thomas Murby, London. Potter, W. D. 1949. U. S. Dept. Agr. SCS-TP-78. Richards, L. A. 1953. Soil Sci. SOC. Am. Proc. 17, 321-323. Rouse, H. (ed. ). 1950. “Engineering Hydraulics: Proceedings 4th Hydraulic Conference, Iowa Institute Hydraulics,” Chapt. 1. Wiley, New York. Russell, E. W. 1938. Imp. Bur. Soil Sci. Harpenden Eng. Tech. Commun. 37. Schultz, H. B., and Carlton, A. B. 1958. Calif. Agr. 12, 1, 13. Schultz, H. B., and Carlton, A. B. 1959. Calif. Agr. 13, 5-6. Sears, P. B. 1935. “Deserts on the March.” Univ. Oklahoma Press, Norman, Oklahoma. Sheng, T. 1961. Joint Comm. Rural Reconstruct. For. Ser. 7, 1-49. Sheppard, P. A. 1947. Proc. Roy. SOC. A188(1013), 208-222. Siddoway, F. H., McKay, H. C., and Klages, K. H. 1958. Idaho Agr. Expt. Sta. Bull. 252. Slater, C. S,, and Hopp, H. 1951. Agron. 3. 43, 1-4. Sneesby, N. J. 1953. Agriculture (London) 60, 263-271. Sobolev, S. S. 1947. Soil Consere. 13, 17-19 (Russian-translated by T. Mills). Staple, W. J., and Lehane, J. J. 1955. Can. 3. Agr. Sci. 35, 440-453. Swineford, A., and Frye, J. C. 1945. Am. 3. Sci. 243, 249-255. Thompson, L. M. 1952. “Soils and Soil Fertility.” McGraw-Hill, New York. Thomthwaite, C. W. 1931. Geogruph. Reu. 21, 633-655. Udden, J. A. 1898. Augustana Library Publ. 1. Vanoni, V. A. 1946. Trans. Am. Soc. Cioil Engr. 3, 67-133. von Karman, T. 1934. 3. Aeronaut. Sci. 1, 1-20. Warn, F. G., and Cox, W. H. 1951. Am. 3. Sci. 249, 553-568. White, C. M. 1940. Proc. Roy. SOC. A174, 322-328. Whitfield, C. J. 1938. 3. Agr. Res. 56, 907-914. Woodruff, N. P. 1954. Kansas Agr. Expt. Sta. Tech. Bull. 77. Woodruff, N. P. 1956. Agron. 3. 48, 499-504. Woodruff, N. P., and Chepil, W. S. 1956. Agr. Eng. 37, 751-754, 758. Woodruff, N. P., and Chepil, W. S. 1958. Trans. Am. SOC. Agr. Eng. 1, 81-85. Woodruff, N. P., and Zingg, A. W. 1952. U.S. Dept. Agr. SCS-TP-112.

302

W. S. CHEPIL AND N. P. WOODRUFF

Woodruff, N. P., and Zingg, A. W. 1955. Trans. Am. Geophys. Union 36, 203-207. Woodruff, N. P., Chepil, W. S., and Lynch, R. D. 1957. Kansas Agr. Expt. Sta. Tech. Bull. 90. Zingg, A. W. 1949. Agr. Eng. 30, 11-13, 19. Zingg, A. W. 1950. U. S. Dept. Agr. SCS-TP-88. Zingg, A. W. 1953a. Iowa State Uniu. Proc. 5th Hydraulic Conf. Bull. 34. Zingg, A. W. 1953b. Trans. Kansas Acad. Sci. 56, 371-377. Zingg, A. W. 1954. Trans. Am. Geophys. Union 35, 252-258. Zingg, A. W., and Whitfield, C. J. 1957. U . S. Dept. Agr. Tech. Bull. 1166. Zingg, A. W., and Woodruff, N. P. 1951. Agron. J. 43, 191-193. Zingg, A. W., Woodruff, N. P., and Englehom, C. L. 1952. Agron. J. 44, 227-230.

PLANT NUTRIENT LOSSES FROM SOILS BY WATER EROSION Harold L. Barrows and Victor J. Kilmer* United States Department of Agriculture, Beltsville, Maryland

I. 11. 111. IV. V. VI. VII. VIII. 1x. X. XI.

Introduction ................................................ Methods and Conditions of Sampling Runoff ..................... Organic Matter Losses ....................................... Nitrogen Losses ............................................. Phosphorus Losses ........................................... Potassium Losses ............................................ Calcium Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Magnesium Losses ........................................... Sulfur Losses ............................................... Interpretation of Runoff Data ................................. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References .................................................

Page 303 305 306 307 309 311 312 313 313 313 315 315

1. Introduction

Modern agriculture is largely dependent upon the maintenance of a favorable equilibrium between losses and gains of available plant nutrient elements in soils. Erosion, cropping, leaching, and volatilization are the four principal pathways whereby nutrient elements are removed from soils. The harvested crop removes only those elements taken up by the crop plant during its period of growth, but leaching losses in humid regions tend to be governed by the degree of mobility of nutrient elements in soils. Volatilization losses are generally thought to be limited to nitrogen and sulfur, but quantitative field data relating to the magnitude of such losses are limited. The removal of elements by erosion is nonselective in the sense that nutrient elements in all forms may be removed by the erosion process. The process tends to be selective in that the organic matter and finer particles of soil relatively high in plant nutrients are more vulnerable to erosion than are the coarser soil fractions. The removal of plant nutrients, either in solution or in colloidal suspension, by water flowing over the surface of the land falls in the category of runoff chemistry. The object O

Present address: Tennessee Valley Authority, Wilson Dam, Alabama. 303

304

HAROLD L. BARROWS AND VICTOR J. KILMER

of this paper is to review the more pertinent work that has been done in the field of runoff chemistry. For a detailed review of the mechanics of rainfall erosion and the basic factors affecting soil loss, the reader is referred to the work of Smith and Wischmeier (1962). Although many papers suggesting methods for controlling erosion were published in the United States prior to 1920, quantitative work relating to water, soil, or nutrients in runoff was not included. Harper (1958) concluded that these earlier workers did not appreciate the economic importance of controlling erosion. One of the earliest approaches to a quantitative evaluation of plant nutrient removal by erosion was the early work of Sampson and Weyl (1918), who analyzed soil taken from eroded and noneroded pastures. The soil from the noneroded pasture contained more than three times as much nitrogen and more than twice as much organic matter as the soil taken from the eroded pasture. The first field plots in the United States specifically designed to measure quantitatively the loss of soil, water, and nutrients from surface runoff were established in Missouri in 1917 (Duley and Miller, 1923). These 1/80-acre plots, located on Shelby loam having a 3.68 per cent slope, were subjected to various management practices. Plots in sod lost 42 pounds of soil per acre per year; fallow plots lost 5195 pounds of soil. The losses of plant nutrients from the sod plots were insignificant, but annual losses per acre from fallow plots were 99 pounds of nitrogen, 47 pounds of phosphorus, 379 pounds of calcium, and 100 pounds of sulfur. Early workers, for the most part, analyzed for total amounts of elements' in eroded material and reported these results in terms of nutrient losses per acre. In considering the present review of this work, the reader should bear in mind that only a small fraction of all nutrients in soils is available to plants. Attempts have been made to determine quantitatively the concentration of soluble ions in runoff, The concentrations are generally very low. It is possible that the low losses reported for soluble cations, as well as phosphorus, may be the result of reabsorption of the ions in solution by the colloidal material in runoff. In order to obtain a reasonable estimate of losses of readily available plant nutrients in runoff, it is necessary to determine ions in both solution and absorbed phases. The distinction between losses of available and total elements is made in the text of this paper. Where losses of available elements are reported, it is SO stated. 1 The elemental system is used in this review, and all data reported as oxides in the original papers have been converted.

NUTRIENT LOSSES BY EROSION

305

Since the eroded material frequently differs in composition from the original soil, the loss of nutrients in runoff may be expressed in terms of an enrichment ratio (E.R.). This is the ratio of the concentration of that element in the runoff to that in the original soil: E.R.

=

Concentration of the element in soil material in runoff Concentration of the element in soil from which runoff originated

II. Methods and Conditions of Sampling Runoff2

The quantitative measurement of runoff requires that the area from which runoff occurs be specifically defined. Such an area may be a watershed with natural boundaries; more commonly, field plots are established. These field plots are, in effect, small watersheds the boundaries of which may consist of corrugated metal strips or low earth dikes. The runoff is collected in a downslope trough, and a pipe carries the runoff to silt boxes and side tanks (Fig. 1 ) . The measuring equipment is designed to collect aliquots of the runoff. This device permits the measurement of runoff occurring under storms of widely varying charac-

FIG. 1. General view of runoff plots, Watkinsville, Georgia. 2 The authors are indebted to 0. E. Hays and R. E. Taylor, Agricultural Research Service, for their assistance in preparing this portion of the manuscript.

306

HAROLD L. BARROWS AND VICTOR J. KILMER

teristics. The amount of runoff collected in open tanks is, of course, corrected for rainfall. The contents of the tanks are thoroughly mixed and sampled. The volume of the sample taken is adjusted to represent a definite percentage of the runoff occurring for any given storm. One milliliter of toluene is added to each sample to retard microbial activity. Precipitated calcium carbonate is added to samples taken for nitrate analysis, and the sample is dried in an oven as soon as possible after collection. The chemical analysis of runoff samples requires somewhat specialized procedures and techniques. Unfortunately, many of the early reports did not include the procedures followed for collecting, preparing, and analyzing the samples. Thus, a critical evaluation of these data is difficult. For a discussion of procedures and techniques used in the analysis of runoff samples, the reader is referred to Jackson (1958). 111. Organic Matter Losses

The importance of soil organic matter is well known. Because of its concentration in the surface soil and its low density, organic matter is among the first of the constituents to be removed through erosion (Slater, 1942), yet it is among the hardest to replace (Slater and Carleton, 1938; Neal, 1939). Applied nutrients are generally less effective on an eroded soil than on a noneroded, friable soil containing adequate organic matter (Lamb et al., 1950). Neal (1939) was able to reduce erosion on Marshall silt loam, 9 per cent slope, by additions of organic material, but he was not able to increase soil organic matter by four annual applications of 16 tons of barnyard manure per acre. Slater and Carleton (1938) found that organic matter was depleted 18 times faster through erosion from a fallow soil than through normal oxidation. They concluded that 9.2 tons of clover hay per acre would have to be applied annually to maintain the soil organic matter at its original level. Knoblauch et aZ. (1942) reported a loss of 337 pounds per acre of organic matter from a 3.5 per cent slope of Collington silt loam protected with a manured cover, whereas the check plots lost 1149 pounds per acre. Hays et al. ( 1948) reported 951 pounds per acre of organic matter lost annually from moderately eroded Fayette silt loam and 668 pounds per acre from severely eroded phase. Bedell et al. (1946) reported a total loss of 555 pounds of organic matter per acre. The eroded material contained about twice as much organic matter as the remaining surface soil from which it came. Free (1956) reported that organic matter was 30 per cent higher in material removed from Honeoye sandy

NUTRIENT LOSSES BY EROSION

307

loam than it was in the soil remaining in the plots. Enrichment ratios for organic matter from Fayette and Almena silt loams of 1.15 and 1.24, respectively, have been reported (Massey et aZ., 1953), as well as E.R.’s of 4 to 4.7 from a Collington silt loam (Neal, 1943, 1944). Although the loss of organic matter is a function of the soil loss, it is not a linear function. Martin (1941) reported that the loss of organic matter depended on the soil losses, but the percentage of the humus in the eroded material decreased as the erosion losses increased. Massey et al. (1953) also observed that when soil loss was high the percentage of organic matter and nitrogen in the runoff was low, but when soil loss was low the percentage was high. One of the first attempts to correlate organic matter losses with soil losses was carried out on a Marshall silt loam by Slater and Carleton (1938). For the period 1930 to 1937 the concentration of organic matter in the original soil dropped 0.002 per cent for each ton of soil lost. Free (1946) found the same relation existing between the organic matter loss and soil removed over a 4-year period from a Honeoye soil. The losses reported for organic matter are critical. It is among the first constituents to be lost through erosion, yet it is among the hardest to replace. Not only is the soil being depleted of one of its most valuable components, but significant quantities of nutrients, such as nitrogen and phosphorus, are removed with the organic matter. The nutrients can be replaced, but they are far less effective on soils in which the organic matter has been lost than they are on soils containing an adequate supply of organic matter. With the trend toward increased fertilizer consumption, additional data are needed relating to specific management practices which will permit a more thorough economic appraisal of organic matter losses caused by water erosion. IV. Nitrogen Losses

The enrichment ratio for nitrogen losses appears, in general, to parallel the E.R. for organic matter losses. This is to be expected, since, in humid region soils, the major portion of soil nitrogen is combined with organic matter. Massey et al. (1953; Massey and Jackson, 1952) studied fertility erosion on two Wisconsin soils. The E.R. for total nitrogen and organic matter removed from Almena silt loam was 1.34 and 1.24, respectively. An E.R. of 1.08 was obtained for nitrogen lost from Fayette silt loam; for organic matter the E.R. was 1.15. Studies made by Neal (1944) on Collington silt loam resulted in a nitrogen E.R. of 5.0 and an organic matter E.R. of 4.7. Ratios of 3.9 and 4.2 for nitrogen and organic

308

HAROLD L. BARROWS AND VICTOR J. KILMER

matter, respectively, have also been reported for Collington sandy loam by Knoblauch et al. (1942). The actual amounts of nitrogen lost under cropping conditions may be relatively small. Hays et al. (1948) reported 42 pounds per acre of nitrogen lost annually from Fayette silt loam planted to oats; however, 90 per cent of this loss came from two intense rains shortly after the grain was seeded. The same soil under corn lost only 2 pounds of nitrogen per acre. Rogers (1941) found yearly nitrogen losses of 32 pounds per acre on Dunmore silt loam planted to either wheat or corn. Collington sandy loam, having a 3.5 per cent slope, lost 67 pounds of nitrogen per acre per year, but when a cover crop was established and manure applied, the nitrogen loss was reduced to 19 pounds (Knoblauch et al., 1942). Earlier work by Duley and Miller (1923) in Missouri showed that a fallow Shelby loam with a 3.68 per cent slope lost 5195 pounds of soil per acre on an annual basis. The eroded material contained 99 pounds of nitrogen, of which 6 pounds was in the nitrate form. A sod cover permitted an annual soil loss of only 42 pounds per acre and negligible losses of nitrogen. Free (1956) reported no difference in the ammonium-nitrogen content of the eroded material and soil from Honeoye silt loam, but the eroded material contained 48 per cent less nitrate than did the plot soil. Attempts have been made to determine the amounts of nitrogen carried in solution in runoff water. As pointed out by Bryant and Slater ( 1948 ) , there are several inherent difficulties encountered in this and other runoff studies: ( a ) solutes in runoff may be derived from either soil or vegetative cover; ( b ) rainwater may contain small amounts of the solute in question; and ( c ) the nature and intensity of rain can affect the amounts and kind of materials removed in solution. Possible volatilization losses and absorption of soluble ions by colloids in suspension also add to the difficulties of obtaining a true measurement of amounts of nitrogen carried in solution in runoff waters. Special techniques are required in such a study. Bryant and Slater (1948), working in New York State, found that 10.4 pounds of dissolved nitrogen per acre was removed annually from a fallow Dunkirk silty loam with a 5 per cent slope, but that a fallow Ontario loam having an 8 per cent slope lost less than a pound of dissolved nitrogen during the same period. Duley (1926) indicated that most of the soluble nitrogen removed is in the organic form, but the annual removal of nitrogen in this form was insignificant. Even lesser amounts of ammonia, nitrate, and nitrite were lost. A small amount of nitrogen is returned to the soil by rain and snow. Buckman and Brady (1960) suggest an average annual net accrual of

NUTRIENT LOSSES BY EROSION

309

5 pounds of nitrogen per acre under a humid-temperate climate. Daniel et al. (1938) reported that 0.4 to 1.5 pounds per acre of nitrate nitrogen was returned annually to Vernon very fine sandy loam in Oklahoma during 1930-1937, and Bryant and Slater (1948) obtained a figure of 8.93 pounds of total nitrogen under New York conditions. The loss of nitrogen is not lineally proportional to the soil loss. Stoltenberg and White (1953) plotted nitrogen loss as a function of total soil loss. When the concentration of solids in the runoff was below 100 pounds per acre-inch, the nitrogen content of the material was 5 times as high as in the soil. The ratio decreased rapidly with increasing concentrations of solids so that the nitrogen content was about double when the total solids were 609 pounds. At higher concentrations of solids there was a small further decrease, and at concentrations of 9ooo pounds per acre-inch, the nitrogen content was about 30 per cent above that in the soil. Bedell et al. (1946) found that the amount of nitrogen removed was more closely associated with total runoff than with total soil losses. Attempts have been made to calculate the total amount of nitrogen lost from various soils and to replace this by additions of fertilizers. Wooley (1943) published tables showing nitrogen losses from the soils of Missouri under different management schemes. These losses varied from none on a well-sodded meadow to 18 per cent for continuous row crop on a 1200-foot, 12 per cent slope. By using this table, and another showing nitrogen removal by various crops, a balance sheet program could be prepared in which soil nitrogen could be maintained. Losses of nitrogen by erosion are probably more serious than losses of any other nutrient. This results from the fact that most of the nitrogen that is lost is combined with the soil organic matter, which is so susceptible to erosion. The losses reported for soluble nitrogen salts in runoff water are exceedingly low and are of little economic importance. However, it is possible that under different conditions the losses of soluble nitrogen salts might be increased considerably. This could occur in runoff from fields where high rates of nitrogen fertilizer have recently been applied. Existing data are insufficient to evaluate the influence of such factors as rates, sources, placements, and times of application relative to the occurrence of runoff. But, it is quite likely that these factors also would influence the rate of soluble nitrogen removed in the runoff water. V. Phosphorus losses

Under ordinary field conditions, phosphorus is one of the least mobile of the plant nutrients. The downward movement of this element through soils is exceedingly slow (Scarseth and Chandler, 1938; Elmer et al.,

310

HAROLD L. BARROWS AND VICTOR J. KILMER

1944; Broadbent and Chapman, 1949). Phosphorus applied to soils remains in the upper few inches of the profile unless mechanically incorporated to lower depths. Organic forms of phosphorus may constitute as much as 65 per cent of the total phosphorus supply (Buckman and Brady, 1960), and as much as 50 per cent of the total phosphorus may be held on clay surfaces in the plow layer of light-textured soils (Scarseth and Chandler, 1938). Thus, phosphorus is very susceptible to losses through erosion of clay and organic matter. Scarseth and Chandler (1938) concluded that 60 per cent of the superphosphate applied over a 26-year period was lost by erosion from a nearly level field of Norfolk loamy sand planted to a cotton, corn, oat, and legume rotation. Ensminger and Cope (1947) observed a 70 per cent loss of phosphorus added to an unlimed, nearly level Norfolk fine sand cropped to cotton. Erosion losses of phosphorus added to limed soil were related to the source of nitrogen applied. Approximately 75 per cent of the phosphorus was lost when nitrogen was applied as ammonium sulfate; the loss was 32 per cent with sodium nitrate. Larger amounts of phosphorus moved into the subsoil under lime or sodium nitrate treatments. Rogers (1942) demonstrated the loss of unreacted phosphorus from a Dunmore silt loam in permanent pasture. He applied triple superphosphate at the rate of 200 pounds per acre followed immediately by a series of l-inch “rains” from a rainfall simulator. The first “rain” removed 9.1 per cent of the applied superphosphate, the second 4.3 per cent. Up to 22 per cent phosphorus loss occurred when “rain” was applied to dry soil immediately after fertilization. Ensminger (1952), who worked with Hartsell fine sandy loam of 2 to 4 per cent slope, considered that phosphorus that could not be accounted for in the surface 16 inches of soil, plus that removed by crops, was lost through the process of erosion. Under a corn and cotton rotation, an average of 63 per cent of the phosphorus applied from various sources was lost. Plots that were enclosed to prevent runoff lost about 20 per cent of the applied phosphorus. Volk (1945), in reporting the results of a study involving cotton grown on Hartsell very fine sandy loam, could not account for 26 per cent of the native and applied phosphorus over a 14-year period. He considered that this unaccounted for phosphorus was removed by erosion. Bedell d a2. (1946) demonstrated the loss of organic phosphorus through processes of erosion. Where corn was grown under prevailing management practices, over 4.5 tons of solids were removed per acre, carrying 20 pounds of phosphorus. Nearly 60 per cent of the phosphorus lost was in organic form.

NUTRIENT LOSSES BY EROSION

311

The concentration of phosphorus in runoff tends to be considerably higher than the concentration in the initial soil. Rogers (1941) found the E.R. of total phosphorus from a wheat-cropped Dunmore silt loam, 20 to 25 per cent slope, to be 1.3; for soluble phosphorus (0.002N H2S04), it was 3.3. Neal (1943, 1944), working with Collington sandy loam having a 3.5 per cent slope, reported E.R.’s for total phosphorus ranging from 1.5 to 3.1. Similar ratios for Almena and Fayette silt loams were 1.9 and 2.2, respectively (Massey et al., 1953). VI. Potassium Losses

Where both total and available potassium losses are reported, the amount of available potassium is, of course, much smaller than the total potassium removed. In this connection one should bear in mind that perhaps 90 to 98 per cent of all soil potassium is in a form that is not readily available to plants. The amounts of potassium removed in solution are generally very limited. Duley (1926) reported that the annual removal of potassium in solution from a Shelby loam ranged from about 1 pound per acre under wheat to about 9 pounds per acre from sod plots. Bryant and Slater (1948) obtained soluble potassium losses of less than 1 pound per acre annually from Ontario and Dunkirk soils. The measured annual soil loss on the latter soil was over 99,000 pounds. It is possible that the low losses reported for potassium in solution may be the result of reabsorption of the ions by collodial material in the runoff. This may be of particular significance if a storm occurs before the added potassium fertilizer has reached equilibrium and if the runoff is not sampled and the solids separated immediately. Under these conditions, an equilibrium would be established between the potassium on the colloidal material and the potassium in the solution that would alter the distribution ratio that existed at the time runoff occurred. Any sample removed and separated after this occurred would be biased in favor of potassium removed with the solids. Little is known about potassium losses from newly fertilized fields, but such losses could be significant when light rains occur or during the first period of more intense storms (Kohnke, 1941). Bedell et al. (1946) reported annual losses of 0.8 to 6.6 pounds of available potassium per acre, the higher loss occurring under corn where conservation management practices were not employed. Hays et al. (1948) determined both total and available potassium losses on Fayette silt loam planted to either oats or corn. Annual losses of total potassium amounted to 854 pounds per acre from oat plots, but were only 26 pounds from plots cropped to corn. The loss of available potassium was 26 and 0.5 pounds

312

HAROLD L. BARROWS AND VICTOR J. KILMER

per acre, respectively. The grain plots permitted much higher losses, but 90 per cent of this loss was caused by two intense storms that occurred shortly after the grain was seeded. Knoblauch et al. (1942) obtained losses of total potassium amounting to 426 pounds annually from Collington sandy loam. Losses were reduced to 98 pounds under a cover crop receiving manure. Yearly losses of total potassium from freshly plowed Shelby loam in Missouri have been as high as 1246 pounds per acre; bluegrass sod reduced the loss to less than 3 pounds. The E.R. for available potassium is much higher than the E.R. for total amounts of that element. Neal (1943, 1944) reported an EX. of 1.4 for total and 5.4 for available potassium. Other workers reported similar findings: E.R.’s obtained for total potassium are 1.07, 3.8, and 7.3; for available potassium, 4.7, 6.7, and 12.6, respectively (Rogers, 1941; Massey et al., 1953; Stoltenberg and White, 1953). Stoltenberg and White obtained the highest E.R., 12.6, under conservation practices. VII. Calcium Losses

Early work by Duley and Miller (1923) showed virtually no removal of total calcium from Shelby loam under sod, but 379 pounds per acre was removed from bare, uncultivated soil. When the soil was cultivated to a depth of 4 inches, the loss per acre rose to 458 pounds (Miller and Krusekopf, 1932). Total calcium removed from a Dunmore silt loam amounted to 77 pounds per acre annually, with an E.R. of approximately 1.0 (Rogers, 1941). Knoblauch et al. (1942) obtained a yearly loss per acre of 101 pounds of calcium from an unprotected Collington sandy loam; protection with a manured cover crop reduced the loss to 25 pounds. Enrichment ratio values were 2.06 and 2.41, respectively. Duley (1926) determined that losses of calcium in solution from a Shelby loam ranged from 11 pounds with a sod cover to 26 pounds per acre from a fallow plot spaded to a depth of 4 inches. A fallow Dunkirk silty clay loam with a 5 per cent slope lost 10 pounds of soluble calcium per acre; when planted to corn, 12 pounds was removed by runoff water. Losses from Ontario loam were about 6 pounds (Bryant and Slater, 1948). Rogers (1942) studied the rate of calcium removal from freshly applied limestone. He applied 2 tons of limestone per acre to a permanent pasture on Dunmore silt loam which was then subjected to l-inch “rains.” Less than 1 per cent of the calcium applied as limestone was removed by the first “rain.” Subsequent “rains” continued to remove comparable amounts.

NUTRIENT LOSSES BY EROSION

313

VIII. Magnesium losses

The limited data relating to loss of magnesium through erosion indicates that negligible amounts of this element are removed in the soluble form. Duley (1926) reported a maximum removal in solution of 2.6 pounds per acre; Bryant and Slater (1948) obtained a loss of 1.33 pounds from fallow Ontario loam and 3.6 pounds per acre from Dunkirk silty clay loam cropped to corn. Reported annual losses of total magnesium are 179 pounds per acre from fallow Shelby loam (Miller and Krusekopf, 1932), and 94 pounds from Dunmore silt loam planted to wheat with an E.R. of 1.39 (Rogers, 1941 ) . IX. Sulfur losses

Only meager data exist on sulfur losses, but these data indicate that relatively significant amounts of sulfur may be carried in solution by runoff water. According to Duley (1926), 11.5 pounds per acre of soluble sulfur was removed from Shelby loam in Missouri plowed to a depth of 4 inches; when plowed to an &inch depth the loss was 26.5 pounds. Bryant and Slater (1948) obtained much lower losses in New York: 1.29 and 4.15 pounds of soluble sulfur removed from fallow Ontario silt loam and Dunkirk silty clay loam, respectively. Duley and Miller (1923) reported losses of 101 pounds of total sulfur per acre from bare, uncultivated Shelby loam. When plowed to depths of 4 and 8 inches, the losses of total sulfur were 47 and 42 pounds per acre, respectively. These results were obtained from the analysis of a composite sample of the annual runoff. X. Interpretation of Runoff Data

Data on nutrient loss through runoff reported prior to 1940 are difficult to interpret. Most of the work conducted during this time consisted of total analyses for the specific elements in the runoff. This generally would include the amounts removed mechanically, in colloidal suspension, and in solution. Often the runoff from each rain would be sampled individually, but the samples would be dried and composited to give one sample for chemical analysis. The concentrations of the various elements in the runoff were multiplied by the total amount of material removed for the year and a figure was obtained giving an indication of the total amount removed annually. It is a misconception to assume that erosional damage is proportional to the pounds of soil removed (Ellison, 1950) or to the total amount of

314

HAROLD L. BARROWS AND VICTOR J. KILMER

nutrients removed. The total amount of organic matter and nutrients removed from highly productive, well-fertilized fields may be considerably larger than the amounts removed from poorer fields (Kohnke and Hickok, 1943), but the percentage losses may be more damaging to the poor fields. Analyses for total nutrients, phosphorus and potassium, for example, also may be misleading, because large portions of these elements are usually in unavailable or slowly available forms. The time at which the loss occurs is also significant. Relatively slight erosion may result in a temporary lowering of the amount of available nutrients which is out of proportion to the amount of soil lost (Rogers, 1941). If this occurs during a critical stage in plant development, growth may be retarded. Papers appearing during the 1940’s demonstrated the complexities involved in the study of nutrient losses through runoff. Hays et al. (1948) reported that 90 per cent of the annual soil loss from a Fayette silt loam occurred during two intense rains shortly after oats were seeded. Rogers (1942) demonstrated that the rate of calcium removal may be fairly constant with repeated rains, but that the rate of phosphorus removal changes with the number of uniform rains occurring after application. Bedell et al. (1946) demonstrated that the amount of nitrogen removed was more closely associated with total runoff than with total solids lost. Other workers demonstrated an inverse relation between rate of erosion and concentration of nutrients lost (Massey et al., 1953; Stoltenberg and White, 1953). More recent attempts to evaluate runoff data have centered around developing mathematical equations which can describe runoff losses. Massey and Jackson (1952) calculated regression equations for the enrichment ratios of organic matter, nitrogen, phosphorus, and potassium from an Almena silt loam, Fayette silt loam, and Miami loam. The equations, based on the data from all plots for three years are as follows:

Yo, = 0.147

+ 0.108~+ 0.1252 +

Y, = 0.193 + 0 . 2 0 8 ~+ 0.1372 Y, = 0.319 0.250~+ 0.0982 Y, = 0.792 + 0 . 5 2 3 ~+ 0.2652 Where:

Y x 2

= log enrichment ratio

= - log tons of solid per acre-inch runoff = - log tons total solids lost per acre

These equations indicate that the nutrients are removed selectively in the following order: organic matter < organic and ammoniacal nitrogen < available phosphorus < exchangeable potassium. The average

NUTRIENT LOSSES BY EROSION

315

values for the enrichment ratios are: organic matter, 2.1; nitrogen, 2.7; available phosphorus, 3.4; and exchangeable potassium 19.3. The equations were tested with nitrogen loss data from a Fayette silt loam. A good comparison was obtained between the actual and calculated amounts of nitrogen removed. Gard et al. (1959) used the method of fitting constants to study runoff from pastures on Grantsburg silt loam. The greatest runoff occurred from January to March, and as the season progressed, more intense storms were required to cause equal runoff. Runoff generally increased with increasing antecedent soil moisture. Other equations have been developed that refine the general runoff equation and more accurately predict runoff losses (Cook, 1946; Moldenhauer and Wischmeier, 1960; Wischmeier, 1959, 1960; Wischmeier and Smith, 1958). XI. Summary

Many data have been accumulated in the field of runoff chemistry during the past forty years. This paper reviewed the more pertinent data relating to water erosion losses of organic matter and plant nutrients from cultivated soils. Several conclusions can be drawn from the foregoing review. Significant losses of organic matter, with the concomitant removal of nitrogen and phosphorus, do occur. Rather large amounts of potassium are removed, but only a small percentage of this is in an exchangeable or plant-available form. Available data indicate that calcium and magnesium losses are of minor importance. Sulfur loss data are meager, but there is an indication that relatively significant amounts of sulfur may be carried in solution by runoff water. Data reporting losses of soluble salts of applied nutrients are insufficient for evaluation of the true economic importance of these losses under modern practices of high-level fertilization. In this regard, we should know the effects of source of fertilizer, rates, placements, ratios, and time of application relative to the occurrence of runoff. In addition, there is a need for further fundamental studies of the physicochemical properties of soils in order fully to understand and to evaluate the true nature of erosion damage. REFERENCES Bedell, G . D., Kohnke, H., and Hickok, R. B. 1946. Soil Sci. SOC. Am. Proc. 11, 522-526. Broadbent, F. E., and Chapman, H. D. 1949. Soil Sci. SOC. Am. Proc. 14, 261-269. Bryant, J. C., and Slater, C. S. 1948. Zowa State Coll. J . Sci. 22, 269-312. Buckman, H. O., and Brady, N. C. 1960. “The Nature and Properties of Soils,” 6th ed. Macmillen, New York.

316

HAROLD L. BARROWS AND VICTOR J. KILMER

Cook, H. L. 1946. Trans. Am. Geophys. Union 27, 726-747. Daniel, H. A,, Elwell, H. M., and Harper, H. J. 1938. Soil Sci. SOC. Am. Proc. 3, 230-233. Duley, F. L. 1926. Soil Sci. 21, 401-409. Duley, F. L., and Miller, M. F. 1923. Missouri Res. Bull. 63. Ellison, W. D. 1950. Land (Friends of the Land) 9, 487-491. Ensminger, L. E. 1952. Soil Sci. SOC. Am. Proc. 16, 338-342. Ensminger, L. E., and Cope, J. T. 1947. J. Am. SOC. Agron. 39, 1-11. Free, G . R. 1946. J. Am. SOC. Agron. 38, 207-217. Free, G. R. 1956. Soil Sci. SOC. Am. Proc. 20, 427-429. Card, L. E., Jacob, W. C., and Van Doren, C. A. 1959. Soil Sci. SOC. Am. Proc. 23, 388-391. Harper, H. J. 1958. Soil Sci. SOC. Am. Proc. 22, 358-366. Hays, 0. E., Bay, C. E., and Hull, H. H. 1948. J . Am. SOC. Agron. 40, 1061-1069. Jackson, M. L. 1958. “Soil Chemical Analysis.” Prentice Hall, Englewood Cliffs, New Jersey. Kilmer, V. J., Hays, 0. E., and Muckenhim, R. J. 1944. J. Am. SOC. Agron. 36, 249-263. Knoblauch, H. C., Kolodny, L., and Brill, G. D. 1942. Soil Sci. 53, 369-378. Kohnke, H. 1941. Soil Sci. SOC. Am. Proc. 6, 492-500. Kohnke, H., and Hickok, R. B. 1943. Soil Sci. SOC. Am. Proc. 8, 444-447. Lamb, J., Jr., Carleton, E. A., and Free, G. R. 1950. Soil Sci. 70, 385-392. Martin, J. P. 1941. Soil Sci. 52, 435-443. Massey, H. F., and Jackson, M. L. 1952. Soil Sci. SOC. Am. Proc. 16, 353-356. Massey, H. F., Jackson, M. L., and Hays, 0. E. 1953. Agron. 1. 45, 543-547. Miller, M. F., and Krusekopf, H. H. 1932. Missouri Res. BuU. 177. Moldenhauer, W. C., and Wischmeier, W. H. 1960. Soil Sci. SOC. Am. Proc. 24, 409-413. Neal, 0. R. 1939. Soil Sci. SOC. Am. Proc. 4, 420-425. Neal, 0. R. 1943. Am. Potato 1. 20, 57-64. Neal, 0. R. 1944. J. Am. SOC. Agron. 36, 601-607. Rogers, H. T. 1941. Soil Sci. SOC. Am. Proc. 6, 263-271. Rogers, H. T. 1942. Soil Sci. SOC. Am. Proc. 7, 69-76. Sampson, A. W., and Weyl, L. H. 1918. U.S. Dept. Agr. BuU. 875. Scarseth, G. D., and Chandler, W. V. 1938. J. Am. SOC. Agron. 30, 361-374. Slater, C. S. 1942. J. Agr. Res. 65, 209-219. Slater, C. S., and Carleton, E. A. 1938. Soil Sd.SOC. Am. Proc. 3, 123-128. Smith, D. D., and Wischmeier, W. H. 1982. Aduan. Agron. 14, 109-148. Stoltenberg, N. L., and White, J. L. 1953. Soil Sci. SOC. Am. Proc. 17, 406-410. Volk, G. W. 1945. 1. Am. SOC. Agron. 37, 330-340. Wischmeier, W. H. 1959. Soil Sci. SOC. Am. Proc. 23, 246-249. Wischmeier, W. H. 1960. Soil Sci. SOC. Am. Proc. 24, 332-326. Wischmeier, W. H., and Smith, D, D. 1958. Trans. Am. Geophys. Union 39, 285291. Wooley, J. C. 1943. Agr. Eng. 24, 377-379.

CREEPING ALFALFAS

D. H.

Heinrichs

Canada Department of Agriculture, Swift Current, Saskatchewan, Canada

I. Introduction ................................................ 11. Types of Root Systems in Alfalfa .............................. A. Rhizomatous Plants ...................................... B. Creeping-Rooted Plants ................................... 111. Physiological Considerations ................................... IV. Breeding for the Creeping-Root Character ....................... V. Genetics of the Creeping-Root Character ....................... VI. Association of the Creeping-Root Character with Other Plant Characters VII. Agricultural Performance of Spreading Alfalfas ................... VIII. The Future of Spreading Alfalfas ............................... References ..................................................

Page 317 319 321 322 325 327 331 332 333 335 336

I. Introduction

The existence of creeping alfalfas has been known at least as far back as the turn of the century. During his trips to Russia and Siberia as Agricultural Explorer for the United States Department of Agriculture (1897-1898, 1906, 1908-1909), N. E. Hansen observed creeping plants in varieties of wild growing, yellow-flowered alfalfa ( Medicago falcata L. ) at several locations. In his account, Hansen (1909) reports that Klingen, a Russian agronomist, found as many as one hundred branches on one plant of M . falcata. He states, “the plants of the variety grow upright on the steppe but if pastured closely, they creep and only the ends are erect.” Hansen mentions that the native M . falcata strains in the Volga region of eastern European Russia endure pasturing, whereas introduced strains of Medicago sativa L. do not. In a later publication (Hansen, 1927), he says this about M . falcata he observed in Orenburg province: “Some people are interested in this variety because of its habit of sprouting from the roots at some distance from the original crown.” Oakley and Garver ( 1913, 1917), Oliver ( 1913), Garver (1922), and Southworth (1921) studied various root systems of alfalfa. The general conclusions arrived at by these workers are perhaps best summarized by Southworth, who states: “The hardiness of alfalfa depends very largely 317

318

D. H. HEINRICHS

on its root system. Plants possessing a branched-root system are much better able to withstand heaving than those having only a single taproot, no matter how great its length may be. Those plants which have the power to produce rooting underground stems are able to renovate themselves, and after the death of the main rootstock are capable of keeping up a separate existence quite independent of the parent rootstock. When alfalfa has the habit of spreading by means of root proliferation, we have a form of spreading and multiplying in a vegetative manner which promises to give the plant greater powers of resistance to cold and also greater powers of recuperation from injury than is possessed by even true rhizomes, and we venture to hope that these properties will render it possible to grow good crops in adverse climatic conditions under which it would be quite impossible to raise common alfalfa.” This statement suggests many possibilities of expansion for the alfalfa crop and has many agricultural implications. Small samples of M . falcata were distributed quite widely to farmers during the early 1920’s from Brookings, South Dakota, where Dr. N. E. Hansen was stationed. Although a few plantings from the original seed distribution were successful, very little of agricultural significance resulted from them, probably because of very low seed yields or lack of follow-up by plant breeders and agronomists. Nevertheless, the foundation for extending alfalfa use into drought-ridden areas had been laid, and a number of plant breeders, as well as farmers, retained an interest in the yellow-flowered alfalfas. After the widespread drought of the 1930’s in the Northern Great Plains region of North America, it was noted that certain M . falcatn plantings had come through the dry period with very little reduction in stand. Four plantings were known in Canada and an undetermined number in the United States. A very successful planting was that of Claude Foster at Coal Springs, South Dakota. The alfalfa in this planting exhibited the creeping-root character to a considerable degree. Very little use under cultivation has been made of the native alfalfas of Russian and Siberian origin within the U.S.S.R. itself. This can be gathered from a statement by Komarov (1945): “Nearly all the wild lucernes within the limits of the U.S.S.R. have up to the present been scarcely tried in cultivation. Their nutritive qualities and the drought resistance of some of the species led us to believe that upon their introduction into cultivation, these species will play a not insignificant role in the problem of reclamation of the desert areas and wastelands of the U.S.S.R.” That the creeping-root character was manifest in some of these species is indicated by certain root descriptions given in the book. Renewed interest in creeping types of alfalfa paralleled the interest

CREEPING ALFALFAS

319

in regrassing and soil conservation in North America ( Aamodt, 1952). It was the apparent need for a persistent legume that started a number of breeding programs in the United States and Canada which undertook to develop more drought-resistant and winter-hardy varieties. Alfalfa breeders at several research institutions began to hybridize plants of M . falcuta and M.sutiva extensively in their breeding programs and discovered spreading plants within the F1 and later generations. Four varieties possessing the spreading habit to a certain degree have been developed as a result of breeding: RHIZOMA released in 1948 (Nilan, 1951), RAMBLER released in 1955 ( Heinrichs and Bolton, 1958), TETON released in 1958 ( Adams and Semeniuk, 1958), and NOMAD released in 1951 (Burlingham & Sons). The breeding program at the Experimental Farm, Swift Current, Saskatchewan, started by Dr. S. E. Clarke (Heinrichs, 1954) did more to stimulate intercst in development of creeping-rooted types of alfalfa than any other recent program. Breeding stock of creeping alfalfa was distributed freely to research institutions in many parts of the world and is being widely used in breeding programs aiming to develop creepingrooted strains. Until very recently, creeping alfalfas were considered to be useful mainly for pastures ( Aamodt, 1952; Heinrichs, 1954; Graumann, 1955, 1958). However, this does not mean that the creeping-root character, if present in high-yielding, quick-recovering hay types, might not add to their usefulness. In recent breeding studies involving creepers of northern origin and high-yielding southern varieties, it has been shown that the creeping-root character is not very closely linked with winter dormancy (Dudley and Hanson, 1961; Edye et al., 1961; Daday, 1962). The character therefore promises to have a much wider application in broadening the usefulness of alfalfa than had been previously believed to be possible. II. Types of Root Systems in Alfalfa

Most of the cultivated alfalfa varieties have a tap-root system. This type of root system is characteristic of varieties originating from M . sativa L. A tap-rooted plant has a rather narrow protruding crown, the tap root penetrating vertically into the ground with branch roots arising at intervals from it. Such plants are unable to spread sideways except to a limited degree by crown expansion as the plant ages. The branch-rooted alfalfa varieties and those with a proportion of both branch and tap roots are of more recent origin. The varieties used in North America differ widely in the extent to which their primary roots are branching (Smith, 1951). These types have variegated flowers and

320

D. H. HEINRICHS

are considered to have originated from natural or artificial crosses between M . satiua and M . falcata; they are often designated as M . media Pers. The majority of such varieties have been developed in northern areas where winters are cold or in areas which are cold because of high elevation. The variety LADAK, a typical variegated variety, with a branched-root system, came from a natural population in the province of Ladakh, India. A branch-rooted plant differs from a tap-rooted plant in that a number of main primary roots arise from the crown instead of only one, and the crown is inclined to be wider. Such plants are able to develop adventitious shoots from the roots. This was demonstrated by Smith (1950), who found that adventitious shoots developed on root segments of LADAK and an M . falcata strain, but not on plants of GRIMM, RANGER, COSSACK, and MONTANA COMMON. Murray (1955) found that adventitious shoots developed also on plants of LADAK, but to a lesser degree than on creeping-rooted plants. Branch-rooted plants are able to withstand heaving during unfavorable winters better than tap-rooted plants. This has been found to be the case in the variety NARRAGANSETT, which is described as having a much-branched root system (Odland and Skogley, 1953). Although the branched-root system in alfalfa can be considered to result in a slight tendency for the plant to spread horizontally, the spread is restricted to the above-surface crown. There are two types of root systems in alfalfa which enable the plant to spread horizontally. These are generally referred to as rhizomatous and creeping-rooted. In the first instance, the spread is attributed to lateral expansion of the low-set crown by short horizontal stems, and in the second the spread is brought about by horizontal roots from which shoots may arise at irregular intervals, Graumann (1955,1958) explains the difference between rhizomatous and creeping-rooted type of spread quite adequately. He also makes it clear that any one strain or variety may have in it a variety of crown and root types. The difference between the two types of spread often is not too clearly demarcated or too well understood, and therefore each will be dealt with under a separate heading. Four types of plants representing different degrees of spread and different root systems are illustrated in Fig. 1. The plants shown in this photograph were excavated in October, 1962, from.an old stand of M . falcata growing on the Experimental Farm, Swift Current, Saskatchewan. This strain is one of N. E. Hansen’s introductions from Siberia, probably the strain known as Semipalatinsk. These plants although differing greatly from one another in ability to spread, have one characteristic in common, and that is the low crown which is located mostly below the soil surface. The low crown, perhaps, is the characteristic in M. falcata

CREEPING ALFALFAS

321

which differentiates it from M. sutivu to a greater extent than the roots themselves. Excellent descriptions and photographs of different crown and root systems and their relationships to types of alfalfa have been

FIG. 1. Plants of M . falcata, probably of the strain Semipalatinsk, excavated in September, 1962, in a 25-year-old field of crested wheatgrass at the Experimetal Farm, Swift Current, Saskatchewan. The alfalfa was an experimental planting seeded the same time as the grass. The photograph shows the deep-set crown of all plants and the different root systems, Left to right: tap root, branch root, rhizomatous root, creeping root.

presented by Garver ( 1922). Garver’s descriptions parallel closely what is said about roots in this article, suggesting that in the study of various alfalfa root systems little progress has been made during the last forty years. A. RHIZOMATOUSPLANTS The rhizomatous type of spread in alfalfa has been described in some detail by Oliver ( 1913) and more recently by Graumann (1958). The spreading of this type of plant is assumed to arise from stems of rootlike appearance which are initiated from the original primary root axis. These rhizomes extend laterally for varying distances, eventually root over part of their length, and emerge from the soil as vegetative stems. Dr.

322

D. H. HEINRICHS

G. G. Moe of the University of British Columbia developed the variety which typifies this type of crown and roots (Nilan, 1951). Smith ( 1955) investigated the underground crown branches ( rhizomes) of a number of varieties. Of the varieties tested, RHIZOMA had the most rhizomes and ARIZONA COMMON the least. There seemed to be a direct relationship between the quantity of rhizomes and the amount of hl. falcatu which figured in the parentage of the varieties. Hanson et al. (1960) express the same thought in Agricultural Handbook No. 177. NOMAD is another new variety with extensive rhizome development ( Aamodt, 1952). It has been observed by Heinrichs (1954) that RHIZOMA and NOMAD do not spread to any extent in the arid climate of the prairie region of Canada. He attributed the lack of rhizomatic spread to inadequate soil nioisture conditions at the surface. Hanson et al. (1960) make the observation that RHIZOMA and NOMAD do not show much tendency to spread in many regions of the United States. The writer is of the opinion that rhizomatic plants will tend to spread quite strongly if climatic conditions are humid and the surface soil is moist for prolonged periods, particularly in the fall. Rooting may occur not only from underground stems, but also from prostrate stems lying on the soil surface. If animals graze and trample rhizomatous plants during such moist periods, the plants will tend to spread even more rapidly than when left ungrazed because the stems may grow more prostrate and thus get tramped into the ground where they can root from the nodes. Rhizomatous plants generally expand in crown width, and the top growth from them is very dense per unit area (Fig. 1 ) . RHIZOMA

B. CREEPINGROOTED PLANTS Truly creeping-rooted plants were described and illustrated by Oakley and Garver (1913), Oakley (1917), Southworth ( 1921), and Garver ( 1922). Recent descriptions and illustrations of this type were presented by Heinrichs ( 1954), Murray (1955), and Heinrichs and Bolton ( 1958). Actually, when creeping-rooted plants showed up in breeding populations resulting from intercrosses between M . falcuta and plants of the varieties LADAK and RHIZOMA, it seemed as though a new discovery had been made. This was the impression because certain hybrids were expressing the spreading habit to a much greater extent than any M . falcata parent. This accentuated spread in hybrid forms appeared to be due to complementary gene effects rather than to hybrid vigor (Heinrichs, 1954). A typical creeping root of a hybrid type is shown in Fig. 2 and of growing plants in Fig. 3. The creeping rootstocks are generally found to be 4 to 8 inches below the surface of the ground. They send up

CREEPING ALFALFAS

323

shoots at intervals, each of which is capable of becoming an independent plant. Some plants are dense creepers, others sparse; that is, certain plants produce more shoots per unit area than others. The sparse type of creeper may be the more desirable type for dryland pastures than

FIG. 2. A typical creeping-rooted plant which shows M . satiua as well as M . fakutu characteristics. The creeping-root development is more distinct than is the case in plant 4 in Fig. 1, which tends to display a combination of rhizomatous and creeping-root development. The pedigree of this plant in terms of varieties is: LADAK x (LADAK X M . f&ai%).

the dense type because in a mixed stand the alfalfa and grass plants will tend to form an interspersed association. Also, there will be less tendency for the stand to become too thick during favorable moisture years, with subsequent lower productive capacity in dry years. Murray (1957) studied the ontogeny of adventitious stems on roots

324

D. H. HEINRICHS

of creeping-rooted alfalfa. She found that the horizontal roots are anatomically true root structures resembling the primary roots. They produce adventitious stems on a much greater range of roots than branch roots. Murray states: “At successive intervals of time in the ontogeny of adventitious stems, there are certain changes which occur within the root during two intimately associated phases of activity. One phase is the initiation and development, at the periphery of the root, of a primordial dome. The other is the differentiation of the subjacent secondary phloem parenchyma cells to form a meristematic zone in which vascular

FIG. 3. Early spring growth showing the spread of strongly creeping-rooted plants, 5 years old. The planting in this nursery originally was 6 feet each way. The plots contained 4 plants of a clonal line, Note that the creeping-rooted plants are encroaching on the noncreepers.

tissues differentiate and connect the peripheral primordia with the cambium of the root.” She further states that the inherent difference in creeping-rooted and noncreeping-rooted plants is in some way associated with the ability of the former to initiate stem primordia on all roots, whereas in the latter this ability is restricted to very few roots and perhaps is seldom brought into function. The writer observed, in the spring of 1962, when severe winter injury occurred in spaced populations, that some branch-rooted plants sent up new shoots from the roots located 4 to 6 inches underground when the main crown had been destroyed as a result of winter injury. Thus, it may be that in noncreeping plants, adventitious shoots are developed mainly under stress conditions.

CREEPING ALFALFAS

325

111. Physiologic01 Considerations

It is a fact that most of the spreading types of alfalfa presently being used as forage crops or for breeding purposes originated in northern areas where the winters are long and cold. It is natural, therefore, that most spreading alfalfas have long dormant periods in their growth cycles and slow regrowth characteristics. The spreading character as such, however, is not restricted to northern types. Oliver (1913) reports rhizome development in native varieties found in northern Africa. He states that the principal functions of the large rhizome may be the storage of water to tide the plants over the very hot and dry summers, because the rhizomes of these plants in northern Africa are evidently formed at the close of the growing season, which is during the winter months, but when brought to a colder climate they form during the autumn months. Aamodt (1952) mentions that wild forms of spreading alfalfas were found in Turkey by H. L. Westover in 1931. These alfalfas were growing in desertlike, denuded fields which the natives call “goat pastures.” Rhizomatous types were also found in Spain and Greece growing along roadsides, in fence rows, and in pastures among grasses. These observations suggest that drought resistance and ability to grow under competition and stress may be the common denominator, rather than cold resistance of the wild forms of spreading alfalfas. Plant breeders in subtropical and tropical climates may be well advised to consider using Mediterranean or Turkish wild alfalfa types in their programs aiming to develop persistent pasture types for their climatic conditions rather than those from northern regions. To the knowledge of the writer there is limited information from physiological studies on spreading alfalfas that relate the rhizomatous or creeping-root characteristic with any clearly defined physiological function. Ludwig (1960) used the technique of “growth analysis” to relate the main environmental factors on the early growth of a creepingrooted clonal Iine selected out of Canadian material. He found that number of plants with adventitious shoots was increased by short days, high light intensities, and high day and night temperatures. The number of adventitious shoots per plant was increased by the same environmental factors. Adventitious shoot production appeared to be directly associated with the growth rate of the roots and the top-to-root ratio, and indirectly with the net assimilation rate. Generally, those environmental factors which increased the net assimilation rate or decreased the top-to-root ratio, increased root growth and the production of ad-

326

D. H. HELNRICHS

ventitious stems. Root segments removed from plants in each environmental treatment and placed in sand in the greenhouse all produced adventitious shoots. This indicates that the environment in which the plants had been growing had little effect on the potential of roots to form adventitious shoots. Since Ludwig’s study was done on a creeping-rooted clone of northern origin the root responses to environmental changes may be directly associated with this particular genotype, That creeping-rooted clones of different growth form may respond differently to environmental conditions in the formation of adventitious shoots, is indicated by results from a recent study by Ludwig (personal correspondence). He found that although in the majority of genotypes daylength had very little effect on the production of adventitious shoots, one clone produced many adventitious shoots in long days but scarcely any in short days, and another produced many shoots on the roots in short days, but scarcely any in long days. The initiation of adventitious shoots, Murray (1957) points out, is likely to involve an interrelationship of many factors. These could be physical ( size and number of primordia), physiological ( chemical stimulus for differentiation and growth and environmental responses), and genetic (an inherent characteristic of the plant). There is no doubt that more knowledge of the nongenetic mechanisms which influence adventitious stem formation would be very useful to the plant breeder in understanding hereditary factors and to the agronomist in forecasting where the creeping type might fit into a particular environment. Winter hardiness relationships among a number of varieties, three of which were of the spreading type, were reported by Heinrichs and Bolton (19%)) Heinrichs (1958, 1959)) and Heinrichs et aZ. (1960). It was quite obvious from the results that increased winter hardiness cannot be attributed to the spreading habit of growth alone. The character must be present in an inherently winter-hardy type. Thus, NOMAD, a spreading variety, was much less persistent in the Prairie Provinces of Canada than the nonspreading varieties VERNAL, GRIMM, and LADAK, which are commonly grown in the area. On the other hand, the creepingrooted variety RAMBLER, which possesses adequate winter hardiness, maintained a better stand than any other variety. This situation was also shown by Clark (1960). That creeping-rooted alfalfas may survive better than noncreepers of similar genotype was noted by Heinrichs (1954). He states: “In creeping-rooted plants, when the main crown was dead, numerous shoots appeared from creeping rootstocks at a considerable distance from the crown. It appeared that killing of the main crown actually stimulated development of aerial shoots from creeping root-

CREEPING ALFALFAS

327

stocks.” It can be reasoned, therefore, that heaving damage or severe overgrazing may similarly stimulate plants to spread. In Australia, Morley et al. (1957) investigated Canadian creepingrooted clones, locally adapted clones, and intercrosses between them for reaction to temperature and day length. They found that clones differed greatly in their reaction to these two environmental factors. The creeping-rooted clones from Canada were very low in winter production because of the winter dormancy characteristic. They state that although introduction of alfalfa breeding material from these northern regions is undoubtedly justified, the concomitant winter dormancy must be eliminated before derived strains can be unreservedly recommended for Australian conditions. In a recent physiological study, Daday (1962) showed that although RAMBLER required short days and low temperatures for best development of creeping roots, the F2 and F3 creeping derivatives from RAMBLER crossed with AFRICAN, HAIRY PERUVIAN, and HUNTER RIVER developed creeping roots satisfactorily when days were longer and temperatures higher. When considering the use of spreading alfalfas in any particular region, one dare not leave out consideration of the environment at their point of origin. The potential real usefulness of the creeping character may never be realized unless it is transferred by breeding to strains physiologically adapted to a specific environment. IV. Breeding for the Creeping-Root Character

Oakley and Carver (1913, 1917) reviewed the agronomic possibilities that lie in the development of new alfalfa strains by hybridizing forms of M. falcata and M . satiua. Superior drought resistance and winter hardiness were considered to be the main desirable characters of M . fulcuta, and they pointed to the role this species has already played in the formation of useful varieties as a result of natural crossing. GRIMMis mentioned as an outstanding example of such a variety. They also envisaged that by the use of the proliferating root character, high-yielding strains may be originated that will be especially resistant to severe climatic conditions and actually aggressive on soils of fairly loose texture. Breeding programs using the principle of combining characters of the two species into more useful varieties were almost at a standstill until after the Great Drought of the 1930’s in the Great Plains area of North America. In Canada such a breeding program was undertaken in 1938 at the Experimental Farm, Swift Current, Saskatchewan, and the aims were outlined by Heinrichs ( 1954). He states that the objective was to combine the winter-hardy and persistence qualities of M . fulcata

328

D. H.HEINRICHS

with good seed-producing qualities of common varieties. The breeding procedure followed was based on determination of the combining ability of individual plants. It was the rediscovery of strongly creeping-rooted types in the hybrid populations that led to concerted efforts to develop a creeping-rooted variety adapted to conditions in the Canadian prairies. Principles of general and specific combining ability were used to evaluate creeping-rooted selections for transmission of this character along with other useful agronomic characters to their progenies. The breeding objective was finally reached in 1955 with the release of the variety RAMBLER ( Heinrichs and Bolton, 1958). Interest of alfalfa breeders of North America in creeping-rooted types of alfalfa was greatly stimulated at the Twelfth Alfalfa Improvement Conference held in Lethbridge, Alberta. At this conference a paper on the subject was presented by Heinrichs (1950) and field demonstrations of spreading plants were shown. Breeding stock of creeping-rooted alfalfa was sent out from the Experimental Farm, Swift Current, Saskatchewan, to a great many institutions in all parts of the world. In addition, the strains of M . falcata and M . falcata X M . sativa origin available at the South Dakota Agricultural Experiment Station, Brookings, South Dakota, were a good source of the creeping character. These could be traced back to N. E. Hansen’s alfalfa introductions from Europe and Asia. At the present time substantial breeding projects, aiming to develop suitable creeping-rooted strains for specific environments, are underway at the following institutions: United States of America Crops Research Division, U.S.D.A. and North Carolina State College, Raleigh, North Carolina. John W. Dudley. Department of Plant Breeding, Cornell University, Ithaca, New York. C. C. Lowe and R. E. Anderson. Nevada Agricultural Experiment Station, Reno, Nevada. H. L. Camahan. Crops Research Division, U.S.D.A. and University of Nebraska, Lincoln, Nebraska. W. R. Kehr. South Dakota Agricultural Experiment Station, Brookings, South Dakota. M. D. Rumbaugh. Kansas State University, Manhattan, Kansas. E. L. Sorenson. Canada Experimental Farm, C.D.A., Swift Current, Saskatchewan. D. H. Heinrichs. Research Station, C.D.A., Lethbridge, Alberta. M. R. Hanna. Genetics and Plant Breeding Research Institute, C.D.A., Ottawa, Ontario. H. A. McLennan. Chile Oficina Estudios Especiales, Casilla 2-P, Santiago. Avendano T. Raul and Ruco E. Osvaldo.

CREEPING ALFALFAS

329

Australia Division of Plant Industry, C.S.I.R.O., Canberra. H. Daday and F. H. W. Morley. Division of Tropical Pastures, C.S.I.R.O., Brisbane. E. H. Hutton and L. A. Edye New Zealand Crop Research Division, D.S.I.R., Christchurch. T. P. Palmer. Great Britain Welsh Plant Breeding Station, Aberystwyth, Wales. W. Ellis Davis. Denmark Forsogsgarden Hojbakkegard. Albertslund pr. Taastrup. H. J. Moller Nielsen.

Methods of breeding for the creeping-root character need to differ little from those used for other quantitatively inherited characters. Heinrichs (1954) states that determination of combining ability of individual plants appears to be a good method for evaluating the capacity to transmit creeping-rootedness to progenies. The best time to score for degree of creeping-root development was in the spring, one year after planting. Creeping-rootedness, however, did not express itself in all plants until two or even three years after planting at Swift Current, Saskatchewan. In a later study of similar material, Morley and Heinrichs (19so) found that general combining ability was more important than specific combining ability, and therefore they concluded that mass selection would be very efficient in breeding for creeping-root. Jones and Hanson (1959) crossed the varieties AFRICAN, BUFFALO, CALNEFDE, and nine North Carolina clones on three Canadian creepers, and in the F2 recovered creeping-rooted plants from crosses with each of the varieties. The North Carolina selections transmitted different amounts of creep to their progenies, indicating the association of complementary factors with the expression of creeping roots. They also observed that the inherent vigor of plants has a bearing on expression of the creeping growth habit. Australian alfalfa breeders are concerned with combining the creeping-root character from Canadian creepers with summer and winter growth habits of subtropical and tropical alfalfas. The writer spent 10 months in Australia during 1961 and was quite impressed with the progress being made at the Fs generation level. The Canadian creepers show creep development in Australia but lack growth vigor, and in Queensland they produce practically nothing. In the F3 generation extremely vigorous creepers were found, as Edye et uZ. (1961) and Daday ( 1962) have shown (Fig. 4). Edye et al. concluded that mass selection for creeping-rootedness in the third and later generations should be an efficient method of breeding. Daday found that the varieties HUNTER RIVER, HAIRY PERUVIAN, and AFRICAN in crosses with RAMBLER produced only 5 per cent creepers in F1 but an average of 41 per cent in the F3 generation resulting from intercrosses of F2 creepers. The range

330

D. H. HEINRICHS

in creep between F3 lines was 0 to 81 per cent (Fig. 4).This compares quite favorably with plantings observed by Heinrichs (1954), who used the northern varieties LADAK and RHIZOMA in his breeding program. In a breeding program initiated at the U S . Regional Pasture Laboratory, University Park, Pennsylvania, Carnahan ( 1959) investigated

FIG. 4. An alfslfa breeding nursery at C.S.I.R.O., Canberra, Australia. The population is second and third generation material from crosses between Canadian creepers and African, DuPuits, and Hunter River selections. The stakes mark the creeping-rooted plants. The photograph was taken near the end of August (end of winter). Note the different degrees of winter dormancy.

methods of evaluating alfalfa for the creeping-root character. He found that creeping root expressed itself with the same expectation in nurseries oversown with timothy as in cultivated nurseries. The work involved in the oversown nurseries was less and, considering that the alfalfa will likely be grown in association with grass, such nurseries may be preferable for evaluating the alfalfa plants for creep development. When considering the breeding programs conducted at widely separated locations, there seems to be unanimity of opinion that fairly rapid progress is possible by following accepted breeding practices based on progeny performance. It also appears obvious that the creeping character must be transferred to varieties adapted to a particular climate before it can display its usefulness in improving persistence.

CREEPING ALFALFAS

331

V. Genetics of the Creeping-Root Character

The variation in type of spreading root systems and the overlapping that appears to occur between one type and another (Fig. 1 ) would tend to lead one to assume that the inheritance of this character might be complex. Heinrichs (1954) proceeded to breed for creeping root on the assumption that the character was a quantitative one. The data he obtained from combining ability tests of creeping-rooted selections seem to justify the assumption, and he concluded that the character appears to be a quantitative one with complementary factors playing a part. Morley and Heinrichs (1960) statistically examined the data obtained on the creeping-root character from a large population of alfalfa grown in experimental blocks at Swift Current, Saskatchewan. Their analysis showed that the estimate of Cov. (full sibs)-Cov. (half sibs) was 0.062, and of Cov. (half sibs) was 0.091, indicating that the genotypic variation was predominantly additive. In a study on estimates of variance in root proliferation in alfalfa, Adams (1959) found that the genotypic variance for extent of creep was mostly nonadditive and not highly heritable. In his analysis, however, Adams did not include the noncreeping fraction of the progenies, and, hence, the results give no information on the inheritance of creeping-root development versus no creep development. Further light on this apparent discrepancy was shed by a study by Heinrichs and Morley (1962). They found that when the data from the entire population were analyzed the genotypic variance was predominantly additive, but when only the creeping fraction was examined the genotypic variance was largely nonadditive. The conclusion drawn from this finding was that the creeping-root character versus tap- or branch-root character is quite highly heritable but the degree of spread is influenced strongly by nonadditive genotypic and environmental factors and perhaps by genes determining vigor of growth, as Jones and Hanson (1959) suggested. Daday (1962), working with a population of a somewhat different background, found considerably more of the variance to be due to nonadditive genetic effects than did Morley and Heinrichs ( 1960) and Heinrichs and Morley (1962). The heritabilities reported for the creeping-root character are quite good, varying from 20 per cent (Morley and Heinrichs, 1960) to 31 per cent (Heinrichs and Morley, 1962), calculated on the individual plot basis. Low heritability, reported by Adams (lQSQ),was substantiated by Heinrichs and Morley (l%2), but applied o d y to degree or extent of spread. Based on the fact that the data from a simple two-way classification into creeping and noncreeping plants gave approximately the

332

D. H. HEINFUCHS

same heritability estimates as those from five scores (1, 2, and 3 representing degrees of creep, and 4 and 5 noncreeping plants with decreasing breadth of crown), Heinrichs and Morley (1962) suggested that laborious scoring systems may not be necessary in assessing plants for creep. At the Eighteenth Alfalfa Improvement Conference at Davis, California, Heinrichs (1962) reported that plants which were scored strongly creeping produced very few more creeping-rooted progenies than those scored only slightly creeping. These results are somewhat at variance with those of Heinrichs ( 1954) and Jones and Hanson (1958), who found the phenotypic correlation between width of plant and per cent plants creeping-rooted to be 0.59 and 0.183, respectively. The genetic picture to date appears to suggest that the creeping-root character is complexly inherited, many modifying genes having an effect on its expression. The fact that practically no 100 per cent creepingrooted lines have been obtained from intercrosses between creepers selected for the character through several generations lends support to this hypothesis. However, the high proportion of additive genotypic variance occurring relative to nonadditive for the character is encouraging to the plant breeder who is concerned with developing creepingrooted strains.

+

+

VI. Association of the Creeping-Root Character with Other Plant Characters

At Swift Current, Saskatchewan, Heinrichs (1954) found the correlation coefficient between winterkilling and degree of creeping-rootedness to be -0.60 in two replicated progeny tests which suffered severe winterkilling during the winter of 1948-1949. In a 2-acre breeding nursery that same year, the correlation coefficient between the percentage of creeping-rooted progenies and vigor was 0.51, significant at the 1 per cent level. This indicated that the creeping-rooted plants suffered less winter injury than noncreepers. In a special study on the association of winter injury with creeping-rootedness, Heinrichs and Morley ( 1960) 0.38. reported the genotypic correlation between the two to be Three characters having a bearing on seed production were checked for association with creeping-rootedness by Heinrichs ( 1954) in eight families. He found no correlation between creeping-rootedness and degree of pod coiling in any family; a significant positive correlation between creeping-rootedness and seed set in one family, a negative one in two families, and none in the remaining five; a significant positive correlation between creeping-rootedness and shatterability in one family, a negative one in two, and none in the remaining five. These results

+

+

CREEPING ALFALFAS

333

indicate that pod type, seed set, and seed shatterability were largely independently inherited from creeping-root. Dudley and Hanson (1961),in a study on interrelationships among characters in F2 progenies from crosses between creeping-rooted and hay type alfalfa clones, found that the interaction of creeper x hay type parents was significant only for the characters of plant width, leaf width, and leaf length. The lack of a significant interaction of the other eight characters indicates that for the majority of the characters in these crosses, general combining ability is of more importance at this stage of the breeding program than specific combining ability. VII. Agricultural Performance of Spreading Alfalfas

There has been a lot of speculation about the usefulness of spreading alfalfas for pasture purposes. Much of what has been said in public or written in popular articles during the last decade has been based on what was known about the persistence ability of yellow-flowered alfalfas in their native habitats. Without much experimental evidence to back them up, agronomists, soil conservationists, and plant breeders reasoned that a low, spreading crown hidden from the ravages of the weather and the grazing animal should persist better than protruding crowned taprooted alfalfas. Recent observations and experimental evidence is confirming these earlier expectations. Although very little has been published on the performance of NOMAD as a grazing plant, the originators (Burlingham Seed Company of Forest Grove, Oregon) claim that it persists well under heavy grazing in certain localities in Oregon and adjacent States. E. R. Jackman, farm crops and range specialist in Oregon, for several decades has been stressing the need for a range and pasture legume in the United States. Since NOMAD was released, he observed it in many field plantings and publicized its success as a pasture legume in the press and farm magazines. In recent years the spreading alfalfas NOMAD, RHIZOMA, and RAMBLER have been evaluated in relation to other varieties at a number of institutions in the United States and Canada. Published data on performance are scarce, but indications are that these creeping varieties are rather low yielding outside the region in which they were developed. Such results should not be misconstrued to mean that creeping varieties are not adapted in certain climatic regions because of the creeping character, but rather because these varieties, as such, are not adapted in these regions whether or not they have creeping rootstocks. This point is well illustrated by results from tests in Australia where

334

D. H. HEINRICHS

has been compared in relation to locally used varieties (Rogers, 1961; Daday et aZ., 1961). RAMBLERdisplayed extremely low vigor at all locations and produced very low forage yields. Australian scientists, Morley .et al. (1957) and Rogers (196l), have recognized that the creeping-root character from Canadian strains will need to be transferred to locally adapted varieties before it will be of much use under Australian conditions. In New Zealand, Palmer (1959) found that RHIZOMA performed rather poorly compared to other varieties, mainly because of its poor winter growth habit. The variety RAMBLER, developed specifically for use on dryland in the Canadian prairies, persists better than other varieties and even yields better in this region, but not in eastern Canada and in western British Columbia, where the rainfall is much greater and soil conditions are more acidic. On the other hand, NOMAD and RHIZOMA, developed in a moister climate, performed relatively better in eastern Canada and certain areas of British Columbia than RAMBLER. Data on these Canadian results have been presented by Heinrichs and Bolton (1958). Clark (1960) reviewed the results from grazing trials with FWMBLER and other varieties conducted at Swift Current, Saskatchewan, where RAMBLER originated. In all trials RAMBLER persisted better than other varieties except the M.falcatu strains. Typical results from a test grazed by sheep from 1959 to 1962 and shown in Table I. It is significant that RAMBLER

RAMBLER

TABLE I Relative Stand of a Number of Alfalfa Varieties during a Series of Very Dry Years at Swift Current, Saskatchewan, after Heavy Grazing by Sheep Adequacy of stand Oct. 6, 1958

"/o Basal ground cover ( point-quadrat method)

(%I

May 1960

May 1961

Aug. 1962

RHIZOMA

94 96 97 97

NOMAD

97

RAMBLER

93 83 77 94

12.9 11.9 2.5 11.6 1.1 26.5 28.2 22.0 26.3

7.6 6.3 2.2 10.1 1.0 16.7 16.8 13.0 15.6

5.9 0.7 2.2 7.0 1.1 10.8 14.8 11.5 10.9

Variety LADAK

CRIMM VERNAL

S.C. Syn. 3513 Semipalatinsk Siberian

persisted as well as Semipalatinsk and Siberian, two strains which are representative of the hardy side of its parentage. Both of these strains set very little seed and shatter the little that does set. This precludes the commercial use of these strains. The writer believes that the creeping-

CREEPING ALFALFAS

335

root character inherent in 65 per cent of plants of RAMBLER is partly responsible for its ability to persist, because it cannot be considered to be quite equal to either Semipalatinsk or Siberian in winter hardiness and drought resistance. The Swift Current strain S.C. Syn. 3513 is more strongly creeping-rooted than RAMBLER and excelled both the M. fukutu strains in persistence. The writer has observed that the soil texture and climate have a bearing on how fast a plant will spread by creeping roots, but there is no evidence to show that a truly creeping-rooted plant will not spread in certain climates or on certain soil types. Spreading of plants will occur in pots in the greenhouse as well as in the fields on sandy or clay soils. Rhizomatous spreading plants, on the other hand, will frequently not creep at all under very dry climatic conditions. VIII. The Future of Spreading Alfalfas

Investigational work indicates that spreading alfalfas have a bright future in making the alfalfa crop even more useful than it is today. Although the idea that the creeping habit offers the greatest possibilities in improving the attributes of the crop for pasture ( Aamodt, 1952; Heinrichs, 1954; Graumann, 1958) is sound, there is no doubt that the character can also add a great deal of usefulness to the hay crop. The selfrejuvenation ability of a creeping type after winter injury, heaving, or drought damage should be as important in a hay crop as in a pasture crop. The breeding work done in Canada, the United States, and Australia clearly indicates that the creeping-root character can be transferred from low yielding, poor seed-setting types of plants to high-producing types. The breeding programs which have progressed beyond the second generation indicate that the creeping character often becomes more pronounced in these later generations than was the case in the creepingrooted parental type. It appears that for maximum expression the character must become genetically associated with the growth vigor inherent in alfalfa varieties adapted to any particular environment. In spite of the wide possibilities that seem to be ahead for the creeping character in alfalfa, it should be stressed that the plant breeder perhaps should concentrate his efforts on developing a pasture type because this is where the greatest need lies. It is the vast dry ranges of so many countries that are crying out for a legume that will increase and stabilize production. Persistence as a requirement in the alfalfa variety should be stressed and restressed. It is so easy to stray away from this requirement because most plant breeders and agronomists use yield as

336

D. H. HEINRICHS

the ultimate yardstick rather than persistence. E. R. Jackman, Soil Conservationist in Oregon, in a recent letter to the writer, said that research persons tend to have a “yield complex and somehow cannot see that persistence comes first. Many will agree that this is generally true. However, it is not necessarily the plant breeder‘s fault, but perhaps just as much the fault of the agronomist, rangeman, or farmer, who often are slow to accept and get new varieties into use where they should be used. The great need in the future will be the proper evaluation of creepingrooted and rhizomatous alfalfas in relation to varieties of similar adaptivity without the spreading habit. This is the area in which much more research needs to be done before a true assessment of the value of the spreading habit of growth will be forthcoming. Research work on the various types of creepers on an individual plant basis requires investigation before the plant breeder will know exactly the type for which he is selecting. The only detailed investigation of this nature was done by Murray (1955). She found that the rate of spreading varied greatly between creeping-rooted clones. There are sparse creepers and dense creepers and all gradations between. Which type to select for will depend on where and how the alfalfa is to be used. Certainly the agronomic possibilities are great indeed. The term creeping alfalfa would be best not used as a synonym for grazing alfalfa because the two have distinct and different meanings. Nevertheless, it seems logical to suppose that the creeping-root character will make a good grazing alfalfa even better. REFERENCES Aamodt, 0. S. 1952. What’s N e w In Crops Soils 5. Adams, M. W. 1959. J. Genet. 56, 395-400. Adams, M. W., and Semeniuk, G. 1958. S . Dakota State Coll. Agr. Expt. Sta. Bull. 469. Burlingham, E. F. & Sons. Forest Grove, Oregon. Nomad Alfalfa. Camahan, H. L. 1959. Agron. J. 51, 625626. Clark, K. W. 1960. Rept. 17th Alfalfa Imp. Conf. pp. 99-106. Daday, H. 1962. Australian J . Agr. Res. 13, 813-820. Daday, H., Mottershead, B. E., and Rogers, V. E. 1961. Australian J . Exptl. Agr. Animal Husbandy 1, 67-72. Dudley, J. W., and Hanson, C. H. 1961. Crop Sci. 1, 59-63. Edye, L. A., Haydock, K. P., and Saunders, A. M. 1961. Conf. Cereal Pasture Plant Breeders C.S.I.R.O. Canberra Australb 1, 31-1-31-9. Gamer, S. 1922. U.S. Dept. Agr. Bull. 1087. Graumann, H. 0. 1955. Soil Conseru. 21, 103-104. Graumann, H. 0. 1958. What’s New Crops Soils 10. Hansen, N. E. 1909. U.S. Dept. Agr. Bull. 150. Hansen, N. E. 1927. S . Dakota State Coll. Agr. Erpt. Sta. Bull. 224.

CREEPING ALFALFAS

337

Hanson, C. H., Garrison, C. S., and Graumann, H. 0. 1960. US. Dept. Agr. Agr. Handbook 177. Heinrichs, D. H. 1950. Rept. 12th Alfalfa Imp. Conf. pp. 59-61. Heinrichs, D. H. 1954. Can. J. Agr. Scl. 34,269-280. Heinrichs, D.H. 1958. Agr. Inst. Reu. May-June. Heinrichs, D. H. 1959. Can. J. Plant Sci. 39, 384-394. Heinrichs, D. H. 1962. Rep. 18th Alfalfa Imprwement Conf., Daois, Calif., pp.

45-50. Heinrichs, D. Heinrichs, D. Heinrichs, D. Heinrichs, D.

H., and Bolton, J. L. 1958. Can. Dept. Agr. Publ. 1030. H., and Morley, F. H. W. 1960. Can. 1. Plant Sci. 40,487-489. H., and Morley, F. H. W. 1962. Can. J. Genet. Cytol. 4, 79-89. H., Troelsen, J. E., and Clark, K. W. 1960. Can. J . Plant Sci. 40,

638-644. Jones, A., and Hanson, C. H. 1959. Agron. J. 51, 710-718. Komarov, V. L. 1945. Flora URSR 11, 129-176. Ludwig, L. J. 1960. M.S.Dissertation, University of New Zealand. Morley, F. H. W., and Heinrichs, D. H. 1960. Can. 1. Plant Sci. 40,424-433. Morley, F. H. W., Daday, H., and Peak, J. W. 1957. Australian J. Agr. Res. 8,

635-651. Murray, B. E. 1955. Ph.D. Dissertation, Cornell Univ., Ithaca, New York. Murray, B. E. 1957. Can. J. Botany 35, 463-475. Nilan, R. A. 1951. Sct. Agr. 31, 123-127. Oakley, R. A,,and Carver, S. 1913. US. Dept. Agr. Bur. Plant Ind. Circ. 115. Oakley, R. A., and Garver, S. 1917. U.S. Dept. Agr. Bull. 428. Odland, T. E., and Skogley, C. R. 1953. Agron. J. 45, 243-245. Oliver, G. W. 1913. U.S. Dept. Agr. Bur. Plant Ind. Bull. 258. Palmer, T. P. 1959. New Zealand J . Agr. Res. 2, 1195-1202. Rogers, V. E. 1961. Australtan J. Exptl. Agr. Animal Husbandry 1, 80-66. Smith, D. 1950. Agron. J. 42, 398-401. Smith, D. 1951. Agron. J . 43, 573-575. Smith, D. 1955. Agron. 1. 47, 588-589. Southworth, W. 1921. Scl. Agr. 1, 5-9.

This Page Intentionally Left Blank

SILICA IN SOILS' J. A. McKeague and M. G. Cline Canada Department of Agriculture, Ottawa, Canada, and Cornell University, Ithaca, New York

I. 11.

III.

IV. V.

Page Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339 Silica in Solid Forms ......................................... 340 A. Forms and Amounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341 B. Stability of Solid Forms . . . . . . . ....................... 345 Silica in Solution ............................................ 353 A. Chemistry of Silica ...................................... 353 B. Dissolution of Solid Forms of Silica and Silicates . . . . . . . . . . . . . . 360 C. Concentrations and Forms of Dissolved Silica in Natural Waters 368 and in Soil Solutions ..................................... D. Interaction of Dissolved Silica with Other Components of Soil . , . 373 Deposition of Silica in Soils . . ............................. 377 A. By Inorganic Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377 B. By Organisms . . . . .................................... 382 Silica in Relation to Kin of Soils ............................. 384 References ..............................................

I. Introduction

Silica in free and combined forms is a dominant component of the solid material of many soils, and dissolved silica is commonly a major solute of soil solutions. Breakdown of primary silicates, translocation of silica in solution, and deposition of secondary silica-containing substances are involved in the development of soils. Ions essential for the growth of plants are released to the soil solution as silicates weather, and these ions may be held against leaching at exchange sites on other silicates. Silica is absorbed in appreciable quantities by some plants and is returned to the surface of the soil as the plants decay. The nature and the transformations of silica and silicates in soil are, thus, fundamental to an understanding of many aspects of soil and plant sciences. The nature of silica and silicate minerals is treated in standard text1 Joint contribution as No, 71 of the Soil Research Institute, Canada Department of Agriculture, Ottawa, and as Agronomy paper No. 802, Cornell University, Ithaca, New York.

339

340

J. A. MC KEAGUE AND M. G. CLINE

books on mineralogy; Lutz and Chandler (1946) and Jackson et al. ( 1948a), among many authors, have discussed these minerals in relation to soils. Jackson et al. (1948b), Jackson and Sherman (1953), Fieldes and Swindale (1954), and Keller (1957) have discussed the weathering of silicate minerals in soils. As pointed out by Nash and Marshall (1956a) and De Vore (1959), however, much more is known about the reactants and the ultimate products of weathering than is known about the transitory intermediate products and the mechanisms of the transformations that occur as minerals decompose and as new substances form, This applies equally to silica in soils. Confusion exists in the literature about such apparently simple matters as: ( a ) the nature of the silica units that exist in soil solutions; ( b ) the concentrations of silica that occur in soil solutions; and ( c ) the effects of various factors, such as soil reaction, on the concentration of dissolved silica in the soil. The chemistry of silica has been summarized by Iler (1955). The nature of dissolved silica in various geological environments has been discussed by Krauskopf ( 1956, 1959) and Siever ( 1957) . The principles revealed by these authors have, however, apparently been overlooked in many studies of soil. The purpose of this paper is to summarize current information about the various forms of silica in soils and about the reactions that they undergo. Silica in solution is emphasized, as this subject has been a source of confusion in the literature on soils and has not been reviewed to the knowledge of the authors. The extensive information concerning solid forms of silica and silicates and their weathering is discussed only briefly. Literature on the synthesis of silicate minerals is mentioned very briefly. Silica and its relationships in different kinds of soil is discussed in very general terms. Literature concerning the effects of silica upon the growth of plants is not reviewed. II. Silica in Solid Forms

The inorganic fraction of soil is derived from the rocks of the earth's crust; the forms of silica in soil include those of the rocks from which soils are formed and all the silica-containing material derived from them. The amounts of silica vary widely according to the nature of the parent material and the transformations it has undergone. More than one hundred relatively common silicate minerals have been listed by Dana (1959), and there are also a variety of important amorphous forms of silica and silicates, Much information is available concerning the chemical and mineralogical compositions of a variety of igneous and sedimentary rocks and

SILICA IN SOILS

341

of various unconsolidated materials, For example, see Clarke ( 1924), Poldervaart (1955), Pettijohn (1957), or any textbook of geochemistry. Comparisons have been made between the chemical composition of rocks and that of soils presumably derived from them; summaries by Marbut (1935) and Mohr and Van Baren (1954) present good examples. Many studies have been made of changes in the mineralogical composition of material from the parent rock to the soil surface. Some of these mineralogical data have been summarized, as in publications of Mohr and Van Baren (1954), Fieldes and Swindale (1954), DHoore (1954), Cady ( 1960), and Sivarajasingham et al. (1962). From the extensive literature on silica and silicates in soil, only a few papers have been selected to illustrate the diversity of forms and amounts of free and combined silica and to indicate the trends in the kinds and amounts of these substances in various soils. Some of the amorphous forms of silica in soils have been stressed, as these have received less attention than crystalline forms.

A. FORMS AND AMOUNTS The amount of silica in soil varies from less than 10 per cent to almost 100 per cent. In some laterites, it is less than 1 per cent (Sivarajasingham et al., 1962). Jackson and associates (1948a) stated that aluminosilicates and quartz often make up 75 to more than 95 per cent of the inorganic material of soils. 1. Crystalline F o m

Smithson (1956) listed the following forms of crystalline silica in soils: ( a ) quartz, ordinary crystalline silica; ( b ) chalcedony, cryptocrystalline quartz; and ( c ) chalcedonite, a secondary form of crystalline silica described by Fieldes and Swindale (1954).Cristobalite, a hightemperature form of crystalline silica, has also been reported in some soils containing volcanic ash ( Rachmat Hardjosoesastro, 1956), and tridymite probably also occurs. Quartz accounts for about 12 per cent of the weight of a hypothetical average igneous rock and about 67 per cent of a hypothetical average sandstone (Clarke, 1924). It may be almost 100 per cent of the inorganic fraction of some soils (Soil Survey Staff, 1960, Profile 39); in other soils, little or no quartz is present. The following estimates (Clarke, 1924) of the average mineralogical composition of rocks give a very approximate indication of the major crystalline forms of silica in rocks from which soil parent material is derived. For igneous rocks: feldspars, 59.5 per cent; amphiboles and pyroxenes, 16.8 per cent; quartz, 12.0 per cent; micas, 3.8 per cent. For

342

J. A. MC KEAGUE AND M. G . CLINE

shale: feldspar, 30.0 per cent; “clay,” 25.0 per cent; quartz, 22.3 per cent. For sandstone: quartz, 66.8 per cent; feldspar, 11.5 per cent; “clay,” 6.6 per cent. These values merely provide some idea of the amounts of crystalline forms that are available as potential sources of silica in soils. If such rocks weather in place, the resistant minerals, such as quartz, are concentrated by loss of some constituents of the associated weatherable minerals. These remain with secondary products of weathering and remnants of weatherable minerals as parent material of soil. If such rocks are disintegrated physically, as during glaciation, the minerals may be sorted to varying degrees before they are deposited. The weathered products of rocks or substances from soils themselves may be transported and sorted before being deposited where soils can form. Thus, the endless processes of weathering and of the geomorphic cycle leave parent materials that range extremely widely in the amounts and kinds of silicon-containing crystalline material. Silica is present in forms and amounts that approximate those of the rocks themselves in some young deposits, such as glacial till ( McCaleb, 1954; Ehrlich et al., 1955). The very resistant primary forms, mainly quartz, have been highly segregated in some (Soil Survey Staff, 1960, p. 109). Secondary forms, such as montmorillonite and kaolinite, have been highly concentrated in others (Soil Survey Staff, 1960, p. 124). Most of the silica has been lost in still others ( Sivarajasingham et aZ., 1962). The potential combinations of kinds of silica-containing crystalline substances in the parent material of soil is as great as the possible combinations of kinds and amounts of primary and secondary silicate minerals found in the lithosphere. Fortunately, a relatively small number is likely to dominate in a given place (Jackson et al., 1948b). Alteration and segregation of silicon-containing substances continue during soil development, in some soils differentially among horizons. Generally, the trend is toward concentration of both primary and secondary forms that are resistant to decomposition. These include mainly quartz, among the primary minerals, and the silicate clay minerals. Amorphous forms commonly occur. 2. Amorphous Forms

Information concerning the amorphous silicon-containing substances in soil is not as complete as that for crystalline materials. The importance of amorphous material in soils and clays has been recognized (Fieldes and Williamson, 1955; Jackson, 1956; Rich and Thomas, 1960; Smithson, 1956), however. Smithson (1956) has listed opaline silica as a common constituent of soils. Opaline silica from plants is nearly 100 per cent of

SILICA IN SOILS

343

some surface layers (Riquer, 1960) and is 1 or 2 per cent of the fine sand and silt fractions of many soils (Bazilevich et al., 1954; Smithson, 1956, 1958, 1959; Beavers and Stephen, 1958; Baker, 1959). Van Rummelen (1953) also reported amorphous silica deposits around root channels in soil; deposition was thought to be due to evaporation of silica-bearing solutions. The term “allophane” has been used for a variety of noncrystalline silica-containing colloidal mineral substances that are common constituents of soils. Such material is known to have important consequences in soils, but it has not been emphasized in the literature, largely because methods are not available for accurate characterization of discrete kinds of substances. The term, as it appears in the literature, has meaning mainly in the perspective of the methods used to characterize the substance studied. Harrassowitz (1926) used the term for weathering products consisting of various hydrous aluminum silicates, or mixtures of hydroxides of aluminum and silicon, the crystallinity of which had not been demonstrated. Ross and Kerr (1934) described it as “an amorphous hydrous aluminosilicate.” White ( 1953) defined it as “any amorphous substance that may be present in clay materials as extremely fine material and which has an indefinite composition.” Jackson (1956) stated that “experience with allophane of soils establishes that co-precipitated iron oxide is an important constituent of amorphous aluminosilicates of soils.” According to Rich and Thomas ( 1960), soil mineralogists consider allophane as amorphous combinations of silica and alumina. The term “allophane” usually implies a substance “amorphous” to X-rays (X-amorphous). It includes some segment of the range from complete disorder to perfect order, which is characteristic of siliceous colloidal material in soils (Jackson, 1956). It implies some hydrous combination of silica and alumina in which properties vary with their proportions ( DeKimpe and Gastuche, 1960), degrees of order, other components in the system, and conditions of the system. Consequently, a variety of properties have been reported for allophane (Ross and Kerr, 1934; Birrell and Fieldes, 1952; Fieldes et al., 1952; White, 1953; Yoshinaga and Aomine, 1962a). Jackson ( 1956) has summarized some “diagnostic” properties of material called “allophane.” Chemical values such as SiO2/Al20, ratio and exchange capacity vary widely. Differential thermal analysis was considered more diagnostic; “allophane” gives a strong endothermic peak at approximately 150°C. (Jackson, 1956; Birrell and Fieldes, 1952; Yoshinaga and Aomine, 1962a). Several attempts have been made to subdivide “allophane” into two or more kinds of substances. For example, Jackson (1956) differentiated “stable allophane” as that which is dissolved little in HCI at pH 3.5 or in

344

J. A. M C KEACUE A N D M. C. CLINE

2 per cent Na2C03 solution from “unstable allophane” soluble in dilute acids and alkalis. Jackson (1956) stated that “unstable allophane” appeared to form quickly by the weathering of permeable volcanic ash; “stable allophane” was thought to be a weathering relic of halloysite and kaolinite. Fieldes (1955) using data of X-ray analysis, differential thermal analysis, infrared absorption spectroscopy, and electron micrography, distinguished two forms of allophane. In “allophane B,” amorphous silica was discrete and the particles were very fine; in “allophane A,” silica and alumina were randomly combined and the particle size was greater. Fieldes considered that with increasing age the clay derived from volcanic ash passes through the sequence: “allophane B,” “allophane A,” metahalloysite, kaolinite. Yoshinaga and Aomine (1962a,b) in a study of certain Ando soils ( Andepts) differentiated an X-amorphous colloid, “allophane,” from an unknown colloid of low crystallinity, which they called “imogolite.” The separation of these substances from crystalline clay and from each other was based upon dispersion of a fraction called “allophane” in acid or alkaline media, dispersion of a fraction called “imogolite” in acid but its flocculation in alkaline media, and flocculation of crystalline clay in acid. They reported the results of studies of election micrographs, X-ray diffraction patterns, infrared absorption, differential thermograms and elemental analysis of presumably pure “allophane.” Yoshinaga and Aomine (1962a) reasoned that much of the allophanic material formerly reported in soils and clays contained crystalline clay minerals and “imogolite” as well as pure allophane. De Kimpe and Gastuche (1960) have reported synthesis of crystalline silicates from gels that could be considered allophanic. Jackson and his associates (Jackson, 1956; Aomine and Jackson, 1959; Hashimoto and Jackson, 1960) have outlined various procedures for the estimation of “allophane” in soils. One of these (Aomine and Jackson, 1959) is based upon a difference in cation exchange capacity of about 100 milliequivalents per hundred grams of “allophane” for samples dispersed at pH 10.5 and pH 3.5. De Kimpe and Gastuche (1960) have reported that high pH and high salt concentration promote charged fourfold coordination of A1 in Si-A1 gels. The effect of pH and other factors upon the apparent cation-exchange capacity of “allophane” has been studied by numerous other workers (Birrell and Gradwell, 1956; Birrell, 1961; Wada and Ataka, 1958) ; appreciable physical adsorption of cations by “allophane” occurs. Palagonite, an amorphous iron-aluminum silicate ( Birrell and Gradwell, 1956), has not been studied as extensively as “allophane” although the amorphous, siliceous soil material usually referred to as “allophane” probably includes palagonite (Jackson, 1956). Results of McKeague and

SILICA IN SOILS

345

Cline (1963b) on the adsorption of monosilicic acid indicate that at least some of the amorphous silica in soils occurs as coatings at the surfaces of iron oxides and other substances. The relatively high silica contents of extracts of soils treated for the removal of “free” iron (Mehra and Jackson, 1960) may indicate amorphous combinations of iron oxide and silica in soils. It is impossible with available methods to separate and to determine quantitatively the various amorphous oxides, amorphous silicates, weakly crystalline substances and crystalline substances of soils, although some progress has been made in characterization of amorphous fractions separated by empirical procedures. The significance of amorphous silicacontaining material in soils has largely been neglected. Sivarajasingham et al. (1962) have emphasized the potential role of such material in the development of crystalline substances and the possible misinterpretations that are a consequence of ignoring it. The fine fraction of many soils contains much X-amorphous material. B. STABILITYOF SOLIDFORMS

1. General Trends The general fate of silica during weathering and soil formation is reasonably well summarized by the statement that silica is lost to drainage waters as the kinds of substances present change to forms increasingly stable in the system. Bases are also lost, and sesquioxides are normally concentrated concurrently. Targul’yan ( 1959), Vinokurov and Bukharayeva (1961), and others have shown that silica is lost relative to sesquioxides even during the very early stages of soil formation from rocks in cold regions. In detail, however, a number of complicating factors temporarily or locally appear to refute the oversimplified statement. Keller (1957) has modified Reiche’s (1945) definition of weathering as “, . . the response of materials within the lithosphere to conditions at and near its contact with the atmosphere, the hydrosphere, and perhaps still more importantly, the biosphere.” Any consideration of weathering and the fate of its products must consider the entire system, including the environmental conditions that control it and all the materials that are present. The systems in which weathering occurs include not only primary silicates as things that are acted upon, but also a great variety of secondary minerals and amorphous substances that are both products of weathering and potential subjects of weathering. It is certain that many compounds, complexes, and adsorption systems are forming and decomposing concurrently. The transformations involved are primarily surface

346

J. A. MC KEAGUE A N D M. G. CLINE

phenomena on a subcrystalline scale. The primary silicates must completely decompose, probably unit cell by unit cell, to permit the drastic change in organization that must occur from primary to secondary silicate. Even gibbsite pseudomorphs after feldspar must have involved complete destruction of each unit cell of feldspar and reorganization of the alumina it contained, or alumina from outside sources, into a new internal structure whose external form is patterned after the feldspar. The system is open, so the opportunity exists, at least momentarily, for silica to be removed in solution. It may, however, be incorporated into new substances as an adsorbed phase, as part of a disorganized amorphous substance of indefinite composition, or as a unit of definite composition that can acquire crystallinity by accretion or by condensation on an existing crystal. Examples of such transformations and some of the numerous possibilities that exist have been described by Sivarajasingham et nl. (1962). Nevertheless the general trend is loss of silica if water passes through a weathering system. Polynov ( 1937) calculated the relative mobilities of elements, as ions or oxides, from the average composition of igneous rocks and the average composition of mineral residues of river water. He found relative mobilities of C1-, SO4--, Ca++, K+, SiOz, Fe203, and A1203 to be 100, 57, 3.0, 1.25, 0.20, 0.04 and 0.02, respectively. Anderson and Hawkes (1958) found similar relative mobilities in a study of the amounts of elements in a stream in relation to the composition of the rocks of the drainage system. If comparable mobilities apply to soils, the percentage of silica should increase during the initial stages of soil development in a material high in primary minerals; during more advanced stages, the percentage of silica should decrease. Many factors can affect the trend in detail at a given site in a soil. Water movement as a medium of transport is, of course, essential. Crompton (1960) has focused attention on the importance of relative rates of weathering and leaching. McKeague and Cline (1963b) have shown that soil material and a variety of solids that are common in soils remove silica from solution, presumably by adsorption, to concentrations far below its theoretical solubility. Noncrystalline material was most effective. The relatively little-known disorganized fraction of soil colloids should provide a major reservoir for retention of silica in the solid phase, at least temporarily. Jackson et al. (1948b) concluded that the rate of weathering of a mineral in a given environment, which is related to the potential rate of release of silica to the soil solution, can be considered a function of the product of intensity and capacity factors. They included temperature, moisture, pH, and Eh among intensity factors and specific surface and nature of the mineral among capacity factors. The “nature

SILICA IN SOILS

347

of the mineral” involves the stability of mineral species in weathering systems.

2. Stability in Relation to Structure and Composition Differences in stability of mineral species, including silicates, is well known ( Goldich, 1938; Pettijohn, 1941 ), Even substances having identical composition but different structures, such as quartz, cristobalite, and opaline silica ( Fournier and Rowe, 1962) have different solubilities. Some of the more important relationships between stability of silicates and their structure and composition are considered here, following Keller‘s (1957) treatment of structure. The basic unit of silicates and silica is the tetrahedron, Si044-. Classification of silicate minerals is based upon the number of 0 atoms of this tetrahedron that are linked directly to Si atoms of other tetrahedra and to other atoms, such as Al, that substitute for Si in tetrahedral coordination. Five possibilities exist: none, one, two, three, or four of the 0 atoms acting as direct bridges between adjacent tetrahedra. Each of these possible structures is represented by a group of silicate minerals. a. Tectosilicates. Each of the four 0 atoms of the Sio4*- tetrahedron is linked to another Si or A1 atom of an adjacent tetrahedron. Each 0 atom is shared by two Si atoms or by one Si atom and one A1 atom, and each Si or A1 atom is surrounded by four 0 atoms. The simplest formula when no substitution is involved is SiO,; thus quartz, tridymite, and cristobalite are tectosilicates. The stability of this continuous threedimensional network of Si and 0 atoms depends partially upon the closeness of packing of the atoms. The density and the closeness of packing increase in the order: opal, cristobalite, tridymite, quartz. As suggested by Iler (1955) and supported by data of Fournier and Rowe (1962.), the solubilities of these substances in water decrease in the order in which they are Iisted. Presumably quartz, the dense form, is also more resistant to physical breakdown than the other forms. Feldspars are tectosilicates in which A1 atoms are substituted for some of the Si atoms; the degree of substitution ranges from about one atom of A1 for every three atoms of Si in albite, orthoclase and microcline to about two atoms of A1 for every two atoms of Si in anorthite. The substitution of Al3+ for Si4+ leaves one excess negative charge; the charge is balanced largely by K + in orthoclase and microcline and by Na+ and C a + + in members of the plagioclase series. The substitution of A1 for Si in tetrahedral positions is not random; Loewenstein (1954) showed that, for any two tetrahedra linked by an 0 bridge, the center of only one of the tetrahedra can be occupied by Al. According to him, the

348

J. A. M C KEAGUE AND M. G. CLINE

center of the other tetrahedron must be occupied by Si or another small ion of electrovalence four or more. Theoretically, feldspars should weather more rapidly than quartz because: ( a ) the relatively large, low-charged K + , Na+, and Ca++ ions that balance the charge deficit are susceptible to chemical attack; they may be replaced by H 3 0 + (Fredrickson, 1951; Garrels and Howard, 1959). Bonds between these ions and 0 are weaker than Si-0 bonds (Pauling, 1948). ( b ) The crystal ionic radius of A13+ is about 0.50A. compared with about 0.41 A. for Si4+ (Pauling, 1948), thus the substituted AP+ does not fit the lattice perfectly, and the strain weakens the structure physically. b. Phyllosilicates. These minerals are two-dimensional networks of silica tetrahedra in which three oxygens of each tetrahedron act as a bridge between it and adjacent tetrahedra. The other oxygen forms a bridge between the Si atom and an octahedrally coordinated Mg or A1 atom. The basic structure is sheets of linked silica tetrahedra and aluminum or magnesium octahedra; the sheets are arranged in various ways. In two-layer phyllosilicates, such as kaolinite and halloysite, the unit is one octahedral layer combined with one tetrahedral layer. In threelayer phyllosilicates, the unit is an octahedral layer between two tetrahedral layers; micas, montmorillonite, and vermiculite are examples. Interlayer cations balance the charge resulting from the substitution of AP+ for Si4+ in tetrahedral layers, and of Mg++, Fe++, etc. for A13+ in octahedral layers. Chlorite is a mineral in which a brucite [Mg (OH)z] layer occupies the interlayer space between the units of three-layer phyllosilicates. The presence of Al( OH), rather than Mg (OH)z in the interlayer space of some soil chlorites has been recognized (Rich and Thomas, lW),and Brydon et al. ( 1961) reported the properties of a “dioctahedral aluminum chlorite” from a soil. The structure is physically strong in two dimensions but weak in the third; thus these minerals break readily into plates. Their chemical stability varies considerably; in general, those with A1 in the octahedral layer are more stable than those with Mg. This may be attributed in part to the greater strength of A1-0 than of the Mg-0 bonds (Pauling, 1948). c. Inosilicates. Only two oxygens of each sio44-tetrahedron serve as direct links to adjacent silica tetrahedra in this group of minerals. The other two oxygens of each tetrahedron are linked to cations such as Ca++, Mg++, and Fe++. The basic structure is, thus, chains of silica tetrahedra linked through metal ions to other chains of tetrahedra. Pyroxenes ( Ca,Mg,Fe) Si03, are common single-chain inosilicates. Amphiboles are double-chain inosilicates; they can be visualized as the

SILICA IN SOILS

349

product of condensation of two adjacent single chains. These chain structures are strong in one direction but weak in two. The numerous Ca-0, Mg-0, and Fe-0 bonds are susceptible to chemical attack. d. Sorosilicutes. Only one oxygen of each Sio4*- tetrahedron is directly linked to an adjacent Si atom; the other three oxygens are bonded to divalent cations. These minerals are not common in soils. e . Nesosilicates. All oxygens of each Sio4‘- tetrahedron are directly linked to a metal cation such as Fe++ or Mg+ +. Each sio44occurs as a unit linked to other tetrahedra by metal cations. Olivine, ( Mg,Fe)8i04, is an important mineral of this group. The metal cation linkages are susceptible to chemical attack; thus, these minerals tend to weather rapidly. These are exceptions; zircon, ZrSi04 is very stable (Pettijohn, 1941) . In many of the silicate structures, there is some degree of substitution of A1 for Si in tetrahedrally coordinated positions and of Mg, Fe, and other elements for A1 in octahedral positions. Balancing cations may also substitute for each other; some substitution of Na+ for K+ is common in orthoclase ( Winchell and Winchell, 1959). Stability generally decreases with increasing amount of substitution and with increasing difference in size of the ions substituted. Stability, in the sense the term is applied to crystalline material, is not strictly applicable to much of the X-amorphous material of soils. Colloidal disorganized fractions commonly have very great specific surface and probably undergo continuing change involving either loss or gain of silica under appropriate conditions. Fieldes (1955) concluded that “allophane” undergoes transformations to crystalline silicates with time. Sivarajasingham et ul. (1962) have also suggested that disorganized material may be a major precursor of crystalline silicate clays. Jackson (1956) recognized differences in stability among kinds of “allophane” presumably depending upon whether its source was volcanic glass or crystalline silicates. Generally, disorganized material is expected to be less stable than crystalline material of similar composition and specific surface. Opal, for example, is more soluble than cristobalite or quartz (Fournier and Rowe, 1962). Volcanic glass weathers rapidly (Mohr and Van Baren, 1954) and typically produces “allophane” (Soil Survey Staff, 1960).

3. Stability in Relation to Particle Size Size of particle is an important factor in stability of silicates and silica through its relationship to specific surface (Jackson et al., 1948b; Iler, 1955). The reactions involved in chemical weathering occur at the surfaces of solids, and specific surface increases with decreasing particle

350

J. A. M C KEAGUE AND M. G. CLINE

size. This is evident in the relatively little silt found in highly weathered material, though both sand and clay may be abundant. The very stable secondary products of weathering are mainly in clay sizes. Primary silicates can occur in sand, silt, or clay particles. In highly weathered material, even quartz has largely been destroyed in silt and clay sizes but remains as sand. Thus, sandy clays and sandy clay loams are common in highly weathered material from old glacial deposits whereas silty textures are common in less weathered material of originally similar character. Coarse particles of primary silicate minerals that are normally considered relatively unstable persist after fine particles of the same or a “more stable” silicate mineral have disappeared. From the data of several authors, including Tedrow ( 1954), McCaleb ( 1954), Ehrlich et al. (1955), Rice et al. (1959), Pawluk (1960), Yassoglou and Whiteside (1960), Cady (1960), and Brydon and Patry ( 1961), it appears that the upper horizons of soils from Wisconsin glacial deposits of humid and subhumid areas have mainly quartz among primary minerals in fractions as fine as coarse clay but have weathered feldspars and altered ferromagnesian minerals in the silt fraction. Relatively unweathered primary silicates are found mainly in the sand fractions of soils that have formed in early Wisconsin drift.

4 . Mineral Stability Series Various authors, including Goldich (1938) and Pettijohn (1941), have published lists of megascopic silicate minerals arranged according to evidence of stability in the zone of weathering. Though such lists differ in detail, they follow comparable patterns and generally reflect the influence of structure and composition discussed in another section. Goldich, for example, listed in increasing order of stability: (1) olivine -a nesosilicate, ( 2) augite-a single-chain inosilicate, ( 3) h o r n b l e n d e a double-chain inosilicate, ( 4 ) biotite-a phyllosilicate having magnesium and iron in the octahedral sheet, ( 5 ) potash feldspar-a tectosilicate having aluminum substituted for one-fourth of the silicon, ( 6 ) muscovite-a phyllosilicate having aluminum in the octahedral sheet, and ( 7 ) quartz-a tectosilicate having no substitution for silicon. The plagioclase feldspars, tectosilicates with aluminum substituted from onequarter to one-half the silicon, were treated separately. All were listed as less stable than potash feldspars, and they were arranged in order of decreasing stability from 100 per cent sodium to 100 per cent calcium as the balancing cations. Jackson et al. (1948b) determined the mineralogical composition of the clay fraction in soil material ranging in age from very young glacial deposits of Quebec, through the various older glacial

SILICA IN SOILS

351

materials of central United States, to the old weathered material of unglaciated areas of the United States, Puerto Rico, and Hawaii. Relative stabilities of minerals in the clay fraction were judged from their presence or absence in the coarse- and fine-clay fractions of A, B, and C horizons of the soils. With due allowance for inherited presence or absence of specific minerals, Jackson et al. summarized the results by tabulating 13 weathering “stages” represented by minerals listed in the order of appearance and disappearance with increasing age: ( 1 ) gypsum, ( 2 ) calcite, ( 3 ) hornblende, ( 4 ) biotite, ( 5 ) albite, (6) quartz, ( 7 ) illite, ( 8 ) interstratified mica intermediates, ( 9 ) montmorillonite, ( 10) kaolinite, (11) gibbsite, (12) hematite, (13) anatase. “Stages” 3 through 10 involve silica. The authors noted that three to five minerals tend to dominate the clay fraction of a given sample, that the “stage” of weathering tends to decrease with depth below the surface, that lack of leaching and accumulation of soluble end products retards the succession of “stages,” that desilication appears to be a primary process from “stage” 7 (illite) to 11 (gibbsite), and that the sequence is sometimes reversed. Many, but not all, studies of the mineralogical composition of soils are consistent with the stability series of Goldich (1938) or others and with the weathering sequence of Jackson et al. (194813). Substantial elements of agreement may be found in many papers that deal with soils in relatively young unconsolidated deposits, including the work of McCaleb (1954), Tedrow (1954), Cann and Whiteside (1955), Ehrlich et al. (1955), Rice et al. (1959), Yassoglou and Whiteside ( 1960), Pawluk (1960), Brydon and Patry (1961), and Gradusov and Dyazdevich (1961). Elements of agreement may also be found in many studies of older, more highly weathered material, such as the numerous studies reviewed by Mohr and Van Baren (1954) and the work of McCaleb (1959), Humbert (1948), Harrison (1933), D’Hoore ( 1954), and others. Many exceptions to the “weathering sequence” of Jackson et al. (1948b) may also be found in the literature. A high proportion of these deal with the advanced “stages” of weathering and especially with soils in residuum of rocks presumably weathered in place. Cady (1960) has emphasized a distinction between weathering of silicates in hard igneous or metamorphic rocks along a sharp front and weathering of similar silicates in unconsolidated material of mixed mineralogy, such as glacial till, volcanic ash, and disintegrated partly weathered rock. In the former, the transformations from primary to secondary minerals of advanced stages, sometimes beyond the stages of silicates, occurs within a few millimeters of the rock. In the latter, transitional secondary silicates are commonly present, and the “stages” may commonly be traced from the

352

J. A. MCKEAGUE AND M. G. CLINF,

C horizon upward. Differences of this kind may be seen in Harrison’s (1933) comparison of weathering of dolerite and granite. Harrison ( 1933), Hardy and Rodriguez (1939), Alexander et al. (1941), DHoore (1954), and others have shown that almost complete desilication may occur within a few millimeters of weathering rock and that resilication of minerals like gibbsite to minerals like kaolinite apparently occurs in some soils derived from such weathered material. Sivarajasingham et al. (1962) have discussed this phenomenon in some detail and have suggested that in some cases apparent resilication may involve destruction of crystallinity of gibbsite and development of order in X-amorphous silica-alumina substances that may be present. In either case, some source of silica is implied. Fieldes and Swindale (1954) have stressed the impact of parent primary silicates on the character of secondary minerals formed during weathering. This is evident in the work of Harrison (1933) and many others. Mohr and Van Baren (1954), however, have shown in their extensive review of the literature that similar rocks may alter to substantially different secondary minerals and dissimilar rocks may alter to similar secondary minerals in different places, presumably under some factor of environmental control. The general case of Fieldes and Swindale is common, but there are many exceptions. Humbert (1948) showed that diabase and granite components within a conglomerate, and consequently in the same general system, weathered to the same kinds of secondary minerals, predominantly kaolin in both cases. This suggests that the character of the system, possibly influenced strongly by the relative availability of silica in solution, exerts a major influence on the character and stability of products of weathering. The “weathering sequence” of Jackson et al. (1948b) is predicated on a kind of norm for environment of soils. The primary silicates of rocks are formed and are stable in systems very unlike those of soils. The authors know of no case in which primary silicates (excluding crystalline silica as forms of quartz and related minerals) were reported to have formed in soil systems. They would not be expected to do so. They would be expected to weather and to contribute silica to soil solutions at relative rates consistent with the properties of structure and composition that have been discussed. Conditions of the open systems involved, including removal of end products, introduction of solutes, temperature, pH, and Eh, as well as properties of the mineral should control absolute rates, which may approach zero. The direction of change, including desilication, should be consistent at rates greater than zero. The secondary silicates, however, are potential products of systems

SILICA IN SOILS

353

like those of soils. General trends of dedication comparable to those postulated by Jackson et al. (1948b) for stages 7 to 11 are undoubtedly effective at the geographic scale of large regions and at time scales commensurate with those of geomorphic cycles. Local and temporary conditions of the open microsystems at the surfaces of secondary particles, however, would be expected to provide opportunities for reversal of trends and for appearance of some one or some combination of secondary silicates and sesquioxides in crystalline or X-amorphous form as “stable” components of the moment. The existence of soils dominated by montmorillonite in local basins where silica and bases accumulate from adjacent landforms dominated by Oxisols would appear to be such a case (Greene, 1945). De Kimpe and Gastuche (1960) have demonstrated that secondary silicates can be synthesized under normal temperatures and pressures. Jackson (1959) gave examples of the influences of various factors of soil formation on the frequency distribution of clay-size minerals in major groups of soils. 111. Silica in Solution

A. CHEMISTRY OF SILICA

Some relevant facts about the chemistry of silica are summarized here as a basis for discussion of silica in soil solutions. Krauskopf (1959) has emphasized that information about the solubility of amorphous silica and the development of a simple valid chemical test (Alexander et al., 1954) for silica in monomeric form have made it possible to base hypotheses of the role of silica in sediments on firm chemical principles. The same can be said of silica in soils. For years it has been assumed in the fields of soil science and geology that the dissolved silica in natural waters is in colloidal form (Reifenberg, 1938; Vilenskii, 1960; Roy, 1945; Krauskopf, 1956) and that a neutral to alkaline reaction is favorable for the loss of silica by leaching (Mohr and Van Baren, 1954; Mason, 1952). Current information strongly indicates that these assumptions are invalid. 1. The Solubilities of Various Forms of Silica

It has been shown (Alexander et al., 1954; Krauskopf, 1956; Greenberg and Price, 1957; Piryutko, 1959) that the solubility of pure amorphous silica at a given temperature and below pH 9 falls within a relatively narrow range. For example, Alexander et al. (1954) obtained values between 120 and 140 p.p.m. at 25°C. and at pH values below 9 whether they started with initially undersaturated or initially supersaturated solutions of monosilicic acid in the presence of amorphous

354

J. A. MCKEAGUE AND M. C . CLINE

silica. They also showed that the solubility of amorphous silica increases sharply with increasing alkalinity above pH 9 due to the fonnation of silicate ions; between pH 2 and pH 9, solubility is almost independent of acidity. Iler (1955) indicated that the solubility of amorphous silica varies linearly with temperature and approaches zero at a temperature slightly below 0" C. It is not possible to state a unique value for the solubility of amorphous silica, even under defined conditions, as Alexander (1957) has shown that the solubility of a given sample depends not only upon conditions like those mentioned above but also upon particle size and the number of S i - O H groups in the internal structure. Thus, different samples of amorphous silica may have somewhat different solubilities. The fact that the solubility of pure amorphous silica is virtually independent of pH within the reaction range of most soils contradicts some hypotheses in soils literature. Correns' data (1949) have been quoted widely in the literature (Mason, 1952; Keller, 1957) to explain various observed phenomena of weathering and soil formation. His data appear to show that the solubility of amorphous silica increases with increasing pH throughout the reaction range from pH 2 to pH 11. Alexander et al. (1954), Krauskopf (1956), and others have shown that this is not true; solubility is almost independent of pH throughout most of that range. Consequently, the hypotheses propounded on the basis of Correns' data must, at least, be scrutinized most critically. Although the solubility of quartz is much lower than that of amorphous silica, the latter is the form precipitated from supersaturated solutions ( Iler, 1955). Solutions undersaturated with respect to amorphous silica but supersaturated with respect to quartz, exist metastably for years (Siever, 1957). According to Siever (1957), crystalline quartz has not been synthesized at temperatures approaching room temperature. Secondary quartz commonly occurs in sediments, however, and it has been reported in soils (Jeffries, 1937; Carroll and Hathaway, 1954; Fieldes and Swindale, 1954). Thus the solubility of quartz is relevant to this paper. Iler (1955) demonstrated the magnitude of the effect of particle size on the solubility of quartz. Using Gardner's (Iler, 1955) value of 6 p.p.m. as the solubility of massive quartz, and an approximate value of 416 ergs/cm.z as the interfacial surface energy between quartz and water, he calculated that the solubilities of quartz particles 10 mp, 5 mp, and 3 mp in diameter would be 28 p.p.m., 120 p.p.m., and 930 p.p.m., respectively. Previously, Lucas and Dolan (1939) had reported an abrupt increase in the solubility of quartz when the particle size was reduced below about 5 p.

SILICA IN SOILS

355

Morey et al. (1962) studied the dissolution of silica from quartz grains in water at 25" C. by rotating polyethylene bottles containing quartz-water suspensions for more than a year and determining dissolved silica periodically. In samples containing quartz grains that had been preheated in an attempt to remove surface strains, the concentration of dissolved silica increased at an accelerating rate for 386 days. It reached a maximum of 80 p.p.m. at that time. The concentration decreased sharply after 386 days, reached a value of 5 p.p.m. after 417 days, and remained at 5 or 6 p.p.m. until the end of the experiment (530 days). The apparent equilibrium concentration at the end of the experiment is similar to some other estimates of the solubility of quartz (Iler, 1955; Van Lier et al., 1960). It is surprising in the light of other observations (Iler, 1955; Siever, 1957; White et al., 1956) that equilibrium was established so rapidly. In samples containing quartz that had not been preheated, the concentration of dissolved silica reached a maximum of almost 400 p.p.m. and remained at values above 200 p.p.m. throughout the remainder of the experiment (Morey et al., 19f32). The preheating treatment had a major effect upon the apparent solubility of quartz; this is consistent with other evidence of a disturbed layer of relatively high solubility at the surface of quartz (Holt and King, 1955). Fournier and Rowe (1962), extrapolating from data obtained at higher temperatures, estimated the solubility of cristobalite at 25" C. to be 27 p.p.m. This is consistent with Iler's (1955) prediction that the solubilities of various forms of SiOz are probably negatively correlated with their densities. Cristobalite has a lower density than quartz. The solubility data discussed apply to various forms of pure silica. These values may not apply to soils, as it has been shown that various impurities markedly affect the apparent solubility of silica (Denny et al., 1937, 1939; King and McGeorge, 1938; Jephcott and Johnston, 1950; Germer and Storks, 1939). These workers measured the dissolution of silica from ground quartz and from silicate minerals, both in the absence of and in the presence of a variety of additives. King and McGeorge (1938) found that the addition of a variety of oxides and hydroxides of polyvalent metal ions depressed the amount of silica in solution. Jephcott and Johnston (1950) found that the effectiveness of various forms of alumina in depressing the dissolution of silica varied widely and was not proportional to their solubilities. Denny et al. (1939) and Germer and Storks (1939) showed that the mechanism by which alumina depresses the solubility of silica is the formation of a coating of hydrated aluminum oxide around the siliceous particles. Adsorption of dissolved silica on alumina was a minor factor in decreasing the amount of silica in solution.

356

J. A. MC KEAGUE A N D M. G. CLINE

Soluble additives may also affect the apparent solubility of silica. Solutes that increase the pH to more than 9 increase the solubility of pure silica due to the formation of silicate ions. The addition of NaCl greatly accelerates the dissolution of silica (Iler, 1955; Van Lier et al., 1960), but at a concentration less than 1 M, it does not affect the equilibrium solubility (Lucas and Dolan, 1939; Greenberg and Price, 1957; Van Lier et al., 1980). Piryutko (1959) observed that the addition of salt solutions to silica gel resulted in temporary supersaturation with silicic acid. Richardson (1959) found that sugars have little influence on the solubility of quartz; he also found that amino acids depress quartz solubility. He attributed the depression to the adsorption of amino acids at the surface of quartz. Siever (1962) found that the apparent solubility of amorphous silica in water from raw peat was 14 p.p.m. He suggested that the reduction in solubility could be due to the adsorption of colloidal organic matter on the surface. 2. Nature and Reactions of Soluble Silica The idea that silica in natural waters is mainly colloidal has persisted in spite of conclusive evidence to the contrary. Dibnert and Wandenbulcke (1923) found that the silica in natural waters they analyzed was not colloidal. Harman (1927) reviewed and extended the evidence that ionic silica exists in alkaline solutions, Roy (1945) explained that the idea that silica sol is the main form of silica in natural waters has been based upon an irrelevant experiment. Iler (1955) summarized evidence based upon the rate of reaction of dissolved silica with molybdic acid, diffusion measurements, electrophoresis, and freezing point depression studies that established the existence of uncharged monomeric silica units, presumably Si( O H ) r , in dilute aqueous solutions. White et al. (1956) and Krauskopf (1956) showed that silica in solution in natural waters is mainly in monomeric form. The nature of monomeric silica in solution is indicated by the ionization constants of monosilicic acid. Although Si( OH)4 is a very weak acid, there is reasonably good agreement among most of the ionization constants reported (Joseph and Oakley, 1925; Roller and Ervin, 1940; Greenberg and Price, 1957; Greenberg, 1958). At room temperature, pK1 is about 9.8 and pK2 is about 12. Association of silicate ions occurs even in very dilute solutions; Roller and Ervin (1940) determined an equilibrium constant of 2200 for the reaction 2( HSi03-) e Si205-H20. Alexander et al. (1954) and Greenberg and Price ( 1957) showed that their data on the effect of pH on the solubility of amorphous silica are consistent with the equilibrium constants quoted if it is assumed that the concentration of Si(OH)4 does not change with pH.

+

SILICA IN SOILS

357

In soil solutions other than those of alkali soils, pH values are less than 9, and Si( OH), is presumably the main form of monomeric silica. Monosilicic acid polymerizes to form colloidal silica when the concentration of SiOa exceeds about 120 p.p.m. at room temperature and near-neutral pH. The reaction is catalyzed by OH- and F- (Iler, 1955) and by silica gel, but not by quartz (Piryutko, 1959). Weyl (1951) proposed that the low solubility of silica is due to incomplete coordination of Si4+ with OH- in Si(OH),. Si4+ has a coordination number of 4 with respect to 0--, but it goes into 6-coordination with small, singly charged anions such as F-. Weyl reasoned that, since OH- is similar in size to F-, Si4+ may tend to go into 6-coordination with OH-. Polymerization of Si( OH)r might involve sharing of OH groups between Si atoms so each Si would be octahedrally surrounded by OH. Then condensation within the polymeric units might occur, resulting in the formation of siloxane bonds and the elimination of water. Iler (1955) proposed a polymerization mechanism involving the combination of a &coordinated silicate ion, HzOSi(OH)6-, with Si( OH)*. Solutions as much as 100 per cent supersaturated with respect to amorphous silica may exist metastably for years at low temperature and in an acidic environment (White et d.,1956). Colloidal silica is a reactive substance due to its negative charge at pH values above about 3.5 (Bolt, 1957) and its hydroxylated surface. It adsorbs cations and polar substances (Iler, 1955), hydrogen bond acceptor groups ( Giles, 1959), and positively charged sols ( Mattson, 1928). Ferric iron or aluminum salt solutions (Hazel d aZ., 1949) and alumina sols (Tamele, 1950; Milliken et al., 1950) react with silica sol to form a mixture of lower pH, presumably due to the acid sites that result when A1 atoms occupy 4-coordinated positions at the surface of the silica structure. These acid sites can retain Na ions in exchangeable form (Milliken et al., 1950). The importance of silica gel as an adsorbent is further evidence of the reactivity of colloidal silica (Cassidy, 1951). It is very unlikely in view of these considerations that colloidal silica would occur free in soil solutions. Quartz or chalcedony might precipitate directly from silica in solution (White et al., 1956; Krauskopf, 1956), but the reaction must proceed extremely slowly. Solutions supersaturated with respect to quartz persist for years. Amorphous silica might be altered gradually to quartz at low temperatures; Carr and Fyfe (1958) found that at elevated temperatures and pressures, changes follow the sequence: amorphous silica, cristobalite, keatite ( a form of silica intermediate between cristobalite and quartz in some physical properties), quartz.

358

J.

A. MC KEAGUE AND M. G. CLINE

Vail (1952) summarized information on the reactions of silicate ions, but the concentrations and the pH values involved are well above those in most soils. Britton (1927) titrated metal salt solutions with NazSiOs and found that precipitation occurred at a pH slightly below that at which the metal hydroxide would be precipitated. Iler (1955) pointed out that the precipitation of metal silicates at low temperatures yields a heterogeneous mixture of colloidal metal hydroxide and silica or of metal ions adsorbed on silica. Dissolved silica might combine chemically with certain organic substances in the soil, although there is no firm evidence that this occurs. According to Iler (1955), Holzapfel concluded that silicic acid occurs in combination with lipoid substances and with cholesterol in animal tissues. The occurrence of silica-organic complexes in plants has been reported (Viehoever and Prusky, 1938; Engel, 1953). Yoshida et al. ( 1959a, 1!362), however, did not find definite evidence of silica in organic combination in rice plants. Clark and Waddams (1957) found that silicic acid does not complex with organic hydroxy acids such as malic, citric, and lactic acids. Richardson’s (1959) studies also indicated little or no tendency for soluble silica to complex with polyhydroxy organic compounds. McKeague ( 1962) reported evidence that silicate, unlike borate, does not complex with certain sugars and that monosilicic acid was not adsorbed by either a peat or a muck sample. Deuel (1960) observed that it has often been postulated that Si-0-C linkages exist in soil, and he cited work that showed that Si-OH groups and freshly formed silicate surfaces combine readily with various organic compounds. The conditions under which some clay minerals may be synthesized at low temperatures from solutions containing silica are known (Hknin, 1954, 1957; Callikre et al., 1956; DeKimpe and Gastuche, 1960), but the mechanisms of the reactions involved are not known.

3. The Colorimetric Determination of Silica Many colorimetric procedures have been outlined for the determination of silica in dilute solutions (Snell et al., 1959). These methods involve reading the absorbance of either a yellow silicomolybdic complex or a blue silicomolybdous complex (molybdenum blue) that is obtained by the reduction of the yellow complex. Govett ( 1961) remarked that the large number of suggested procedures indicates dissatisfaction with the methods; he pointed out that the dissatisfaction is due to lack of understanding of the chemistry of the colored complexes formed. The purpose of this section is to outline briefly some of the pertinent results of Stricklands studies (1952a, b, c ) of silicomolybdic acids and the appli-

SILICA IN SOILS

359

cation of these results to the choice of an analytical method for silica, as discussed by Strickland ( 1 9 5 2 ~ )and Govett ( 1961). Strickland (1952b) found that when a molybdate solution is added to a solution of silicic acid, as in the colorimetric determination of silica, either one or the other, or both members of a pair of yellow silicomolybdic complexes may form. The conditions that determine which of these complexes develops were studied. In the visual region of the spectrum, the absorption of beta silicomolybdic acid is almost twice that of the alpha complex. The beta complex changes spontaneously to alpha silicomolybdic acid; the change is retarded by excess molybdate, low temperature, and low electrolyte concentration. The reduction of these silicomolybdic acids results in the formation of three possible silicomolybdous complexes ( molybdenum blue ) , The absorption spectrum of one of the reduced complexes differs markedly from those of the other two. It is apparent that the absorbance of a solution of either silicomolybdic acid or silicomolybdous acid depends not only upon the concentration of Si present, but also upon the nature of the complexes involved and hence upon the procedure used in color development. Formerly it was assumed that only one silicomolybdic complex and one silicomolybdous complex were formed in colorimetric determinations of silica. Govett ( 1962) extended Strickland's (1952a, b, c ) work and outlined a procedure that is thought to eliminate errors and confusion resulting from mixtures of alpha and beta silicomolybdate complexes in various proportions as well as problems arising from polymerization of the molybdate reagent. The method is based upon the development of beta silicomolybdic acid under conditions such that its change to the alpha complex is slow. The use of methods such as this should increase the accuracy and convenience of the colorimetric determination of silica. Data from older procedures are probably reliable, however, if uniformity of technique was maintained and if frequent comparisons were made with the standard. Alexander et a2. (1954) outlined a simple procedure for determining whether or not all the silica in a solution is in monomeric form. In this procedure, an acidified ammonium molybdate solution is added to the sample solution, and the absorbance at 410 mp is read every half minute until it ceases to change. Alexander et al. claim that at least 98 per cent of the monosilicic acid present reacts with molybdic acid within 2 minutes. Disilicic acid takes about 10 minutes to react completely, and polysilicic acids react more slowly. The validity of this simple test has been confirmed ( Iler, 1955; Krauskopf, 1956; O'Connor, 1961).

360

J. A. MC KEAGUE AND M. G. CLINE

B. DISSOLUTION OF SOLIDFORMS OF SILICAAND SILICATES 1. As Affected by the Nature of the Surface The dissolution of quartz is greatly affected by the apparent presence at the surface of a form of silica that is more soluble than quartz. There is evidence that ground feldspars and other silicates have a surface of similar nature ( Kitto and Patterson, 1942). Such surfaces have been referred to by various authors as disturbed, amorphous, strained, highsolubility, and disorganized. Because the dissolution of silica from some soils follows a pattern similar to that of the dissolution of silica from ground quartz, the literature concerning these surfaces is summarized. Studies of the dissolution of quartz in water have yielded evidence of a disturbed surface. Lucas and Dolan (1939) showed that repeated extractions of quartz powder with water yielded successively less silica in solution. This could be due either to complete dissolution of the finer particles during the earlier extractions or to a more soluble form of silica at the surface of the particles. These authors, as well as King and McGeorge (1938) and Kitto and Patterson (1942), reported that when quartz powder is added to water, an initial stage of rapid dissolution is followed by a prolonged period of slow increase of the concentration of silica in solution. This suggested the possibility of a readily soluble amorphous layer at the surface of quartz. Clelland et al. (1952) and Paterson and Wheatley ( 1959) interpreted similar results as indicating a “disturbed” surface layer. Clelland and Ritchie (1952) concluded that the surface material was not amorphous as it did not dissolve as rapidly as silica gel. Measurements of the rate of dissolution of ground quartz in NaOH solutions indicate the presence of a layer of high solubility at the surface (Holt and King, 1955; Van Lier et al., 1960; Bergman and Paterson, 1961).Pretreatment of quartz with HF or NaOH to remove the surface material results in consistent solubility values that are lower than those obtained for the untreated material (Van Lier et al., 1960). The number of monolayers of SiOz equivalent to that in the layer of high solubility varied from about five to one ( Bergman and Paterson, 1961). Density measurements ( Clelland and Ritchie, 1952; Paterson and Wheatley, 1959) indicate that a layer of lower density than quartz occurs at the surface of quartz particles. X-ray diffraction studies ( Nagelschmidt et al., 1952; Dempster and Ritchie, 1953) of quartz ground to different sizes were interpreted in the same way. The apparent quartz content dropped from 100 for the unground material to 58 for the fraction less

SILICA IN SOILS

361

than 1 micron (Nagelschmidt d al., 1952). Etching of this fraction with H F restored the apparent quartz content to more than 90 per cent. Dempster and Ritchie (1953) and Ganichenko et al. ( 1960) studied the effect of fineness of grinding on the proportion of quartz that would undergo the alpha to beta quartz inversion at 574” C. The percentage decreased with increase in time of grinding and with specific surface. After the ground quartz had been treated with HF, differential thermal analysis indicated almost 100 per cent quartz. The surface material was apparently not ordinary quartz as it did not undergo the alpha to beta quartz inversion. Forms of silica that give the X-ray diffraction pattern of quartz but that do not display the normal thermal effect at the alpha to beta inversion temperature have been reported in clays ( McDowall and Vose, 1952) and in soils (Fieldes, 1952). Gibb ef n2. (1953) studied the surface of quartz, before and after HF treatment, by electron microscopy and electron diffraction. They postulated that a “vitreous skin” at the surface of the particles gradually blends into a crystalline core. Heavens (1953) and Talbot and Kempis (1960), however, deduced from similar studies that there is not a disturbed layer of measurable thickness at the surface of ground quartz particles. Holt and King (1955) found that monosilicic acid is adsorbed by alkali-extracted samples of quartz powder, but not by untreated samples. Alkali-treated samples, after equilibration with monosilicic acid, had the same solubility in water as untreated samples. The presence of an “adsorbed” layer of monosilicic acid is indicated also by results of a study of the exchange between surface Si atoms and radioactive Si31(OH)4 in solution. Holt and King concluded that the commonly observed rapid initial solubility of quartz is due to the desorption of adsorbed silicic acid. They reasoned that Si ( OH)4 must be coordinated to the surface, but that condensation does not occur. Weyl (1953) suggested that the disturbed surface of quartz is brought about by a rearrangement of surface Si and 0 atoms in such a way that the strong electric field around Si4+ is screened by the more polarizable 0-- ions that are displaced to the surface. This rearrangement of atoms is considered by Weyl to extend to an appreciable depth within the crystal. If this were so, the surfaces of silicate minerals should be distorted in a similar manner. It has been observed (Lucas and D o h , 1939; Kitto and Patterson, 1942) that some silicates display the same anomalous solubility effects as quartz, and Nash and Marshall (1956b) have postulated that the surface of feldspar is somewhat distorted.

362

J. A. MC KEAGUE AND M. G . CLINE

2. Mechanism of Dissolution of Silicates Nash and Marshall (1956a) stated that the reactions by which the most common minerals in the earths crust, the feldspars, are decomposed, are among the least understood of chemical changes. A similar statement could be made regarding the dissolution of many other silicates. The lack of detailed knowledge about these common reactions is, perhaps, due to the complex structures of silicates and to the fact that these minerals do not dissolve congruently, rather than to lack of study. It is known that the abrasion pH of silicate minerals is in the alkaline range ( Keller, 1957); thus, the accelerated decomposition of these minerals in acid solutions is reasonable from consideration of the mass law. Graham (1941b) stated that the hydrolysis of feldspars in carbonated water proceeds about ten times as rapidly as in pure water. Graham (1940, 1941a) demonstrated that acid clay is an effective agent in the breakdown of a variety of aluminosilicates. Nash and Marshall (1956a) found that the dissolution of members of the plagioclase series and of microcline was greater in 0.01 N HC1 than in 0.01 N NH&l and that both solutions were much more effective than water in dissolving these feldspars. Nutting ( 1945) reported that many silicate minerals would dissolve completely in 0.01 to 0.4 per cent acid if enough acid were added to neutralize the bases in the minerals and if enough water were added to dissolve the released silica. The acid environment adjacent to plant roots is favorable for the dissolution of silicates (Keller, 1957). Murata (1946) noted that some minerals dissolve completely in strong acids and yield gelatinous silica; other minerals release insoluble silica. Those that dissolve completely fall within two structural classes: (1) minerals containing silicate radicals of small molecular weight, such as orthosilicates like olivine, and ( 2 ) minerals with continuous Si-O-Si frameworks incorporating a sufficient number of aluminum or iron atoms. Apparently the acid attacks Fe-0 or A1-0 linkages yielding small units. Minerals that release insoluble silica have Si-0-Si structures of large dimensions, little substitution of Fe for Si and an A1 to Si ratio less than 2:3. Many inosilicates, e.g., amphiboles; phyllosilicates, e.g., kaolinite; and tectosilicates, e.g., orthoclase, fall into this group. Gastuche et al. (1960) found that when kaolin is treated with HCl, Si and A1 atoms go into solution at the same rate until the solution becomes saturated with silica. Then A1 alone dissolves without the structural framework breaking down. These authors noted that the rate of dissolution of aluminosilicates in acid is proportional to their A1

363

SILICA IN SOILS

contents and reasoned that the dissolution of silica is the limiting process involved. Garrels and Howard (1959) briefly reviewed some of the early work on dissolution of silicate minerals in water at low temperatures. They cited work of Correns and von Engelhardt which showed that subjecting ground adularia to a continuous supply of water brought about continuous release of K, Al, and Si in “ionic” forms. The ratio of K, Al, and Si in the leachate was 10.4:1:3.6 rather than the theoretical ratio of 1:1:3; thus the outer layer of the residual particles must have been relatively rich in the most insoluble component, aluminum. Garrels and Howard reported that Tamm, and Correns and von Engelhardt found evidence by alternately titrating a feldspar suspension with HCl and KOH that the exchange of H+ or K + at the feldspar surface is essentially reversible. Nash and Marshall (1956b) studied the reactions of albite surfaces with salt solutions. They found that NH4C1 was much more effective than MgCIz in dispIacing Na+ from albite; NH4+ was fixed at the surface and MgC12 was not effective in displacing it. The dissolution of silica from albite bore no relation to the amount of Na+ released; more silica was dissolved at pH values of 4 and 7 than at pH 8.62. The authors outlined the differences in behavior between feldspar and other exchange surfaces. They interpreted their results as indicating that a zone of lattice adjustment extends from the unchanged feldspar toward the surface. The essential cellular structure of the feldspar, although somewhat disturbed or loosened, was thought to be still present in this zone. Their results suggested that the coordination relationships of feldspars are preserved in this disturbed outer zone, which apparently becomes increasingly accessible to cations and solvent molecules with nearness to the surface. Garrels and Howard (1959) prepared suspensions of muscovite and adularia smaller than 200-mesh, titrated the suspensions with KCl, and recorded pH as a function of KCl concentration and temperature. The results indicated that these minerals react with water to produce a surface film in which H + had displaced K+, that the reaction is reversible, and that it occurs rapidly. Data for feldspar at pH values below 10 were consistent with the simple reaction mechanism KOr HzO e H O r + O H - + K + ; above pH 10, the K + concentration exceeded the OH- concentration, so some other anion (possibly silicate) or negative surface must have been involved. At 25” C., H-feldspar was favored over K-feldspar until ( K + ) / ( H + ) exceeded 1oB or loxo;H-mica was favored over K-mica until ( K + ) / ( H + ) exceeded lo7 or 108. With increasing temperature, the H+ forms were less strongly favored. The

+

364

J. A. MCKEAGUE AND M. G. CLINE

authors considered the hydrolysis reaction as the first step in the changes of K-feldspar to mica and of mica to kaolin. The energy that would be released in the exchange of H + or K+ would account for a major part of the energy involved in the weathering of feldspars and mica. De Vore ( 1959) discussed in detail the nature of the ( O O l ) , ( O l O ) , and (100) surfaces of feldspar and showed that chains of tetrahedra from these surfaces could combine to form mica sheets without further transformation. He cited no direct evidence of chain-type breakdown of feldspars but reasoned from energy considerations that it is theoretically feasible. According to De Vore, chain-type breakdown would simplify the explanation of mica formation from feldspars, as due to the ordered arrangement of Si and A1 atoms (Loewenstein, 1954; De Vore, 1959), A1 and Si tetrahedra would be transported to the site of mica crystallization together. Hauser (1952) outlined Jenny’s ideas on the surface weathering of orthoclase. Jenny considered that HzO is attracted to unsatisfied surface A1 and Si atoms. According to Jenny’s hypothesis, polarization of these HzO molecules would be so strong that H + ions would be released, leaving Si-OH and Al-OH groups at the surface, The released H + ions would be attracted to surface 0 atoms giving -OH groups. Ion exchange between K+ ions of the lattice and H+ ions of water would occur. The solution would become alkaline, and surface A1 atoms would tend to coordinate more OH- ions and assume octahedral coordination. The surface would become unstable, and groups of silica tetrahedra and alumina octahedra would be liberated. Fredrickson ( 1951) proposed a mechanism of weathering of albite in which protons from “crystalline” HzO would migrate into the structure of albite and displace Na+ ions. The breakdown of the structure was thought to be due to high local concentration of charge around the small H+ ions. This mechanism has been criticized by McConnell (1951) because of several doubtful assumptions involved. Hydrothermal studies have yielded information on the dissolution of silicates, but the results cannot be interpreted readily in terms of what occurs at low temperatures. Morey and Hesselgesser ( 1951), Morey and Chen (1955), and Morey and Fournier (1961) found that treatment of feldspar at temperatures of 300” C. or more yielded solutions containing alka1i:silica:alumina in higher-than-theoretical ratios. In one experiment (Morey and Chen, 1955), analcite formed from albite; in another (Morey and Fournier, 1961), albite was partially altered to boehmite, paragonite, and amorphous material, and microcline was partially altered to muscovite. Brindley and Radoslovich (1956) studied the alteration of albite

SILICA IN SOILS

365

crystals and powder in 0.1 N HCl under hydrothermal conditions. The albite crystals altered partially to boehmite, but they retained their sizes and shapes. Alumina accumulated relative to silica in the altered crystals; thus, even in this acid environment, silica was lost in solution more readily than alumina. No evidence was detected of structural modification of the feldspar residue or of preferential orientation of the alteration product in relation to the feldspar. Fredrickson and Cox ( 1954a) studied the hydrothermal decomposition of albite; samples removed during the run from the bottom of the bomb contained nearly three times as much solids as did fractions removed from the top. This was thought to indicate that particles large enough to settle relatively rapidly had been released by albite. Electron micrographs of material from the cooled samples indicated 0.1-to 5micron fragments of distorted albite, a gel, and zeolite-like crystals. From a similar study with quartz, Fredrickson and Cox (1954b) concluded that fragments of various sizes separated from the crystal. Mosebach (1957) strongly criticized the conclusions of Fredrickson and Cox. They had assumed that withdrawals of samples had negligibIe effects upon pressure, temperature, and nature of phases within the bomb. Mosebach‘s calculations indicated that these assumptions were invalid. Secondary precipitation may have occurred on sampling. Mosebach‘s (1957) results indicated that during hydrothermal treatment quartz is dispersed at least to Si( OH), and probably partly to ions. Within the upper few feet of the earth’s crust, biological factors undoubtedly influence the decomposition of silicate minerals (Jacks, 1953; Iler, 19.55; Keller, 1957). Keller (1957) stated that plant roots, acting as a continuing source of H+ ions, play a part in the weathering of silicates. He noted that primitive plants are more effective than higher plants in weathering fresh rock, and he suggested that this may be due to stronger acidity adjacent to these plants or to chelating agents released by them. In parts of the Soviet Union, fertilization with silicate bacteria, Bacillus diceus, is a common practice (Cooper, 1959). According to Cooper, Alexandrov, who isolated the species on a nitrogenfree medium, believes that B . siliceus is an autotroph that derives its energy from decomposition of aluminosilicates in soil. Webley et al. (1W),using a plate method, found that a Pseudomonas species from soil produced 2-ketogluconic acid that dissolved powdered silicates, presumably by chelation. It is clear that the mechanisms of dissolution of silicates are still not fully understood, Several of the hypotheses mentioned are highly speculative, and results of experiments are subject to a variety of interpretations. Garrels ( 1959) emphasized that the “alteration of silicates

366

J. A. MCKEACUE AND M. G . CLINE

takes place through whole sequences of incongruent solutions.” The nature of the units released as silicate minerals decompose is still in question. Results of Correns and von Engelhardt (Garrels and Howard, 1959), Armstrong (1940), and Gastuche et al. (1960) suggest that the dissolution products are in true solution. Other work (Fredrickson and Cox, 1954a, De Vore, 1959) suggests that lattice fragments of various sizes as well as ions are released. Some quantitative information is available concerning the initial reactions of alkali ions of feldspars with water and with salt solutions (Nash and Marshall, 1956a, b; Garrels and Howard, 1959); less is known about the release of Si and A1 from these minerals. The extension of hydrothermal studies to lower temperatures may yield useful information. AND FORMS OF DISSOLVED SILICAIN N A W L C. CONCENTRATIONS WATERS AND IN SOIL SOLUTIONS 1. Silica in Natural Waters The nature and the concentrations of dissolved silica in various natural waters have been thoroughly discussed recently by Siever (1957) and by Krauskopf (1956, 1959); thus, only a brief summary is included here. Clarke (1924) assembled a vast amount of data on the composition of dissolved matter in natural waters. Silica was determined gravimetrically, so the values quoted may include silica from colloidal-sized silica and silicate fragments as well as silica in true solution. His calculated approximate average composition of river waters of the world indicates that silica accounts for 11% per cent of the dissolved solids. The compositions of the Nile, Amazon, and Mississippi rivers were similar to the world average; near the mouths of these rivers, the silica concentrations were 17, 11, and 11 p.p.m., respectively. Clarke stressed the wide differences in concentrations of silica among samples from different rivers and among samples from different parts of the same river. Silica accounted for a large part of the dissolved solids in tributaries rising in igneous rocks; downstream the proportion of silica was much lower, and in some streams, the concentration was lower. Kobayashi (1960) determined silica in the waters of 225 rivers in Japan. The mean dissolved silica concentration (exclusive of that in suspended particles ) of those that flowed through regions of sedimentary rocks was 10.4 p.p.m.; those flowing through regions of volcanic rocks averaged 45.1 p.p.m. Murata (1940) noted that volcanic ash is a good source of dissolved silica. Siever ( 1957) summarized published data concerning silica in natural waters; the vast majority of analyses reported for river waters showed

SILICA IN SOILS

367

SiOz concentrations less than 35 p.p.m., and only two were supersaturated with respect to amorphous silica. Groundwater samples reported contained little SiOBin solution in the upper zones of the earth’s crust but as much as 50 or 60 p.p.m. in deep formations. Dapples (1959) suggested that this increase may be due to the increase of temperature with depth. Data reviewed by Siever showed that surface ocean water commonly contained SiOz at concentrations less that 1 p.p.m., and the highest value reported for deep ocean water was 12 p.p.m. Data for some alkaline lakes indicated high concentrations ( u p to 300 p.p.m.) of dissolved silica. White et al. (1956) found dissolved silica at concentrations of several hundred parts per million in some hot spring waters. With increasing distance from the spring, and thus lower temperature, the concentration of silica in solution decreased. White et al. reported that water somewhat less than 100 per cent supersaturated with respect to amorphous silica exists metastably at low temperature and low pH. In general, their results on the concentration and the polymerization of silica in hot spring water are consistent with the laboratory results of Alexander et al. (1954) and of Krauskopf (1956) on the solubility of silica. The concentration of dissolved silica in natural waters may be affected by salt content or by pH changes within the weakly acidic to weakly alkaline reaction range of most waters. Kovda et aE. (1958) found dissolved silica concentrations of 30 to 40 p.p.m. in slightly acidic runoff water in the foothills of the Amur area, 5 to 20 p.p.m. in nearly neutral water on lower terraces and 2 to 10 p.p.m. in saline surface water on the lowland. Bien et al. (1958) considered that salts in ocean water as well as material in suspension are involved in the inorganic precipitation of silica from river water entering the sea. There is little doubt that most of the dissolved silica in natural waters is in true solution as monosilicic acid (Siever, 1957, 1962; Krauskopf, 1956, 1959). This has been shown by analyses of waters both by the method of Alexander et al. (1954) for determining monomeric silica and by a method for determining total silica. Hypotheses involving the assumption of colloidal silica in natural waters are, in general, not valid. There are, however, exceptions to these generalizations; for example, Iwasaki et 02. (1951) found that much of the silica in pool water from a hot spring was in colloidal form. These authors also cited evidence that an appreciable portion of the silica in river water they analyzed was present as colloidal-sized mineral fragments and that some colloidal silica was present. This is at variance with most of the results cited by Roy (1945) and by Krauskopf (1959). Krauskopf ( 1956) reasoned that colloidal silica and silicate fragments could exist temporarily in waters

368

J. A. MCKEAGUE AND M. G. CLINE

undersaturated with respect to amorphous silica because the dissolution of these substances occurs slowly.

2. Silica in Soil Solutions Siever (1957) commented on the difficulty of getting a clear picture of the concentration of dissolved silica in soil waters. He attributed this to the difficulty of sampling soil solutions and to the lack of good analytical data. Many data on the composition of soil solutions are available, but few of these include silica values. Some of the silica data are from gravimetric determinations and are difficult to interpret as they include any colloidal-sized silica or silicate fragments that were not removed by filtration. The soils literature ( Polynov, 1937; Reifenberg, 1938; Mohr, 1944; Kubiena, 1953; Vilenskii, 1960; Karpachevskii, 1960, among many) commonly assumes that silica in soil solutions is colloidal. No valid direct evidence has been found to support this contention among either older data or recent information (Raupach, 1957; McKeague and Cline, 1963a). Lysimeter studies have yielded some data on the silica concentrations of soil solutions. Joffe (1940) reported on a seven-year study of the composition of the lysimeter leachates that passed through the Al, the Al and A2, and the Al, A2, and B1 horizons of a Gray-Brown Podzolic soil in New Jersey. The average annual translocation of silica, in pounds per acre, through specific horizons was: Al-8.8; Al and A2-2.8; Al, A2, and B1-l.6. From the data, it is possible to calculate that the SiOz concentrations of solutions that passed through the Al horizon, the Al and A2 horizons, and the Al, A2, and B1 horizons averaged 2.5, 3.5, and 7.0 p.p.m., respectively. Demolon and Bastisse ( 1942) found SiOz concentration of 9.4 p.p,m. in the leachate from the A horizon of a silty soil and 6.0 p.p.m. in the solution that had passed through both the A and the B horizons. The concentration of soluble silica in the leachate from a young granitic soil was 6.3 p.p.m. Remezov (1958) analyzed the leachate from a quartziferous sandy loam overlying a silty clay layer 1 to 2 meters below the surface. Over a four-year period, the concentration of Si ranged from 0.2 to 2.0 p.p.m. and the concentration of A1 varied from 5 to 24 p.p.m. Though those figures indicate Si:A1 ratios widely different from the 4 : l values calculated from Clarke’s (1924) data for the average of river waters of the world, lysimeter results suggest that the concentration of dissolved silica in soil solutions is low and is similar to that of river waters. Results of Harrison’s (1933) gravimetric analyses of silica in percolating water and river water of British Guiana support this conclusion. Silica concentrations of 8.5, 12.0, and 34.5 p.p.m. for drainage water from a downslope sequence of soils in the Karelian

SILICA IN SOILS

369

Isthmus also are consistent with the preceding generalization ( Rozhnova and Schastnaia, 1959). Joffe (1949) and Ponnamperuma (1955) cited evidence that poor drainage and low Eh are associated with relatively high concentrations of silica in solution, This may be due to the reduction of iron that is combined with silica and, hence, the liberation of silica. High silica concentrations may be found in solutions of strongly alkaline soils. Kelley and Brown (1939) reported that the silica content of a 1:5 (soi1:water) extract of an alkaline surface soil from Oregon was 12.5 milliequivalents per liter (as SiOs--) or 375 p.p.m. (as SiOz). The pH of the extract was 10.4; this is well above the reaction range of most soils, but it is not unique among soils of the world. Kelley and Brown found that the sum of anions in milliequivalents balanced the sum of cations in milliequivalents only when the soluble silica was considered as SO3--; this is not consistent with the ionization constants of monosilicic acid ( Iler, 1955). Determinations of dissolved silica in aqueous extracts of soils have yielded information about factors that affect the concentration of silica in soil solutions. Results from such studies must be interpreted carefully, however, as the method of preparing the extract has a marked effect upon the data obtained. For example, McKeague and Cline ( 1963a) using medium- and coarse-textured samples, showed that equilibrating soil-water mixtures by shaking resulted in concentrations of dissolved silica several times as high as those obtained in similar soilwater mixtures that were equilibrated by perfusing or standing. Precise solubility data are not obtained by treating soil with water under specified conditions; according to Garrels (1959) the breakdown of silicate minerals in soil would take place through sequences of incongruent solutions. Raupach (1957) published Hutton's data concerning the silica content of saturation extracts of a large number of soil samples from Australia. The samples varied widely in texture, pH, and other properties. A plot of dissolved silica concentration against pH showed that, in general, silica decreased with increasing pH. Raupach stated that the average concentrations of dissolved silica in saturation extracts at pH 6 and pH 9 were 6 X lo-' M (36 p.p.m.) and 1 X M ( 6 p.p.m.), respectively. This relationship between pH and silica concentrations of soil extracts is supported by data of Whitney and Peech (1952), Woodruff (1954), Raupach ( 1957), and McKeague and Cline (1963a). Woodruff determined silica in solution in 1 per cent beidellite suspensions adjusted to different degrees of base saturation. The Si concentration increased from

370

J. A. MCKEAGUE AND M. G . CLINE

2.4 p.p.m. at 100 per cent base saturation to 18.1 at 0 per cent base saturation. Whitney and Peech (1952) determined the silica concentrations of dialyzates of a montmorillonite clay suspension adjusted to different degrees of Na saturation. The dissolved silica concentration decreased from 0.74 x M (44.4 p.p.m.) at pH 4.41 to about M (2.4 p.p.m.) when the pH of the clay phase was slightly 0.4 X above 8. Raupach (1957) obtained similar results in a similar experiment; he extended the data to pH 11 and found that the low point in the dissolved silica vs. pH curve occurred at pH 9. Hutton's data (Raupach and Piper, 1959) indicate, however, that the silica concentration of saturation extracts of soils continues to decrease as the pH increases beyond 9. McKeague and Cline (1963a) found that when samples of a given soil material were extracted with solutions of varying pH, the concentration of silica in solution was negatively correlated with pH. For a loam-textured sample, concentrations of silica in solution were: 8.7, 2.7, and 1.5 p.p.m. for extracts at pH values of 5.2, 7.3, and 8.3, respectively. Data of Baird (1952), MacLeod ( 1962), and McKeague and Cline (1963a), however, indicated that there is no general relationship between pH and silica concentration of aqueous extracts of different soils. McKeague and Cline (1963a) reported concentrations of monomeric silica averaging about 200 p.p.m. in suspensions of electrodialyzed 2:1 clays that had stood at room temperature for several years. These suspensions were supersaturated with respect to amorphous silica. The relationship between pH and dissolved silica concentration of extracts of a soil has important implications in soil genesis. It has generally been assumed that neutral to alkaline conditions are favorable for losses of silica from soils and that silica is lost slowly under acid conditions because it is relatively insoluble (Nikiforoff et al., 1948; Joffe, 1949; Mohr and Van Baren, 1954; Keller, 1957). This is the basis for the idea that neutral to alkaline conditions are favorable for laterite formation (Joffe, 1949; Harder, 1952; Mason, 1952). Woodruff (1954) clearly pointed out the fallacy of these ideas. Actually, the greater concentrations of silica in solution at low pH is consistent with the fact that lateritic soils are commonly acid. Baird (1952) determined the silica dissolved from various soils by boiling 1:2 soil-water suspensions for 5 minutes, by carrying out a soxhlet extraction for 6 hours, and by doing two successive 6-hour soxhlet extractions. The 5-minute boiling treatment yielded concentrations of silica in solution ranging from 3 to 47 pep.". with different soils. With a 1:4 soilwater ratio, a 6-hour soxhlet extraction yielded dissolved silica concentrations from 4 to 195 p.p.m. A second extraction released amounts of

SILICA IN SOILS

371

silica ranging from slightly less than half as much to almost as much as the first extraction. Silica concentration was positively correlated ( r = 0.71) with specific surface. Results of MacLeod (1962) and of McKeague and Cline (1963a) for a limited number of soils extracted at room temperature do not support this relationship. It is probable that specific surface of sample, like pH of extracting solution, is only one of several factors that control the concentration of silica in a soil extract. Several studies have shown that concentration of silica in solution is positively correlated with soil-to-solution ratio (Woodruff, 1954; Reifenberg and Buckwold, 1954; McKeague and Cline, 1963a). Woodruff used the clay fraction of a Putnam silt loam subsoil; the dissolved SiOz concentration of a 1 per cent suspension was 6.9 p.p.m.; of an 8 per cent suspension, 31.1 p.p.m.; and of a 64 per cent suspension, 52.3 p.p.m. Reifenberg and Buckwold (1954) extracted soil with 1/3 M phosphate solutions; as the soil-to-solution ratio was increased, the concentration of silica in solution increased, but the release of silica per gram of soil decreased. McKeague and Cline ( 1963a) prepared different soil-water ratios of several soil samples and determined dissolved silica periodically. Concentrations of silica in solution increased with increasing soil to solution ratio. In low-ratio mixtures, concentrations of dissolved silica continued to increase gradually even after 2 weeks; in high-ratio mixtures of some soils, the dissolved silica concentrations reached a maximum value after about 2 days and then declined to almost constant values. The effects of various salt solutions on the dissolution of silica from soils and clays have been studied. Reifenberg and Buckwold (1954) found that phosphate solutions effected a significant release of silica from soils and clay minerals; silica concentrations ( determined gravimetrically) in some of the extracts exceeded 2.00 p.p.m. Solutions of sodium chloride, nitrate, sulfate, borate, hypobromite, and acetate were little different from pure water in their silica-dissolving capacities. Low and Black (1947, 1950) observed that when phosphate solutions were added to kaolinite, phosphate was fixed and silica was released. Other substances, such as 8-hydroxyquinoline, that form stable complexes with aluminum had similar effects. Low and Black (1947) considered that any substance that would reduce the activity of aluminum would bring about dissolution of clay and, hence, release of silica. Presumably chelating substances formed from organic matter in soil play such a role. McKeague and Cline (1963a) extracted soil samples with 0.01 and 0.001 M solutions of several salts and with water. Neutral salts such as NaCl had little effect upon the dissolution of silica. Low silica concentrations were found in the NaHC03 extracts;

372

J. A. MCKEAGUE A N D M. G. CLINE

this is considered to be due to the pH increase effected. Phosphate solutions brought about a marked increase in dissolution of silica from soil samples but not from quartz; concentrations of dissolved silica were highest in phosphate solutions of low pH. The concentration of dissolved silica in soil-water mixtures and presumably in natural soil solutions increases with increasing temperature ( McKeague and Cline, 1963a). Limited data indicated an almost linear relationship; however, coefficients were not the same for all soil samples. Thus, it would not be possible to predict silica concentrations of natural soil solutions from data obtained by extracting soil samples at elevated temperatures. Frankart et al. (1960) stated that the “solubility” of silica in aluminosilicates increases linearly with temperature from about 20 p.p.m. at 20°C. to about 40 p.p.m. at 100°C. Presumably other factors would be involved in determining the apparent solubility-nature of mineral, size of particles, etc. McKeague and Cline (1963a) studied the effect of evaporation on the concentration of dissolved silica in soil-water mixtures. Suspensions containing 20g. of soil and 400ml. of water were shaken for a week, dissolved silica was determined, the suspensions were allowed to evaporate at room temperature to one-tenth or less of their original moisture contents, and dissolved silica was again determined. Results showed that during the period of evaporation, from one-half to three-fourths of the dissolved silica was removed from solution. All the silica that remained in solution was in monomeric form, and the concentrations were not greater than one-fourth that of a solution saturated with respect to amorphous silica. The results were interpreted, in the light of other experimental data ( McKeague and Cline, 1963b), as indicating that during the evaporation of water from a soil, dissolved silica is adsorbed at certain particle surfaces. Thus the concentration of silica in soil solution probably remains relatively low even at moisture contents well below field capacity. McKeague and Cline (1963a) studied the rate of dissolution of silica from three soil samples by adding enough water to saturate subsamples, allowing them to stand at room temperature for varying times, centrifuging, and determining silica in solution. Within 5 minutes, the concentrations increased to about half those found after 10 days. Concentrations of dissolved silica continued to increase rapidly for about an hour, but the rate of increase became progressively slower with time. This suggests that when water enters a soil, the concentration of dissolved silica increases rapidly and approaches a maximum value before the soil solution has penetrated far into the soil. Repeated extraction of

SILICA IN SOILS

373

soil samples with new aliquots of water results in decreasing concentrations of silica in successive aliquots, but the rapid initial dissolution of silica continues with each extraction ( McKeague and Cline, 1963a). In general, the results of these studies of the dissolution of silica from soil samples are similar to the results of the studies of the dissolution of silica from quartz and from silicate minerals. Many of these data are consistent with the hypothesis that disorganized surfaces, probably partially composed of adsorbed monosilicic acid, govern the apparent solubility relationships of silica in soils. Evidence is strong that most of the silica in soil solutions is in the form of monosilicic acid ( McKeague and Cline, 1963a). Dissolved silica in extracts of a wide variety of soil-water mixtures equilibrated in a variety of ways was present in all cases as monomeric silica according to the test of Alexander et al. (1954). Concentrations of dissolved silica reported in soil solutions (Joffe, 1940; Demolon and Bastisse, 1942; Raupach, 1957) are far below the solubility of amorphous silica; thus silica sol should not be stable in soils. Demolon and Bastisse (1942, 1944, 1945) and Bastisse (1949, 1960), however, consider that silica occurs in drainage waters and in soil solutions in the form of “pseudosoluble” complexes with iron and other metals. This interpretation, based largely upon the results of experiments with synthetic iron-silica preparations, is questionable, Concentrations of iron in solutions of aerated soils are so low that iron-silica complexes could not account for a significant proportion of the dissolved silica in soils. The solubility product of Fe( OH), is approximately (Latimer, 1952), so at pH 6.0, the concentration of Fe+++ would be of the order of moles per liter. OF DISSOLVED SILICAWITH OTHERCOMPONENTS D. INTERACTION OF SOIL

1 . Adsorption of Dissolved Silica by Soils and Other Solids Data from the literature of industrial chemistry indicates that dissolved silica in natural waters [probably Si( OH),] is adsorbed by certain substances that may occur in soils. Behrman and Gustafson’s (1940) summary of the subject shows that successful methods of removal of silica from industrial waters involve the use of freshly precipitated iron, aluminum, or magnesium hydroxides. Naturally occurring oxides, such as limonite and bauxite, were found to have little capacity for removing silica; freshly precipitated CaC03 and dolomitic lime were said to be even less effective. Betz et al. (194Oa) found that drying or aging freshIy precipitated aluminum hydroxide resulted in a marked decrease of its capacity to adsorb dissolved silica.

374

J. A. hdC KEAGUE AND M. G . CLINE

The adsorption of dissolved silica from river water is sensitive to pH. Betz et al. (1940a) found that freshly precipitated aluminum hydroxide was most effective in the pH range 8.3 to 9.1. Lindsay and Ryznar (1939) recommended a pH between 8.3 and 8.7. Magnesium hydroxide adsorbed silica most effectively at pH 10.1 (Betz et al., 1940b). The optimum reaction for adsorption of silica by freshly precipitated ferric hydroxide was reported to be 9 (Behrman and Gustafson, 1940). Temperature affects the removal of silica from solution by the adsorbents mentioned. Freshly precipitated aluminum hydroxide adsorbed progressively less silica as the temperature was increased throughout the range 23°C. to 95°C. (Betz et al., 1940a). Temperature had the opposite effect on the removal of dissolved silica from water by MgO (Betz et al., 1940b). At 95°C. the operational temperature, 1 g. of MgO removed from solution as much as 0.3 g. of SiOp (Betz d al., 1941). Clearly, some solids are very effective in removing silica from dilute solutions; the mechanism by which the process occurs is, however, not known with certainty. Betz d al. (1940b) showed that the removal of silica from solution by MgO is not a stoichiometric reaction involving the formation of MgSi03. Data for the removal of silica from solution by aluminum hydroxide (Lindsay and Ryznar, 1939) and MgO (Betz et al., 1941) follow Freundlich‘s adsorption isotherm. According to Trapnell ( 1955), however, proof that a reaction conforms to this isotherm would require data showing that the heat of adsorption declines logarithmically with degree of surface coverage; such data are not available for reactions involving the adsorption of silica from solution. The increase in adsorption of silica by MgO at higher temperatures is not consistent with the fact that adsorption reactions are, theoretically, exothermic. Adsorption may, however, increase with temperature over a certain interval (Trapnell, 1955). According to McBain (1950), adsorption from solution is usually not greatly affected by temperature. Giles and MacEwan (1957) and Giles et al. (1960) pointed out that adsorption of solutes from solution is a complex, little-understood phenomenon. Cassidy’s ( 1951) operational definition of adsorption would certainly include the removal of dissolved silica from industrial waters by aluminum hydroxide, MgO, etc. He stated (Cassidy, 1951, p. l ) , “a substance is said to be adsorbed if its concentration . . . in a boundary region is higher than in the interior of the contiguous phases. The definition of adsorption rests upon measurements of concentration, and carries no implication of mechanism.” Cassidy noted that in cases involving adsorption from solution, evidence of adsorption is usually obtained by showing that the solution has become less concentrated. This meaning is intended by “adsorption” in this paper.

SILICA IN SOILS

375

Krauskopf's (1956) results on the interaction of dilute solutions of monomeric silica with various solids indicate that added solids have little effect upon the concentration of dissolved silica. He found that suspended solids such as calcite, hydrous iron oxide, kaolinite, and montmorillonite did not change the concentration of silica in true solution. Krauskopf (1959) noted that aluminum oxide removes silica from solution and he suggested that the apparent specific effect of aluminum may indicate that A13+ reacts slowly with silicic acid to form clay minerals. This may be so, but Behrman and Gustafson (1940) and McKeague and Cline (1963b) have shown that aluminum compounds are certainly not the only substances that have marked capacity to remove silica from dilute solutions. McKeague and Cline (1963b) tested the capacities of a variety of solids for removing monomeric silica from solutions at concentrations below 100 p.p.m. They found that freshly precipitated hydroxides of polyvalent metal ions such as Fe3+, A13+, Ni++, and C o f f as well as ground brucite were most effective; some soil samples and iron oxide minerals were moderately effective; alkaline earth carbonates and the silicate minerals tested were ineffective in removing silica from solution. The results indicated that an adsorption reaction was involved, as removal of silica from solution increased as the surface of the added solid was increased. Data on the interaction of monosilicic acid with a ferruginous oxide clay sample from an Oxisol followed Freundlichs adsorption isotherm. The approximate adsorption equation at pH 4.8 and 25°C. was, y = 160 Co.O; at pH 8 and 25°C. it was, y = 740 CO.',where y is micrograms of SiOz adsorbed per gram of soil, and C is the equilibrium concentration of monosilicic acid ( as SiOz p.p.m. ). Adsorption of dissolved silica by all the soil samples tested increased with increasing pH at least up to pH 9. For example, 2-g. samples of soil were added to 30ml. of monosilicic acid solutions buffered at various pH values. With Vergennes clay and an initial SiOz concentration of 28 p.p.m., the final concentrations of SiOz were 23.2, 11.0, and 6.8 p.p.m. at pH values of 6.2, 7.7, and 9.0, respectively. The authors interpreted the results as indicating that a pH-dependent adsorption reaction plays a major role in controlling the concentration of silica in soil solutions and the adsorption of Si( OH) is one of the reactions contributing to the development of amorphous coatings on soil particles. Evidence was presented to support the contention that, although the adsorption of silica is pH dependent, deposition of silica at the site of a pH increase in soil would not be appreciable. Results of experiments of Bien et al. (1958) on the deposition of silica from river water entering the sea indicated that dissolved silica may be

376

J. A. MCKEAGUE AND M . G . CLINE

adsorbed by suspended matter in river water and that this may be flocculated by electrolytes in ocean water. The decrease in concentration of “dissolved silica when river water was mixed with sea water could not be explained solely by the dilution involved, and it occurred too rapidly to be due only to biological processes such as the growth of diatoms. Thus, evidence from a variety of sources shows that dissolved silica is adsorbed by soil material and by substances similar to some of those occurring in soils. Since the ratio of solid to solution is high in soils, the influence of adsorption on the concentration of dissolved silica in soils is probably greater than that suggested by results of experiments ( McKeague and Cline, 1 W b ) involving low adsorbent-to-solution ratios.

2. Reactions of Dissolved Silica with Other Solutes in Soil Some general discussion of the reactions of dissolved silica was included in the section dealing with the chemistry of silica. This section is limited to a review of the meager literature concerning the interaction of dissolved silica with other components of soil solutions. Literature concerning possible reactions of silica sol in soils (Mattson, 1928; Reifenberg, 1938; Kubiena, 1953) is not considered because evidence already presented indicates that silica sol is not a common component of soil solutions. Furthermore, Khan ( 1960) showed, contrary to Reifenberg’s (1938) contention, that silica sol does not peptize iron oxide. The combination of dissolved silica with aluminum in soil solution was suggested (Raupach, 1957) as a possible explanation of the observed negative correlation between pH and silica concentration of soil extracts. Raupach and Piper (1959) considered that this view was supported by results of Okamoto et al. (1957)) who studied the effect of pH on the precipitation of silica by various concentrations of aluminum sulfate. Okamoto et al. found that A1 at 20 p.p.m., the lowest concentration used, reduced the concentration of silica in solution from 35 p.p.m. to about 15 p.p.m. when the pH was in the range of 8 to 9; silica in solution increased as the pH was varied above or below this range. The usual soluble A1 concentration found in extracts of near-neutral soils is, however, less than l p.p.m. (Magistad, 1925; Frink, 1960) rather than 20 p.p.m. McKeague’s (1962) data on the concentrations of dissolved aluminum and silica in extracts of various soils and on the interaction of very dilute solutions of aluminum and silica indicate that aluminum in soil solutions has little influence on the concentration of dissolved silica. A possible explanation of the results of Okamoto et al. (1957) is that aluminum hydroxide precipitated from the highly supersaturated solutions and that Si( OH)a was adsorbed by or coprecipitated with it.

SILICA I N SOILS

377

Direct information concerning the interaction of dissolved silica with other solutes in soils is not available. The possibility of combination of Si( O H ) with soluble organic substances was mentioned previously. A consideration of the solubility products of calcium silicates ( Greenberg et al., 1960) shows that in most soils precipitation of these compounds is highly improbable. In general, it appears that the chemistry of dissolved silica in soils involves heterogeneous reactions between solutions and solid surfaces rather than homogeneous reactions in solution. IV. Deposition of Silica in Soils

Dissolved silica in soils is an active solute; it is not a passive component that is inevitably lost in drainage water soon after it is released by the weathering of primary rocks. Silica is deposited in soils, probably mainly by adsorption on surfaces of various kinds of inorganic compounds, including silica, silicates, and nonsiliceous substances, and by the decay of plants that accumulate silica.

A. BY INORGANIC MECHANISMS 1. As SiOz Precipitates and Accretions Secondary quartz is frequently found in soils (Jeffries, 1937; Nikiforoff and Alexander, 1942; Fieldes and Swindale, 1954; Carroll and Hathaway, 1954; Breeze, 1960) and in sediments (Gilbert, 1949; Pettijohn, 1957; Siever, 1957, 1959), but it is not known whether this quartz was formed directly from silica in solution or whether it is an organized product of amorphous silica. According to Siever (1957), formation of quartz from solution at low temperature must take place extremely slowly. Breeze (1960) considered that the development of secondary quartz overgrowths on primary quartz grains is an important mode of deposition of silica in soils. Amorphous silica is apparently deposited by inorganic means in some soils. Kovda et al. (1958) found white siliceous powder and opaline incrustations on pebbles to a depth of 6 feet in “meadow soils” of an arid area. They attributed this partly to deposition of silica due to evaporation of incoming drainage waters. Szabolcs and Darab (1958) stated that an accumulation of amorphous silica is frequently observed on the Hungarian plains, especially in areas of solonetzic soil; in some cases it occurs as a white dust coating the soil. The proportion of biologically deposited silica in this dust was not indicated. Vamos (1961) considered that the accumulation of amorphous silica in surface horizons of degraded alkali soils is related to flooding. He reasoned that microbial activity following flooding could cause anaerobic conditions and sulfide

378

J. A. MC KEAGUE A N D M. G. CLINE

formation. With the return of aerobic conditions, sulfides would be oxidized, producing sulfuric acid; this strong acid would attack silicates and liberate silica, Generally, it would appear that precipitation of silica, in the sense that soil solutions became saturated with respect to amorphous silica and deposit silica in discrete bodies large enough to be identified, would require special circumstances. It has been shown ( McKeague and Cline, 1963a) that silica in the liquid phase of soil suspensions is in monomeric form and is maintained at concentrations far below the solubility of polymeric silica even after evaporation to the most concentrated suspension from which liquid could be extracted conveniently. Silica was removed from solution, apparently by adsorption on the solid phase of soil as water was evaporated. Presumably, if the liquid phase was evaporated to air dryness, the silica concentration of some soil solutions might exceed the solubility of polymeric silica. At the concentrations measured in these experiments, the amount of silica left in the most concentrated final solution was approximately 3 x 10-2 mg. per cubic centimeter of soil of about 1.3 bulk density. Some, perhaps most, of this would have been “adsorbed as drying continued if the trends of the experiment had continued. The amount precipitated in a single drying cycle would be extremely small at any single site in a soil. Nevertheless, this may be a mechanism by which siliceous powder and opaline incrustations, such as those described by Kovda et al. (1958) and Szabolcs and Darab (1958), are formed in soils. It would require that the layer be subject to wetting and drying cycles, that silica that is redissolved upon rewetting not be completely removed by leaching, and that nuclei of silica formed in this manner would have greater affinity for silica in solution than would other components of the solid phase. It appears likely that either crystalline or amorphous nuclei might grow slowly by accretion to sizes that can be detected. This might occur even at concentrations of the soil solution below the solubility of amorphous silica. Incrustations at depths not normally subject to wetting and drying cycles must involve some other mechanism. It is possible that, in zones where water stands or moves very slowly over large distances, concentrations of silica may reach the solubility of polymeric forms. Data on the silica concentration of groundwater ( Siever, 1957; Dapples, 1959), however, do not support this.

2. As a Cementing Agent Certain soil horizons are commonly assumed to be, and probably are, cemented at least partially by silica. This conclusion is based more,

SILICA IN SOILS

379

however, upon conjecture than upon experimental data. Large areas of soils with pans thought to be cemented by iron, silica, silica and iron, or silica and lime, occur in the United States (Nikiforoff, 1937; Soil Survey Staff, 1960). Indurated horizons cemented in part by an agent that is soluble in concentrated alkali are called “duripans” by the Soil Survey Staff (1960, p. 55); the cement is presumed to be silica or an aluminum silicate. Three kinds of duripans are mentioned: (1) those that are cemented partially by CaC03 as well as by silica, ( 2 ) those that are presumably cemented by alternating layers of iron and silica, and (3) those that occur in albic (A,) horizons of certain Humods (Humic Podzols) and are thought to be cemented by silica or by aluminum and silica. Nikiforoff and Alexander (1942) concluded from chemical and petrographic studies of a hardpan from California that it was cemented by iron oxide and silica that had been liberated by weathering at the point of cementation. Winters (1942) considered that a type of pan, later to be called a fragipan, in the Red-Yellow Podzolic soil region was silica cemented; he had no direct evidence of this, Knox (1957) investigated the basis of the rigidity of fragipans; treatment of clods from several soils with a variety of solvents indicated silica as a probable cement of only one of the fragipans studied. Mohr’s (1944) theory of the formation of laterite from volcanic ash involves the assumption that silica is precipitated at depth to form a siliceous hardpan that restricts drainage. Details of the cause of silica precipitation and of the nature of the siliceous pan were not given. Breeze (1960) suggested that overgrowths of secondary quartz on primary quartz particles may act as cementing agents. Cementation of sedimentary rocks by silica may be related to deposition of silica as a cement in soils. Pettijohn (1957) concluded that silica is the most common cement of sedimentary rocks; quartz is the form usually involved, but opal and chalcedony also occur as cements (Goldstein, 1948; Gilbert, 1949). Secondary quartz often occurs as enlargements on detritial quartz grains. According to Pettijohn (1957) the mode of origin of silica cement in sandstone is not yet clearly established. Siever (1959) considered that quartz may form directly from silica in interstitial water in sediments or that it may be biologically precipitated as amorphous silica, dissolved, and deposited as quartz. Silcrete, which occurs as a prominent surface feature of some old landscapes in arid regions of Australia and South Africa, is a striking example of silica deposition. Williamson (1957) reported that the SiOz percentage of silicified sedimentary rock in Australia, including silcrete, ranges from 95 to more than 99. Whitehouse (1940) described layers of

380

J. A. M C KEAGUE AND M. G. CLINE

silica-cemented material, known locally as “billy” if sandy or as “porcellanite” if clayey, beneath laterite in Australia. Jessup (1960) has concluded that silicified layers within “pallid zones” of profiles presumed to have once contained laterite are comparable to the silica-cemented layers described by Mohr ( 1944). According to Stephens ( 1957), silcrete occurs as a subsurface feature of some lateritic soils in more humid regions as well as in arid regions of Australia. He postulated a connected genetic history for the formation of laterite and silcrete, silica being retained at ever shallower depths in the profile with increasing dryness of the environment. The mechanism of formation of silcrete is not clear although several suggestions have been made ( Stephens, 1957; Williamson, 1957; Jackson, 1957; Jessup, 19so). Silcrete could form by cementation of sand by upward-moving, silica-bearing water. Exposure of the silicified zone of a laterite profile by erosion could result in surface sikrete. Jackson (1957) considered the fact that cappings of silcrete occur in arid areas but not in humid areas as good evidence that they are products of an arid climate rather than the result of a former more humid climate.

3. As a Component of Secondary Silicates It is generally accepted that a variety of layer silicates are synthesized in soils, and something is known of the soil-forming factors that are favorabIe for the formation of such minerals (Jenny, 1941; Keller, 1957; Barshad, 1959). Opinions vary considerably, however, about the relative importance of clay formation in situ as a factor in development of zones of apparent clay enrichment. Relative accumulation of clay due to losses of other soil constituents and clay translocation are other potential contributing mechanisms. Barshad ( 1955) outlined methods of calculating the amount of clay formed in the various horizons of a soil. Rich and Thomas (1960) recently discussed the origin of clay minerals in soils. No attempt is made here to review this vast subject; it is pertinent, however, that some of the silica that is released to solution during the decomposition of primary silicates is probably a source of silica of some secondary layer silicates found in soils. It was indicated in a previous section that there is uncertainty about the nature of the units released as primary silicate minerals weather; thus, it is not surprising that details are not known of the mechanisms of formation of secondary siliceous substances in soil. Clay minerals have been synthesized from dilute solutions at low pressure and at temperatures of 100°C. or less (Hknin, 1954, 1957; Callikre & al., 1956; DeKimpe and Gastuche, 1960). The nature of the silicon and aluminum units that combine is not known, although Calliere et al. (1958) considered that

SILICA IN SOILS

381

precipitation of aluminum hydroxide and subsequent fixation of silica is involved. Rich and Thomas (1960) stated that many studies indicate that clay minerals form by an ionic mechanism. Barshad (1955) suggested that the formation of clay minerals from soluble and colloidal materials is catalyzed by other aluminosilicates; he proposed that the reactants are adsorbed at specific sites on the surface of an aluminosilicate and are consequently oriented in a manner related to the structure of the mineral. Condensation of the adsorbed material through dehydration could produce more or less regularly linked silica tetrahedra and alumina octahedra. Hauser (1952) thought that clays form by the alteration of silica and alumina or magnesia gels; De Vore (1959) suggested that surface chains separated from primary silicates may combine to form clays. It may be that the apparent controversy about the mechanism of formation of secondary silicates in soils is largely a matter of semantics. If the component solutes are adsorbed at solid surfaces and have some degree of molecular organization at the time of adsorption and during dehydration, it could be argued either that secondary silicates form from substances in true solution or that they form by rearrangement of amorphous material at particle surfaces. There is need for investigation of the molecular arrangement of adsorbed material on particle surfaces in soil and for a study of the possible catalytic effects of surfaces on the formation of secondary silicates. Amorphous silicon-containing substances as well as crystalline layer silicates form in soils. Much of the colloidal material in some soils is amorphous (Jackson, 1956; Toth, 1955; Rich and Thomas, 1960). Allophane, an amorphous hydrous combination of silica and alumina, has been discussed, and palagonite, an amorphous iron-aluminum silicate ( Birrell and Gradwell, 1956) has been mentioned. The adsorption of monosilicic acid by ferric oxides and hydroxide ( McKeague and Cline, 1963b), and the appreciable dissolution of silica from soil material treated for freeiron removal with a dithionite-citrate solution buffered at pH 7.3 (Mehra and Jackson, 1960) support the contention (Jackson, 1956) that amorphous combinations of iron and silica occur commonly in soils. It may be that amorphous or weakly crystalline calcium silicates are formed in some soils. Taylor and Howison (1956) discussed the relationships between tobermorite, a hydrated calcium silicate, and the clay minerals. According to these authors, tobermorite minerals vary widely in crystallinity from perfect crystals to amorphous material formed in the setting of Portland cement. Such minerals have not been reported in soils to the knowledge of the authors; however, they may form in some alkaline soils. According to Greenberg et al. (1960), the sohbility prod-

382

J. A. MC KEAGUE AND M. G . CLINE

ucts of colloidal hydrated calcium silicates at 25°C. are: ( C a + + ) ( H2Si04--) = and ( C a + + ) ( H3Si04-)Z= 10-s,5. From a consideration of these solubility products, the ionization constants of monosilicic acid, and the concentrations of silica in soil solutions, one would expect the precipitation of hydrated calcium silicates only in strongly alkaline soils. The hypothesis that gibbsite and other insoluble oxides and hydroxides form as direct weathering products of basic rocks in tropical regions and are subsequently resilicated to form kaolinite and other siliceous substances has been discussed recently by Sivarajasingham et nZ. (1962). These authors reasoned that contrasting crystal lattices of gibbsite and its presumed resilication products implies complete destruction of the crystal framework and, in effect, reactions involving independent units of aluminum and silicon. They also stressed the probability that disorganized A1-0-Si amorphous material may be a precursor of the “resilicated mineral observed. Results of adsorption studies (McKeague and Cline, 1963b) indicate the feasibility of silication of hydrous oxides of iron and aluminum in soils. The extent of resilication presumably depends, among several factors, upon the chemical composition and degree of crystallinity of the various adsorbents in soils and upon the concentration of dissolved silica maintained in soil solution.

B. BY ORGANISMS Probably one of the major modes of deposition of silica in the terrestrial environment is through the action of higher plants. Russell (1950) stated that probably two times as much silica is cycled annually through plants as is lost in drainage water. The silica content of different plants and of various parts of the same plant are highly variable (Russell, 1950; Iler, 1955; Lovering, 1959; Yoshida et al., 1959a,b). The wood of most temperate-zone trees contains essentially no silica, whereas the hollow stems of bamboo contain masses of almost pure silica gel (Iler, 1955) and many tropical trees contain several per cent silica on a dry-weight basis. About 10 per cent of the dry weight of some species of Equisetum (horsetail) and of some grasses is silica (Lovering, 1959). When such vegetation decays, much silica is returned to the surface of the soil. Amorphous silica in the form of opal is apparently the main silicon compound in plants. Baker (1959) defined opal phytoliths as “minute bodies of isotropic silica which have been precipitated as unwanted material or as reinforcements of cell structures in some plants, including grasses, sedges, reeds and some woods.. . .” Opaline phytoliths in plant tissues and in soils have been studied for more than 100 years (Smithson, 1958);they sometimes make up as much as 1 to 2 per cent of the weight

SILICA IN SOILS

383

of the surface soil under grass (Beavers and Stephen, 1958; Baker, 1959). Riquer (1960) described an unusual soil formed in basaltic ash in the Reunion Islands; a horizon 5 to 30 cm. in thickness beneath the organic horizon was composed of nearly pure opal phytoliths. Information is meager about silicon in plants in forms other than opal. Lanning et al. (1958) found alpha quartz as well as opal in lantana. Viehoever and Prusky (1938) and Engel (1953) reported evidence of the presence of silica-organic complexes in plants. Some Russian scientists believe that clay minerals are synthesized within plants (Jacks, 1953; Iler, 1955). Some attempts have been made to estimate the dynamics of addition and removal of biologically deposited silica in soil. Novorossova (1951) estimated that 50 to 60 tons of silica is returned to a hectare of soil by litter of spruce forest in lo00 years. She suggested that this may be a factor in the silica enrichment of the A2 horizon of a Podzol. Riquer (1960) reported few opal phytoliths in A2 horizons of Podzols of temperate regions; he suggested, however, that some of the chalcedony and quartz present may have been derived from phytoliths. Baker (1959) reasoned from the silica content and the dry matter production of grass and from the percentage of phytoliths in some surface soils in Australia that the duration of the cycle of these opaline bodies must be about 1000 years. Lovering (1959), considering the high silica content of many tropical trees, suggested that the depletion of silica and the enrichment of sesquioxides in lateritic soils may be effected to a considerable extent by the accumulation of silica by vegetation and the removal of fallen litter by erosion before the silica is released. As an example, he considered a tropical forest of silica-accumulating plants averaging 2.5 per cent silica and 16 tons dry weight growth per acre per year. In 5000 years, 2000 tons of silica would be removed from an acre of soil by the forest. Lovering calculated this to be equivalent to the silica in an acre foot of basalt. He stressed the fact that data were not available that would make possible a calculation of the amount of silica removed by erosion of fallen litter relative to the amount of silica returned to the soil by decaying vegetation. Some lower plants are highly effective in depositing silica; for example, diatoms deposit opaline silica from water that is far below saturation with respect to amorphous siIica (Iler, 1955; Lewin, 1961). Freshwater diatoms inhabiting the capillary water of soils contribute their siliceous remains to some mineral soils ( Smithson, 1956). Layers consisting largely of diatom remains occur in some organic soils (Dawson, 1956). The mechanism by which diatoms deposit silica from very

384

J. A. MC KEAGUE AND M. G . CLINE

dilute solutions is not understood; Lewin (1961) outlined several opinions on the subject. Her studies indicate that adsorbed cations such as Fe3+ and A13+ act as a protective coating. She found, however, that the silica of living cells was less soluble than that of killed cells; this finding suggested that some additional mechanism, as well as adsorption of cations, is involved in preventing the dissolution of the opaline silica of diatoms in undersaturated solution. Deposition of amorphous silica in soils occurs also through the agency of lower animals. Smithson (1959) stated that opaline sponge spicules occur widely in soils; some of these are derived from in situ deposition in wet areas, others are transported. It is apparent that the biological deposition of silica proceeds on a grand scale in soils. As pointed out by Siever (1957) this is a method by which amorphous silica may be deposited from very dilute solutions. Biological deposition of amorphous silica as well as adsorption of dissolved silica by various solids contributes toward maintaining the generally low concentrations of silica in soil solutions, V. Silica in Relation to Kinds of Soils

The orders of the new system of soil classification of the United States are used as the basis of this discussion. The orders are classes of the highest category and have been formed to group together those soils whose properties indicate that the kinds and strengths of processes tending to develop horizons have been similar (Soil Survey Staff, 1960). Neither form nor content of silica, as such, is diagnostic of the orders. Both are related in varying degrees to properties of soils and to soilforming processes. Both are covariants of criteria of some classes. Exceptionally high concentrations of resistant minerals, including quartz, and of volcanic glass in sand fractions are diagnostic of some classes at the suborder, great group, and subgroup levels. “Allophane,” “weatherable minerals,” and the proportions of oxide and 1:l silicate clays are also used directly to differentiate specific classes at high categorical levels. Mineralogy as such, however, is used as a diagnostic criterion mainly at the level of soil families. Nevertheless, forms, amounts, and properties of silica and silicates are related through soil properties and soil genesis to kinds of soil differentiated at high levels in the system. Entisols include many soils that have been called Alluvial soils, Regosols, Lithosols, Tundra soils, and Low-Humic-Gley soils in the United States (Soil Survey Staff, 1960, p. 105). They are very young pedogenically and either lack genetic horizons or have only the beginnings of such horizons. Consequently, the forms and amounts of silica in them

SILICA IN SOILS

385

are inherited directly from the parent material and range very widely. At one extreme, “Quarzopsamments” are required to have more than 95 per cent quartz or other resistant minerals that do not weather to release iron or aluminum. Many of these are very old geologically, but they are young pedogenically in the sense that their inherited mineralogy prevents differentiation of distinctive horizons. At the other extreme are deposits of fresh alluvium rich in primary silicates but without horizons because of extreme youth. The order can also include soil in highly weathered material low in silica and high in sesquioxides if the deposit is so young that the original depositional structure has not been destroyed. It can also include soils very high in silicate clays such as montmorillonite if well expressed genetic horizons have not formed and if the deposit lacks the evidence of self-mixing or self-displacement diagnostic of Vertisols. Vertisols include swelling clays that have been variously called Grumusols, Regur, and Black Cotton soils among several names (Soil Survey Staff, 1960, p. 124). They must contain more than 35 per cent expanding-lattice silicate clay, and generally they have more than 40 per cent. Montmorillonite is the dominant clay mineral in most cases (Simonson, 1954). They may, however, contain significant amounts of other silicates, including both silicate clays and primary minerals. Quartz may be a significant component. The diagnostic properties, however, are those that are evidence of self-dispIacement or self-mixing, and these are associated with expanding-lattice silicates. Some of these silicates have been inherited in an alluvial material; some have formed from country rock or sediment during weathering. There is some evidence that some that are found in basins may have formed by synthesis of expanding-lattice clays under the influence of water rich in bases and silica from adjacent uplands. Greene (1945) has described soils that qualify as Vertisols in such basins. Znceptisols include many soils that have been called Brown Forest, Tundra, Ando and certain Gley soils and Sols Bruns Acides, Lithosols, and Regosols (Soil Survey Staff, 1960, p. 136). They are soiIs that represent pedogenic youth in the sense that they have one or more distinctive horizons, but these are restricted to those that are believed to form quickly. They are soils of humid regions where moisture is available for weathering, but they cannot be extremely weathered. Consequently, the forms and amounts of silica are nearly all inherited. They may be inherited from young deposits like glacial drift, from residuum of material weathered in place, or from transported weathered material. Consequently, silica in the clay fraction may include that in kaolinite, montmorillonite, micas, or allophane as dominant constituents if the soil does

386

J. A. M C KEAGUE AND M. G . CLINE

not have diagnostic horizons believed to require substantial periods of time for development. The Inceptisols cannot, however, have major proportions of oxide clays unless they are associated with significant amounts of weatherable silicates. The sand fraction may be dominated by quartz, but the soil cannot be so quartzose that distinctive horizons cannot form. In areas of recent glaciations, Inceptisols commonly are dominated by mixtures of quartz, primary silicates, and 2 : l lattice silicate clays in varying proportions. In unglaciated areas without volcanic deposits 1 : l lattice silicate clays are more prominent and primary weatherable silicates are less prominent. A special suborder, Andepts, is recognized for soils dominated by allophane in the clay fraction and volcanic glass in the sand fraction. Essentially all Inceptisols have been subjected to some degree of weathering and leaching and have lost some silica. Losses relative to parent rock have been very large in some developed in residuum of rock weathered in place. Losses have been small in many Inceptisols of glaciated regions. Silica may be a cementing agent in “duripans,” which occur in some varieties. Aridisols include most soils that have been called Desert, Red Desert, and Reddish Brown soils and Sierozems and Solonchak as well as some of the Regosols and Lithosols of dry climates (Soil Survey Staff, 1960, p. 156). They are “primarily soils of dry places.” The rate of weathering and leaching of products of weathering are restricted by lack of moisture. Consequently, inherited forms and amounts of silica are conspicuous. The order is restricted, however, to materials that are not so quartzose that distinctive horizons cannot form. Generally, the Aridisols have varying combinations of quartz and a variety of primary silicates in the sand fraction with one or more of the silicate clays dominant in the clay fractions. Volcanic glass and allophane may be major constituents. If oxide clays are present in major amounts, they must be associated with significant amounts of weatherable silicates or 2:1 lattice clays. Contrary to popular concepts (Joffe, 1949, pp. 468469; Mohr and Van Baren, 1954, p. 38), silica is no more soluble at the pH values common in nonalkali Aridisols than in more acid soils of more humid regions. Nevertheless, cementation of pans, presumably by silica or an aluminosilicate (Soil Survey Staff, 1960) is common ( Nikiforoff, 1937). It is possible that such cementation may occur as a result of drying cycles and limited movement of water, the possible consequences of which are discussed in Section IV. MoZZisols include most soils that have been called Chernozem, Brunizem, Chestnut soils, Reddish Prairie soils, and the associated Planosols and Humic Gley soils, as well as some others (Soil Survey Staff, 1960, P. la),The most significant distinguishing feature is the thick dark

SILICA IN SOILS

387

surface horizon characteristic of soils of the steppes. Soils that have such surface horizons but base status indicative of an advanced stage of leaching or oxide clays indicative of an advanced stage of weathering, as well as some others, are excluded. Consequently, the degree of weathering is limited. Silica is commonly present as a mixture of weatherable primary silicates, quartz, and silicate clays. Their proportions may range widely. Some have lost a high proportion of weatherable primary silicates but are rich in 2:l lattice silicate clays. In many, silicate clays have been concentrated in the B horizon, at least partly by illuviation, and quartz is correspondingly concentrated in the surface soil. Kaolinite may be an important constituent, but in some Mollisols it appears to be mainly inherited. Silica has probably been removed to some degree from essentially all Mollisols. The amount lost ranges from very little in some Mollisols of very recent deposits to at least the amounts involved in the destruction of most of the weatherable primary minerals and formation of silicate clays of the 2 : l lattice type in certain members that were formerly called planosols. The surface horizon is likely to contain plant opal from grasses, which tend to segregate such particles (Bazilevich et aZ., 1954; Beavers and Stephen, 1958; Baker, 1959). Solonetzic varieties also commonly have siliceous powder above the alkaline B horizon, which Raupach (1957) suggested might be due to deposition of silica from solution at a site of abrupt increase of pH. Other explanations appear more plausible ( McKeague, 1962). Spodosols include soils that have been called Podzols, Ground-Water Podzols, and Brown Podzolic soils. Their diagnostic criterion is a “spodic” horizon, which can be grossly described as an illuvial concentration of free iron, humus, or both, usually with aluminum in a form comparable to “allophane” (Soil Survey Staff, 1960, p. 192). The parent material is usually siliceous, and quartz in sand or silt fractions is usually the main form of silica. DeConinck and Laurelle (1960) and many others have discussed the influence of texture and mineralogy on development of Podzols. If an A2 horizon is present, it has been stripped of most weatherable material other than that in coarse fractions; Cady (1960) has indicated that weathering is very intensive in it. Material is removed in accordance with its susceptibility to chelation ( Swindale and Jackson, 1956) by processes discussed by Bloomfield ( 1955),Delong and Schnitzer ( 1955), Flach ( 1960), and others, concentrating quartz and comparable substances not subject to chelation (Crompton, 1960). Riquer ( 1960) has discussed Spodosol-like soil in the humid tropics in which plant opal dominated the A? horizon. Within the spodic horizons, weatherable silicates may be preserved by coatings of sesquioxides (Cady, 1960; Flach, 1960).

388

J. A. MCKEAGUE AND M. G. CLINE

Alfisols include most soils that have been called Noncalcic Brown, Gray-Brown Podzolic, and Gray-Wooded soils as well as some Planosols and Half-Bog soils (Soil Survey Staff, 1960, p. 202). They have an illuvial concentration of silicate clay in the B horizon. Cady (1960)suggested that textural B horizons may be favored by the presence of montmorillonite. Pennington and Jackson ( 1947) suggested that its fineness is a major factor in its mobility. Many AEsols, however, contain little montmorillonite. AIfisols have base saturation patterns or values that indicate they have not been extremely leached. They are, consequently, mainly not in highly weathered material; many are in relatively late PIeistocene deposits. In many Alfisols, the clays are inherited 2:1 lattice types. Most Alfisols contain appreciable amounts of weatherable primary silicates in sand and silt fractions. Usually, the degree of weathering decreases with depth ( McCaleb and Cline, 1950; Cady, 1960; Yassoglou and Whiteside, 1960).Though they may be high in quartz, they cannot be so quartzose that adequate silicate clay is not available to form the illuvial concentration that is diagnostic. Reifenberg ( 1938), Kubiena (1953),Duchaufour ( 1960), and many others have attributed the mobility of silicate clay in soils of this kind to the action of silica sol as a protective colloid. The literature of chemistry (Iler, 1955) and geology (Krauskopf, 1959; Siever, 1957) as well as studies of soil extracts (McKeague and Cline, 1983a) strongly indicate that polymeric silica could not be a major factor. Ultisols include most soils that have been called Red-Yellow Podzolic and Reddish Brown Lateritic soils as well as some very acid Humic Gley, Low-Humic Gley and Ground-Water Laterite soils in the United States (Soil Survey Staff, 1960, p. 226). Like the Alfisols, they have illuvial silicate clay accumulations in the B horizons, but unlike the AIfisols, their pattern of base saturation is indicative of intense leaching. This is commonly associated with at least moderately strong weathering. The soil must have enough weatherable primary silicates or 2:l lattice silicate clays to be distinctly detectable. This is a criterion for differentiation from Oxisols. Normally the amounts are small, but some members have moderately large amounts of feldspar or mica in sand or silt fractions; others have large amounts of 2:l lattice clays (Soil Survey Staff, 1 W ) . Kaolinite is commonly the dominant clay mineral ( McCaleb, 1959). Quartz is dominant among all constituents in some members, but as in the Alfisols, it cannot be so high that too little clay is present to satisfy requirements of clay concentration in the B horizon. If an A2 horizon is present, quartz has been concentrated in it by loss of other constituents. Oxisols include most of the soils that have been called Latosols and Ground-Water Laterite soils (Soil Survey Staff, 1960, p. 238). They

SILICA IN S O I t S

389

represent the most highly weathered soil materials and those in which combined silica is lowest. To this end, they have been defined as lacking detectable 2:l lattice clays other than vermiculite or allophane and as having almost no vermiculite or weatherable primary silicates in the diagnostic horizon. That horizon must have a significant proportion of oxide clays relative to 1:l lattice silicates. These, therefore, are soils that have lost a high proportion of the combined silica of original rock materials. They can, however, contain as much as 85 per cent quartz in sand and silt fractions. One suborder, the Acrox, is so dominated by oxide clays that the soil of the diagnostic horizon in suspension either has an electrically positive charge or is electrically neutral. These represent the extreme case of depletion of silica among all mineral soils. The Nipe series of Puerto Rico is an example (Soil Survey Staff, 1960, p. 92). REFERENCES Alexander, G. B. 1957. J. Phys. Chem. 61, 1563-1564. Alexander, G. B., Heston, W. M., and Iler, R. K. 1954. J . Phys. Chem. 58, 453-455. Alexander, L. T., Hendricks, S. B., and Faust, G. T. 1941. Soil Sci. SOC. Am. P ~ o c 6, . 52-57. Anderson, D. H., and Hawkes, H. E. 1958. Geochim. Cosmochim. Acta 14, 204210. Aomine, S., and Jackson, M. L. 1959. Soil Sci. Soc. Am. Proc. 23, 210-214. Armstrong, L. C. 1940. Am. Mineralogist 25, 810-820. Baird, G. B. 1952. Ph.D. Thesis, Cornell Univ., Ithaca, New York. Baker, G. 1959. Australian I. Botany 7 , 64-87. Barshad, 1. 1955. In “Chemistry of the Soil” (F. E. Bear, ed.), pp. 1-52. Reinhold, New York. Barshad, I. 1959. “Clays and Clay Minerals,” Proc. 6th Natl. Conf. Clays Clay Minerals 1957, pp. 110-132. Pergamon, New York. Bastisse, E. M. 1949. C m p t . rend. 229, 521-522. Bastisse, E. M. 1960. Ann. Agron. A l l , 441-467. Bazilevich, N. I., Kalashnikova, R. A., and Yarilova, Y. A. 1954. Tr. Pochv. Inst. Akad. Nauk S S S R 44, 262-280. Beavers, A. H., and Stephen, I. 1958. Soil Sci. 86, 1-5. Behrman, A. S., and Custafson, H. 1940. I n d . Eng. Chem. 32, 468-472. Bergman, I., and Paterson, M. S. 1961. 1. Appl. Chem. 11, 369-376. Betz, L. D., Noll, C. A,, and Maguire, J. J. 1940a. Znd. Eng. Chem. 32, 13201323. Betz, L. D., Noll, C. A., and Maguire, J. J. 1940b. Ind. Eng. Chem. 32, 13231329. Betz, L. D., Noll, C. A., and Maguire, J. J. 1941. I n d . Eng. Chem. 33, 814-820. Bien, G. S., Contois, D. E., and Thomas, W. H. 1958. Geochim. Cosmochim. Acta 14, 35-54. Birrell, K. S. 1961. I. Soil Sci. 12, 307-316. Birrell, K. S., and Fieldes, M. 1952. J . Soil Sci. 3, 156-166. Birrell, K. S . , and Gradwell, M. 1956. J . Soil Sci. 7 , 130-143.

390

J. A. MCKEAGUE AND M.

G. CLINE

Bloomfield, C. 1955. J. Sci. Food Agr. 8, 841-851. Bolt, G. H. 1957. J . Phys. Chem. 81, 1168-1189. Breeze, B. F. 1980. Australian J. Sci. 23, 18-20. Brindley, G. W., and Radoslovich, E. W. 1958. “Clays and Clay Minerals,” Proc. 4th Natl. Conf. Clays Clay Minerals 1955, pp. 330-338. Natl. Acad. Sci. Natl. Res. Council Publ. No. 456. Britton, H. T. S. 1927. J . Chem. Soc. 1927, 425-438. Brydon, J. E., and Patry, L. M. 1981. Can. I. Soil Sci. 41, 169-181. Brydon, J. E., Clark, J. S., and Osbome, V. 1981. Can. Mineralogist 6, 595-809. Cady, J. G. 1980. Trans. Intern. Congr. Soil Sci. 7th Congr. Madison, Wisconsin, 1960 4, 418-424. Callihre, S., Hbnin, S., and Esquevin, J. 1958. Trans. Intern. Congr. Soil Sci. 6th Congr. Paris, 1956 B, 159-183. Cann, D. B., and Whiteside, E. P. 1955. Soil Scl. SOC. Am. Proc. 19, 497-501. Carr, R. M., and Fyfe, W. S. 1958. Am. Mineralogist 43, 908-917. Carroll, D., and Hathaway, J. C. 1954. “Clays and Clay Minerals,” PTOC. 2nd Natl. Conf. Clays Clay Minerals 1953, pp. 171-182. Natl. Acad. Sci. Natl. Res. Council Publ. 327. Cassidy, H. G. 1951. “Adsorption and Chromatography.” Interscience, New York. Clark, E. R., and Waddams, J. A. 1957. Nature 180, 904-905. Clarke, F. W. 1924. U.S. Geol. Suru. Bull. 770, 1-841. Clelland, D. W., and Ritchie, P. D. 1952. J. Appl. Chem. 2, 42-48. Clelland, D. W., Cumming, W. M., and Ritchie, P. D. 1952. J. Appl. Chem. 2, 31-41. Cooper, R. 1959. Soils Fertilizers 22, 327-333. Correns, D. W. 1949. “Einfuhrung in die Mineralogie.” Springer, Berlin. Crompton, E. 1980. Trans. Intern. Congr. Soil Sci. 7th Congr. Madison, Wisconsin, 1960 4, 408-412. Dana, J. D. 1959. “Dana’s Manual of Mineralogy,” 17 ed. (revised by C. S. Hurlbut, Jr.). Wiley, New York. Dapples, E. C. 1959. Soc. Econ. Paleontologists Mineralogists. Spec. Publ. 7, 3851. Dawson, J. E. 1956. Aduan. Agron. 8, 377-401. DeConinck, F., and Laurelle, J. 1980. Trans. Intern. Congr. Soil Sci. 7th Congr. Madison, Wisconsin, 1960 4, 157-184. DeKimpe, C., and Gastuche, M. C. 1980. Trans. Intern. Congr. Soil Sci. 7th Congr. Madison, Wisconsin, 1960 4, 512-522. De Long, W. A., and Schnitzer, M. 1955. Soil Sci. Soc. Am. PTOC. 19, 380-383. Demolon, A., and Bastisse, E. 1942. Compt. rend. 215, 188-190. Demolon, A., and Bastisse, E. M. 1944. Ann. Agron. 14, 285-298. Demolon, A., and Bastisse, E. M. 1945. Ann. Agron. 15, 30-57. Dempster, P. B., and Ritchie, P. D. 1953. J. Appl. Chem. 3, 182-192. Denny, J. J., Robson, W. D., and Irwin, D. A. 1937. Can. Med. Assoc. J. 37, 1-11. Denny, J. J., Robson, W. D., and Irwin, D. A. 1939. Can. Med. Assoc. J . 40, 213-228. Deuel, H. 1980. Trans. Intern. Congr. Soil Sci. 7th Congr. Madison, Wisconsin, 1960 1, 38-53. D’Hoore, J. 1954. Publ. Inst. Natl. Etude Agron. Congo Belge Ser. Sci. 82. De Vore, G. W. 1959. “Clays and Clay Minerals,” Proc. 6th Natl. Conf. CEays Clay Minerals 1957, pp. 26-41. Pergamon, New York.

SILICA IN SOILS

391

Dibnert, F., and Wandenbulcke, F. 1923. Compt. rend. 176, 1478-1480. Duchaufour, P. 1960. “Prkcis de P6dologie.” Masson, Paris. Ehrlich, W. A., Rice, H. M., and Ellis, J. H. 1955. Can. J . Agr. Sci. 35, 407-421. Engel, W. 1953. Planta 41, 358-390. Fieldes, M. 1952. Nature 170, 366-367. Fieldes, M. 1955. New Zealand J. Sci. Technol. B37, 336-350. Fieldes, M., and Swindale, L. D. 1954. New Zealand J . Sci. Technol. B36, 140154. Fieldes, M., and Williamson, K. I. 1955. New Zealand J. Sci. Technol. B37, 314335. Fieldes, M., Swindale, L. D., and Richardson, J. P. 1952. Soil Sci. 74, 197-205. Flach, K. W. 1960. Ph.D. Thesis, Cornell Univ., Ithaca, New York. Foumier, R. O., and Rowe, J, J. 1962. Am. Mineralogist 47, 897-902. Frankart, R., Gastnche, M. C., and Fripiat, J. J. 1960. Puhl. Inst. Natl. Etude Agron. Congo Belge Ser. Sci. 86. Fredrickson, A. F. 1951. Bull. Geol. SOC. Am. 62, 221-232. Fredrickson, A. F., and Cox, J. E. 1954a. Am. Mineralogist 39, 738-750. Fredrickson, A. F., and Cox, J. E. 195413. Am. Mineralogist 39, 886-900. Frink, C. R. 1960. Ph.D. Thesis, Cornell Univ., Ithaca, New York. Ganichenko, L. G., Egorov, M. M., Kiselev, V. F., Krasil’nikov, K. G., and Khodakov, G. S. 1960. Proc. Acad. Sci., USSR Phys. Chem. Sect. (English Transl.) 131, 275-278. Consultants Bureau Enterprises Inc., New York, 1960. (Translation-1960. Dokl. Akad. Nauk S S S R 131, 597-600.) Carrels, R. M. 1959. Res. Geochem. pp. 25-37. Carrels, R. M., and Howard, P. 1959. “Clays and Clay Minerals,” Proc. 6th Natl. Con!. Clays Clay Minerals 1957, pp. 68-88. Pergamon, New York. Gastuche, M., Demlon, B., and Vielvoye, L. 1960. Bull. SOC.Chim. France 1960, 60-70. Germer, L. H., and Storks, K. H. 1939. lnd. Eng. Chem. Anal. Ed. 11, 583-592. Gibb, J. G., Ritchie, P. D., and Sharpe, J. W. 1953. J. Appl. Chem. 3, 213-218. Gilbert, C. M. 1949. 1. Geol. 57, 1-17. Giles, C. H. 1959. In “Hydrogen Bonding” (D. Hadzi, ed.), pp. 449-460. Pergamon, New York. Giles, C. H., and MacEwan, T. H. 1957. “Electrical Phenomena and Solid-Liquid Interface,” Proc. 2nd Intern. Congr. Surface Activity, London, 1957, 3, 457-461, Buttenvorths, London. Giles, C. H., MacEwan, T. H., Nakhwa, S. N., and Smith, D. 1960. I. Chem. SOC. 1960, 3973-3993. Goldich, S. S. 1938. J. Geol. 46, 17-58. Goldstein, A., Jr. 1948. J . Sediment. Petrol. 18, 108-125. Govett, G. J. S. 1981. Anal. Chim. Acta 25, 69-80. Gradusov, B. P., and Dyazdevich, G. S. 1961. Sooiet Soil Sci. (English Transl.) 1961, 749-756. Am. Inst. Biol. Sci., Washington, D.C. 1962. (Translation1961. Pochuooedenie 1961(7 ) , 59-66.) Graham, E. R. 1940. Soil Sci. 49, 277-281. Graham, E. R. 1941a. Soil Sci. 52, 291-295. Graham, E. R. 1941b. J. Geol. 49, 392-401. Greenberg, S. A. 1958. J. Am. Chem. SOC. 80, 6508-6511. Greenberg, S. A., and Price, E. W. 1957. J . Phys. Chem. 61, 1539-1541. Greenbcrg, S. A., Chang, T. N., and Anderson, E. 1960. 1. Phys. Chem. 64, 1151-1157.

392

J. A. MCKEAGUE AND

M.G. CLINE

Greene, H. 1945. Soil Sci. SOC. Am. Proc. 10, 392-395. Harder, E. C. 1952. In “Problems of Clay and Laterite Genesis,” pp. 35-84. Am. Inst. Mining and Metallurgical Eng., New York. Hardy, F., and Rodriguez, G. 1939. Soil Sci. 48, 361-384. Harman, R. W. 1927. J. Phys. Chem. 31, 816-825. Harrassowitz, H. 1928. Fortschr. Geol. Palaeontol. 4( 14), 253-588. Harrison, J. B. 1933. “The Katamorphism of Igneous Rocks Under Humid Tropical Conditions.” Imp. Bur. Soil Sci., Harpenden, England. Hashimoto, I., and Jackson, M. L. 1980. “Clays and Clay Minerals,” Proc. 7th Natl. Conf. Clays Clay Minerals 1958, pp. 102-113. Pergamon, New York. Hauser, E. A. 1952. In “Problems of Clay and Laterite Genesis,” pp. 100-108. Am. Inst. Mining and Metallurgical Eng., New York. Hazel, F., Schock, R. V., Jr., and Gordon, M. 1949. J. Am. Chem. SOC. 7, 22582257. Heavens, 0. S. 1953. Acta Cryst. 8, 571-572. HCnin, S. 1954. Trans. Intern. Congr. Soil Sci. 5th Congr. Leopolduille, Belgian Congo, 1954 2, 505-508. Hhnin, S. 1957. Agrochimica 1, 328-337. Holt, P. F., and King, D. T. 1955. 1. Chem. SOC. 1955, 773-779. Humbert, R. P. 1948. Soil Sci. 85, 281-290. Iler, R. K. 1955. “The Colloid Chemistry of Silica and Silicates.” Cornell Univ. Press, Ithaca, New York. Iwasaki, I., Katsura, T., and Tarutani, T. 1951. Bull. Chem. SOC. Japan 24, 227230. Jacks, G. V. 1953. Sci. Progr. (London) 41, 301-305. Jackson, E. A. 1957. I. Australian Inst. Agr. Sci. 23, 198-208. Jackson, M. L. 1956. “Soil Chemical Analysis-Advanced Course.” M. L. Jackson, Madison, Wisconsin. Jackson, M. L. 1959. “Clays and Clay Minerals,” Proc. 6th Natl. Conf. Clays Clay Minerals 1957, pp. 133-143. Pergamon, New York. Jackson, M. L., and Sherman, G. D. 1953. Aduan. Agron. 5, 219-318. Jackson, M. L., Pennington, R. P., and Mackie, W. Z. 1948a. Soil Sci. SOC. Am. PrOC. 13, 139-145. Jackson, M. L., Tyler, S. A., Willis, A. L., Bourbeau, G. A., and Pennington, R. P. 1948b. J . Phys. Colloid Chem. 52, 1237-1260. Jeffries, C. D. 1937. Soil Sci. 43, 357-388. Jenny, H. 1941. “Factors of Soil Formation.” McGraw Hill, New York. Jephcott, C. M., and Johnston, J. H. 1950. Arch. Id.Hyg. Occupational Med. 1, 323-340. Jessup, R. W. 1980. J. Soil Sci. 11, 108-113. Joffe, J. S. 1940. Soil Sci. 50, 57-63. Joffe, J. S. 1949. “Pedology.” Pedology Publ., New Brunswick, New Jersey. Joseph, A. F., and Oakley, H. G. 1925. J . Chem. SOC. 127, 2813-2818. Karpachevskii, L. 0. 1960. Suuiet Soil Sci. (English Trawl.) 1960, 493-500. Am. Inst. Biol. Sci., Washington, D.C. 1981. (Translation-1980. Pochuuuedenie 1980(5), 43-52.) Keller, W. D. 1957. “The Principles of Chemical Weathering.” Lucas, Columbia, Missouri. Kelley, W. P., and Brown, S. M. 1939. 1. Am. SOC. Agron. 31, 41-43. Khan, D. H. 1960. I . Sci. Food Agr. 11, 133-138.

SILICA IN SOILS

393

King, E. J., and McGeorge, M. 1938. Biochem. J. 32, 417-425. Kitto, P. F., and Patterson, H. S. 1942. J . Id.Hyg. Toxicol. 24, 59-74. Knox, E. G. 1957. Soil Set. SOC. Am. Proc. 21, 328-330. Kobayashi, J. 1980. Be?. Ohara Inst. Landwirtsch. Biol. Okayama Uniu. 11(3), 313-358. Kovda, V. A., Zimnovets, B. A,, and Amchislavskaya, A. G. 1958. Soviet Soil Sci. (English Transl. ) 1958, 471-479. Am. Inst. Biol. Sci., Washington, D.C. 1959. (Translation-1958. Pochuouedenie l958( 5 ) , 1-11.) Krauskopf, K. B. 1958. Geochim. Cosmochim. A d a 10, 1-28. Krauskopf, K. B. 1959. SOC. Econ. Paleontologists Mineralogists Spec. Publ. 7 , 4-19. Kubiena, W. L. 1953. “The Soils of Europe.” Murby, London. Lanning, F. C., Ponnaiya, B. W. X., and Crumpton, C. F. 1958. Plant Physio2. 33, 339-343. Latimer, W. M. 1952. “The Oxidation States of the Elements and their Potentials in Aqueous Solutions,” p. 224. Prentice Hall, Englewood Cliffs, New Jersey. Lewin, J. C. 1981. Geochim. Cosmochim. Acta 2, 182-198. Lindsay, F. K., and Ryznar, J, W. 1939. Ind. Eng. Chem. 31, 859-881. Loewenstein, W. 1954. Am. Mineralogist 39, 92-98. Lovering, T. S. 1959. Bull. Geol. Soc. Am. 70, 781-800. Low, P. F., and Black, C. A. 1947. Soil Sci. Soc. Am. Proc. 12, 180-184. Low, P. F., and Black, C. A. 1950. Soil Scl. 70, 273-290. Lucas, C. C., and Dolan, M. E. 1939. Can. Med. Assoc. J . 40, 128-134. Lutz, H. J., and Chandler, R. F. 1948. “Forest Soils.” Wiley, New York. McBain, J. W. 1950. “Colloid Science.” Heath, Boston. McCaleb, S. B. 1954. Soil Sci. 77, 319-333. McCaleb, S. B. 1959. Soil Sci. Soc. Am. Proc. 23, 184-170. McCaleb, S. B., and Cline, M. G. 1950. Soil Sci. 70, 315-328. McConnell, D. 1951. Bull. Geol. SOC. Am. 62, 700-701. McDowall, I., and Vose, W. 1952. Nature 170, 386. McKeague, J. A. 1962. Ph.D. Thesis, Cornell Univ., Ithaca, New York. McKeague, J. A,, and Cline, M. G. 1983a. Can. J . Sod Sci. 43, 70-82. McKeague, J. A., and Cline, M. G. 1983b. Can. J . Soil Sci. 43, 83-98. MacLeod, L. B. 1982. Ph.D. Thesis, Cornell Univ., Ithaca, New York. Magistad, 0. C. 1925. Soil Sci. 20, 181-225. Marbut, C. F. 1935. “Soils of the United States.” Atlas of American Agriculture, 111. U.S. Govt. Printing Office, Washington, D.C. Mason, B. 1952. “Principles of Geochemistry.” Wiley, New York. Mattson, S. 1928. Soil Sci. 25, 289-311. Mehra, 0. P., and Jackson, M. L. 1980. “Clays and Clay Minerals,” Proc. 7th Natl. Conf. Clays Clay Minerals 1958, pp. 317-327. Pergamon, New York. Milliken, T. H., Jr., Mills, G. A,, and Oblad, A. G. 1950. DCcussions Faraday SOC. 8, 279-290. Mohr, E. C. J. 1944. “The Soils of Tropical Regions with Special Reference to the Netherlands East Indies” (H. L. Pendleton, translator). Edwards, Ann Arbor, Michigan. Mohr, E. C. J., and Van Baren, F. A. 1954. “Tropical Soils.” Interscience, New York. Morey, G. W., and Chen, W. T. 1955. Am. Mineralogist 40, 998-1000. Morey, G. W., and Fournier, R. 0. 1981. Am. Mineralogist 46, 888-899.

394

J. A. MC KEAGUE AND M.G. CLINE

Morey, G. W., and Hesselgesser, J. M. 1951. Econ. Geol. 46, 821-835. Morey, G. W., Foumier, R. O., and Rowe, J. J. 1982. Geochim. Cosmochim. Acta 26, 1029-1043. Mosebach, R. 1957. J. Geol. 65, 347-383. Murata, K. J. 1940. Am. J. Sci. 238, 588-598. Murata, K. J. 1948. Bull. US. Geol. Sun]. 950, 25-33. Nagelschmidt, G., Gordon, R. L., and Griffin, 0. G. 1952. Nature 169, 539-540. Nash, V. E., and Marshall, C. E. 1958a. Missouri Uniu. Agr. Expt. Sta. Res. Bull. 613. Nash, V. E., and Marshall, C. E. 195813. Missouri Uniu. Agr. Expt. Sta. Res. Bull. 614. Nikiforoff, C. C. 1937. Soil Sci. 43, 105-125. Nikiforoff, C. C., and Alexander, L. T. 1942. Soil Sci. 53, 157-172. Nikiforoff, C. C., Humbert, R. P., and Cady, J. G. 1948. Soil Sci. 65, 135-153. Novorossova, L. E. 1951. Pochuouedenie 1951(2), 115-118. Nutting, P. G. 1945. Science 101, 819-821. O’Connor, T. L. 1981. 1. Phys. Chem. 65, 1-5. Okamoto, G., Okura, T., and Goto, K. 1957. Geochim. Cosmochim. Acta 12, 123-132. Paterson, M. S., and Wheatley, K. 1959. 3. Appl. Chem. 9, 231-240. Pauling, L. 1948. “The Nature of the Chemical Bond,” 2nd ed. Comell Univ. Press, Ithaca, New York. Pawluk, S. 1980. Trans. Intern. Congr. Soil Sci. 7th Congr. Madison, Wisconsin, 1960 4, 314-322. Pennington, R. P., and Jackson, M. L. 1947. Soil Sci. SOC. Am. Proc. 12, 452-457. Pettijohn, F. J. 1941. J. Geol. 49, 610-625. Pettijohn, F. J. 1957. “Sedimentary Rocks,” 2nd ed. Harper, New York. Piryutko, M. M. 1959. Bull. Acad. Sci. U S S R Diu. Chem. Sci. (English Transl.) 1959(3 ) , 355-380. Consultants Bureau Enterprises Inc., New York, 1959. (Translation-1959. 120. Akad. Nauk S S S R Otd. Khim. Nauk 1959(3 ) , 379-384.) Poldervaart, A. 1955. Geol. SOC. Am. Spec. Papers 62, 119-144. Polynov, B. B. 1937. “The Cycle of Weathering” (A. Muir, translator). Murby, London. Ponnamperuma, F. N. 1955. Ph.D. Thesis, Comell Univ., Ithaca, New York. Rachmat Hardjosoesastro, R. 1958. 1. Soil Sci. 7, 185-188. Raupach, M. 1957. C.S.I.R.O. (Australia) Soil Publ. 9, 35 pp. Raupach, M., and Piper, C. S. 1959. Australian J. Agr. Res. 10, 818-831. Reiche, P. 1945. Uniu. New Mexi. Publ. Geol. 1. Reifenberg, A. 1938. “The Soils of Palestine” (C. L. Whittles, translator). Murby, London. Reifenberg, A., and Buckwold, S. J. 1954. J. Soil Sci. 5, 108-118. Remezov, N. P. 1958. Soviet Soil Sci. (English Trawl.) 1958, 589-598. Am. Inst. Biol. Sci., Washington, D.C.1980. (Translation-1958. Pochuouedenie 1958(8 ) , 1-12. ) Rice, H. M., Forman, S. A,, and Patry, L. M. 1959. Can. 3. Soil Sci. 39, 165-177. Rich, C. I., and Thomas, G. W. 1980. Aduan. Agron. 12, 1-39. Richardson, E. 1959. J. Appl. Chem. 9, 371-378. Riquer, J. 1980. Trans. Intern. Congr. Soil Sci. 7th Congr. Madison, Wisconsin, 1960 4, 425-431. Roller, P. S., and Ervin, G . E., Jr. 1940. 3. Am. Chem. SOC. 62, 461-471.

SILICA IN SOILS

395

Ross, C. S., and Kerr, P. F. 1934. U.S. Geol. Sum. Prof. Paper 185G, 135-148. Roy, C. J. 1945. Am. J. Sci. 243, 393-403. Rozhnova, T. A., and Schastnaia, L. S. 1959. Soobt Soil Sci. (English Tram,!.) 1959, 15-24. Am. Inst., Biol. Sci., Washington, D.C. 1960. ( Translation-1959. Pochoooedenie 1959( 1), 19-29.) Russell, E. J. 1950. “Soil Conditions and Plant Growth,” 8th ed. Longmans Green, London. Siever, R. 1957. Am. Mineralogist 42, 821-841. Siever, R. 1959. SOC. Econ. Paleontologists Mineralogists Spec. Publ. 7, 55-79. Siever, R. 1962. J. Geol. 70, 127-150. Simonson, R. W. 1954. I. Soil Sci. 5, 275-288. Sivarajasingham, S., Alexander, L. T., Cady, J. G., and Cline, M. G. 1982. Adoan. Agron. 14, 1-60. Smithson, F. 1956. J. Soil Sci. 7, 122-129. Smithson, F. 1958. J. Soil Sci. 9, 148-154. Smithson, F. 1959. J. Soil Sci. 10, 105-109. Snell, F. D., Snell, C. T., and Snell, C. A. 1959. “Colorimetric Methods of Analysis,” Vol. 2A. Van Nostrand, Princeton, New Jersey. Soil Survey Staff. 1960. ‘‘Soil Classification, A Comprehensive System, 7th Approximation,” U S . Govt. Printing Office, Washington, D.C. Stephens, C. G. 1957. Proc. Australian Conf. Soil Sci. 2nd Conf. Vol. I, Part 1. C.S.I.R.O. ( Australia ) , Melbourne. Strickland, J. D. H. 1952a. J. Am. Chem. SOC. 74, 862-867. Strickland, J. D. H. 195213. J. Am. Chem. SOC. 74, 888-871. Strickland, J. D. H. 1952c. 1. Am. Chem. SOC. 74, 872-876. Swindale, L. D., and Jackson, M. L. 1956. Tram. Intern. Congr. Soil Sci. 6th Congr. Paris, 1956 E, 233-239. Szabolcs, I., and Darab, K. 1958. Acta Agron. 8, 213-235. Talbot, J. H., and Kempis, E. B. 1960. Nature 188, 927-929. Tamele, M. W. 1950. Discussions Faraday SOC. 8, 270-279. Targul’yan, V. 0. 1959. Sooiet Soil Sci. (English Trawl.) 1959, 1287-1298. Am. Inst. Biol. Sci., Washington, D.C. (Translation-1959. Pochoouedenie 1959( 11), 37-48. ) Taylor, H. F. W., and Howison, J. W. 1956. Clay Minerals Bull. 3( l6), 98-111. Tedrow, J. C. F. 1954. Soil Sci. SOC. Am. Proc. 18, 479-481. Toth, S. J. 1955. In “Chemistry of the Soil“ (F. E. Bear, ed.), pp. 85-106. Reinhold, New York. Trapnell, B. M. W. 1955. “Chemisorption.” Butterworths, London. Vail, J. G. 1952. ‘“Soluble Silicates,” Vol. 1. Reinhold, New York. Vamos, R. 1981. Agrokem. Talajton 10, 53-66; 1961. Soils Fertilizers 24, 337 ( abstr. ) . Van Lier, J. A., De Bruger, P. L., and Overbeek, J. T. G. 1980. J. Phys. Chem. 64,1675-1682. Van Rummelen, F. F. F. E. 1953. Landbouw Buitenz. 25, 225-230. Viehoever, A., and Prusky, S. C. 1938. Am. J. Pharm. 110, 99-120. Vilenskii, D. G. 1960. “Soil Science” (A. Birron and Z. S. Cole Trans].). Office of Technical Services Translation 60-51197 of “Pochvovedenie,” 3rd ed. 1957. U.S. Dept. Commerce, Washington, D.C. Vinokurov, M. A., and Bukharayeva, L. G. 1961. Swiet Soil Sci. (English Transl.) 1961, 591-597. Am. Inst. Biol. Sci., Washington, D.C. 1962. (Translation-1981. Pochoooedenie 1961(6 ) , 1-10.)

396

J. A. MC KEAGUE AND M. C. CLINE

Wada, K., and Ataka, H. 1958. Soil Plant Food (Tokyo) 4, 12-18. Webley, D. M., Dd, R. B., and Mitchell, W. A. 1960. Nature 188, 766-767. Weyl, W. A. 1951. Penna. State Coll. Mineral Id.Expt. Sta. Bull. 57. Weyl, W. A. 1953. In “Structure and Properties of Solid Surfaces” (R. Gomer and C. S. Smith, eds.), pp. 147-184. Univ. Chicago Press, Chicago, Illinois. White, W. A. 1953. Am. Mineralogist 38, 634-642. White, D. E., Brannock, W. W., and Murata, K. J. 1956. Geochim. Cosmochim. Acta 10, 27-59. Whitehouse, F. W. 1940. Univ. Queensland Papers Dept. Geol. 2 ( 1 ) . Whitney, R. S., and Peech, M. J. 1952. Soil Sci. SOC. Am. Proc. 18, 117-122. Williamson, W. 0. 1957. Am. J. Sci. 255, 23-42. Winchell, A. N., and Winchell, H. 1959. ‘‘Elemen,ts of Optical Mineralogy,” 4th ed., Pt. 11. Wiley, New York. Winters, E. 1942. Soil Sd.SOC. Am. Proc. 7 , 437-440. Woodruff, C. M. 1954. “Clays and Clay Minerals,” Proc. 2nd Natl. Conf. Clays Clay Minerals 1953, pp. 120-123. Natl. Acad. Sci. Natl. Res. Council Publ. 327. Yassoglou, N. J., and Whiteside, E. P. 1960. Soil Sd.SOC.Am. Proc. 24, 396-407. Yoshida, S., Ohnishi, Y., and Kitagishi, K. 1959a. Soil Plant Food (Tokyo) 5( 1). 23-27. Yoshida, S., Ohnishi, Y., and Kitagishi, K. 1959b. Soil Plant Food (Tokyo) 5(3), 127-133. Yoshida, S., Ohnishi, Y., and Kitagishi, K. 1962. Soil Sci. Plant Nutr. (Tokyo) 8(3), 15-21. Yoshinaga, N., and Aomine, S . 1962a. Soil Sci. Plant Nutr. (Tokyo) 8 ( 2 ) , 6-13. Yoshinaga, N., and Aomine, S . 1962b. Soil Sci. Plant Nutr. (Tokyo) 8(3), 22-29.

AUTHOR INDEX Numbers in italics indicate the pages on which the complete references are listed.

A Aamodt, 0. S., 30, 32, 114, 319, 322, 325, 335, 336 Aberg, E., 31, 33, 114 Adams, A. F. R., 74, 76, 118 Adams, M. W., 319,331,336 Afnkyan, E. K., 135, 154 Agerberg, L. S., 132, 154 Ahlgren, H. L., 30, 32, 114 Alban, L. A., 148, 152,154 Alexander, C. W., 86, 92, 94, 114 Alexander, G. B., 353, 354,359,387, 373, 389 Alexander, L. T., 341, 342, 345, 346, 349, 352, 377, 379, 382, 389, 395 Alexander, M.,145, 148, 154 Allaway, W. H., 148, 149, 157 Alley, H. P., 195,200, 209,210 Allison, L. E., 147, 154 Amchislauskaya, A. G., 367, 377, 378, 393 Amin, J. V., 149, 154 Anderson, D. H., 346, 389 Anderson, D. T., 277, 278, 299 Anderson, E., 377, 381, 391 Anderson, J. R., 192, 210 Aomine, S . , 343, 344, 389, 396 Armbrust, D. V., 233, 234, 293, 295, 298, 300 Armstrong, L. C., 366,389 Army, T. J., 81, 114,281,301 Askew, H. O., 150,154 Aspinall, D., 10, 104,114,115 Ataka, H., 344, 396 Atkinson, H. J., 138, 138, 139, 145, 154, 159

B Bagnold, R. A., 221,242, 245, 299 Baird, G. B., 370, 389 Baker, G., 343, 382, 383, 387, 389 Ballard, L. A. T., 8, 115 Bambergs, K., 150, 154 Banerjee, D. K., 127,154 Barber, S . A., 133, 134, 135, 152, 156, 159

Barbier, G., 130, 145, 146,154,159 Barnes, T. W., 38, 40, 116 Bamette, R. M.,128, 149, 157, 158 Barshad, I., 380, 381, 389 Bastisse, E. M.,368, 373, 389, 390 Bateman, H. P., 46,117 Bates, C. G., 241,283, 299 Baughman, N. M.,132, 154 Bay, C. E., 308, 308,311,314,316 Bazilevich, N. I., 343, 387, 389 Bear, F. E., 139, 142, 144, 148,155, 156 Beavers, A. H., 343, 383, 387, 389 Beckwith, R. S., 143, 151,154,158 Bedell, G. D., 306, 309, 310, 311, 314, 315 Beeson, K. C., 132, 136, 143,154, 157 Behrens, R., 195, 209 Behrman, A. S., 373, 374,375,389 Beijerinck, M. W., 148, 154 Bendken, L. E., 54,117 Benedek, J., 147,156 Bennett, C. F., 83, 93, 115 Bennett, H. H., 232, 299 Berger, K. C., 130, 132, 136, 137, 145, 154, 157, 158 Bergman, I., 380,389 Bernard, R. L., 172, 210 Betz, L. D., 373, 374, 389 Bhatti, H. M.,138,159 Bien, G. S., 387, 375,389 Biggar, J. W., 130, 154 Birrell, K. S., 343, 344, 381,389 Bisal, F., 237, 300 Biswas, T. D., 139, 154 Biswell, H. H., 200, 209 Bjerrum, J., 128, 154 BIack, A. L., 277, 288,300 Black, C. A., 371, 393 Black, J. N., 19, 21, 41, 81, 83, 87, 88, 91, 92, 93, 99, 100, 101, 103, 107, 115

Blackman, G. E., 19, 53, 65, 75, 83, 115, 116 Blaser, R. E., 74, 115 Bloomfield, C., 133, 137, 149, 154, 155, 157, 387, 390

397

398

AUTHOR INDEX

Boawn, L. C., 152, 158 Bobikyan, R. A., 135,154 Bobko, E. V., 130,155 Bodrov, V. A., 241, 299 Bohmont, D. W., 195,209 Boischot, P., 144, 155 Boken, E., 144, 153,159 Bolt, G. H., 357, 390 Bolton, J. L., 319, 322, 326, 328, 334, 337 Bond, G., 61, 115 Bourbeau, G. A., 340, 342, 346, 349, 350, 351, 352, 392 Bower, C. A., 130,156 Boyle, L. W., 183, 209 Boyle, W. D., 206,209 Bradfield, R., 257, 299 Bradford, G. R., 134, 155 Brady, N. C., 74,115,308,310,315 Brannock, W. W., 355, 356, 357, 387, 396 Bray, R. H., 82, 116, 127, 154 Breeze, B. F., 377, 379, 390 Breland, H. L., 82, 116 Bremner, J. M., 133, 155 Brewing, 0. H., 282,300 Brewster, J. F., 135, 156 ' Brill, G. D., 306, 308, 312, 316 Brim, C. A., 34, 116 Brindley, G. W., 384, 390 Britton, H. T. S., 358,390 Broadbent, F. E., 134,155,310,315 Bromfield, S. M., 146, 151, 155 Bronsart, H. v., 147, 155 Brougham, R. W., 61, 88,95,101,115 Brown, A. L., 126, 155 Brown, B. A., 53,115 Brown, G., 125,155 Brown, J. C., 143, 151, 155 Brown, S. M., 389, 392 Browning, G. M., 286,299 Brunt, D., 218,219, 299 Bryant, J. C., 308, 309, 311, 312, 313, 315 Brydon, J. E., 348, 350, 351, 390 Buckman, H. O., 308,310,315 Buckwold, S. J., 371, 394 Bukharayeva, L. G., 345,395 Bula, R. J., 49, 115 Burkser, E. S., 139, 155

Burlingham, E. F., 319, 336 Burriel, F., 139, 155 Butkevich, K. P., 148, 151,159 Butler, J. R., 140, 155 Byers, H. G., 139, 159

C Cable, D. R., 195, 209 Cabom, J. M., 241,284, 299 Cady, J. G., 341, 342, 345, 346, 349, 350, 351, 352, 370, 382, 387, 388, 390, 395 Callihre, S., 126, 155, 358, 380, 390 Camp, A. F., 143,155 Cann, D. B., 351,390 Carleton, E. A., 306, 307, 316 Carlton, A. B., 287, 289, 301 Camahan, H. L., 330,336 Carr, M. H., 121, 155 Carr, R. M., 357, 390 Carroll, D., 354, 377, 390 Cassidy, H. G., 357, 374, 390 Chabannes, J., 130, 154 Chandler, R. F., 340, 393 Chandler, W. V., 309, 310,316 Chang, T. N., 377, 381, 391 Chapman, H. D., 310,315 Chapman, W. H., 109,116 Charles, A. H., 6, 115 Chen, W. T., 364, 393 Chepil, W. S., 221, 222, 223, 224, 225, 227, 228, 229, 230, 233, 234, 235, 236, 237, 238, 239, 240, 241, 242, 244, 245, 246, 247, 248, 249, 251, 252, 253, 254, 255, 256, 257, 258, 259, 280, 281, 262, 253, 284, 265, 287, 288, 289, 270, 271, 272, 273, 274, 276, 277, 278, 279, 281, 282, 284, 290, 291, 292, 293, 294, 295, 296, 298, 299, 300, 301, 302 Childers, J. F., 86, 116 Chipman, E. W., 73, 115 Chippindale, H. G., 103, 115 Christensen, P. D., 142, 144, 155 Clapp, C. E., Jr., 135, 155 Clark, E. R., 358, 390 Clark, J. S., 348, 390 Clark, K. W., 326, 334, 336, 337 Clarke, F. W., 341, 366,388, 390

AUTHOR INDEX

Clarke, W. S., 86, 116 Clelland, D. W., 360, 390 Clements, F. E., 3, 4, 7, 27, 40, 105, 115 Cline, M. G., 341, 342, 344, 345, 349, 352, 368, 369, 370, 371, 375, 376, 378, 381, 382, 388, 395 Coleman, D., 139, 155 Coleman, N. T., 135,155 Coley, W. C., 74, 75, 116 Contois, D. E., 367, 375, 389 Cook, H. L., 315,316 Cooper, R., 364, 390 Cope, J. T., 310, 316 Cornelius, D. R., 199, 209 Correns, D. W., 354, 390 Cowling, D. W., 62, 116 Cox, J. E., 364, 366, 391 Cox, W. H., 248,301 Crompton, E., 346, 387, 390 Crowe, G. B., 178, 210 Crumpton, C. F., 383, 393 Cumming, W. M., 360,390

399

Demolon, A., 368, 373, 390 Dempster, P. B., 360, 361, 390 103, DeMumbrum, L. E., Jr., 127, 128, 129, 155 346, Denny, J. J., 355, 390 372, Desjardins, J. G., 145, 154 393, Deuel, H., 358, 390 DeVore, G. N., 129, 155, 340, 364, 366, 381, 390 de Wit, C. T., 6, 12, 13, 42, 43, 45, 115, 117 Dhawan, C. L., 139, 156 D’Hoore, J., 341, 351, 352, 390 Diehl, H., 135, 155 Diknert, F., 356, 391 Dion, H. G., 125,155 Dobritskaya, Y. I., 131, 155 Dolan, M. E., 356,260, 361, 393 Donald, C. M., 10, 14, 15, 19, 21, 22, 27, 31, 33, 54, 55, 56, 57, 58, 59, 60, 61, 71, 72, 77, 83, 87, 88, 89, 92, 97, 98, 104, 105, 106,115,117, 118 Dopler, T. F., 139, 155 Dorph-Petersen, K., 150, 155 Douglas, F. D., 199, 210 D Dragsdorf, R. D., 128,156 Daday, H., 319, 327, 329, 334, 336, 337 Drake, M., 74, 75, 11 6 Dalton, F. H., 147, 155 Drapala, W. J., 71, 115 Dana, J. D., 340,390 Drozdova, T. V., 134, 157 Daniel, H. A., 247, 248, 300, 301, 309, Dubachova, T. D., 130,155 316 Duchaufour, P., 388, 391 Daniel, L. J., 135, 156 Duda, J., 135,156 Danielson, R. E., 81, 117 Dudley, J. W., 319, 331, 333, 336 Dapples, E. C., 367, 378, 390 Duley, F. L., 275, 300, 304, 308, 311, Darab, K., 377, 378, 395 312,316 Darrow, R. A., 195, 209 Duff, R. B., 364, 396 Davidson, J. L., 27, 87, 88,92, 93, 115 Dungan, G. H., 16, 48, 86, 69, 70, 72, Davies, E. B., 145, 155 73, 115, 116, 117 Davies, G. J., 10, 115 Dunklee, D. E., 130, 157 Davis, F. S., 178, 193, 210 Durroux, M., 144, 155 Dawson, J. E., 134, 155, 383,390 Dyazdevich, G. S., 351,391 DeBruger, P. L., 355, 356,360,395 DeConinck, F., 387, 390 E de Groot, A. J., 144, 155 Eaton, F. M., 129, 156 DeKimpe, C., 343, 344, 358,380, 390 Edye, L. A., 319, 329,336 Delas, J., 144, 155 Egorov, M. M., 361, 391 Ehrlich, W. A., 342, 350, 351, 391 Delmas, J., 144, 155 Einstein, H. A., 222, 224, 300 DeLong, W. A., 150,155, 387,390 Demias, C., 144, 155 Elanan, P., 137, 144,157,159 Elberse, W. T., 77, 118 Demlon, B., 362, 366, 391

400

AUTHOR INDEX

Elder, W. C., 195, 199,209 Elgabaly, M. M., 128, 127, 128, 129, 156 Ellis, J. H., 342, 350, 351, 391 Ellison, W. D., 313,316 El-Samni, El-Sayed Ah, 222, 224, 300 Elwell, H. M., 195,209,309,316 Emel'yanova, M. P., 134,157 Engel, W., 358, 383, 391 Engledow, F. L., 67,116 Englehom, C. L., 242,291,302 Ennick, G. C., 42,115 Ensminger, L. E., 310, 316 Erdmann, M. H., 31,33,116 Ervin, G. E., Jr., 356,394 Esquevin, J., 126,155,358,380,390 Evans, L. T., 133,156 Evans, S. T., 77,117

F Faust, G. T., 352,389 Fellows, M., 129,156 Fenster, C.R., 277, 278, 281, 300 Ferber, A. E., 287, 300 Ferguson, W. S., 150, 156 Ferris, H. F., 67,68,117 Fieldes, M., 340, 341, 342, 343, 344, 349, 352, 361, 377, 389, 391 Filippov, A. I., 130, 155 Fireman, M., 130, 154 Fisher, C. E., 195, 209 Flach, K. W., 387, 391 Fly, C. L., 248, 300 Forman, S. A., 350,351,394 Forsee, W. T., Jr., 140,156 Fortmann, H. R., 143,156 Foster, R. L., 247, 301 Foumier, R. O., 347, 349, 355, 384, 391, 393, 394 Foy, C. L., 181,210 Frankart, R., 372, 391 Frederickson, A. F., 348, 361, 364, 366, 391 Free, G. R., 306,307, 308,316 French, C. E., 143,156 Frink, J. R., 376, 391 Fripiat, J. J., 372, 391 Frush, H. L., 135,156 Frye, J. C., 248, 301 Fujimoto, C. K., 139, 142, 144, 150, 156

Fujinaga, T., 131, 156 Fyfe, W. S., 357, 390

G Gaddum, L. W., 149,158 Gall, 0. E., 126, 149, 157, 158 Gallego, R., 139,155 Gamble, E., 137, 159 Ganichenko, L. G., 361, 391 Gard, L. E., 315, 316 Garrels, R. M., 348, 363, 364, 366, 369, 391 Garrison, C. S., 322, 337 Gamer, S., 317, 321, 322, 327, 336, 337 Gastuche, M. C., 343, 344, 358, 362, 366, 372, 380, 390, 391 Geiger, R., 218, 219, 240, 300 Gerei, L., 147, 156 Germer, L. H., 355, 391 Gerretsen, F. C., 146, 156 Gibb, J. G., 361, 391 Gibbs, 0. E., 124, 128,156 Gieseking, J. E., 127, 158 Gilbert, C. M., 377, 379,391 Gilbert, F. A., 136,156 Giles, C. H., 357, 374, 391 Giles, G. R., 136, 145, 146, 154, 159 Gleen, H., 146, 147, 156 Goldich, S. S., 121, 158, 347, 350, 351, 391 Goldschmidt, V. M., 121, 122, 123, 136, 138, 156 Goldstein, A., Jr., 379, 391 Goldstein, S., 238, 300 Goodall, D. W., 12,27,63,64,116 Gordon, M.,357, 392 Gordon, R. L., 360, 361,394 Goto, K., 107, 109, 117,376, 394 Govett, G. J. S., 358, 359,391 Gradusov, B. P., 351,391 Gradwell, M., 344, 381, 389 Graham, C. A., 199,209 Graham, E. R., 362,391 Graumann, H. O., 319, 320, 321, 322, 335,336, 337 Gray, B., 74, 75, 116 Gray, L., 143, 154 Gray, L. F., 135, 156 Greb, B. W., 277,286,300

AUTHOR INDEX

401

Heintze, S. G., 125, 133, 144, 155, 156 Hemstock, G. A., 148, 156 Hendricks, S. B., 352, 389 Henin, S., 126, 155, 358, 380, 390, 392 Henriksen, A., 144, 156 Herriott, J. B. D., 62, 116 Hervey, D. F., 199, 210 Hesketh, J. D., 86, 95, 116 Hesselgesser, J. M., 364, 394 H Heston, W. M., 353, 354, 359, 367, 373, Haas, A. R. C., 139, 156 389 Halvorson, H. O., 137,158 Heydemann, A., 127, 156 Hammer, K. C., 143, 154 Hibbard, P. L., 126, 138,156 Hanley, F., 47, 116 Hickok, R. B., 306, 309, 310, 311, 314, Hansen, N. E., 317, 336 315, 316 Hanson, C. H., 319, 322, 329, 331, 333, Hide, J. C., 266, 300 337 Hill, A. C., 139,148,156 Hanson, H., 3, 4, 7, 27, 40, 103, 105, 115 Hdton, J. L., 162, 210 Hanson, W. D., 17, 34,116 Himes, F. L., 133, 134, 135,156 Harder, E. C., 370, 392 Hinish, W. W., 143, 156 Harder, H., 130, 156 Hinson, K., 17, 34, 116 Hardt, G., 267, 300 Hochster, R. M., 147, 156 Hardy, F., 352, 392 Hodgson, G. L., 19, 65, 116 Harlan, H. V., 109, 116 Hodgson, J. F., 126, 127, 128, 129, 133, Harman, R. W., 356, 392 152, 156, 159 Harmer, P. M., 150, 158 Hodgson, J. M., 170, 171, 210 Harper, H. J., 282, 300, 304, 309, 316 Hoff, D. J., 65, 116 Harper, J. L., 3, 5, 12, 14, 25, 27, 69, Hofner, W., 137, 158 116 Holliday, R., 10, 12, 13, 14, 16, 39, 116 Harrassowitz, H., 343, 392 Holmes, R. S., 151, 155 Harris, V. C., 174, 179, 199, 209 Holstun, J. T., Jr., 173, 178, 210 Harrison, C. M., 31, 33,116 Holt, N. B., 135, 156 Harrison, J. B., 352, 368,392 Holt, P. F., 355, 380,381, 392 Harvey, W. A., 196, 210 Hoon, R. C., 139, 156 Hashimoto, I., 344, 392 Hopkins, E. S., 264, 267, 289, 300 Hatcher, J. T., 130, 156 Hopp, H., 240, 301 Hathaway, J. C., 354, 377, 390 Howard, P., 348, 363, 366, 391 Hauser, E. A., 364, 381,392 Howison, J. W., 381, 395 Hauser, E. W., 183, 209 Hozumi, K., 8, 26, 116 Havis, L., 266, 300 Hull, H. H., 306,308, 311, 314, 316 Hawkes, H. E., 127,158, 346,389 Humbert, R. P., 352, 370, 392 Haydock, K. P., 319, 329,336 Hunt, I. V., 150, 159 Haynes, J. L., 14, 19, 22, 80, 116 Hunt, T. F., 14, 117 Hays, 0. E., 306, 307, 308, 309, 311, 312, Hurley, U. K., 112, 116 314, 316 Hurwitz, C., 147, 155 Hazel, F., 357, 392 Hyder, D. N., 195, 210 Heavens, 0. S., 361, 392 I Heinecke, A. J,, 86, 116 Heinrichs, D. H., 319, 322, 326, 327, Ignatieff, V., 156 328, 329, 330, 331, 332, 334, 335, Iizuka, H., 284, 300 Iler, R. K., 340, 347, 349, 353, 354, 355, 337

Green, J. O., 62,116 Greenberg, S. A., 353, 356,377, 381, 391 Greene, H., 353, 385, 392 G r a n , 0. G., 360,361,394 Grimes, D. W., 78, 81, 116 Gdyakin, I. V., 145, 156 Gumbel, E. J., 214, 232, 300 Gustafson, H., 373,374, 375,389

402

AUTHOR INDEX

Kalinske, A. A., 224, 300 Kamprath, E. J., 144, 158 Kanwar, J. S., 139, 158 Karpachevskii, L. O., 368, 392 Karper, R. E., 78, 79, 116 Kasanaga, H., 87, 95, 96,116 Katalymov, M. V., 137, 157 Katsura, T., 387, 392 Kee, N. S., 133, 149, 157 Keller, W. D., 340, 345, 347, 354, 362, 364, 370, 380, 392 Kelley, W. P., 128, 157, 389, 392 J Kempen, H. J., 181,210 Jackob, J. A., 46, 11 7 Kempis, E. B., 331, 395 Jacks, G. V., 235,300, 364,383,392 Kennedy, W. K., 116 Jackson, E. A., 380, 392 Jackson, M. L., 127, 128, 129, 155, 306, Kerr, H. D., 193, 210 307, 311, 312, 314, 316, 340, 341, Kerr, P. F., 343, 395 342, 343, 344, 345, 346, 349, 350, Khan, D. H., 376,392 351, 352, 381, 387, 388, 389, 392, Khodakov, G. S., 381, 391 Kiesselbach, T. A., 88,69, 116 393, 394, 395 Kilmer, V. J., 309, 316 Jacob, W. C., 315, 316 King, D. T., 360, 361, 392 Jamison, V. C., 128, 157, 257, 299 King, E. J., 355,360, 393 Jansen, L. L., 162, 210 Kira, T., 8, 10, 11, 12, 13, 23, 24, 26, Jantti, A., 81, 116 116, 117 Jarvis, R. H., 47, 116 Kiselev, V. F., 381, 391 Jeffries, C . D., 354, 377, 392 Kissinger, H. E., 128, 156 Jeffries, H., 224, 300 Kitagishi, K., 358,382, 396 Jenny, H., 156, 380, 392 Kitto, P. F., 360, 361, 393 Jenny, M., 126, 128, 129, 157 Jensen, H. L., 132, 157 Klages, K. H., 277, 301 Jensen, M. E., 65, 81, 117 Klinger, B., 199, 210 Jephcott, C. M., 355, 392 Klingman, D. L., 178, 193, 195, 197,209, Jessup, R. W., 380, 392 210 Joffe, J. S., 368, 369, 370, 373, 388, 392 Knoblauch, H. C., 306, 308, 312,316 Joham, H. E., 149,154 Knott, J. E., 132, 157 Johnson, C. M., 71, 115 Knox, E. G., 379, 393 Johnson, I. J., 31, 33, 114 Kobayashi, J., 366, 393 Johnston, J. H., 355,392 Kohnke, H., 308, 309,310,311, 314,315, Jones, A., 329, 331, 337 316 Jones, G. B., 137, 157 Kolodny, L., 306, 308, 312, 316 Jones, H. W., 126,157 Komarov, V., L., 318, 337 Jones, L. H. P., 125, 131, 157 Kosegarten, E., 150, 157 Jones, M. G., 52,116 Kovda, V. A., 367, 377, 378,393 Jordan, J. V., 136, 157 Koyama, H., 8,23,24,26,116 Jorge, N. H., 49, 51, 116 Kozlowski, T. T., 81, 114 Joseph, A. F., 358, 392 Kramer, J. J., 86, 11 6 Kramer, P. J., 81, 116 K Krasil’nikov, K. G., 361, 391 Kabata, A,, 138, 157 Krauskopf, K. B., 122,157, 340, 353, 354, Kalashnikova, R. A., 343, 387, 389 356, 357, 358, 359, 364, 367, 373, 382, 383, 388,389,392 Ippen, A. T., 222, 225, 300 Irwin, D. A., 355, 390 Isbell, H. S., 135, 156 Ishibashi, M., 131, 156 Ivanov, D., 140, 158 Iwaki, H., 19, 83, 90,91, 116 Iwasaki, I., 367, 392 Iyer, J. G., 139, 156

403

AUTHOR INDEX

356, 357, 359, 366, 367, 375, 388, 393 Krusekopf, H. H., 312, 313,316 Kubiena, W. L., 368,376,388,393 Kubota, J., 121, 132, 137, 139, 140, 148, 149, 152, 154, 157 Kurtz, T., 82, 116 Kuwamoto, T., 131, 156

1 Lal, K. N., 136, 157 Lamb, J., Jr., 306, 316 Lamm, C. G., 132,157 Lane, D. E., 279, 300 Lang, A. L., 16, 66, 69, 72, 73, 115, 116 Langan, L. N., 132,157 Langham, W. H., 247, 301 Lanning, F. C., 383, 393 Larson, R. E., 178, 195, 209,210 Larson, W. E., 50, 116 Lather, W. M., 373, 393 Laurelle, J., 387, 390 Lazar, V. A., 132, 157 Lee, J. A., 72, 110, 116 Leeper, G. W., 125, 139, 157 Lees, H., 133, 155 Lehane, J. J., 241, 286, 301 Lemon, E. R., 148,149,157 Leonard, C. D., 143,159 Leonard, 0. A,, 196,210 Levick, R., 137, 138, 139, 159 Lewin, J. C., 383,384,393 Lewis, A. H., 150, 156 Leyden, R. F., 143,157 Lindsay, F. K., 374,393 Linehan, P. A., 61, 116 Lipsett, J., 76, 117 Loewenstein, W., 347, 364, 393 Lovering, T. S., 382, 383,393 Lovvom, R. L., 137, 159 Low, P. F., 148,156,371, 393 Lowe, J., 61, 116 Lucas, C. C., 354, 356,360, 361, 393 Ludwig, L. J., 325, 337 Lundblad, K., 137, 144, 157, 159 Lutz, H. J., 340, 393 Lyles, L., 274, 280, 281, 300, 301 Lynch, R. D., 235, 236, 271, 281, 282, 302

M McBain, J. W., 374, 393 McCaleb, S. B., 342, 350, 351, 388, 393 McCalla, T. M., 235, 254, 259, 2433, 266, 281, 301 McCarty, M. K., 197, 210 McCloud, D. E., 86,92,94,114 McClung, A. C., 135, 155 McConnell, D., 364, 393 McCool, M. M., 150, 157 McDowall, I., 381, 393 MacEwan, T. H., 374, 391 McGeorge, M., 355, 380, 393 McHargue, J. S., 141, 157 MacKay, D. C., 73, 115 McKay, H. C., 277, 301 McKeague, J. A., 344, 346,358, 368, 369, 370, 371, 372, 373, 375, 376, 378, 381, 382, 387, 388, 393 McKenzie, J. O., 137, 157 Mackie, W. Z., 340, 341, 392 MacLean, A. J., 135, 154 MacLeod, L. B., 370, 371, 393 McNaught, K. J., 150, 157 McWhorter, C. G., 172, 173, 178,210 Magistad, 0. C., 376, 393 Maguire, J. I., 373, 374, 389 Makhonina, G. I., 137, 159 Malin, J. G., 235, 301 Malihska, E., 135, 156 Malquori, A., 150, 157 Malyuga, D. P., 132, 157 Mann, H. H., 38,40,116 Mann, P. J. G., 125, 133, 155, 157 Manskaya, S. M., 134, 157 Marbut, C. F., 341, 393 Marion, P. T., 195, 209 Marsh, G. C., 150, 159 Marshall, C. E., 124, 128, 156, 340, 361, 362, 363, 366, 394 Martin, A. E., 133,157 Martin, J. P., 146, 157, 259, 28.5, 266, 301, 307, 316 Martin, S. C., 195, 210 Martini, M. L., 109, 116 Mason, B., 121, 123, 157, 353, 354, 370, 393 Massey, H. F., 152, 157, 307, 311, 312, 314, 316 MBtB, F., 147, 156

404

AUTHOR INDEX

Mather, K., 7, 116 Mathews, 0. R., 290, 301 Mattson, S., 357, 376, 393 Matveeva, T. V., 130, 155 Maunsell, P. W., 150, 154 Meadors, C. H., 195, 209 Mederski, H. J., 65, 116, 150, 157 Mehra, 0. P., 345, 381, 393 Melsted, S. W., 82, 116, 127, 154, 158 Metzger, W. H., 266, 300 Middleton, G. K., 109, 116 Midgley, A. R., 130, 157 Miller, E. E., 49, 115 Miller, J. H., 181, 210 Miller, M. F., 304, 308,312, 313,316 Miller, M. H., 145, 157 Millikan, C. R., 147, 150,157,158 Milliken, T. H., Jr., 357, 393 Mills, G. A,, 357,393 Milne, A., 2, 3, 116 Milne, R. A., 221, 222, 239, 240, 241, 245, 300 Milthorpe, F. L., 10, 79, 115,116 Mitchell, R. L., 120, 136, 138, 140, 148, 158, 159

Musgrave, R. B., 86, 95,116 Musick, J. T., 78, 81, 116 Myers, H. E., 285, 301 Myers, L. F., 76, 117

N Naftel, J. A,, 130, 158 Nagelschmidt, G., 360, 361, 394 Nair, C. K. N., 134,155 Nakhwa, S. N., 374, 391 Nash, V. E., 340, 361, 362, 363, 366,394 Neal, 0. R., 308,307, 311, 312, 316 Neal-Smith, C. A., 77,115 Nelson, C. E., 53,54, 117 Nelson, J. L., 127, 152, 158 Nelson, 0. E., 17, 117 Nijhawan, S. D., 139, 158 Nikiforoff, C. C., 370, 377, 379, 386, 394 Nilan, R. A., 319,322,337 Noll, C. A., 373, 374,389 Northmore, J. M., 134, 158 Novorossova, L. E., 383, 394 Nozdrunova, E. M., 150,158 Nutting, P. G., 362,394

Mitchell, W. A., 364, 396 Mitra, S. P., 127, 158 0 Mitscherlich, E. A., 10, 117 Oakley, H. G., 356,392 Mohr, E. C. J., 341, 349, 351, 352, 353, Oakley, R. A., 317,322,327,337 368, 370, 379, 380,386,393 Oblad, A. G., 357, 393 Moldenhauer, W. C., 315, 316 OConnor, T. L., 359, 394 Monsi, M., 82, 87, 95, 96, 103, 116, 117 Odland, T. E., 320, 337 Moore, David P., 135, 155 Ogawa, H., 11, 12, 116 Moreland, D. E., 162, 210 Ohlrogge, A. J., 17, 117, 145, 157 Morey, G. W., 355, 364,393,394 Ohnishi, Y., 358, 382, 396 Morley, F. H. W., 327, 329, 331, 332, Oka, H. I., 107, 117 334, 337 Okamoto, G., 376, 394 Morrow, C. E., 14, 117 Okura, T., 376, 394 Morton, H. L., 195, 209 Oliver, G. W., 317, 321, 325, 337 Mosebach, R., 364, 394 Olsen, F. R., 31, 33, 117 Moss, H. C.,248, 301 Olson, R. V., 130, 158 Mott, G. O., 53, 117 Orchiston, H. D., 76, 118 Mottershead, B. E., 334, 336 Ortega, E., 143, 158 Mouat, M. C. H., 75,117 Osbome, V., 348, 390 Muckenhim, R. J., 309, 316 Overbeek, J. T. G., 355, 356,360, 395 Mulder, G., 147, 158 P Murata, K. J., 355, 356, 357, 362, 366, 367, 394, 396 Page, A. L., 128, 129, 158, 159 Murray, B. E., 320, 322, 323, 326, 336, Palmer, A. E., 264, 289,300 Palmer, T. P., 334, 337 337

AUTHOR INDEX

405

Papadakis, J. S., 34, 117 Pronin, M. E., 140, 158 Parker, K. W., 195, 210 Prusky, S. C., 358,383, 395 Parkinson, G. R., 234, 301 Puckridge, D. W., 18, 19,22, 83, 117 Parks, R. Q., 130, 158 Parks, W. L., 135, 158 Q Paterson, M. S., 360,389, 394 Quastel, J. H., 125, 147, 156, 157 Patry, L. M., 350,351,390,394 Patterson, H. S., 300, 361,393 R Paul, G. W., 150,157 Rachmat Hardjosoesastro, R., 341, 394 Pauling, L., 348, 394 Pawluk, S., 350, 351, 394 Radoslovich, E. W., 364, 390 Peach, M., 133, 152, 159 Randhawa, N. S., 139, 158 Peak, J. W., 327, 334,337 Rao, H. G. G., 147,158 Pedziwilk, Z., 135, 156 Rao, M.S. S., 138, 157 Peech, M., 126, 158 Raupach, M., 368, 369, 370, 373, 378, Peech, M. J., 369, 370, 396 387, 394 Peek, T. C., 259, 265, 301 Reestman, A. J., 12, 117 Pendleton, J. W., 16, 34, 46, 48, 68, 89, Reeve, R., 133, 157 70, 72, 73, 101, 115, 116, 117 Reiche, P., 345, 394 Penman, H. L., 81, 117 Reifenberg, A., 353, 368, 371, 376, 388, Pennington, R. P., 143, 156, 340, 341, 394 342, 340, 349, 350, 351, 352, 388, Remezov, N. P., 308, 394 Reuszer, H. W., 152, 158 392, 394 Reuther, W., 137, 158 Perekaljskii, F. M., 70, 117 Perici, E., 150, 157 Rice, H. M., 342, 350, 351, 391, 394 Perkins, A. T., 128, 156 Riceman, D. S., 77, 117, 137, 157 Rich, C. I., 342, 343, 348, 380, 381, 394 Peters, E. J., 178, 193, 210 Richards, L. A., 2-52, 301 Peters, R. A., 193, 210 Richardson, E., 358,358, 394 Petersen, M. L., 54, 117 Pettijohn, F. J., 341, 347, 349, 350, 377, Richardson, J. P., 343, 391 Richardson, P. W., 127, 158 379, 394 Ridgman, W. J., 47, 116 PbwWB, T. L., 248, 301 Riecken, F. F., 139, 158 Philipson, T., 130, 158 Phillip, J. R., 87, 93, 115 Ringwood, A. E., 121, 127,158 Riquer, J., 343, 383, 387, 394 Phillips, W. M., 182, 210 Ritchie, P. D., 360, 361, 390, 391 Pickworth Farrow, E., 84, 117 Piper, C. S., 143, 145, 158, 370, 376, 394 Roberts, J. L., 31,33,117 Robins, J. S., 53, 54,117 Piryutko, M. M., 353, 350, 357, 394 Robinson, E. D., 195,209 Poldervaart, A., 341, 394 Polynov, B. B., 229, 301, 348, 368, 394 Robinson, P. R., 5, 53,117 Robinson, W. O., 139, 159 Ponnaiya, B. W. X.,383, 393 Ponnamperuma, F. N., 148,158,369,394 Robson, W. D., 355, 390 Rodriguez, G., 352, 392 Porter, K. B., 65, 81, 117 Rogers, H. T., 308, 310, 311, 312, 313, Potter, W. D., 232, 301 314, 316 Powers, W. L., 136, 157 Rogers, L. H., 149,158 Prakash, D., 127, 158 Rogers, V. E., 334, 336, 337 Prendergast, J. J., 112, 117 Roller, P. S., 358, 394 Price, E. W., 353, 356, 391 Ross, C. S., 343, 395 Prill, R. C., 139, 158 Rossiter, R. C., 17, 117 Prince, A. L., 144, 158

406

AUTHOR INDEX

Rovira, A. D., 151, 158 Rowe, J. J., 347, 349, 355, 391, 394 Roy, C. J., 353, 356, 367, 395 Rozhnova, T. A., 369, 395 Russell, E. J., 382, 395 Russell, E. W., 263, 301 Russell, M. B., 81, 117 Rytikova. M. N., 150,158 Ryznar, J. W., 374, 393

s Saeki, T., 82, 87, 95, 96, 103, 117 Sakai, K. I., 64, 106, 107, 109, 117 Sakazaki, N., 11, 12, 116 Salisbury, H. F., 150,155 Salter, P. J., 79, 80, 117 Sampson, A. W., 304, 316 Samuel, G., 146, 158 Sanchez, C., 144, 158 Sandell, E. B., 121, 158 Santhirasegaram, K., 47, 50, 51, 52, 70, 117 Sarata, U., 150, 158 Satyanarayan, Y., 139, 156 Saunders, A. M., 319, 329,336 Sayre, J. D., 14, 19, 22, 80, 116 Scarseth, G. D., 309, 310, 316 Scharrer, K., 137, 158 Schastnaia, L. S., 369, 395 Schlichting, E., 144, 158 Schnitzer, M., 134, 137, 158, 159, 387, 390 Schock, R. V., Jr., 357, 392 Schultz, A. M., 200, 209 Schultz, H. B., 287, 289, 301 Schwarzenbach, G., 128, 154 Sears, P. B., 232, 235, 301 Sears, P. D., 62, 117 Sedletskii, I. D., 140, 158 Seif, R. D., 34, 101, 117 SeifFert, H. H., 150,158 Semeniuk, G., 319, 336 Shafer, N. E., 193, 210 Shapter, R. E., 54, 74, 118 Sharpe, J. W., 361, 391 Shaw, B. T., 130,158 Shaw, W. C., 162, 163, 172,210 Shelfond, V. E., 3, 115 Shemyakina, A. F., 150, 158 Sheng, T., 287,288, 301

Sheppard, P. A., 220,301 Sherman, D. G., 139, 142, 144, 150, 156 Sherman, G. D., 150, 158, 340, 392 Shinozaki, K., 10, 12, 13, 117 Siddoway, F. H., 225, 233, 234, 274, 277, 281, 282, 290, 293, 295, 298, 300, 301 Siever, R., 340, 354, 355, 356, 366, 367, 368, 377, 378, 379, 384, 388, 395 S i k , B. J., 143, 159 Sillen, L. G., 128, 154 Simon, J., 199, 210 Simonson, G. H., 139, 158 Simonson, R. W., 385,395 Singh, G. D., 70, 315 Sisodia, U. S., 70, 115 Sivarajasingham, S., 341, 342, 345, 346, 349, 352, 382, 395 Skerman, V. B. D., 146,155 Skogley, C. R., 320, 337 Slater, C. S., 240, 301, 306, 307, 308, 309, 311, 312, 313, 315, 316 Slatyer, R. O., 81, 117 Sletten, W. A., 65, 81, 117 Slife, F. W., 46, 117 Smith, C. B., 143, 156 Smith, D., 49, 108, 115, 117, 319, 320, 322, 337, 374, 391 Smith, D. D., 304, 315,316 Smith, H. F., 67,68, 117 Smith, P. F., 137, 158 Smith, R. J., Jr., 172, 173, 177, 178, 210 Smithson, F., 341, 342, 343, 382, 383, 384, 395 Sneesby, N. J., 288, 301 Snell, C. A,, 358, 395 Snell, C. T., 358, 395 Snell, F. D., 358, 395 Sneva, F. A., 195,210 Sobolev, S. S., 287, 301 Sonder, L. W., 200, 210 Southworth, W., 317, 322, 337 Sowell, W. F., 17, 117 Specht, A. W., 137, 158 Spencer, W. F., 127, 158 Sprague, H. B., 5,53, 67, 68, 117 Sprague, M. S., 194, 210 Stahler, L. M., 101, 117 Staniforth, D. W., 46, 49, 116, 117, 118 Staple, W. J., 241,286, 301

407

AUTHOR INDEX

Stapledon, R. G., 111, 117 Starkey, R. L., 137, 158 Steenbjerg, F., 144, 147, 150, 153, 159 Stenberg, M., 144, 159 Stephen, I., 343, 383, 387, 389 Stephens, C. G., 380, 395 Stem, W. R., 25, 54, 55, 56, 57, 58, 59, 60, 87, 88, 98, 101, 102, 117, 118 Stewart, I., 143, 159 Stoltenberg, N. L., 309, 312, 314, 316 Storks, K. H., 355, 391 Stradaioli, G., 150, 157 Strickland, J. D. H., 358, 359, 395 Stringfield, G. H., 34, 118 Strong, L. E., 137, 159 Suggs, D. D., 206,209 Sund, J. M., 199, 210 Suneson, C. A., 109, 110, 111, 118 Sutherland, A. J., 150, 155 Svanberg, O., 137, 144, 157, 159 Swaine, D. J,, 123, 136, 138, 140, 148, 159 Swift, R. W., 143,156 Swindale, L. D., 340, 341, 343, 352, 377, 387, 391, 395 Swineford, A., 248, 301 SylvCn, N., 108, 109, 118 Szaboles, I., 377, 378, 395

T Talbot, J. H., 361, 395 Tamele, M. W., 357, 395 Targul'yan, V. O., 345,395 Tarutani, T., 367, 392 Taylor, G. A., 143, 156 Taylor, H. F. W., 381, 395 Teakle, L. J. H., 132,159 Tedrow, J. C. F., 350, 351,395 Templeman, W. G., 53, 75,115 Thomas, G. W., 342, 343, 348, 380, 381, 394 Thomas, W. H., 367,375,389 Thompson, J. T., 183, 209 Thompson, L. M., 269, 301 Thomthwaite, C. W., 213, 234, 301 Thorp, J., 137, 159 Tiffin, L. O., 151, 155 Tiller, K. G., 124, 127, 128, 129, 133, 140, 152, 156, 159

Timmons, F. L., 182, 200, 206, 207, 210 Timonin, M. I., 146, 151, 159 Titlyanova, A. A., 137, 159 Toth, S. J., 139, 142, 143, 144, 148, 155, 156, 157, 158, 381, 395 Trapnell, B. M. W., 374, 395 TrocmC, S., 145, 146, 154, 159 Troelsen, J. E., 326, 337 Tmmble, H. C., 54, 74, 118 Truog, E., 132, 137, 141, 145, 154, 157, 159 Tschirley, F. H., 195, 209 Turekian, K. K., 121, 155 Tyler, S. A., 340, 342, 346, 349, 350, 351, 352, 392 Tyulin, A. F., 134, 159 Tyuryukanov, A. N., 137, 159

U Udden, J. A., 248, 301 V Vail, J. G., 358, 395 Vamos, R., 377, 395 Van Baren, F. A., 341, 349, 351, 352, 353, 370, 386, 393 van den Bergh, J. P., 77, 118 Van Doren, C. A., 315,316 Van Lier, J. A., 355, 356, 360, 395 Vanoni, V. A., 246, 301 Van Rummelen, F.F.F.E., 343, 395 Verma, R. P., 222,225, 300 Viehoever, A., 358, 383,395 Vielvoye, L., 362, 366, 391 Viets, F. G., 140, 152, 158, 159 Vilenskii, D. G., 353, 368, 395 Vinogradov, A. P., 123, 138, 139, 159 Vinokurov, M. A., 345,395 Vlasyuk, P. A., 146, 151, 159 Voigt, A., 135, 155 Volk, G. W., 310, 316 von Karman, T., 218, 219, 220, 234, 301 Vose, W., 361, 393

W Wada, K., 344, 395 Waddams, J. A., 358,390 Wahhab, A., 136, 159 Wain, R. L., 143, 150, 159 Walker, J. M., 152, 159

408

AUTHOR INDEX

Walker, T. W., 74, 75, 76, 117, 118 Wallace, A., 145, 159 Wandenbulcke, F., 358, 391 Wam, F. G., 248, 301 Wame, L. G. G., 11, 16,118 Wassermann, V. D., 8, 118 Watkin, B. R., 62, 118 Watson, D. J., 86, 87, 91, 98, 118 Watson, S. J., 150, 156 Wear, J. I., 144, 159 Weaver, J. E., 3, 4, 7, 27, 40, 103, 105, 115 Weber,C. R.,46,49,117,118 Webley, D. M., 364,396 Wehrmann, J., 150, 158 Weihing, R. M., 68, 69, 116 Wells, D. J., 62, 116 Went, F. W., 89,118 Weyl, L. H., 304, 316 Weyl, W. A., 357, 361,396 Wheatley, K., 360, 394 Whetstone, R. R., 139, 159 White, C. M., 224, 225, 229, 301 White, D. E., 355,356, 357, 396 White, J. L., 135, 158, 309, 312, 314, 316 White, W. A., 343, 367, 396 Whitehouse, F. W., 379, 396 Whiteside, E. P., 350, 351, 388, 390, 396 White-Stevens, R. H., 150, 159 Whiffield, C. J., 281, 288, 301, 302 Whitney, R. S., 369, 370, 396 Whittig, L. D., 128, 129,158,159 Whyte, R. O., 235, 300 Wiggans, R. G., 65,118 Wiklander, L., 144, 158 Wilcox, L. V., 129, 156 Wikerson, J. A., 181, 210 Willard, C. J., 172, 210 Williamson, K. I., 342, 391 Williamson, W. O., 379,380, 396 Willis, A. L., 340, 342, 34.8, 349, 350, 351, 352, 392 Willis, W. O., 50, 116

Willoughby, W. M., 76,118 Wills, B. C., 143,159 Wilsie, C. P., 31, 33, 114 Wilson, C. M., 137,159 Wilson, J. H., 150,157 Wilson, J. W., 61, 96, 118 Winchell, A. N., 349, 396 Winchell, H., 349, 396 Winsor, H. W., 137,159 Winter, A., 199, 210 Winters, E., 379, 396 Wischmeier, W. H., 304, 315, 316 Wittmuss, H., 279, 300 Witts, K. J., 96, 118 Woodhouse, W. W., Jr., 137, 159 Woodruff, C. M., 369,370, 371,396 Woodruff, N. P., 235, 238,241,242,246, 247, 249, 271, 274, 277, 278, 280, 281, 282, 283, 284, 289, 290, 291, 292, 293, 295, 296, 298, 300, 301, 302 Wooley, J. C., 309, 316 Wooten, H. H., 192,210 Wooten, 0. B., 172, 173, 210 Wooten, 0. B., Jr., 178, 210 Wright, J. R., 134, 137, 138, 139, 159 Wright, M. J., 199, 210

Y Yang, T. T., 139,159 Yarilova, E. A., 148,159 Yarilova, Y. A., 343,387,389 Yassoglou, N. J., 350,351, 388, 396 Yoshida, S., 358, 382, 396 Yoshinaga, N., 343,344,396 Yudintseva, E. V., 145,156 Z Zende, G. K., 128,159 Zimnovets, B. A,, 387, 377, 378, 393 Zingg, A. W., 220, 232, 234, 242, 274, 281, 283, 284, 289, 291, 293, 295, 301, 302

SUBJECT INDEX A Abortion, 199 Acrolein, 202 Acrylaldehyde, 202 Agropyron repens, 181 Agrostemmu githago, 25 Albite, 351, 364-365 Alfalfa, 46, 164, 187, 193,317-338 creeping, 317-338 creeping-root breeding, 327-330 genetics of creeping-root, 331333 performance of spreading, 333-335 physiology of spreading, 325330 types of root systems, 319-324 Alfisols, 388 A l h gramineum, 207 Allium canudense, 164, 200 Allium vineale, 164, 200 Allophane, 343-344, 349, 384, 385, 386, 389 Aluminum, 130, 376 Amaranthus retroflexus, 187 Amaranthus a nos us, 192 Ambrosia spp., 165 3-Amino-1,2,4-triazole, see Amitrole Amitrole, 190, 191, 192, 195, 205, 208 Ammonium sulfate, 53 Amphiboles, 341 Anatase, 351 Anion exchange, 130 Annual bluegrass, 188 Anthoranthum odoratum, 77 Apple, 86, 191, 192 Arisols, 386 Artemisa, 195 Artemisa sp., 200 Asparagus, 287 Atrazine, 186 Australia lucerne, 39 Azide, 146, 147 Azuki bean, 24

B Bacillus silicus, 365 Bacteria, 146, 151-152 Bahiagrass, 196 Bamboo, 382 Bananas, 28

409

Barley, 1, 10, 42-44, 47, 109-110, 193, 281 Barnyardgrass, 173, 177 Basalt, 120 Beet, 97 Bermudagrass, 61, 112, 181 Biological yield, 9, 14, 15 Biotite, 127, 350, 351 Birdsfoot trefoil, 193 Bitterweed, 164, 199 Blackberries, 190 Blackjack oak, 195 Blueberries, 190, 192 Bluegrass, 197 Boneset, 199 Boron, distribution in soil, 137-140 geochemistry of, 122-124 reactions with soil, 129-131, 132, 135 Bracken fern, 190 Bromus catharticus, 71 Bromus mudritends, 5,69, 77 Bromus rigidus, 5, 69 Bromus sp., 12 Broomcom, 281 Brucite, 127 Buckwheat, 83, 90

C Calcite, 351 Calcium, 127, 130, 131 erosion losses, 304, 312 Calcium nitrate, 53 Camelinu, 44 Canada thistle, 170, 171, 181 Carbon dioxide, 8 competition for, 4, 6, 84 Carpetweed, 187 Carrots, 4 Carya sp., 197 cassia spp., 199 Castanea, 82 Castorbeans, 288 Cation exchange capacity, 75 Cattails, 205 Cauliflower, 79, 80 Cenchus spp., 192 Centaurea calcitrapa, 1 Cereal-legume mixture, 35

410

SUBJECT INDEX

Chalcedonite, 341 Chalcedony, 341,357,383 Chaparral, 195 Chelates, 137 Chemical weed control, 161-210 costs and benefits, 166-168 Chenopodium album, 187 Chernozem, 386 Chickweed, 188, 190 2-Chloro-4,6-bis( isopropylamino) -s-triazine, see Simazine 9- ( p-Chlorophenyl )-1,l-dimethylurea, see Monuron Cholesterol, 358 Chondrilla iuncea, 76 CIPC, 173, 177, 178, 181, 186, 188, 189, 190 Cirsium arvense, 170 Clay, 257, 342, 344, 380-381, 385 Clover, 5, 13, 27, 38, 40, 41, 45, 99, 107, 145, 164 Clover-grass competition, 52-62, 74-77, 99 Coastal bermudagrass, 193, 194 Cobalt, adsorption of, 126-129, 133 availability, 143-152 distribution in soil, 136-140 geochemistry of, 121-124 Cocksfoot, 112 Coconut palm, 28, 38 Companion crop, 193 Competition, see plant competition, 1-118 Competition-Density Effect, 11 Competition index, 36-38 Convolvulus sp., 101 Copper, availability, 143-152 distribution in soil, 136-140 forms of in soil, 126-129, 132-135 geochemistry of, 121-124 Copper sulfate, 207 Cork oak, 28 Corn, 13, 16, 17, 24, 26, 28, 34, 46, 49, 50, 66, 68, 72-74, 95, 101, 164, 172, 178, 181, 274, 281, 308, 310, 311, 312, 313 Cotton, 28, 81, 274, 281, 291, 310 weed control, 164, 174, 179, 180, 181, 199 Cover crop, 46 Crabgrass, 187

Cranberries, 190, 191 Crested wheatgrass, 196 Cristobalite, 341, 349, 357 Crop rotation, 170-171 Crotalaria, 181, 288 Crotaloria mucronata, 164 Crotalaria spectabilis, 164 Cyptomeria, 82 Curly dock, 191 Curly top virus, 165 Currants, 190 Cuscuta spp., 164 Cynodon dactybn, 86, 92, 112, 181 Cyperus rotundus, 172

D 2,4-D, 170, 171, 190, 191, 195, 196, 197, 199,200,205,206 4-(2,4-DB), 192, 193, 195 Dactylis glomerata, 111 Dalapon, 186, 190, 191, 192, 193, 195, 205, 206 Delphinium barbeyi, 199 Delphinium spp., 199 Diatoms, 383 Dichlorobenil, 190 2,6-Dichlorobenzonitrile, 190 2,4-Dichlorophenoxyacetic acid, see 2,4-D 4-( 2,4-Dichlorophenoxy) butyric acid, see 4( 2,4-DB) 3-( 3,4-Dichlorophenyl)-1,l-dimethylurea, see Diuron 3,4-Dichloropropionanilide, see DPA 2,2-Dichloropropionic acid, see Dalapon Digitaria spp., 187 N,N-Dimethyl-2,2-diphenylacetamide, 173 3,5-Dimethyltetrahydro- 1,3,5,2H-thiadiazine-2-thione, 186 3,5-Dinitro-o-cresol, 190 4,6-Dinitro-o-sec-butylphenol,see DNBP Diphenamide, 173 Diuron, 181, 186, 190, 191 DMTT, 186 DNBP, 183, 186, 189, 190 DNC, 190 Dodder, 164, 181 Dolomite, 123 Dormancy, 8

411

SUBJECT INDEX

DPA, 177, 178 Drought resistance, 325

E Echinochoh crusgalli, 173 Eleusine indicn, 187 Endothall, 175 Entisols, 384-385 EPTC, 172, 193 Equisetum sp., 382 Eragrostis lehmanniana, 196 Erosion, see Water or Wind Erosion control, 182 Ethyl N,N-di-n-propylthiolcarbamate, see EPTC Eupatorium perfoliatum, 199 European red mite, 191 Evening primrose, 199 F Fallow, 279, 304, 306 chemical, 182 Feldspar, 124, 341, 342, 347-348, 361, 362 Fenac, 181 Fescue, 104 Festuca pratensis, 103 Flax, 44, 172 Fodder crops, 16, 32 Foxtail, 46, 49, 50 Fragipan, 379

G Gabbro, 120 Genetics, competitive ability, 106-109, 111 Genotype, 15, 16 Genotype-density interaction, 16, 17, 34 Gibbsite, 351, 382 Goosegrass, 187, 190 Grain yield, 15 Granite, 121 Grapes, 188-189 Crass, 5, 29, 45, 47, 98, 274, 275, 382 Grass-clover competition, 52-62, 74-77, 99 Gray speck disease, 146 Growth analysis, 325 Gypsum, 351

H Halloysite, 127, 131, 344, 348 Harding grass, 196 Hausmanite, 125 Hectorite, 127 Hebnium sp., 199 Helenium tenuifolium, 164 Helianthus, 82 Hematite, 351 Henbit, 188, 190 Hickory, 197 Hornblende, 351 Horsenettle, 192

I Illite, 351 Imgolite, 344 Impatiens balsamina, 24 Inceptisol, 385-386 Iodine, 134 Iron, distribution in soil, 136-140 forms in sod, 125-126, 128 geochemistry of, 121-123 Irrigation, 180 Isopropyl N-( 3-chlorophenyl ) carbamate, see CIPC

J Johnsongrass, 172, 189

K Kale, 87 Kaolin, 127, 130, 362 Kaolinite, 131, 344, 348, 351, 382, 385, 387 Knotweed, 187, 188 Kochia, 288

L Lagurus ooatus, 77 Lambsquarters, 187 Lamium amplexicaule, 188 Lantana, 383 Larkspur, 199 Leaf area, fluctuations, 90-95 Leaf area index, 86-88, 96 Leaf arrangement, 95-99 Leaf canopy, 85-90, 96 Legume, 29, 35, 181, 192, 274, 275, 310 Lehmann lovegrass, 196

412

SUBJECX INDEX

Lespedeza, 196, 197, 199 Light, competition for, 4, 5, 7, 8, 18, 19, 29, 47, 58-60, 78, 82-85, 86, 87, 88, 95-99, 101, 103, 104, 105 Lime, 141, 142, 267 Limestone, 123, 312 Little-leaf disease, 147 Lolium italicum, 103 Lolfum perenne, 77, 104 Lolium rigidum, 9, 22, 72 Lucrene, 39, 41, 108, 318

M Magnesium, 121, 313 Manganese, availability, 141-152 distribution in soil, 136-140 forms of in soil, 125-126 geochemistry of, 121, 123 Manganite, 125 Mangolds, 12 Manure, 145, 306, 312 MCPA, 172 Meadow fescue, 112 Medicago fakata, 317, 318, 319, 320, 322, 327,328,334,335 Medicago media, 320 Medicago satiua, 77, 317, 319, 320, 321, 327, 328 Mesquite, 195 Methyl bromide, 186, 188 2-Methyl-4-chlorophenoxyacetic acid, 172 Mica, 341, 348, 385 Micronutrients, chemistry of in soils, 119159 distribution of in soil, 136-140 factors affecting availability, 140-152 forms of in soil, 124-135 geochemistry of, 120-124 Microorganisms, 235, 265 availability of soil nutrients, 145-148, 149 Miscanthus, 82 Mollisols, 386387 Mollugo uerticflhta, 187 Molybdenum, distribution in soil, 136-140 geochemistry of, 121-124 reactions with soil, 131, 132 Montmorillonite, 126, 127, 128, 129, 348, 351, 370, 385

Monuron, 186, 191 Muscovite, 127, 363, 364

N Net assimilation rate, 85 Nitrate, 199 Nitrogen, competition for, 16, 29, 41, 45, 49, 50, 53, 56, 59, 60, 61, 62, 7176, 78, 104, 177 erosion losses, 304, 307-309, 314 Nontronite, 127, 131 Nurse crop, 6. 46 Nutrients, competition for, 4, 5, 18, 29, 47, 70-77 Nutrient losses, see water erosion Nutsedge, 172, 181

0 Oats, 16, 28, 42-44, 48, 125, 145, 146, 151, 193, 308,310,311 Oetwthera sp., 199 Oliver, 28 Orchardgrass, 111 Oyzopis miliacea, 196 7- Oxabicyclo- (2.2.1 ) heptane- 2,3- dicarboxylic acid, 175 Oxisols, 388-389 Oxygen, competition for, 4 , 7

P Palagonite, 344 Pasture, 28, 38, 39, 93, 94, 111 weed control, 192-200 Peach, 191 Peanuts, 81, 177, 182 Peas, 13, 145, 181 Peat, 144 Perennial ryegrass, 9, 16, 39, 53 Phalaris, 105, 106 Phahris tuberosa, 104 Phaluris tuberosa var. stenoptera, 196 Phaseolus chysanthos, 24 Phlogopite, 127 Phosphate, 125, 176-177 Phosphorus, 6, 76-77 erosion losses, 304, 309-311, 314 Photosynthesis, 59, 86, 93 Phragmites, 82 Pigweed, 187 Pine, 86

SUBJECT INDEX

Pineapple, 182 Pinus, 82 Pinus densiflora, 24 Plantain leaf buttercup, 199 Plant arrangement, 63-70 Plant community, equilibrium, 39-45 influence of density, 9-17 Plant competition, 1-118 competitive ability, 106-112 factor interaction, 103-106 grass and clover, 52-62,74-77,99 height, 99-103 index, 36-38 interspecific, 28-38 intraspecific, 46-62 leaf area, 90-95 leaf arrangement, 95-99 leaf canopy and growth, 85-90 light, 82-85,88, 87, 88, 95-99, 101,

103, 104 nature of, 2-9 nutrients, 70-77,103, 105 plant arrangement, 63-70 plant density, 9-28 water, 77-82 Plant density, 9-28,29 Poa annu, 188 PodzoI, 134, 137,138,386,387 Poison ivy, 165, 192 Polygonurn aoiculare, 187 Pondweed, 207 Post oak, 195 Potamogeton spp., 207 Potassium, 74-75,77 Potassium, erosion losses, 311-312,314 Potato, 10, 181 Prosopis, 195 Pseudomonas sp., 365 Pyrolucite, 125 Pyrophyllite, 127 Pyroxenes, 341

413

Ragweed, 165, 199 Ranunculus sp., 199 Rape, 10 Raspberries, 190 Red clover, 48, 109, 193 Red pine, 24 Redtop, 30 Relative growth rate, 23-25 Respiration, 59,86,87 Rhizoctoniu s o h i , 182 Rhizosphere, 146, 151-152 Rhus radicans, 165 Rhyolite, 120 Rice, 64,107,287,358 weed control, 169, 173, 176-177,180 Rumex c r i s p , 191 Runoff, data interpretation, 313-315 sampling methods, 305-306 Russian thistle, 165 Ryegrass, 13,38,40,61, 104, 112

S Sage brush, 195, 196,200 Salsoh kali var. tenuifoliu, 165 Sandburs, 192 Sclerotium rolfsii, 182 Sedges, 190 Sesone, 183,186, 187,188,190 Setarta species, 49 Sewage effluent, 145 Shale, 122, 123 Shinnery oak, 195 Sierozem, 386 Silica, chemistry of, 353-359 deposition of in soil, 377-384 kinds of soils, 384-389 solid forms, 340-353 solution forms, 360-377 Silicates, 122, 128 Silt, 257 Silvex, 199, 206 Simazine, 186, 190, 191,194 Q Skeleton weed, 76 Quackgrass, 181 Quartz, 341, 342, 347, 349, 350, 351, Smartweed, 49 SMDC, 186 354-355,357, 360, 383 Smilograss, 196 Quercus spp., 195 Sodium 2,4-dichlorophenoxymethyl sulR fate, see Sesone Sodium-N-methyldithiocarbamate, 186 Radish, 24

414

SUBJECr INDEX

Soil, micronutrient elements in, 119-159 silica in, 339-396 Soil rot fungus, 182 Solanum carolinense, 192 Sorghum, 78, 81, 175, 182, 274, 281, 291 Sorghum halapense, 172 Southern blight fungus, 182 Soybeans, 13, 17,24,46,49, 152 weed control, 164, 171, 175, 178, 181 Spiny pigweed, 192 Spring wheat, 170, 171 Spodosols, 387 Stellaria media, 188 Stem rust, 182 Strawberry, 89, 187-188, 192 Strip cropping, 289-291 Stubble mulching, 277 Subterranean clover, 9, 17, 19-22, 25, 51, 87, 90, 91, 92, 100 Sudangrass, 71, 281 Sugarbeet, 87, 165, 176 Sulfides, 121 Sulfur, 74 erosion losses, 304, 313 Sunflower, 288 Superphosphate, 310

T 2,4,5-T, 191, 195, 196, 197, 199, 206 Talc, 127 Tanveed, 199 TCA, 172, 175, 194, 195 2,3,5,6-Tetrachlorotetraphthalicacid, 190 Thiocarbamates, 173 Timothy, 30, 112 Toluene, 146 Tomatoes, 173 Touch-me-not, 24 Tourmaline, 124 Trichloroacetic acid, see TCA 2,4,5-Trichlorophenoxyacetic acid, see 2,4,5-T 2- ( 2,4,5-Trichlorophenoxy) propionic acid, see Silvex Trifoliicm subterraneum, 8, 22 Triticum vulgare, 22 Turnip, 24 Typha spp., 205

U Urine, 60 Utisols, 388

V Vermiculite, 127 Vertisol, 385 Vetch, 28 vicia spp. 164 Volatilization, 303

W Water, competition for, 4, 5, 6, 18, 29, 47, 77-82, 104, 105 Water erosion, 303-316 calcium losses, 312 nitrogen losses, 307-309 organic losses, 306-307 phosphorus losses, 309-311 potassium losses, 311-312 sampling methods, 305-306 Water plantain, 207 Water systems, 200-208 Weed control, 161-210 cultivated crops, 170-185 fruits and tree nuts, 185-192 pastures and rangelands, 192-200 water systems, 200-208 Weeds, losses caused, 163-165 Wheat, 1, 7, 8, 13, 17, 18, 27, 28, 49, 50, 66, 67, 68, 75, 83, 145, 170, 182, 274, 277, 308, 311 Wheat straw, 264 White clover, 53, 61, 101, 109, 196 White persicaria, 10 Wild garlic, 164, 200 Wild onion, 164, 200 Wild vetch, 164 Willow, 284 Windbreaks, 283-291 Wind erosion, 182, 211-302 control, 270-291 cycle of, 229-249 equation, 291-296 equilibrium forces, 222-229 soil properties, 249-270 surface wind, 216-222 Winter hardiness, 326 Winter wheat, 279, 280

SUBJECT INDEX

Y Yellow-flowered alfalfa, 317

Z Zea mays, 22, 68

Zinc, adsorption of, 126-128, 132-133 availability, 143-152 distribution in soil, 136-140 geochemistry of, 121-124

415

SUBJECT INDEX

This Page Intentionally Left Blank

Y

Yellow-flowered alfalfa, 317

Z Zea mays, 22, 68

Zinc, adsorption of, 126-128, 132-133 availability, 143-152 distribution in soil, 136-140 geochemistry of, 121-124

415